CROP STEERING ACADEMY — EDITION 2026-10-20-r4 All 24 lessons have captioned videos with locally generated Kokoro af_heart narration, AI-generated teaching illustrations, native diagrams and captures of the real application. Teaching images and screenshots do not verify a physical installation. Written using ASD-STE100 principles. Not certified or represented as full formal conformance. All numerical crop settings are examples. Online practice is not hardware certification. A1 — Before the buttons: water, roots and air Look inside a pot first: roots need water and air, water leaves in three ways, and one probe measures only one small spot. Start inside the pot This course starts in a pot, not on a screen. A pot, or container, holds the growing material and the roots. The part of the pot where the roots live is called the root zone. Look at the cutaway on the screen. It shows three things sharing that space: solid material, water and air. Every decision you later make in the app changes how much water and air sit in the root zone. So you need a clear picture of it first. Here is the first question of the course. What do the roots need from the pot? Hold your answer in mind. Most people say water. That is half right. On the next slide you will find the other half. It changes how you think about watering. Inside the root zone Picture a handful of gravel. Each stone is solid, but there are gaps between the stones. The material in a pot works the same way. Growers call the growing material around the roots substrate. Two common substrates are coco, made from shredded coconut husk, and stone wool, made from rock spun into fibres. Look at the cutaway. The solid pieces are the substrate. The gaps between those pieces are called pore spaces. Each pore space holds either water or air. Roots need both. They take up water, and they also need oxygen. Oxygen is the gas in air that keeps root cells alive and working. In the picture, some spaces hold water and some hold air. Now imagine filling every space with water. There would be no room left for air, and the roots would run short of oxygen. So more water is not always better. Good substrate does three jobs, and the picture shows all three. It holds some water in its spaces. It lets extra water run out through the holes in the bottom of the pot. Growers call that drainage. And as water drains away, air moves back into some of the spaces. That mix of held water and returning air is what roots need. Next, you will follow water through a pot step by step. Follow the water Picture a spare pot of substrate with no plant in it, sitting in a tray. Pour water on top. Follow the water in the picture. It enters the substrate at the top. Some of it stays, held in the pore spaces. The rest runs through and drains out of the bottom into the tray. For example, pour in one litre. If a quarter of a litre runs into the tray, then three quarters of a litre stayed in the pot. Over the next hours, the water held in the pot goes down. The picture shows three ways it leaves. First, roots take water in and the plant uses it. Growers call this plant uptake. Second, water at the surface turns into vapour and drifts into the air. That is evaporation. Third, some water keeps draining out of the bottom for a while after watering. All three lower the amount of water in the pot. This matters every time you read a number later. Suppose a pot lost half a litre since the morning. The plant did not necessarily use half a litre. Part may have evaporated, and part may have drained. The picture keeps all three paths side by side for that reason. When water in a pot falls, ask which of the three paths it took before you decide what the plant used. Pause before adding water Here is a real situation. A pot stays very wet all day. The leaves droop and look unhealthy. Many beginners think the plant looks thirsty, so they add more water. Look at the three clues in the picture: a very wet pot, unhealthy leaves, and a question about the cause. Think about what you learned on the last two slides about water filling the pore spaces. You have two choices on the screen. The first is to add more water straight away. The second is to stop and find the cause before you change anything. Pick one now, before the answer appears. Which would you choose? Check first. The picture lists three checks. Look at the crop itself: the leaves, the stems and, where you can, the roots. Check the equipment that delivers water, and make sure it stops when it should. Then look at the reading. A sensor is a device that measures something, such as wetness, and the number it reports is called a reading. Each check takes a few minutes and puts nothing at risk. Why check first? Different problems can look the same. A very wet pot with drooping leaves may mean the roots are short of oxygen, because water fills the spaces. More water would make that worse. Look at the three boxes in the picture. The reading, the equipment checks and what you see at the roots each tell part of the story. Put the three together, then decide. What steering and sensing can do Think of steering a car. You make small turns to keep it heading where you want. Crop steering works the same way. You adjust growing conditions to guide how the crop grows. The picture links the two: adjust the conditions, influence the growth. This course steers with one condition, irrigation, which means supplying water to the crop. The app you will learn, called Crop Steering, is software that helps you plan and run that irrigation. With irrigation you control three things, shown on the screen. When water arrives. How much arrives each time. And how long you wait before the next delivery. The picture also shows what changes how much water the crop needs: light, temperature and humidity. Humidity is the amount of moisture in the air. For example, on a hot, bright, dry day a plant uses more water than on a cool, dim, damp one. Steering has a direction. You can manage water to encourage more leafy growth, or to encourage flowering. Later lessons name these two directions and show how to choose between them. Look at the third label in the picture: keep roots working. Whichever way you steer, the roots must stay healthy, so drying the pot as hard as possible is never the goal. Steering guides growth, and the result still depends on the whole grow. Now the limit of sensors. A thermometer in your kitchen tells you the kitchen temperature, not the bedroom. A moisture sensor works the same way. The part pushed into the substrate is called a probe, and in this course sensor and probe mean the same device. Look at the region marked around the probe in the picture. That small spot is all it measures. Another pot in the same row can be wetter or drier. So you use three kinds of evidence together, shown as three boxes here. Sensor readings give you numbers from one small spot. Equipment checks tell you whether water is really reaching each pot. Crop observations tell you how the plants themselves are doing. When all three agree, you can act with confidence. When they disagree, find out why before you change anything. You will practise in the demo. The demo is a practice copy of the app. It looks and behaves like the real thing, but it runs no real equipment, so a mistake costs nothing. Later, when you set up a real system, start with one zone. A zone is a group of plants that get their water together, at the same time. Starting with one zone keeps any early mistake small. Time to check yourself. Look at the cutaway on the screen. Point to the solid substrate. Point to the water in the pore spaces. Point to the air. Then name one thing a single probe cannot tell you. For example, it cannot tell you whether a pot at the far end of the zone is dry. In the next lesson you will put a number on how wet the substrate is. That number is called volumetric water content, or VWC for short. It describes how much of a sample is water. You will work it out with drawn boxes before you ever see it in the app. INDEPENDENT PRACTICE Draw a pot of substrate. Label the solid material, water and air. Then explain why one probe reading cannot tell you how wet every pot in the zone is. Compare your reasoning: Pore spaces hold water or air, and roots need both. A probe measures one small spot, so other pots in the zone can be wetter or drier. SOURCES Course evidence and production notes: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/PRODUCTION-AND-SOURCES.md METER: soil volumetric water content: https://metergroup.com/measurement-insights/soil-moisture-sensors-how-they-work-why-some-are-not-research-grade/ Athena Commercial Handbook: source guide and background (not a software specification): https://growershub.athenaag.com/athena-handbook-v41iwrqr Athena: Precision Irrigation Strategy: https://support.athenaag.com/hc/en-us/articles/25975395644315-Precision-Irrigation-Strategy A2 — VWC: what does 60% actually mean? Use ten equal boxes to understand volumetric water content, turn a percentage into litres, and learn what a VWC number does not tell you. Recall before the reveal Last lesson you looked inside a pot. The root zone, the part of the pot where roots live, holds three things. Look at the question on the screen. Before the answer appears, try to name all three. Saying them out loud helps you remember. This whole lesson builds on those three parts. Here is a hint. One part is solid, one is a liquid and one is a gas. Take a moment now, before the answer appears. What are the three parts inside the pot? Six out of ten The three parts are solid material, water and air. The solid material is the substrate, such as coco or stone wool. Water and air share the pore spaces, the gaps between the solid pieces. Look at the picture. It draws all three parts as ten equal blocks. Now we want one number that says how wet the substrate is. First, one word. Volume means how much space something takes up. Picture the sample, the bit of substrate you are measuring, split into ten equal boxes. Then ask how many of those boxes the water would fill. That share is called volumetric water content, or VWC. It is the water volume divided by the total sample volume. You then multiply by one hundred to make a percentage, which means a number out of every hundred. Look at the picture. Water fills six of the ten equal boxes. Two boxes are solid material and two are air. Now do the sum step by step. Six divided by ten is zero point six. Zero point six times one hundred is sixty. So this sample is at sixty percent VWC. Sixty out of every hundred parts of its volume is water. The boxes make the sum visible. Real substrate does not look like that. The water spreads through thousands of tiny pore spaces between the solid pieces. It does not sit in six neat blocks. Use the boxes to picture the share of water, not where the water sits. The share is what the VWC number tells you. Put litres beside the fraction Now turn the percentage into litres. A litre is a measure of volume, written as a capital L. Take a bigger sample: five litres of substrate at sixty percent VWC. Sixty percent as a decimal is zero point six. Five times zero point six is three. So the sample holds three litres of water. The picture checks it the other way. Three divided by five, times one hundred, is sixty percent. The picture compares two ideas. One side shows three litres of water already inside the sample. The other side shows a common beginner mistake: reading sixty percent and pouring in another three litres. The number describes what is already there. It is not an amount to add. Three more litres in a five litre pot would mostly run straight out of the bottom as drainage. What VWC does not tell you Back to the sixty percent sample. Sixty percent of its volume is water, so forty percent is something else. Look at the forty percent marked in the picture. Many people get the next question wrong. Is all of that forty percent air? Think about the three parts from the start of the lesson before you answer. No, it is not all air. The forty percent holds the solid material as well as the air. In our ten boxes, two were solid and two were air. So VWC tells you how much water there is, but not how much air. To know the air share, you also need to know how much space the solids take up. That depends on the substrate. You might read VWC like a car's fuel gauge, as if sixty percent means sixty percent of the plant's water is left. It does not work that way. Some water sits loosely in the bigger spaces, and roots take it up readily. Some clings tightly in the smallest spaces, and roots struggle to pull it out. The water roots can actually use is called plant-available water. Substrates hold water differently, so the same VWC can mean different conditions for the plant. Look at the probe in the picture. It measures only the small region of substrate around it. Two things make its reading trustworthy. First, position and contact. The probe must sit firmly in the substrate with no gaps of air around it. Second, calibration. Calibration means checking the probe's number against a known amount of water, and correcting it if they differ. For example, a probe that shows sixty percent when the true value is fifty-five would mislead you every day. Position matters across the zone too. Remember, a zone is a group of plants watered together. A probe in a pot close to where the water comes in may read wet. Meanwhile, a pot at the far end of the row is dry. Look at the two pots in the picture. Later lessons show how to choose probe positions that represent the whole zone. Teach it back Now apply what you know. A probe shows forty percent VWC. Look at the question on the screen. Does forty percent always mean the substrate is too dry and needs water? Decide before you go on. No. Forty percent can mean different things in different substrates. Coco and stone wool hold water differently, so forty percent in one is not the same as forty percent in the other. The picture lists what you use to judge it: your substrate, readings you have checked against your own pots, your crop, and today's conditions. There is no single VWC target that suits every system. Before you choose any target, make sure the meaning is clear. Look at the unfinished sentence on the screen. Sixty percent VWC means what? Say your answer in one sentence. Try to use the words water, volume and sample. Then continue and compare. Here is one good answer. About sixty percent of the measured sample's volume is water. The picture shows it again: six of ten equal boxes are water. Six divided by ten, times one hundred, is sixty percent. If your sentence said the same thing in other words, you have it. The picture shows two limits side by side. First, sixty percent is not the amount of plant-available water. Some of that water is held too tightly for roots to use. Second, one probe describes its own small spot, not every pot in the zone. Keep both limits in mind each time you read a VWC number. You will hear plenty of advice about VWC from other growers and online. Use the two questions on the screen to test it. What does the number actually measure? And in which substrate and pot was it measured? Advice that answers both is advice you can compare with your own grow. The next lesson uses a kitchen sponge. You will see what fully soaked means, which growers call saturation. You will see how much water stays after the extra drains away. And you will meet a reference number in the app, and learn why changing it adds no water. INDEPENDENT PRACTICE Imagine ten equal boxes of substrate. Four contain water. What is the VWC? What fills the other six? Compare your reasoning: 40% VWC. The other six boxes hold solid material and air. They are not all air, and the 40% is not all usable water. SOURCES Course evidence and production notes: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/PRODUCTION-AND-SOURCES.md METER: soil volumetric water content: https://metergroup.com/measurement-insights/soil-moisture-sensors-how-they-work-why-some-are-not-research-grade/ A3 — The sponge, the pot and the Post-it Use a kitchen sponge to understand saturation and container capacity, then see why changing a setting in the app adds no water. Start with a kitchen sponge Last lesson, VWC was the share of a sample that is water. Now picture a kitchen sponge. Dunk it in a sink full of water, then lift it out. It drips. Look at the sponge on the screen. Water has filled the spaces inside it. Growers call this state saturation. Saturation means the pore spaces are filled with water. Notice what has not changed. The sponge material is still there. Water filled the spaces, but it did not replace the solid parts. Substrate behaves the same way. When it is saturated, water fills the spaces between the solid pieces, and those pieces still take up room. So a saturated sample is below one hundred percent VWC. For example, if solids take up one fifth of the volume, water can fill at most eighty percent. Let it drip, without squeezing Now hold the sponge over the sink. Do not squeeze it. Let the extra water drip out on its own. After a while the dripping stops. Look at the sponge in the picture. It still feels wet, but some of its spaces now hold air again. Squeezing would force out water the sponge normally keeps, so you leave it alone. The picture shows two different states side by side. One is soaked and dripping: saturation. The other is wet after draining: water held, with air back in some spaces. A pot goes through the same two states after a big watering. Roots need oxygen, so the second state, with some air, is the one that lets them breathe. Keep these two states separate in your mind. From the sponge to the pot Now move from the sponge to a real pot. Take a spare container of substrate with no plant in it. Water it until it is saturated. Then let the extra water drain out of the bottom on its own. Do not press or squeeze the substrate. Look at the picture. It is the same idea as the sponge, now in a pot. When the draining stops, look at what is left. The cutaway shows water still held in the spaces, and air back in some of them. The amount of water the pot holds at that point has a name. Growers call it container capacity. It is the water your substrate holds in that container once the extra has drained away. It is lower than saturation. Container capacity depends on the setup. A tall pot drains more water from its upper part than a short, wide pot of the same substrate. So two pots of the same coco can hold different amounts. Different substrates hold different amounts too. The two labels in the picture sum it up: the substrate and the container both matter. That is why you check your own pots rather than borrow someone else's number. Now the app. It needs numbers to compare readings against. Each one is called a reference value. One reference value is named field capacity. The name comes from farming in open fields, but in the app it stands for your container capacity. The picture shows the order: first the container check, then the number. The app can also update this number by itself from peaks. A peak is the highest reading, usually right after watering. Check which way yours is set. To trust your container check, repeat it the same way each time. Use the same substrate, the same container, the same starting wetness and the same way of draining. There is no single drainage time that suits every setup, so write down how long you let it drain. If two repeats give close results, your method is working. Then compare the probe with what you can see and weigh. Look at the probe in the picture. It must sit firmly in the substrate, in a sensible spot. Weigh the container when saturated, and again after draining. A litre of water weighs one kilogram. So a pot that is three hundred grams lighter lost about three hundred millilitres. A millilitre, written mL, is one thousandth of a litre. If plenty drained but the reading barely moved, check the probe's placement and contact. Now a worked example. After draining, a one litre substrate sample holds six hundred millilitres of water. First put both in the same unit. One litre is one thousand millilitres. Six hundred divided by one thousand is zero point six. Times one hundred, that is sixty percent VWC. So in this example the container capacity is sixty percent. Your own pot may give a different number. A Post-it cannot add water Imagine writing sixty percent on a sticky note, a Post-it, and putting it beside the drained sponge. The note describes the sponge. Look at the picture: a wet sponge and its note. Now cross out sixty and write seventy. Has the sponge gained a single drop of water? No. Writing on the note changes the note, and nothing else. The app works like that note. A value you type into the app is called a setting. The field capacity reference is one such setting. It tells the app which number to compare readings against. Look at the picture. The note now reads seventy, and the sponge is unchanged. Change the setting from sixty to seventy, and the pot holds the same water as before. The new number can change what the app does next, but the edit itself adds no water. Change the note: what happens? Test the idea on a pot. You change the field capacity reference in the app from sixty to seventy percent. Look at the question on the screen. What happens to the water already in the pot at that moment? Nothing happens to the water. You changed the note, not the pot. The picture shows it: the note reads seventy, and the sponge is the same. For the water to change, water must physically enter the pot by watering. Or it must leave by drainage, evaporation or plant uptake. A changed setting can change what the app decides later, but it moves no water by itself. Remember that sixty percent was a worked example. It came from six hundred millilitres in a one litre sample. Your own substrate and container may give fifty-two percent, or sixty-five. The picture puts the example beside a reminder: your pot needs its own check. Use the number from your own container check, not one from a lesson. One more trap. Remember the peak, the highest reading, usually right after watering. At that moment the substrate is close to saturation and still draining. So the peak usually sits above container capacity. A single sharp spike can also be a faulty reading. Look at the two questions on the screen. Before you use any peak as your reference, compare it with your drained container check. Now bring it all together with the sponge. Picture three things in order. First, the sponge soaked and dripping. Second, the same sponge wet after the extra water drained away. Third, a note beside it with a number. The picture shows the first two. They are the sponge's real, physical states. The note is different. It holds no water. It describes the sponge. In a pot, saturation and container capacity are physical states of the substrate. The field capacity reference is a setting in the app, like the note. It is useful when it comes from a careful container check. Look at the picture: the real sponge on one side, and the note on the other. In the next lesson the pot starts to dry. After watering, water leaves by plant uptake, evaporation and drainage, so the readings fall. You will learn two ways to measure that fall. As you go, keep the sponge and its note separate in your mind. One is the water. The other is a number about the water. INDEPENDENT PRACTICE Over a sink, soak a kitchen sponge, lift it dripping, then let the extra water drain without squeezing. Describe what changed. Write a number on a note beside it, then change the number. Did the sponge gain water? Explain how this maps to a pot and a setting in the app. Compare your reasoning: You tell saturation apart from container capacity, and explain that changing a setting adds or removes no water. Your own pot needs its own repeatable container check. SOURCES Course evidence and production notes: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/PRODUCTION-AND-SOURCES.md METER: soil volumetric water content: https://metergroup.com/measurement-insights/soil-moisture-sensors-how-they-work-why-some-are-not-research-grade/ A4 — Dryback without the confusing maths Use a bank balance to understand dryback, then describe a fall in water content two ways: in percentage points and as relative dryback. Think of a bank balance In lesson A three you watched water fill a pot and drain. Now we follow what happens next, as the pot slowly dries. Start with something familiar. Your bank balance is sixty dollars in the morning. By evening it is forty-five dollars. Fifteen dollars has gone. You know how much left, but not yet where it went. Water in a pot behaves the same way. Look at the line on screen. Remember, VWC is the share of the sample volume that is water. The probe read sixty percent VWC earlier and forty-five percent later. Water left the pot between those two readings. Growers call this fall in water content dryback. Some of that water went into the plant. Some evaporated from the surface, and some drained out. Dryback counts all of them together. Always compare two readings from the same probe, because each probe samples its own small spot. Next, we measure the size of the fall. First find the size of the fall Back to the bank balance. Sixty minus forty-five is fifteen dollars. Water content works the same way, with two new words. The first reading in a comparison is called the starting reading. Here it is sixty percent VWC. The ending reading is forty-five percent VWC. Subtract the end from the start. Sixty minus forty-five is fifteen. Look at the bar marked drop on screen. Because we subtracted one percentage from another, we do not call the answer fifteen percent. We call it fifteen percentage points. A percentage point is the plain gap between two percentages. Each reading also needs the time it was taken. That recorded time is called a timestamp. Say the readings came at eight in the morning and two in the afternoon. Write those timestamps beside both readings. Without them, nobody can tell whether the fall took one hour or one day. Next, we compare the fall with where it started. Then compare the fall with the start Fifteen points sounds like one fixed amount. But picture two pots. One starts at sixty percent, the other at thirty percent. Losing fifteen points is a quarter of the first pot's water. It is half of the second pot's water. So growers also compare the fall with the starting reading. Divide the fall by the starting reading. Then multiply by one hundred to turn the answer into a percentage. This result is called relative dryback. Look at the fraction on screen, with fifteen on top and sixty below. Fifteen divided by sixty is zero point two five. Zero point two five times one hundred is twenty-five. So the relative dryback is twenty-five percent. The same fall now has two correct descriptions: fifteen percentage points, or twenty-five percent relative dryback. Always divide by the starting reading, because that is the amount the fall is compared with. Next, you try one with the steps laid out. Your turn, with the steps supplied Now you try, with the steps laid out. Look at the two numbers on screen. The starting reading is eighty percent VWC. The ending reading is sixty percent. Step one: subtract the end from the start. Step two: divide that fall by the starting reading, eighty. Step three: multiply by one hundred. Pause the video now and write both answers. Here they are. Eighty minus sixty is twenty, so the fall is twenty percentage points. Twenty divided by eighty is zero point two five. Times one hundred gives twenty-five percent relative dryback. Notice something useful. The first example also gave twenty-five percent, though its fall was only fifteen points. Different falls can give the same relative dryback when the starting readings differ. That is why every answer needs its label, points or relative. Next, you work one with no steps shown. Try it without the steps This time there are no steps on screen. The starting reading is fifty percent VWC, taken at seven in the morning. The ending reading is forty percent, taken at eleven. Work out both descriptions, then pause. Here is the working. Fifty minus forty is ten, so the fall is ten percentage points. Ten divided by fifty is zero point two. Zero point two times one hundred is twenty, so the relative dryback is twenty percent. Now write the full record as a grower would. Fifty percent at seven, forty percent at eleven, same probe. Ten points, or twenty percent relative dryback. Another person can check every step from that one line. Keep one idea from lesson A two in mind. Forty percent VWC tells you how much water is in the sample. Judge how easily roots can use it from your substrate and crop. Next, we see how the app uses relative dryback. Read the setting before entering a number Now connect this to the app. In lesson A three you met three ideas. A setting is a value you type into the app. A reference value is a number the app compares readings against, like the Post-it. The peak is the highest water-content reading the probe recorded. One setting is called morning dryback. It tells the app how far the pot should dry below its peak before the first watering of the day. Lesson B two explains that morning period. The app reads this setting as relative dryback, measured from the peak. Look at the three numbers on screen. The peak is sixty percent VWC. The setting is ten percent. Ten percent of sixty is six points. Sixty minus six is fifty-four. So the app's reference is fifty-four percent VWC. If you read ten as points, you would expect fifty percent and misjudge when watering starts. Before typing any number, read the help text beside the setting to learn which kind it is. Next, we recall the whole method. Recall the method Time to pull the lesson together. Cover the screen and say the method aloud. First, find two readings from the same probe, each with its timestamp. Second, subtract the ending reading from the starting reading. That gives the fall in percentage points. Third, divide that fall by the starting reading and multiply by one hundred. That gives relative dryback. Look at the three boxes on screen. The first holds what the probe observed. The second holds what you calculated from those observations. The third holds the reference the app works out from its setting. Keep these three apart in your notes. A setting changes the reference, not the water in the pot, like the Post-it in lesson A three. Try one more aloud. Seventy-five percent falls to sixty. That is fifteen points. Fifteen divided by seventy-five is zero point two, so twenty percent relative dryback. The next lesson moves from the water itself to the salts dissolved in it. INDEPENDENT PRACTICE Calculate both dryback descriptions for 70% to 56% VWC. Then find the VWC reference for a 10% morning dryback setting from a 50% peak. Compare your reasoning: 14 percentage points; 20% relative dryback; 45% VWC reference. Name the starting reading used in each relative calculation. SOURCES Crop Steering: GROW_PLANS.md: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/GROW_PLANS.md Crop Steering: training/crop-steering-from-zero/WORKBOOK.md: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/WORKBOOK.md Athena Commercial Handbook: source guide and background (not a software specification): https://growershub.athenaag.com/athena-handbook-v41iwrqr Athena: Precision Irrigation Strategy: https://support.athenaag.com/hc/en-us/articles/25975395644315-Precision-Irrigation-Strategy A5 — EC: what it measures, and which EC you are looking at Use salted water to understand EC, then label each reading by where it came from: feed, bulk, pore-water, runoff or lab extract. Start with salted water In lesson A four you tracked how much water is in the pot. Now we ask what is dissolved in that water. Stir a spoon of table salt into a glass of water. The salt seems to vanish, but it is still there, split into tiny charged pieces. We call this a dissolved salt. Plant food works the same way. Nutrients are the minerals a plant needs to grow, such as nitrogen, potassium and calcium. Growers supply them as salts dissolved in water. Those charged pieces let electricity pass through the water. Look at the two beakers. The right one has more dissolved pieces, and its reading shows higher conductivity. How easily current passes is called electrical conductivity, or EC for short. The device that measures it is called a meter. EC tells you roughly how much salt is dissolved in total. It cannot tell you which nutrients are there, or how much of each. Next, we follow the water to see where EC gets measured. Ask where the reading came from The same water travels through several places, and EC can be measured at each one. Follow it on screen from left to right. First, the tank: the container where nutrient solution is mixed before watering. The EC of that solution is called feed EC. Second, the substrate in the pot. A probe buried there measures the wet substrate as a whole: water, solid pieces and air together. That is called bulk EC. Third, the water held in the pore spaces between the solid pieces. This is the water roots actually touch. Its EC is called pore-water EC. Fourth, collected samples. Water that drains out of the bottom of the pot after watering is called runoff. A laboratory extract is different. A lab mixes substrate with water using a written method, then measures the liquid. All of these can use the same unit, yet each measures a different thing. So a label always goes beside the number. Next, we label two real readings. Worked example: label two readings First, the unit. Conductivity is measured in siemens, named after an engineer. Plant solutions conduct only weakly, so we use thousandths of a siemens, called millisiemens. The meter measures across a set distance of one centimetre. So the full unit is millisiemens per centimetre, written mS slash cm. Now two readings. A meter dipped in the tank reads two point zero millisiemens per centimetre. Label it feed EC. The second comes from a TEROS twelve. That is one specific probe model, made by a company called METER. You push its three metal needles into the substrate. It measures water content, temperature and EC at the same spot. Its EC output is bulk EC. Here it reads zero point eight millisiemens per centimetre. Look at the two bars on screen. Bulk EC reads lower partly because the probe also senses solid pieces and air, which carry little current. So label each reading first, then compare like with like. Next, we see how software turns bulk EC into an estimate. A calculated value needs a method Pore-water EC is closest to what roots experience. But you cannot easily push a meter between substrate pieces. So some software works it out instead. It feeds the bulk EC reading into a formula. The formula also needs other readings from the same probe, such as water content and temperature. A number worked out this way is called an estimate. It is calculated, not measured directly. Look at the chain on screen. On the left is the direct bulk reading. In the middle is the formula. On the right is the estimated pore-water EC. For example, a bulk reading of zero point eight goes in, and a larger number comes out. That is because pore water carries the salts without the solid pieces and air. Every formula has a range where it works well, and these formulas are usually less reliable in very dry substrate. Find the app's note on which formula it uses, and keep the word estimate on your record. Next, you finish some labels yourself. Your turn: finish the labels Your turn. Three samples are on screen, and you name each one. Sample one is the solution in the tank, ready to go to the pots. Sample two is a cup of water caught from the pot's drain hole after watering. Sample three is the water held in the spaces between substrate pieces. Pause now and say all three labels aloud. Here are the answers. Sample one is feed EC. Sample two is runoff EC. Sample three is pore-water EC. One more detail about runoff. Its EC depends on how and when you collect it. Water caught in the first minute can read differently from the last drops. So write the collection method beside the result. For example: all the drainage from one watering, caught in one cup. Use that same method every time you want to follow a trend. Next, a fresh record with real numbers. Try a fresh record Here is a fresh record. The tank reads one point seven millisiemens per centimetre. The drainage cup reads two point four. Start with the labels. The tank solution is feed EC. The drainage cup is runoff EC. Look at the two bars. Two point four minus one point seven is zero point seven, so runoff is higher by zero point seven. Before you explain that gap, check two things about the measurements. First, check the meter. Dip it in a standard solution, a bottle sold with a known EC, and confirm the meter shows that number. Second, check temperature. Warm water conducts better than cold water, so the same solution reads higher when warm. Most meters correct for this automatically to a standard temperature, usually twenty-five degrees Celsius. This correction is called temperature adjustment. Confirm both readings used it. With the meter checked and both readings adjusted, you can compare the two numbers fairly. Next, we pull the lesson together. Recall what the labels mean Cover the screen and name the places EC is measured: feed, bulk, pore-water, and collected samples such as runoff or a lab extract. Now look at the four words on screen. Every EC reading needs all four. Type: which of those places it came from. Unit: for example, millisiemens per centimetre. Method: how it was measured or collected, and whether it is an estimate. Context: the crop, the substrate and the time. Keep EC and VWC apart in your head too. VWC tells you how much of the sample is water. EC tells you how easily that water carries current, which is a clue to dissolved salts. For how much water or food the roots can actually use, combine these readings with your substrate, crop and checks. With all four labels in place, you can compare like with like. The next lesson asks whether you can trust a reading at all. INDEPENDENT PRACTICE Label these new samples: supply solution at 1.9 mS/cm; direct TEROS 12 output at 0.6 mS/cm; water extracted by a stated laboratory procedure at 2.7 mS/cm. Explain why these numbers cannot identify an individual deficient nutrient. Compare your reasoning: Feed EC; bulk EC; extract EC. EC gives conductivity information, not the concentration of each nutrient. SOURCES METER: Soil electrical conductivity, measurement types and conversions: https://metergroup.com/education-guides/soil-electrical-conductivity-the-complete-guide-to-measurements/ METER TEROS 11/12 manual: measurements, installation and calibration: https://publications.metergroup.com/Manuals/20587_TEROS11-12_Manual_Web.pdf Crop Steering: training/crop-steering-from-zero/WORKBOOK.md: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/WORKBOOK.md A6 — Can you trust that reading? Check every reading with three questions: is it the right probe, is it recent, and is the measurement reliable? A reading is a dated local observation You now know what VWC and EC mean. This lesson asks a different question: can you trust the number in front of you? Think of a thermometer in the hallway. It cannot tell you the temperature in the bedroom. Now think of a photo taken yesterday. It cannot show the room as it looks today. A probe reading has both limits. It describes one small spot, at one moment. Look at the three checks on screen. One: is this the right probe, in the place you think it is? Two: is the reading recent? Three: is the measurement itself reliable? Ask all three every time, because each one can fail on its own. A sensible-looking number can come from the wrong pot, from four hours ago, or from a badly placed probe. We take the three questions one at a time. Next: how old is the reading? Worked example: how old is it? Lesson A four introduced the timestamp: the time a reading was taken. To find a reading's age, subtract its timestamp from the time now. Look at the clock on screen. It is fourteen hundred, which is two in the afternoon. Probe A last sent a reading at ten hundred, ten in the morning. Fourteen minus ten is four, so that reading is four hours old. Probe B last sent a reading at thirteen fifty-nine. That is one minute ago. Growers call how recent a reading is its freshness. A reading too old to describe the pot now is called stale. Probe A is stale. Probe B is fresh. Watch for one trap. When a probe stops sending, many graphs keep drawing a flat line at the last value. A flat line can mean steady conditions, or no new readings at all. The timestamp tells you which. Next, we check that the probe is the one you think it is. Check identity and units Next, identity. First, one new word. A room is a space with its own lights. It holds one or more zones, and each zone is a group of plants watered together. Every probe should carry a physical tag naming its room and zone. Look at the tag on the probe and the name in the app. They must match. If the app says room two, zone one, the probe should sit in room two, zone one. A swapped cable can make one zone's probe report under another zone's name. Matching tag to screen is called checking probe identity. Then check units. Some probes send VWC as a fraction instead of a percentage. The unit is cubic metres of water per cubic metre of substrate. A value of zero point four five means forty-five hundredths. Multiply by one hundred to get forty-five percent VWC. Reading it as zero point four five percent makes it one hundred times too small. That is called a scale error. Next, we look at the probe in the pot. A fresh reading can still be inaccurate A reading can be fresh and correctly named, yet still wrong. Look at the probe in the pot on screen. In lesson A two, calibration meant checking how a probe's output relates to the real quantity. A probe set up for one substrate can misread in another, so the calibration must suit your substrate. Placement matters too. Placement means where in the pot the probe sits. A probe right beside the spot where water enters reads wetter than the far side of the pot. Then there are air gaps. The probe's needles must be in firm contact with the substrate. A pocket of air around a needle makes the probe read drier than the pot really is. Growers call these pockets air gaps. For example, two identical pots might read fifty percent and forty-one percent only because one probe has an air gap. Push each probe in fully, and firm the substrate around it. Then record where and how you installed it. Next, you practise the age check, with one more step. Your turn: calculate, then check Your turn. It is fifteen thirty, half past three in the afternoon. The last valid reading arrived at fifteen twenty-five. Valid means the probe sent a real number, not an error message. Pause and work out the age. Fifteen thirty minus fifteen twenty-five is five minutes. So the reading is fresh. Now look at the three rows on screen. Age is answered. Identity is checked, because the tag matches the app. The third row is still open: placement and calibration. To close it, look up the installation record. It should show where the probe sits, that it was pushed in firmly, and which calibration it uses. Freshness tells you the reading is recent, and this record tells you the measurement is sound. With all three rows ticked, you can use the number. Next, two probes that disagree. Try a changed scenario Now two probes in the same zone disagree. It is sixteen twenty. Probe A shows forty-eight percent, timestamped twelve twenty. Probe B shows forty-three percent, timestamped sixteen nineteen. Work out the ages first. Sixteen twenty minus twelve twenty is four hours, so A is stale. B is one minute old. So these two numbers describe different moments. Probe A may have stopped reporting. Check in this order. First, identity: do both tags match the screen? Second, communications. That means the cable or wireless link that carries readings from the probe to the app. A loose plug or a dropped wireless connection stops new readings arriving. Third, physical evidence in the pot: placement, air gaps, and how heavy and wet the pot feels when lifted. Treat the gap between forty-eight and forty-three as unexplained until those checks are done. Next, what to do when sensing fails. Recall the three questions Say the three questions aloud: right probe, recent reading, reliable measurement. Sometimes sensing fails and you cannot fix it straight away. Every site should have a written fallback procedure. That is a plan for watering safely without trusted probe readings, for example by lifting pots to judge their weight. Your site also has a responsible operator: the person in charge of irrigation decisions there. Tell them what failed and follow the procedure with them. Look at the three boxes on screen. First, a fault record: which probe, what you saw, and when. Second, independent evidence, meaning a check that does not rely on the faulty probe, such as lifting a pot. Third, the local fallback procedure itself. Keep the app's safety settings where they are and fix the sensing. Changing them to force water hides the fault and can harm the crop. The next lesson shows how to read these readings on a graph. INDEPENDENT PRACTICE At 09:45, probe A last reported 55% at 06:45. Probe B reports 50% at 09:44. Explain what is known, what remains unknown and your first three checks. Compare your reasoning: A is three hours old; B is one minute old. Current zone uniformity is unknown. Check identity and timestamps, then communications, then placement, air gaps and calibration. SOURCES METER TEROS 11/12 manual: measurements, installation and calibration: https://publications.metergroup.com/Manuals/20587_TEROS11-12_Manual_Web.pdf Crop Steering: USER_GUIDE.md: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Crop Steering: training/crop-steering-from-zero/WORKBOOK.md: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/WORKBOOK.md B1 — Read water and EC on the same graph Read a time graph with two axes, find the shots, and tell a measured line from a planning line and a draft. Start with time Think of a car dashboard. The speedometer and the fuel gauge sit on one panel, but each one tells you something different. A graph does the same job for the root zone. A graph is a picture of readings over time. Each reading becomes a point, and a line joins the points. This graph has two lines. The small key beside it, called the legend, tells you which line is which. The blue line is VWC, the water content you met in lesson A two. The orange line is EC, the conductivity you met in lesson A five. Now look along the bottom edge. This edge is the time axis. An axis is a ruler printed along one edge of a graph. Read time from left to right, like the pages of a diary. Moving right means later. For example, six in the morning sits to the left of noon, and noon sits to the left of six in the evening. The shaded band is night, when the lights in the grow room are off. The probes keep measuring through the night, so both lines carry on across the band. Next, you read the blue line against its own ruler. Read water content on the left The blue VWC line has its own ruler on the left edge. This left axis is marked in percent. Its scale, the spacing of the numbers, might run from thirty percent at the bottom to seventy percent at the top. To read the line, pick one point on it. Look straight across to the left axis to read its value. Then look straight down to the time axis to see when it was measured. Try it at the marked point. Straight across, the left axis reads sixty percent. So at that moment, sixty percent of the sampled volume was water. The line here is solid because it joins real readings from the probe. Growers call this the measured line. A smooth line can still be old. Lesson A six showed that every reading has an age, the time since the probe last reported. Before you trust the line, make three checks. Confirm which probe drew it. Read the time of its last reading. Look for gaps where readings are missing. If the last reading is four hours old, the line describes four hours ago. To know about now, you need a fresh reading. Add EC and a second axis Remember the salted water from lesson A five. More dissolved salt lets electrical current pass more easily, and EC measures how easily it passes. The orange EC line uses a second ruler on the right edge, called the right axis. It is marked in millisiemens per centimetre, written mS slash cm. Read an orange point by looking straight across to the right axis, never the left. Next, ask where this EC came from. A line on a graph is only as good as the device behind it. Follow the orange line back to the probe that supplied it. A reading from the supply tank describes the solution before it reaches the plants. To learn about the water around the roots, you need a probe in the root zone. Lesson A five gave you three names. Feed EC is measured on the solution you supply from the tank. Bulk EC is measured by a probe in the wet substrate, across the whole mix around it. Pore-water EC describes only the water in the spaces between the solid pieces. Check the line's name in the legend to see which one your graph shows. Now look where the blue and orange lines cross. Picture the speedometer needle and the fuel needle pointing the same way. Would that mean your speed equals your fuel? Ask the same question here. Does a crossing mean VWC and EC are equal? No. Each line is drawn against its own ruler. At the crossing, the blue line might read fifty percent on the left axis. The orange line might read two point five mS slash cm on the right axis. They meet on the screen only because of how the two rulers are spaced. Read each line on its own axis, and follow how each one changes over time. So never subtract one line from the other. Taking EC away from VWC is like taking litres away from kilometres per hour. The two numbers measure different things in different units, so the answer means nothing. Compare VWC with VWC, and EC with EC. Next, you add the moments when water was given. Shots, plans and measurements One watering, from start to stop, is called a shot. Each shot appears on the graph as a small mark at its start time, called an event marker. The decision to water comes from the Controller, the part of Crop Steering that decides when to water. Lesson C one explains it fully. In the app, a filled dot marks a shot that started. A ring marks a shot the Controller held back. After a dot, look for the blue line to step up as water reaches the probe. Between shots, the line slopes down again. The orange line does not have to move with the blue one. After a shot, EC can rise, fall or stay level. Mixing in the substrate, drainage, temperature and the plant all affect it. So when the two lines change together, treat it as a clue to check, rather than a cause you have found. Think of a journey plan beside the trip you actually drove. The graph can show both. The dashed line is the planning reference: the path the app's settings aim for. The solid line is the measured line: what the probe recorded. For example, the plan might say sixty percent at noon while the probe read fifty-eight. Each line shows its own number. Keep the two apart, even where they overlap. Through the night; move a threshold Follow the dashed planning line into the shaded night. It keeps going to the next morning, like a route on a map that covers the whole journey. This lets you see the plan for a full day and night in one view. The dashed line shows what the settings plan. Tomorrow, the solid line will show what really happened. The dashed orange EC line does something simpler at night. The app sets no separate EC aim for the night. Instead, it draws a straight join from the last daytime EC value in the plan to the next morning's value. For example, it might join two point eight in the evening to three point zero the next morning. Read that join as a link between two settings, not a prediction of salts. Now a new idea. A threshold is a trigger line: when the reading crosses it, the Controller may act. This graph has a night-time emergency threshold for VWC. If the root zone dries below it overnight, the Controller may give a protective shot. Watch a practice edit in the demo. The draft threshold changes from thirty-five percent to thirty-nine percent, a rise of four percentage points. A draft is a proposed change that is not saved yet. The saved line at thirty-five percent stays on screen for comparison. The measured blue line does not move at all. Explain what changed Test yourself. After that edit, one thing changed. Was it the past water content, the draft threshold, or a shot booked for tonight? Only the draft threshold changed. A draft stays on your screen until you save it. Saving is a separate step that sends the value to the system. Even a saved threshold books no shot. Think of a fridge alarm limit: it marks trouble, not the temperature you aim for. The Controller checks its readings and safety rules each time before it waters. Now read the whole graph in one pass. Point to the time axis. Point to the left VWC axis and the right EC axis. Find one shot marker. Find the dashed planning line. Then find the time of the last reading, so you know its age. In the next lesson, you split the watering day into four parts, labelled P zero to P three. INDEPENDENT PRACTICE Open the practice graph. Drag the example emergency threshold from 35% to 39%. Which line moves? Can this draft change yesterday’s measured readings? Compare your reasoning: Only the draft threshold line moves. The measured line stays exactly where it was: a setting cannot rewrite recorded readings. SOURCES Course evidence and production notes: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/PRODUCTION-AND-SOURCES.md METER: soil volumetric water content: https://metergroup.com/measurement-insights/soil-moisture-sensors-how-they-work-why-some-are-not-research-grade/ Athena Commercial Handbook: source guide and background (not a software specification): https://growershub.athenaag.com/athena-handbook-v41iwrqr Athena: Precision Irrigation Strategy: https://support.athenaag.com/hc/en-us/articles/25975395644315-Precision-Irrigation-Strategy B2 — P0 to P3: build the day one phase at a time Split the grow-day into four phases, learn what starts and ends each one, and tell forecast shots from shots that really ran. Four phases across the grow-day Think of a working day: arrive, warm up, keep going, then go home. A grow room has a daily rhythm too, set by its lights. Lights-on is the time the grow lights switch on. Lights-off is the time they switch off. One grow-day runs from one lights-on to the next lights-on. For example, with lights on at six in the morning, the grow-day lasts twenty-four hours, until six the next morning. The app splits each grow-day into four parts called phases. A phase is one stage of the day with its own job and its own rules. Look at the strip. The four phases are labelled P zero, P one, P two and P three: wait, refill, maintain, then dry back overnight. Knowing the order lets you read any day on the graph. Next, the first phase. Wait first: P0 morning dryback At lights-on, every zone starts in P zero, the morning wait. No water is given, so the root zone keeps drying, like a sponge draining before you top it up. Lesson A four called this fall dryback. P zero aims for a dryback target. A target is the value a phase is trying to reach. P zero ends at the first of three events. The dryback reaches its target. Or the VWC falls to the P two re-water threshold, which you meet on slide four. Or the time limit runs out; the app calls that limit the maximum wait. For example, lights-on is six in the morning, the maximum wait is ninety minutes, and drying is slow. Six o'clock plus ninety minutes is half past seven, so P zero ends by then at the latest. That is why lights-on alone cannot tell you when the first shot comes. Refill: P1 ramp-up The second phase is P one, the refill. The app calls it ramp-up. The Controller gives a series of shots, one watering each, with a pause between them. Each shot lifts the VWC a step, like filling a bath one bucket at a time. The refill aims for the P one target. For example, P zero ended at fifty-four percent and the target is sixty. That is six percentage points to climb. If each shot lifts the reading about two points, six divided by two gives about three shots. Some water drains, and the probe sees only part of the pot, so the real count can differ. To know how many shots ran, read the record of shots, not the drawing. Next, the long middle of the day. Maintain: P2 and the re-water threshold The third phase is P two, maintain. Think of cruise control: the car holds its speed instead of speeding up. P two holds the VWC near the top of the refill. It uses a threshold called the re-water threshold. Each time the VWC falls to that line, the Controller can give one shot. Some screens label it the re-water point; it is the same line. For example, the refill ends at sixty percent and the re-water threshold is fifty-seven. That is a three-point fall. If the zone dries two points an hour, three divided by two gives one and a half hours between shots. The planning chart draws shots at that spacing. A drawing of shots the plan expects is called a forecast. If the zone dries more slowly, fewer shots happen, and that is normal. Worked example: a 06:00–18:00 lighting day Now put the phases on a clock. Our example room has lights on at six in the morning and lights off at six in the evening. Six in the morning plus twelve hours is six in the evening, so the lights are on for twelve hours. At six in the morning, each zone enters P zero and waits. When P zero ends, P one refills, then P two maintains. Those hand-overs depend on the readings, so the clock cannot give their times. At lights-off, six in the evening, every zone moves to P three, whatever phase it was in. The Controller can also start P three a little earlier, so the night's dryback finishes by morning. P three lasts until six the next morning, twelve hours later, when a new grow-day starts. Expect phases by their order, not by fixed clock times. Overnight: P3 and the Rescue floor The fourth phase is P three, the overnight dryback. The lights are off, and normally no water is given. The root zone dries slowly, like a sponge left on the sink overnight. One protection stays awake. In lesson B one you dragged a night-time emergency threshold. Growers call that line the emergency floor. In the app it is labelled Rescue floor, and this course uses that name from here on. If the VWC falls below the Rescue floor overnight, the Controller can give a rescue shot. A rescue shot is a small protective shot that stops the root zone drying too far. For example, with the floor at thirty-five percent, a reading of thirty-six gets no shot. A reading of thirty-four is below the floor, so a rescue shot can follow. Treat the floor as an alarm line, never as a level to aim for each night. Your turn: explain another day Pause and fill the gaps. The morning wait is P zero, and the maintain phase is P two. Now picture a room lit from eight in the morning. Eight in the morning plus twelve hours is eight in the evening, so lights-off is at eight in the evening. P zero starts at eight in the morning. P three runs from eight in the evening to eight the next morning. The P one and P two times stay unknown until the readings decide them. Last, separate two lists of shots. A forecast shows the shots the plan expects. The event record lists the shots that really ran, with their times. When you check what happened, read the event record. The next lesson shows how a slider tilts these settings toward one kind of growth. INDEPENDENT PRACTICE For a fictional 07:00–19:00 lighting day, order the four phase cards and give one job per card. Then explain why a forecast of six P2 shots does not promise six shots. Compare your reasoning: P0 from 07:00, then P1, P2, and P3 from 19:00 at the latest to 07:00. P2 shots follow the readings, so the count can differ. The Rescue floor is an alarm line, not a nightly target. SOURCES Crop Steering: GROW_PLANS.md: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/GROW_PLANS.md Crop Steering: USER_GUIDE.md: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Athena Commercial Handbook: source guide and background (not a software specification): https://growershub.athenaag.com/athena-handbook-v41iwrqr Athena: Precision Irrigation Strategy: https://support.athenaag.com/hc/en-us/articles/25975395644315-Precision-Irrigation-Strategy B3 — Vegetative and generative: a bias, not a magic switch Use a dimmer picture to learn vegetative and generative steering, then calculate how the slider blends two endpoints in a profile. Think of a dimmer Think of a dimmer switch. It does more than switch a light on or off. It slides smoothly between the two. Keep that picture for this lesson. A plant puts its effort into two broad kinds of growth. Vegetative growth builds roots, stems and leaves. Generative growth builds flowers and fruit. Growers can nudge that balance through watering. A planned nudge toward one side is called a steering bias. Bias is your intention, separate from the plant's age. A young plant and a flowering plant can each get either bias. For example, a grower may give a flowering plant a vegetative bias for a week to help it grow more leaf. The plant decides how it responds, so you watch the crop as well as the settings. Next: how watering does the nudging. How watering nudges the plant How does watering nudge the plant? Think of the sponge again. Smaller drybacks keep the root zone more evenly wet. In suitable conditions, that tends to support vegetative growth. Larger drybacks let it dry further between shots, which tends to support generative growth. For example, compare a zone drying from sixty to fifty-seven percent each morning with one drying from sixty to fifty-one. The first fall is three points, a gentler, more vegetative pattern. The second is nine points, a more generative pattern. EC, light and climate also push the balance. So do the substrate, the plant variety and its growth stage. These are tendencies, so check the crop as well as the numbers. Next: how the app stores a set of numbers for each direction. Endpoints, profiles and the slider Now the dimmer becomes a setting in the app. Picture two complete sets of numbers, one for each end of the range. Each set is called an endpoint. The vegetative endpoint holds the settings you want for a fully vegetative day. The generative endpoint holds the settings for a fully generative day. Together, the two endpoints are saved as a profile, so you can reuse them. Between them sits the slider, a control you drag from zero to one hundred. Look at the scale. Zero uses the vegetative endpoint. One hundred uses the generative endpoint. Fifty sits halfway, so each setting lands halfway between its two endpoint values. The app calls this mixing a blend. You choose the endpoint values yourself, because the app cannot work them out from your pot size. Next, a blend worked step by step. Worked example: halfway Work one setting through. Take the maximum wait from lesson B two, the longest P zero can last. Say the vegetative endpoint is thirty minutes and the generative endpoint is fifty. First, find the difference: fifty minus thirty is twenty minutes. Next, turn the slider into a fraction: fifty out of one hundred is one half. Then take that fraction of the difference: half of twenty is ten. Last, add it to the vegetative endpoint: thirty plus ten is forty minutes. The app then rounds the result to the setting's step, the smallest change that setting accepts. With a one-minute step, forty stays forty. A result of forty point four would round to forty. The same four steps work for every setting in the profile. Next, you try one. Your turn, with the steps supplied Your turn, with the steps laid out. Keep the endpoints at thirty and fifty minutes, and set the slider to twenty-five. Pause and work it out before you listen on. The difference is still twenty minutes. Twenty-five out of one hundred is one quarter. One quarter of twenty is five. Thirty plus five is thirty-five minutes. Now a special case. Suppose both endpoints are thirty minutes. The difference is zero, and any fraction of zero is zero. So the setting stays at thirty wherever the slider sits. Existing set-ups often start this way, because both endpoints copy your current settings. If the slider seems to change nothing, check whether the endpoints are equal. Try a new pair on your own Now try one on your own. The endpoints are twenty and sixty minutes, and the slider is at seventy-five. Work it out first, then check. The difference is sixty minus twenty, which is forty minutes. Seventy-five out of one hundred is three quarters. Three quarters of forty is thirty. Twenty plus thirty is fifty minutes. Notice the trap: seventy-five is a position on the slider, not a time. It never goes into the answer as minutes. The blend tells you the setting the app will use. Whether that setting suits your plants is a separate question, which you answer by watching the crop. Check the profile before you trust it Last, three checks before you trust a slider change. First, a profile can be shared. Several zones or weeks can use the same profile. Edit it once, and every one of them changes. So before you edit, look at which zones use it. Second, check each EC value in the profile. Lesson A five showed that feed EC and pore-water EC are different measurements, so confirm which one each field means. Third, moving the slider makes a draft, like the draft threshold in lesson B one. The plan on screen updates at once. Nothing changes in the grow room until the plan is saved and switched on, steps a later lesson covers. Now explain the dimmer aloud, then try the practice. The next lesson works out how much water each shot delivers. INDEPENDENT PRACTICE On paper, blend fictional endpoints of 40 and 80 minutes at slider 25. Repeat with both endpoints at 40. Then explain why pot size alone cannot choose these endpoints. Compare your reasoning: 40 + 0.25 × 40 = 50 minutes; then 40 minutes. You choose endpoints from your crop, substrate and readings; pot size alone cannot set them. SOURCES Crop Steering: GROW_PLANS.md: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/GROW_PLANS.md Grodan: Introduction to crop steering; substrate-specific manufacturer guidance: https://www.grodan.com/siteassets/downloads/downloads-na-101/grow-guide-2023/introduction-to-crop-steering.pdf Athena Commercial Handbook: source guide and background (not a software specification): https://growershub.athenaag.com/athena-handbook-v41iwrqr Athena: Precision Irrigation Strategy: https://support.athenaag.com/hc/en-us/articles/25975395644315-Precision-Irrigation-Strategy B4 — How much water did each plant get? Treat drippers as small taps, calculate water per plant and per zone from flow and runtime, then separate delivery from runoff and uptake. A dripper is a small tap Picture a garden tap left running. The water that comes out in a set time is its flow rate. In a grow room, each plant is fed by drippers. A dripper is a small nozzle on a thin pipe that lets water out slowly at the plant. Its flow rate is given in litres per hour, written L slash h. Four litres per hour means four litres would come out in one full hour. Look at the picture. Each example plant has two drippers, each rated at four litres per hour. The zone has twelve plants. So each plant gets two times four, which is eight litres per hour. The printed rating is a starting figure. A catch test measures the real flow: you collect one dripper's water in a cup for a timed period and measure it. Next, flow becomes a shot volume. Worked example: thirty seconds A shot runs for a set time, called its runtime. Our example shot has a runtime of thirty seconds. The flow rate is per hour, so first count the seconds in an hour. Sixty seconds times sixty minutes is three thousand six hundred seconds. Each plant gets eight litres per hour. Multiply eight by thirty to get two hundred and forty. Divide by three thousand six hundred. The answer is about zero point zero six six seven litres. Multiply by one thousand to get millilitres: about sixty-six point seven millilitres per plant. That is less than a small glass of water. Keep the unrounded figure for the next step, so later answers stay accurate. Scale from one plant to twelve So far you have worked out one plant. The zone has twelve plants, all fed the same way. Multiply the per-plant amount by twelve. Use the exact fraction, one fifteenth of a litre, rather than the rounded figure. One fifteenth times twelve is zero point eight litres for the whole zone. Per plant tells you what one pot received. Per zone tells you what the whole group used, which matters for the tank. Now double the runtime to sixty seconds. At the same flow, the water doubles too: zero point eight times two is one point six litres per zone. To total a whole day, add up only the shots in the event record, the shots that really ran. A shot's share of the pot How big is a shot compared with the pot? Take a separate example: a forty-five-second shot on the same two drippers. Eight times forty-five is three hundred and sixty. Divided by three thousand six hundred, that is zero point one litres, or one hundred millilitres per plant. Each pot holds five litres of substrate. Zero point one divided by five is zero point zero two. Times one hundred is two percent of the substrate volume. You might expect the VWC to rise two points. Often it rises less, or unevenly, because some water drains out and the probe samples only one spot. Use two percent as a size guide, and read the probe to see the real response. Your turn: complete the calculation Your turn. Each plant has one dripper rated at three litres per hour, and the shot runs for sixty seconds. The zone has twelve plants. Pause and work it out before you listen on. One dripper at three litres per hour gives three litres per hour per plant. Three times sixty is one hundred and eighty. Divide by three thousand six hundred: zero point zero five litres, or fifty millilitres per plant. Times twelve plants is zero point six litres per zone. Last, note where the three litres per hour came from. Was it the printed rating, or a catch test? Write that beside the answer, so the next reader knows how firm the number is. Use the time that actually ran Sometimes a shot does not run as long as asked. The Controller has a safety limit on how long one shot may last. The time asked for is the requested runtime. The time that actually ran is the effective runtime. For example, a shot requests one hundred and twenty seconds, but the limit is sixty seconds. The effective runtime is sixty. Each plant gets eight litres per hour. Eight times sixty, divided by three thousand six hundred, is about zero point one three three litres. That is about one hundred and thirty-three millilitres. Using the requested time would give about two hundred and sixty-seven millilitres, double the real figure. So always calculate with the effective runtime. The app shows both times side by side. Recall: delivery, runoff, uptake Last, three words that sound alike but mean different things. Delivery is the water the drippers supplied. Runoff, from lesson A five, is water that drains out of the pot. Uptake is water that enters the plant. A flow-times-runtime calculation gives delivery only. Try one more. Two drippers at three litres per hour run for ninety seconds, with fifteen plants. Two times three is six litres per hour per plant. Six times ninety is five hundred and forty. Divide by three thousand six hundred: zero point one five litres per plant. Times fifteen plants is two point two five litres per zone. That is water delivered. To learn how much drained, measure the runoff. Neither number tells you the plant's uptake. The next lesson looks at the equipment that makes each shot happen. INDEPENDENT PRACTICE A fictional zone has 18 plants and two 2 L/h drippers per plant. The effective runtime is 75 seconds. Calculate litres and millilitres per plant, then litres per zone. Explain why the result is not uptake. Compare your reasoning: About 0.0833 L = 83.3 mL per plant; 1.5 L per zone. This is delivery at the stated flow. It does not measure runoff or uptake. SOURCES Crop Steering: GROW_PLANS.md: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/GROW_PLANS.md Crop Steering: training/crop-steering-from-zero/WORKBOOK.md: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/WORKBOOK.md Athena Commercial Handbook: source guide and background (not a software specification): https://growershub.athenaag.com/athena-handbook-v41iwrqr Athena: Precision Irrigation Strategy: https://support.athenaag.com/hc/en-us/articles/25975395644315-Precision-Irrigation-Strategy C1 — What you need: the smallest useful system Plan one zone by five jobs: plan, measure, supply, switch and protect. Meet Home Assistant, entities and the two Crop Steering parts. Five jobs, one small zone In lesson A one you met the zone. It is a group of plants that get their water together, through one pipe and one on and off control. This lesson plans the equipment for one zone. Look at the single zone on this card. You build that one zone first and prove it waters correctly. Only then do you add a second. A mistake in a small system costs one zone, not a whole room. Now picture watering by hand. You decide when and how much. You push a finger into the pot to feel how wet it is. You fill a watering can. You tip it, then stop. And you keep a towel handy for spills. An automatic system needs the same five jobs. This course calls them plan, measure, supply, switch and protect. The five labels on this card name them, and the rest of this lesson takes them one at a time. Host, software and probes Start with the plan job. It needs a computer that stays switched on day and night. On that computer runs Home Assistant: free home-automation software that collects readings from sensors and switches equipment on and off. The computer that runs it is called the host. This card shows the two kinds of host the course supports. One is a Raspberry Pi, a small, low-cost computer about the size of a credit card. The other is a small PC with an Intel or AMD processor, the chip design most desktop computers use, known as x eighty-six sixty-four. You need only one. Like a laptop, the host needs three basics, listed on this card. It needs the correct power supply, because a weak charger can make it restart at random. It needs suitable storage: the memory card or disk that holds its software and saved data. And it needs a reliable network connection to your router, the box that links your devices to each other and to the internet. The host lessons name the exact parts. Setup has two alternative host lessons. Lesson C two A installs Home Assistant on a Raspberry Pi. Lesson C two B installs it on an x eighty-six sixty-four PC. As the card shows, both arrive at the same place: Home Assistant running and ready for the next step. Follow the lesson that matches your computer and skip the other. If someone else built the system and you only run it, you skip setup and go straight to the daily-use lessons. Inside Home Assistant, every reading and every control gets its own name tag, like the labels on a switchboard. One tag might be a moisture reading. Another might be the on and off control for the water. Home Assistant calls each named reading or control an entity. Crop Steering joins Home Assistant through an integration: a package that teaches Home Assistant a new job. The integration stores your room settings and shows the Crop Steering screens. The Controller, the watering software from lesson B one, is a separate program. It reads those settings and entities and decides when to water. Follow the path on this card from entity to integration to Controller. You need both Crop Steering parts. Now the measure job. Lesson A one introduced the probe, a device that measures one small spot. The moisture probe is your system's wetness check, so choose it with care. Pick a model made for your substrate. Place it in a pot that behaves like the rest of the zone. Before buying, check the maker's specification for the three items on this card. They are the exact model, its connection type, meaning how it sends readings, and the units it reports. Some probes send their readings in a wiring language that Home Assistant cannot read on its own. They need a bridge: a small box that translates the probe's signal into one Home Assistant understands. Follow the card from the probe, through the bridge, to a Home Assistant entity. Confirm compatibility from the maker's list of supported probes for that exact bridge, rather than from matching wire colours. EC readings need a location label too. Lesson A five named several kinds of EC, and this card compares three of them. Feed EC measures the nutrient solution in the tank. Bulk EC measures the wet substrate as a whole, water and solid together. Pore-water EC describes the water held in the pore spaces between the pieces of substrate, and it is usually estimated from bulk EC. When you choose an EC probe, read which of these it actually reports. A reading belongs to the place where its sensor sits. The card shows a tank sensor and a root-zone sensor side by side. Suppose you rename the tank sensor's reading root-zone EC. It still measures the tank, because the name changed and the sensor did not move. Keep every label true to where its sensor really is, so later decisions use the right number. Air readings complete the measure job. Two readings help explain why plants drink more on some days. One is air temperature. The other is relative humidity: how much moisture the air holds, compared with the most it could hold. Think of washing on a line. It dries fast on a hot, dry day and slowly on a cool, damp one. The card shows both readings feeding into water demand. Next comes the supply job: how water reaches the pots. Draw the actual water path Trace the water from where it starts, the way you would follow a garden hose back to the tap. Each zone has a zone valve: a tap that opens and closes when it receives an electrical signal. Water needs pressure, meaning push, to flow out of the drippers. A pressurised supply, such as a building's water pipe, may have enough push to feed the zone valve directly. A tank of stored water usually does not. It needs a pump: a motor-driven device that pushes water along the pipe. The card shows both routes. Many diagrams show a pump, but you add one only when your supply needs the extra push. If your route does need a pump, it also needs protection made for that pump. For example, dry-run protection switches the pump off when the tank runs empty, because a pump running without water can overheat and fail. The card pairs the pump with its protection, and the pump maker's instructions list exactly what it requires. Now follow the rest of the path on the card. A filter catches grit that would block the drippers. A pressure regulator holds the push steady, so each dripper delivers its rated flow. A drainage path carries runoff away. Backflow protection stops water running backwards into the supply. Siphon protection stops water draining out of the tank by itself through the pipe after the pump stops. A qualified plumber confirms these parts suit your equipment and your local rules. Safe switching and proof Now the switch job. Control hardware is the device that turns a pump or valve on and off when the software asks. A relay is a common example: a switch operated by an electrical signal. A pump draws a burst of extra power for a moment as it starts, like the extra shove needed to get a loaded trolley rolling. That burst is called the starting load. As the card shows, choose control hardware rated for both the starting load and the running load. The big number printed on a household smart plug often describes simpler loads, such as heaters, so check its rating for motors. Mains electricity is the high-voltage supply behind your wall sockets and switchboard, and wiring it wrongly can kill. A licensed electrician does any mains wiring for your system. This course never asks you to open or wire mains equipment, as the card reminds you. Your part is to plan, label and check. Two people turning the same tap send mixed instructions: one opens it while the other closes it. Equipment works the same way. An output is anything the system switches, such as a pump or a valve. Give each output exactly one owner, meaning the one system allowed to switch it. The card shows one owner linked to one output. An old irrigation timer must not also switch that output. So before the new Controller switches anything, list every system that can already switch your pump and valves. Include old timers, other home-automation rules and manual switches. Then choose one owner for each output and remove the others, following the checklist on this card. The last job is protect, and the card shows its three parts. First, a physical shutdown: a real switch or valve, close at hand, that cuts power or water directly. Second, independent protection, such as dry-run protection, that works even if the software fails. Third, a safe place to catch test water. The Crop Steering screens also have a Watering switch, which tells the Controller whether it may start shots. Keep Watering OFF until your system passes its checks, and use the physical shutdown in an emergency, because it stops the equipment itself. During testing, you measure what really arrives. Put collection cups under the drippers and time the run with a stopwatch, as in the catch test from lesson B four. Then send the stop command and watch the outlet. The water must physically stop. The card shows the cups, the stopwatch and the outlet. A dashboard is the screen of readings and buttons you use to watch the system. When a dashboard icon changes to off, the software is reporting what it believes happened. To know the water really stopped, look at the outlet itself, as the card shows. Essentials first, extras later Once one zone works, extra instruments add useful information. The card lists the first ones. A tank level sensor shows how much water is left. Feed EC and pH show what is in the solution, where pH describes how acidic or alkaline it is. Tank temperature shows how warm the solution is. Later you might add more probes and a verified flow meter, which measures the water passing through the pipe. Automated nutrient dosing, thermal cameras and artificial intelligence are optional extras. Your first zone works without them. Build the basics, prove them, then add one item at a time. Your own layout Now apply all of this to your own space. On paper, sketch your water route like a simple road map. Label the supply, the pump if you need one, the zone valve, the drippers and the drainage path, in the order the water travels. The card shows that order. Beside the water route, draw the measure and switch side. Add the probe, its bridge if it needs one, the host and the Controller. Draw a line from the Controller to each output it switches, and check that each output has only one line. Then mark the physical shutdown. The card shows both sides together. Now circle every item you cannot name exactly, for example a pump with no model number. As the card says, each circle is a question to answer before you buy or wire anything. Check exact models against the course's equipment and setup guide before you buy. Builders then pick a host lesson: C two A for a Raspberry Pi, or C two B for an x eighty-six sixty-four PC. Both host routes then join the same steps, shown on this card. Install the host and both Crop Steering parts. Connect one zone's probes and outputs in the software. Prove that the water physically starts and stops. Only then add the next zone. If you only run a system someone else built, go straight to lesson D one. There you find your room and zone in the app. INDEPENDENT PRACTICE On paper, sketch your planned water path from supply to drainage. Add the probe, host, Controller and physical shutdown. Draw one line from the Controller to each output. Circle anything you cannot name exactly. Do not switch any real equipment. Compare your reasoning: Every one of the five jobs appears, each output has exactly one owner line, the physical shutdown is marked, and every unknown is circled for an installer or supplier to answer. SOURCES Course evidence and production notes: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/PRODUCTION-AND-SOURCES.md METER: soil volumetric water content: https://metergroup.com/measurement-insights/soil-moisture-sensors-how-they-work-why-some-are-not-research-grade/ Athena Commercial Handbook: source guide and background (not a software specification): https://growershub.athenaag.com/athena-handbook-v41iwrqr Athena: Precision Irrigation Strategy: https://support.athenaag.com/hc/en-us/articles/25975395644315-Precision-Irrigation-Strategy C2a — Give Home Assistant a home: Raspberry Pi Install Home Assistant OS on a Raspberry Pi, protect the card's old files, and open Home Assistant in a browser. One computer keeps the system running Lesson C one named the host: the computer that runs Home Assistant, the software that collects readings and switches equipment. This lesson builds the host on a Raspberry Pi, the credit-card-sized computer. Lesson C two B builds it on a PC instead, and you need only one. Every computer needs an operating system: the base software that starts it and runs its programs, as Windows does on a laptop. You will install Home Assistant OS, an operating system made only to run Home Assistant. Look at the picture on screen. Readings come in, messages go out, and one always-on host sits in the middle. If you already have a working Home Assistant host, check it and skip ahead to lesson C three. Keep pumps and valves disconnected while you set up. Lay out the parts first Before you start, lay the parts out on a table, as the screen shows. You need a Raspberry Pi four or Pi five with at least two gigabytes of memory, the working space it uses while running. You need a storage card: a micro S D card, the small memory card used in cameras and phones, of at least thirty-two gigabytes. The official guide suggests one marked A two, a speed rating printed on the card. You need a card reader for your everyday computer. You need the power supply made for that Pi model, because phone chargers and computer USB ports often cannot supply enough power. Last, you need an Ethernet cable: the network cable with a clip-on plug that joins the Pi to your router. The official guide requires it for installation. Check the current official guide before you buy, because its list can change. Example: choose the spare card Writing the software is different from copying a file. Home Assistant OS comes as an image: one file that is an exact copy of a ready-to-run storage card. Writing the image copies it onto your card and replaces everything already on it. The tool for this is Raspberry Pi Imager, a free program from the makers of the Pi that runs on your everyday computer. Now imagine Imager lists two places to write: a spare sixty-four gigabyte card and a one terabyte photo drive. Look at the two choices on screen. You choose the sixty-four gigabyte card, because it is the spare you checked and backed up. You leave the photo drive alone. In Imager, choose OS, then Other specific-purpose OS, then Home assistants and home automation, then Home Assistant. Pick the build for your Pi four or Pi five, then choose the checked card as storage. Your turn: the last check before Write Pause before you press Write. Say this sentence out loud: I am replacing everything on this card, and any files I need are saved somewhere else. Add the Pi model and the matching image you chose. If any part is uncertain, sort it out first. Then select Next, then Write, and wait until Imager says the write has finished. Select Finish, then eject the card, meaning tell your computer it is safe to remove. The screen shows the next three steps in order. Push the card into the Pi while the Pi has no power. Plug the Ethernet cable from the Pi into the same router as your everyday computer. Connect the power supply last. Open Home Assistant in a browser Your everyday computer now needs to find the Pi. Every device on your network has a network address, like a house number on a street. Your web browser uses it to reach that device. Wait a few minutes after power-on. Then type homeassistant dot local into the browser, as the screen shows. If nothing opens, add colon eight one two three to the end. That number is a port: a numbered door on the host where Home Assistant answers. You can also use the Pi's numbered network address, listed in your router's settings, followed by colon eight one two three. When the page opens, check it belongs to this new Pi and not an older system. You will then see onboarding: the short set of welcome screens Home Assistant shows on its first start. Lesson C three walks through them. A working page shows the host is reachable; the probes and water get their own checks later. Keep a way back Now try a different case. A card holds old research files with no other copy. Before writing, you copy those files somewhere safe and identify the card again. For your real installation, keep a written record, as the screen shows. Note the Pi model, the card size, the Home Assistant version and the address that worked. If no page appears after five minutes, check the power supply and the network cable. Then follow the official troubleshooting steps, which may mean writing the card again, perhaps on a different card. With this record, another person could rebuild your host. Lesson C three picks up at onboarding, then adds the Crop Steering software, accounts and backups. INDEPENDENT PRACTICE Prepare a Pi setup sheet naming the exact Pi model, power supply, card and its size, where any old files were saved, and the address that opened Home Assistant. Compare your reasoning: The sheet names the Pi model, its power supply, the checked card, the backup location, the Home Assistant version and the working address. SOURCES Home Assistant: Raspberry Pi installation: https://www.home-assistant.io/installation/raspberrypi/ One-zone equipment and staged commissioning guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/EQUIPMENT-AND-SETUP.md C2b — Give Home Assistant a home: dedicated PC Install Home Assistant OS on a dedicated x86-64 PC, using a live Ubuntu USB stick, after checking the disk and protecting its files. Another home for Home Assistant Lesson C one named the host: the computer that runs Home Assistant, the software that collects readings and switches equipment. This lesson builds the host on an x eighty-six sixty-four PC. That is a computer with an Intel or AMD processor, the chip design most desktop computers use. Lesson C two A uses a Raspberry Pi instead, and you need only one. A popular choice is a NUC, a palm-sized box PC. Its small shape alone does not make it suitable, so the next slide checks the details. The PC must be dedicated, meaning it does this one job and nothing else. The card shows that one host is enough. If you already have a working Home Assistant host, check it and skip ahead to lesson C three. Keep pumps and valves disconnected while you set up. Check the computer before changing it Every computer needs an operating system: the base software that starts it and runs its programs, as Windows does on a laptop. You will install Home Assistant OS, an operating system made only to run Home Assistant. First check the exact machine against the official guide for x eighty-six sixty-four computers. Look at the three checks on this card. One, the processor must be sixty-four-bit, a design almost every PC from the last ten years has. Two, the PC must start with UEFI: the built-in startup program that runs the moment you press power. You open its settings by pressing a key during startup, and many people still call it the BIOS. There, turn off Secure Boot, a feature that only lets approved operating systems start. Three, the disk must use five-hundred-and-twelve-byte logical sectors, the size of the small blocks it stores data in. Home Assistant OS cannot start from drives that only use four-kilobyte blocks. Installing replaces everything on the disk, so back up its files and write down each setting you change. Example: the USB stick and the disk have different jobs Picture a workbench with a borrowed toolbox and a permanent cabinet. The toolbox is a USB stick of at least eight gigabytes holding live Ubuntu. Ubuntu is a free operating system. Live means it runs straight from the USB stick without being installed, and it leaves no trace when you remove the stick. The cabinet is the PC's internal disk, which will hold Home Assistant OS for good. The card shows both jobs. Start the PC from the USB stick and choose Try Ubuntu, not Install Ubuntu. Plug in an Ethernet cable, the network cable that joins the PC to your router, so Ubuntu can reach the internet. From the official Home Assistant installation page, download the Home Assistant OS disk image: one file that is an exact copy of a ready-to-run disk. The next slide restores it onto the internal disk. Your turn: identify both devices Now the step that erases the disk. In Ubuntu, open Show Applications, then the program called Disks. Select the internal disk, open the three-dot menu and choose Restore Disk Image. Pick the downloaded file, then Start Restoring. Before you press the final Restore, fill in the blanks on this card. The USB stick is this device. The destination is that disk. Its backup is stored there. Identify each one by the name and capacity shown on your screen, never by a name copied from someone else's guide. If Ubuntu reports the destination is busy, you probably started from the wrong device, so fix the startup choice. When the restore finishes, shut down and remove the USB stick when Ubuntu asks. Then set UEFI to start from the internal disk. A capable office PC may still be unavailable Now decide for a different machine. An office PC passes every hardware check from slide two. But it still holds current work files with no backup, and its owner has not agreed to give it up. The card shows the result: capable hardware and unprotected data, so you stop before erasing. First get the owner's agreement and a working backup, or choose another suitable host. While you wait, you can keep learning in the demo, the practice copy of the app from lesson A one, which runs no equipment. Check first start and write down the result Start the PC from its internal disk, with the Ethernet cable plugged in, and give it a few minutes. The first start needs internet, because it downloads the latest Home Assistant. Every device on your network has a network address, like a house number on a street. Your web browser uses it to reach that device. On your everyday computer, type homeassistant dot local into the browser. If nothing opens, add colon eight one two three. That number is a port: a numbered door on the host where Home Assistant answers. When the page opens, check it is this PC. You will see onboarding: the welcome screens Home Assistant shows on first start. As the card shows, record the PC model, disk, backup location, changed settings, version and address. Lesson C three continues from onboarding. INDEPENDENT PRACTICE Prepare a dedicated-PC installation record naming the live USB stick, the destination disk and its capacity, the backup location, every UEFI setting you changed and the address that opened Home Assistant. Compare your reasoning: The record names the USB stick, the checked destination disk, the backup location, changed UEFI settings, the Home Assistant version and the working address. SOURCES Home Assistant: Generic x86-64 installation: https://www.home-assistant.io/installation/generic-x86-64/ One-zone equipment and staged commissioning guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/EQUIPMENT-AND-SETUP.md C3 — Install the software and protect access Finish onboarding, install both Crop Steering parts at the same version, confirm the heartbeat and keep a backup you can restore. Finish onboarding, then read an entity ID Your host lesson ended at onboarding: the welcome screens Home Assistant shows on its first start. Complete them now. Create your own account, meaning a username and password that belong to you alone. Write down how you would get back in if you forgot the password, and keep that recovery note private and safe. Set your timezone, the local clock region, so lights-on times match the real clock. Now recall an entity: one named reading or control inside Home Assistant. Each entity has an exact written address, called an entity ID. The card shows one: sensor dot practice underscore moisture. One probe can supply several entity IDs, for example moisture, temperature and battery. The entity ID tells software where to look, and the physical label on the probe tells you where it really is. Two Crop Steering parts do different jobs Crop Steering arrives in two parts, as lesson C one showed. The integration is the package that adds Crop Steering to Home Assistant: it stores your room settings and shows the Crop Steering screens. The Controller is a separate program that reads those settings and decides when to water. Both parts carry one version number, and you install the same number of each. This is called version matching, and it lets each part understand the other. You get the integration from HACS, short for Home Assistant Community Store: a free extra for Home Assistant that downloads community-made integrations. HACS does not install the Controller, which comes from a different place on slide four. The card shows all three pieces. Throughout setup, keep the Watering switch OFF. Remember, that switch tells the Controller whether it may start shots. Example: install the integration and create a room Now install the integration, using the links in the Crop Steering installation guide. If HACS is not installed yet, install it first by following its own guide. Open the guide's HACS link for this project and select Download. Then restart Home Assistant. Restarting closes and reopens Home Assistant's software so it loads the new part, and it takes a minute or two. Next, open the guide's Crop Steering setup link. Choose manual setup, type a room name and enter how many zones the room has. For practice, the card uses a room called Practice Room with one zone. This creates the room's record in software only. The real probes and valves are connected in lesson C four. Add and start the separate Controller The Controller installs from a different place, the app store. On Home Assistant OS, an app is a separate program that runs beside Home Assistant on the same host. The app store is the screen where you install apps, and it only lists apps from sources it knows. An app repository is a web address that adds another source to that list. Use the guide's Add the app repository link. Then install Crop Steering Controller at the same version as the integration. Review its options and start it, with Watering still OFF. The card shows two settings that keep it running. Start on boot starts the Controller whenever the host starts, and it is on by default. The Supervisor is the part of Home Assistant OS that manages apps. Watchdog, which you turn on, lets the Supervisor restart the Controller if it stops. Your turn: the screens open but no heartbeat arrives Open Crop Steering in the Home Assistant sidebar, the menu down the left side of the screen. Next, check that the Controller is talking. A running Controller regularly sends Home Assistant a short status message saying it is alive. This message is called the heartbeat. Its age tells you how recently the Controller reported, as a reading's age did in lesson A six. The Controller sends it using an access key that the Supervisor hands over automatically, so you never type or paste one. Now the case on the card: the Crop Steering screens open, but no heartbeat arrives. The screens come from the integration, so check the Controller app. Is it installed, started and at the same version? If an older Controller already runs this room, finish its documented upgrade first, so that one Controller owns the room. Keep a recovery copy you can actually use Last, make a way back. A backup is a saved copy of your Home Assistant setup that can rebuild it on new hardware. Make one in Settings, then System, then Backups. Home Assistant normally encrypts backups: it locks them with an encryption key, a long code needed to open them. It offers that key as a download called the emergency kit. As the card shows, keep the emergency kit and a backup copy away from this host. A USB stick in a drawer works. A backup kept only on the host is lost if its storage fails. Finally, compare Home Assistant's clock with your phone. Then move on to lesson C four, with Watering still OFF. INDEPENDENT PRACTICE Prepare a software record with both installed version numbers, the room name, the heartbeat age, how you would recover your account, and where the backup copy and emergency kit are kept. Compare your reasoning: The record shows matching version numbers, the intended room, a fresh heartbeat, an account-recovery method and an off-host backup with its emergency kit. SOURCES Crop Steering user guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Crop Steering installation, upgrade and rollback: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/INSTALL.md One-zone equipment and staged commissioning guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/training/crop-steering-from-zero/EQUIPMENT-AND-SETUP.md C4 — Tell the software which zone is which Map one zone's probes and valve, enter per-plant sizing, and wait for the Controller's Mapping acknowledged receipt with Watering OFF. A contact name needs the right number A phone contact can have the right name and still dial the wrong number. Software works the same way. Mapping means telling Crop Steering which Home Assistant entities belong to each zone: this probe, this valve. First add each probe and valve control to Home Assistant through its own supported setup, because Crop Steering only maps entities that already exist. Check each entity ID, the exact written address from lesson C three, against the physical label on the equipment, as the card shows. Then open Crop Steering, select the room, and open the Equipment menu, then Setup. Keep the Watering switch OFF, so the Controller starts no shots, and keep the physical shutdowns your installer set up in place. Match the real plumbing and measurements Setup asks how water reaches the room, called the supply type. Choose Zone valves only when each zone valve is fed straight from a pressurised supply. Choose A pump, then zone valves when water flows only while a pump runs. Map a pump only if one really exists, and give every zone its own valve. Then look at the probes, as on this screen. Check two things for each moisture probe. First the units: VWC should read in percent. Second the last update: the time of its newest reading, which should be recent. For each EC probe, note whether it gives feed, bulk or pore-water EC. If an optional reading does not exist, leave it unmapped rather than borrowing another number. Example: twelve plants, not one giant pot Next comes sizing: the numbers that turn runtime into litres. Setup asks for four numbers per zone, and all except the plant count are per plant. Imagine twelve plants, each in three litres of substrate, each with two drippers measured at three litres per hour. Enter plant count twelve, substrate per plant three litres, drippers per plant two, and flow per dripper three litres per hour. The card shows the zone delivery. Twelve plants times two drippers is twenty-four drippers. Twenty-four times three litres per hour is seventy-two litres per hour. Keep the per-plant box at three litres, not thirty-six. Thirty-six litres is the zone's total substrate, and it would tell the software each plant has a giant pot. Check the clock and the receipt Two checks come before saving. First the clock. Write down the timezone, the room's lights-on and lights-off hours in Crop Steering, and the times on the real light timer. They must agree, because Crop Steering does not switch the lights; it only uses those hours to plan the day. Second, the receipt. Choose Review configuration, read the list of changes and problems, then choose Save configuration. Each save creates a new revision: a numbered edition of the setup. Saving is like posting a letter. The Controller then reads the new revision and sends back an acknowledgement, its receipt, which the card shows as Mapping acknowledged. Wait for that receipt for this revision. A fresh heartbeat shows the Controller is alive, and only the acknowledgement shows it accepted your setup. Your turn: three faults in one record Now review a record yourself. Eight plants, four litres per plant, one dripper each, measured at two point five litres per hour. Eight plants times one dripper is eight drippers. Eight times two point five litres per hour is twenty litres per hour. The arithmetic is right, but the card shows three faults. The moisture entity belongs to the next room. The lights-on and lights-off hours are blank. And the revision is saved but not acknowledged. Each fault must be fixed before Watering can be turned on. You can practise this kind of review in the demo, with no equipment connected. Read the accepted setup back After the acknowledgement arrives, read the setup back. Open Overview, the main status screen, then Selected room, its view of one room. Next open Equipment, then Probes. Compare each reading, its units and its age with your written setup sheet. Check the sizing, the measured flow, the light hours and the one owner of each output. A state that reads unknown means the software cannot tell, so treat it as a fault to fix, not as OFF. The card shows the accepted revision and the live values. Lesson C five checks the physical protections, and lesson C six proves the water starts and stops. Until those pass with the responsible people present, Watering stays OFF. INDEPENDENT PRACTICE Review the eight-plant example, calculate the zone delivery, and list how to fix the wrong-room probe, the blank lights hours and the missing acknowledgement. Compare your reasoning: You calculate 20 L/h, keep 4 L substrate per plant, and fix the wrong-room probe, the blank lights hours and the missing acknowledgement before Watering is turned on. SOURCES Crop Steering user guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Crop Steering installation, upgrade and rollback: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/INSTALL.md One-zone equipment and staged commissioning guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/training/crop-steering-from-zero/EQUIPMENT-AND-SETUP.md C5 — Make the physical equipment safe to test Learn what the Watering switch can and cannot stop, check how each probe and pump is connected, and give every output one controller before any water flows. Watering OFF is a software block, not isolation Think of a stop request to a bus driver. Only that driver hears it. Crop Steering’s Watering switch works like that request. Each room has a Watering switch at the top of its screen, and you have kept it OFF since lesson C three. When Watering is OFF, the Controller starts no shots in that room. That includes rescue shots and test shots. A test shot is a shot you ask the app to run on demand. A shot that is already running stops within a few seconds. That is a software block. It does not cut electricity, and it cannot stop an old timer or a hand switch. Physical isolation means cutting power or water with a real device. Examples are a switch that cuts the electricity supply, or a closed hand valve on the water line. Electrical isolation is the power part of that. Look at the cues on screen. Before anyone works on a pump, valve or wiring, agree the approved electrical and water shutdown, and who controls it. Next, you check how each probe connects. Use the probe’s documented connection A plug that fits is like an envelope that fits a letterbox. The letter inside can still be wrong. So you check four things before you connect a probe. Supply voltage is the electrical push the probe needs to run, measured in volts. Too much can destroy it. The signal protocol is the wiring language from lesson C one: how the probe sends its readings. The interface is whatever the probe plugs into. Often that is a bridge, the translator box. The interface must speak the probe’s language. Firmware is the small program built into the probe or bridge. Some probes need a particular firmware version. Look at the cues on screen. For example, before you connect a TEROS twelve, read its maker’s sheet. Confirm its voltage, protocol and the bridges it supports. Then follow its placement and calibration steps. Never guess a wire’s job from its colour, and never connect an unknown signal straight to the metal pins on a computer board. For a first install, choose a documented, assembled route. Next, check that a working reading measures the right thing. Example: a number can be fresh and irrelevant A reading can be recent and still answer the wrong question. Picture a thermometer hanging in the tank, not in the pot. Here is a made-up example. Remember: an entity is one named reading or control inside Home Assistant. This entity shows twenty-two. It updated two minutes ago, so it is fresh. Its name mentions Room A. But its unit is degrees Celsius, and the sensor sits in the tank. Moisture in Room A is VWC, a percentage of the substrate’s volume, measured in the pot. So this number cannot stand in for moisture. Look at the three labels on screen. Check four things together: quantity, units, real location and age. The same care applies to EC. Remember from lesson A five: bulk EC is measured through the substrate. Pore-water EC describes the water between the pieces. A bulk EC reading does not become pore-water EC because a field asks for it. A wrong reading makes the Controller water for the wrong reason. Next, look at the pump. Check what starts and stops the pump Remember from lesson C one: a pump draws a burst of extra power as it starts, like the shove that gets a loaded trolley rolling. That burst is the starting load. A rating is the most a device is designed to handle safely. It is printed on its label. A smart plug’s headline rating usually describes a heater, not a motor. A competent installer is a qualified person, such as a licensed electrician or plumber. That person checks the real pump, its control hardware, the ratings and the protection it needs. Mains is the high-voltage power behind your wall sockets, and this course gives no mains-wiring instructions. The installer also plans for faults. Dry-run means the pump runs with no water and overheats. Deadheading means the pump pushes against a closed valve, with nowhere for water to go. Backflow and siphon, from lesson C one, mean water running backwards or draining on its own. Look at the cues on screen. Before any test, agree what will be energised, meaning receive power, where water comes out, and how the test ends. Next, find any second controller. Your turn: two controllers know the same valve Two people steering one car is dangerous, even when one of them is resting. A valve works the same way. Here is the case. An old timer is still wired to a valve. You have also mapped that valve into Crop Steering, and Watering is OFF. Remember from lesson C four: mapping tells Crop Steering which entity is which valve. The old timer is a competing controller, a second system that can also open the valve. Look at the comparison on screen. The left side can still act. The right side is OFF, but that does not stop the left. So the valve is not isolated. Write down both routes to water. Then choose one owner for the output, and have the installer disable or remove the other route. Crop Steering may still do read-only observation. That means it reads values but sends no commands. Agree how the plants still get water while you sort this out. Two owners can fight, double-water or leave a valve open. Next, close the physical checks. Close the physical checks before the wet test Before you leave a house, you check the doors, the oven and the taps. This slide is that last walk round. A wet test is the first time water really flows through the new system. First, trace each probe to its Home Assistant reading. Then trace each control entity back to the real valve or pump. Where the design allows, the installer confirms each output with a low-risk indicator, such as a small lamp, in place of the real load. Never use a live crop as the first wiring test. Look at the three cues on screen. Then write the setup record: interface, ratings, protections, one owner and the physical shutdown. For example: Zone one valve, owned by Crop Steering, old timer removed, water shutdown at the labelled hand valve. An open question keeps this stage unfinished. Move to lesson C six only with an approved catch-test method and safe collection ready. Lesson C six runs that first careful water test. INDEPENDENT PRACTICE Explain the old-timer conflict in your own words. Then fill in a physical setup record for one zone: probe interface, output owner, ratings, protections and physical shutdown. Compare your reasoning: The record names one owner per output, the competent installer’s rating and protection checks, and an agreed physical shutdown. Any blank line keeps C6 on hold. SOURCES Crop Steering user guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Crop Steering installation, upgrade and rollback: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/INSTALL.md One-zone equipment and staged commissioning guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/training/crop-steering-from-zero/EQUIPMENT-AND-SETUP.md C6 — First water: catch it, measure it, stop it Commission one zone in two stages: an installer-run equipment test with Watering OFF, then a supervised Controller cycle once every readiness check passes. Two tests answer two questions Before a new building opens, someone checks the pipes and the wiring. For your system, that check is called commissioning. It means proving each part works safely before you rely on it. It has two stages. Look at the cues on screen. Stage one is the equipment test. It asks: does water go where it should, and does it stop? Stage two is the supervised Controller cycle. It asks: does Crop Steering water correctly on its own while someone watches? In stage one, the Watering switch stays OFF. Remember from lesson C five: OFF blocks every Controller shot, test shots included. So the competent installer runs the valve or pump with the maker’s approved procedure. The test has a bounded duration, meaning a fixed length of time agreed in advance. For example, thirty seconds, with cups under the drippers and the physical shutdown within reach. Next, watch the whole water path. Watch the whole water path Picture filling a bath while you check every other tap in the house. You want one tap running and no others. With the installer present, remove only the isolation this one test needs. Leave everything else shut off. Look at the diagram on screen. Water flows from the supply, through the valve, to the drippers. Remember from lesson B four: a dripper is a small tap at each plant. Place a catch cup, a small measuring cup, under several drippers. Use one cup per dripper, not one cup for all. Time the run with a stopwatch. While it runs, watch three things. The intended drippers should flow. Other zones should stay dry. The pump and valve should switch in the agreed order. Then stop the test the agreed way, and keep watching the drippers until the water stops. Also check for siphon, from lesson C one: water that keeps draining after the stop. A screen icon reports what the software sent, so your eyes on the dripper are the proof. Next, turn a cup into a number. Example: a cup becomes a flow measurement A cup tells you millilitres in a number of seconds. Your setup from lesson C four asks for litres per hour, written L slash h. Here is how to convert, step by step. Suppose one dripper fills its cup with thirty millilitres in thirty seconds. Step one: change millilitres to litres. Thirty divided by one thousand is zero point zero three litres. Step two: scale up to an hour. An hour has three thousand six hundred seconds. Zero point zero three times three thousand six hundred is one hundred and eight. Step three: divide by the thirty seconds the test ran. One hundred and eight divided by thirty is three point six. So this dripper delivers three point six litres per hour. Look at the three steps on screen. Now say three cups held thirty, thirty-three and thirty-seven millilitres. They work out to three point six, three point nine six and four point four four litres per hour. One total would hide that difference. These are example numbers, so never enter them for a real zone. Next, you try one yourself. Your turn: finish the catch record Now you try it. A dripper fills twenty-four millilitres in thirty seconds. Pause and work it out. Twenty-four divided by one thousand is zero point zero two four litres. Times three thousand six hundred is eighty-six point four. Divided by thirty is two point eight eight litres per hour. Write that on the record with the dripper, the runtime and the conditions. Look at the cues on screen. Suppose its neighbours gave about three point six. That gap is uneven delivery, meaning some plants get less water than others. Look for a clog, a leak or low pressure. Fix it, then repeat the catch. Then the installer tests the independent protections, such as dry-run protection, with the maker’s method. Remember from lesson C five: deadheading means a pump pushing against a closed valve. Never fake a fault by pulling wires or power while the pump runs, or by shutting valves against it. That can wreck the pump or flood the room. Next, decide whether a good catch is enough. A good catch is one part of readiness A pilot runs a whole checklist, not only the fuel check. Your system needs the same. The readiness checklist is the list of checks that must all pass before stage two. Look at the cues on screen. The items are: correct room and outputs, fresh probe readings and acknowledged mapping. Then measured flow, water seen stopping and tested protections. Then lights-on and lights-off hours, timezone, approved local targets and recovery access. Remember from lesson C four: acknowledgement is the Controller’s receipt for a saved setup. Now decide a case. The catch and the stop both passed. You entered the measured flow and pressed Save configuration. But Mapping acknowledged has not appeared. Keep Watering OFF. Find out why the Controller has not accepted the revision, fix it and wait for the receipt. Turning Watering on will not repair a setup problem. Next, the first supervised cycle. Supervise the first enabled cycle A learner driver’s first trip has an instructor in the next seat. Stage two is that first trip. Once every readiness item passes, the responsible operator turns Watering ON for the selected room. Enabling Watering lets Crop Steering start any shot it decides is due, including rescue shots. So this is real watering, not a harmless check. Use approved local settings, not this course’s example numbers. Keep the catch cups and the shutdown ready. Look at the cues on screen. Watch the right drippers run, other zones stay dry and the water stop. Then check the event record, which lists shots that happened. Check the next action, the Controller’s note of what it plans next. If water keeps flowing when it should not, use the physical shutdown first. Then turn Watering OFF. Write down what happened and the agreed final state, for example Watering ON, with the operator checking each shot today. Repeat both stages for each new output. After the quiz, lesson C seven shows how to move zones over from an older controller, one at a time. INDEPENDENT PRACTICE Calculate the flow for 18 mL caught in 20 seconds. Then review a setup that is saved but not acknowledged, and write your stage-two decision and final-state record. Compare your reasoning: 0.018 × 3600 ÷ 20 = 3.24 L/h. Keep Watering OFF until Mapping acknowledged appears. Record the real delivery, the observed stop and the agreed final state. SOURCES Crop Steering user guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Crop Steering installation, upgrade and rollback: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/INSTALL.md One-zone equipment and staged commissioning guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/training/crop-steering-from-zero/EQUIPMENT-AND-SETUP.md C7 — Move from Growlink one zone at a time Record what your existing Growlink system does, then hand over one zone at a time, with one controller per output and a checked way back. Moving house starts with a labelled inventory Before you move house, you label every box while it is still in the old house. This lesson is about that kind of move. Growlink is a commercial irrigation controller sold by another company. Some growers already water with it. Migration means moving control of your watering from one system to another, here from Growlink to Crop Steering. If you have no Growlink system, follow along on paper. Look at the cues on screen. First, observe and change nothing. Pick one room and one zone. Write down its probes, valves, pump and the controller in charge today. Also write who runs it and who can recover it if it fails. For example: Room A, Zone one, one moisture probe, one valve, shared pump, owned by Growlink, run by the head grower. A working Growlink probe may still be unable to connect to Home Assistant. The current operator keeps caring for the crop while you plan. Next, keep the old records safe. Keep the records in their original context Photos of the old house help you set up the new one. Old records do the same job here. An export is a file a program saves so you can open the data somewhere else, such as a spreadsheet. Some products also offer an API, a doorway that lets other software fetch data automatically. Use only what your own Growlink account actually shows. If it offers neither, copy settings and history by hand or with screenshots. Keep the original files and their dates unchanged. Look at the cues on screen. Write down the timezone, lights-on and lights-off times, and substrate per plant. Add plant and dripper counts, measured flow, targets and how each value was measured. For example, a value labelled EC needs a note of where it was measured: tank, runoff or substrate. Crop Steering does not read Growlink files. You will enter checked values yourself on the Setup screen from lesson C four. Next, match old names to real equipment. Example: one old zone, several new addresses A contact called Mum on a new phone is only useful if the number is right. Names in two systems work the same way. Each Home Assistant entity has a friendly name, the readable label people see, such as Room A Zone one valve. It also has an entity ID, the exact address software uses, from lesson C three. Imagine Growlink calls a valve Room A, Zone one. Your Home Assistant now has an entity with that same friendly name. That match proves nothing about the real device. Someone may have typed the label onto the wrong valve. Look at the cues on screen. Trace the actual valve and probe with the installer. Confirm the interface, units and data age, as in lesson C five. Write the exact new entity ID beside the old record. Mark anything you cannot confirm as unknown. Then use lesson C four’s mapping and sizing checks to enter values. Old targets may mean something different in Crop Steering’s fields. Next, plan the handover. Agree the handover and the way back When a night nurse hands over to a day nurse, they agree the exact moment. Only one is in charge at a time. A handover is that moment for a zone, when control of its valve moves from Growlink to Crop Steering. Plan it with the responsible operator and the installer. Agree the time, which old controls get disabled, and how the plants get water during the work. Look at the diagram on screen. Each physical output has one controller at a time. Crop Steering may observe read-only, as in lesson C five, but it must never command a valve that Growlink still runs. Also write a rollback. A rollback is a checked set of steps to put the old arrangement back if a check fails. For example: make sure Crop Steering can no longer run Zone one, re-enable its Growlink schedule, then watch one shot start and stop. A rollback never means leaving both schedules running in case one fails. Next, try a review yourself. Your turn: the name matches, the method is unknown Now review a different record yourself. Pause after each fact and decide. Fact one: the old sensor is labelled EC, and nobody knows how it was measured. The Crop Steering field expects pore-water EC, the salts in the water between substrate pieces. Fact two: the valve’s friendly name matches, but no one has traced it to the real valve. Look at the cues on screen. Both gaps stay open. A familiar-looking screen cannot close them. So park both mappings. That means leave them out for now and list them as unresolved. Investigate the EC method, and trace the valve with the installer. Meanwhile the zone stays with its current, checked owner, here Growlink. Once both gaps are closed, finish the checks from lessons C three to C five. Then plan lesson C six’s two stages. A guess here could feed the Controller a wrong number or switch the wrong valve. Next, prove the zone. Prove one zone before expanding A builder proves one room is watertight before plastering the next. Do the same with zones. At the agreed handover, follow the approved shutdown and ownership steps. Then run lesson C six’s commissioning on this zone. Remember the two stages: an installer-run equipment test with Watering OFF, then a supervised Controller cycle. Look at the cues on screen. Check acknowledged mapping, fresh readings, measured flow and water stopping. Check approved targets and light timing too. Supervise the first enabled cycle, and record its event and the final state. If any check fails, keep the new output off and follow your rollback. Expand only when this zone’s record is complete. For example, Zone one passes on Monday. Zone two still needs its own catch, stop and supervised cycle, even if it looks identical. Keep the old records and the rollback until every zone is proven. After the quiz, lesson D one opens the Overview screen you will use every day. INDEPENDENT PRACTICE Build a one-zone Growlink handover sheet: original records, physical labels, checked new entity IDs, one owner per output, crop-care cover, commissioning gates and rollback steps. Compare your reasoning: The sheet keeps the original records, links each new entity ID to traced equipment, names one owner per output and lists checked rollback steps. Zone 1 is proven before any other zone moves. SOURCES Crop Steering user guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Crop Steering installation, upgrade and rollback: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/INSTALL.md One-zone equipment and staged commissioning guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/training/crop-steering-from-zero/EQUIPMENT-AND-SETUP.md D1 — Find your room, zone and next action Find a zone by its full address, check how old each report is, and read the Controller’s reason for what it does next. A zone needs its full address Think of a block of flats. Flat two in one building is not flat two in the building next door. Zones work the same way. Remember from A one: a zone is a group of plants watered together. A room is a space with its own lights, and it holds one or more zones. Open the isolated demo, the practice copy that runs no equipment, and select Overview, the screen for checking rooms and zones. Its first tab, All zones, shows a card for every zone, grouped by room. The second tab, Selected room, shows one room at a time. Across the top, each room has a small button with its name, called a room chip. Select a chip, then select a zone card. That opens the zone page: one page about one zone. Zone two in Room A and Zone two in Room B are different addresses. Next, you check how fresh the information on that page is. Check when each source last reported Yesterday’s weather report cannot tell you if it is raining now. Every number on screen has an age too. Remember from C one: Home Assistant collects readings and switches equipment. Each named reading or control inside it is an entity, and each entity keeps the time of its own last report. Look at the top bar, the strip across the top of every page. Beside the room chips, it shows how long ago the Controller last reported for this room. That regular message from the Controller, saying it is running, is called a heartbeat. A probe has its own last report time. The Equipment screen, on its Probes tab, shows it for every probe. Here is an example. The top bar says the Controller reported one minute ago. The moisture probe last reported at nine in the morning, and it is now one in the afternoon. From nine to one is four hours, so that moisture reading is four hours old. Check each source’s own age before you trust its number. Next, read the reason the Controller gives. Read the reason beside the state When a bus is late, the board at the stop usually says why. The zone page has the same kind of board. On the zone page, find three things. The phase, which you met in B two. The last shot. And the panel called Why slash next action. That panel gives the Controller’s reason for what it is doing now, and the condition it is waiting for next. Here is an example. The zone is in P two. The last shot was at nine twelve. The panel says it is waiting for moisture to fall to the re-water threshold of sixty percent. The probe reads sixty three percent. Sixty three minus sixty is three points still to dry, so no shot is due yet. That pause is the rule working as designed. The Controller’s records say what it decided; to see that water reached the plants, you look at the pots. Next, you put these pieces together in a worked example. Worked example: Training East, Zone 2 Now add the two controls in the top bar. In this fictional paper example, the Training East chip is pressed, so Training East is the selected room. Zone two is in P two. Its last shot was at nine twelve. Why slash next action says it is waiting for moisture to fall. Now look at the top bar. Watering is the room’s on and off permission for the Controller. When Watering is off, the Controller starts no shots in that room, rescue shots included. Stop all turns Watering off, then switches every pump and valve of that room off, and reads each one back. Both controls act on the selected room only, Training East. One more idea: a shared pump is one pump that feeds more than one room. If Training East shares its pump with Training West, Stop all also stops that pump, so Training West loses water too. The review box warns you about this. Read it before you confirm. Next, you meet a state the screen cannot confirm. Your turn: finish an honest record A fuel gauge showing dashes instead of a number does not mean the tank is empty. It means the gauge cannot tell. The app does the same. When it cannot confirm a state, it shows the word unknown. For example, the top bar reads Watering unknown when that switch is missing or not answering in Home Assistant. Now practise in the demo. Select Room B, then Zone one. Find the pump state and its report age. Write them down exactly as displayed. If a state says unknown, write unknown. Writing off instead would claim something nobody confirmed. Then check the source, meaning the real device behind the display. The link between a name on screen and a real device is called a mapping. It is set up under Equipment, Setup. You check a mapping with the person who installed it, by reading the setup, rather than pressing buttons to see what moves. Next, find another zone on your own. Find another zone independently Now find a zone without the example, like finding a friend’s flat with only the address. Choose any other demo zone. Write six things: the room, the zone number, the phase, the last shot time, the report ages and the next action. Then choose where the app opens. Open Settings and help, then Settings, then Display, then Default view. The default view is the page this browser shows first when you open the app. For example, choose Selected room if you check one room every morning. This choice lives only in this browser. Another person’s phone keeps its own choice, and nothing about watering changes. A saved bookmark still opens the page it points to. Return to Overview and say the selected room’s name aloud before you read any card. That habit stops you reading one room’s numbers while thinking of another. The last slide checks what you remember. Recall the address and control scope Close your notes and answer three questions aloud, the way a pilot repeats instructions back. First: which room do Watering and Stop all act on? The selected room, the one whose chip is pressed. Second: what does unknown tell you? That the app cannot confirm the state, so you record unknown and check the source. Third: after Stop all, how do you know the water really stopped? You look at the outlets and drippers. Off on the screen is a report from software. Remember from C one: the physical shutdown is the hand-operated way to cut water or power at the equipment itself. If water keeps flowing when it should not, use the physical shutdown first, then investigate. Tomorrow, repeat the navigation with a different demo address. Work on real equipment comes after your site’s responsible person has shown you around it. Next lesson, you change one target safely. INDEPENDENT PRACTICE Identify another demo zone and explain its next action without changing irrigation. Compare your reasoning: Record room, zone, phase, last shot, report ages and next action; keep any unknown as unknown. SOURCES User guide: navigation, top bar and zone page: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Current source: selected-room safety controls: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/frontend/src/components/room-controls.tsx Current source: browser default view: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/frontend/src/components/landing-preference.tsx D2 — Change one target and see the curve move Change one target the safe way: read who owns it, draft, review, apply, read back, then restore. Begin by reading who owns the targets Before you edit a shared document, you check whether someone else is editing it. Targets work the same way. Remember from B two: a target is a comparison value the Controller’s rules use. Open Irrigation, then its first tab, Current day. Choose a zone from the Zone list. Under the tabs, read the line that starts Target source. It names who owns today’s targets. Manual room targets means you set them here. Automatic setpoints means the app can adjust some targets by itself. Setpoint is another word for target. Armed schedule means a grow plan owns them. A grow plan is a dated timetable of targets, which you build in the next lesson. When a plan owns the targets, this page shows them read only. You change them in Calendar, the planning screen. For practice, pick a demo room whose Target source is not armed schedule. Two owners writing the same target would undo each other’s work. Next, look at the chart beside the targets. Tell the three lines apart A recipe card, your pencil notes on it, and the meal you actually cooked are three different things. The Current day chart shows three matching lines. Remember from B one: a dashed line is a planning reference and a solid line is measured. The saved reference shows the target values stored now. The draft shows a change you have proposed but not sent. As in B one, a draft is an unsaved proposal. Here it lives only in this browser until you apply it. The dark solid lines show what the zone’s probes recorded. Use the legend to match each name to its line. Here is an example. The saved re-water threshold sits at sixty percent. You drag the draft to sixty one. The draft line moves up one point. The saved line stays at sixty. The recorded moisture line does not move at all, because it is history. Mixing these up makes people think an edit changed the plants. Next, you learn what the setting you are about to edit actually does. Learn what the setting means A smoke alarm is not the temperature you set the heater to. It acts only when something goes wrong. Select the question mark beside a setting. This is the field help: a short note on what the setting does, when it acts and what it changes. For this exercise, find the P three setting labelled Rescue floor. You met it in B two, where growers also call it the emergency floor. The help calls it the overnight safety level of water content. In P three, the Controller gives a rescue shot each time moisture is below it. So it is a night alarm, not a level to aim for every night. Next, read its unit and its step. The unit is percent VWC. The step is the smallest change the setting accepts. We call it the allowed step. If the allowed step is one point, thirty can become thirty one, but not thirty point four. Write down the starting value before you touch anything. You need it to put things back. Next, you make the change. Worked example: 30 becomes a 31 draft Think of checking your online shopping basket before you pay. In this fictional paper example, the saved Rescue floor is thirty percent and the allowed step is one point. Thirty plus one step is thirty one. Type thirty one, or drag the line. The draft line moves to thirty one while the saved line stays at thirty. Now select Review changes. A review box opens and names the room the change affects. It lists the setting with its old value and new value: thirty, then an arrow, then thirty one. Read every row against what you meant to do. The button at the bottom says Apply one change. Apply is the only step that sends the value to Home Assistant, where the Controller reads its targets. In the isolated demo, Apply changes sample data only. If a row is wrong, select Back to editing instead. Next, you check that the change was accepted. Check the receipt, then restore After you pay, you check the receipt. After Apply, you check the readback. Readback means the app reads the value back from Home Assistant and shows what was actually accepted. While it works, the button says Applying and verifying. If Home Assistant refused a value, the box says Some changes were not applied, and gives the reason. That value stays in your draft for review. If the result is unclear, for example because the connection dropped, find out what was accepted before you try again. When the readback shows thirty one, put things back. Restoring means returning to the starting value you wrote down. Thirty one minus one step is thirty. Use the same path: draft thirty, Review changes, Apply, then read back thirty. Practising the way back matters as much as the way forward. Next, try it in a different zone. Try another demo zone A recipe that worked in one kitchen still needs checking in another. Choose a different zone, ideally in another demo room. Use its own starting value and its own allowed step, rather than copying thirty and thirty one. Before you edit, predict which line will move. Then make one step, review, apply, read back, restore, and read back again. One more idea matters here. The Zone list also has an option called Room lights and shared targets. A value set there applies to every zone in the room. A value set on one zone overrides the room value, for that zone only. Example: the room re-water threshold is sixty percent, and Zone three has its own value of fifty eight. Zone three uses fifty eight. Every other zone uses sixty. Check which scope you are editing before you apply. Next, recall the whole journey. Recall the complete change journey Teach the shopping journey to someone else without looking at the screen. The draft proposes a change in this browser. Review changes shows the room, the setting, the old value and the new value. Apply sends it to Home Assistant. Readback shows what was accepted. Restoring uses the same four steps to go back. Then name the lines. The saved reference shows the stored target. The draft shows your proposal. The dark solid lines show what the probes recorded. Finally, say who owns the targets. If Target source says armed schedule, a grow plan owns them, and you change them in Calendar. Check that the demo value is back where it started, and note both readback values in your practice record. A clean record shows the next person that nothing was left half changed. Next lesson, you build that grow plan in Calendar. INDEPENDENT PRACTICE Change the demo Rescue floor by one allowed step in a second zone, then restore it. Compare your reasoning: Both changes have correct scope, reviewed values and readback; distinguish references from actual history. SOURCES User guide: Current day and action effects: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Planning guide: reference lines and manual drafts: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/GROW_PLANS.md Current source: shared target ownership: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/frontend/src/pages/irrigation.tsx Current source: editing and review: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/frontend/src/pages/strategy.tsx D3 — Plan days and weeks without losing control Build a dated timetable of targets in Calendar, then keep Review & save, Arm plan, activation and the Watering switch apart. A timetable points to a settings card A school timetable says which lesson happens on which day. The lesson notes live in a separate folder. A grow plan works the same way. A grow plan is a dated timetable of targets for each zone. In Irrigation, open the second tab, Calendar. Calendar shows that timetable. Each day in it points to a profile. A profile is a named bundle of settings, like one folder of lesson notes. Before you change anything, read two things. First, the plan’s status. Draft means you can still edit it. Armed or active means the Controller is about to use it, or is using it, so editing is locked. Second, Target source, which you read in D two. In the demo, start with a plan whose status is draft. Editing a plan the Controller is following would change a live crop’s targets. Next, look inside a profile. A profile holds two endpoints Remember the dimmer from B three. An endpoint is the set of values at one end of the slider: vegetative at zero, generative at one hundred. A profile stores both endpoints. Under Calendar, open the section called Calendar and profiles. It has three tabs: Steering balance, Endpoint profiles and Recipes. Open Endpoint profiles. Look at the two columns, one for each endpoint. One profile can serve many days and many zones. A run of days in one zone that uses one profile is called a schedule block. For example, a profile named Bulk might serve days one to fourteen in Zone one, and days one to twenty one in Zone two. That is two blocks sharing one profile. Change the profile, and both blocks change. If one zone needs different values, duplicate the profile first, then edit the copy. A shared edit can change a zone you never meant to touch. Next, work out a blend. Worked example: a 2%–4% blend Halfway between two numbers sits in the middle. Shot size is how much water each P two shot gives, as a share of the substrate volume, as in B four. In this fictional profile, the P two shot size is two percent at the vegetative endpoint. At the generative endpoint it is four percent. The difference is four minus two, which is two. Half of two is one. Add one to two, and the halfway blend is three percent, before rounding to the setting’s step. The probe may rise by a different amount. Now give it dates. Open Steering balance and set the Zone start date. That date is day one for this zone. Each zone can have its own start date. If Zone one starts on the first of March, day seven is the seventh of March. Assign the profile to days one to seven, then read the blocks either side. Every day needs exactly one block, with no gaps and no overlaps. A day with no block has no targets, and the plan holds that zone. Next, store the plan and switch it on. Store first, then request activation Booking a flight, checking in and boarding are separate steps. A grow plan has the same kind of steps. First, Review and save. The app checks the draft for problems, then stores it in Home Assistant. A saved plan is still a draft. Second, Arm plan. Arming is a request: start using this plan at the next local lights-on. The Arm plan button stays unavailable until the Controller reports that it supports plans. Third, activation. Activation is the moment the Controller takes over the plan’s dated targets. Example: you arm at three in the afternoon, and lights-on is at six in the morning. From three in the afternoon to six the next morning is fifteen hours, so activation comes fifteen hours later. Check the status the Controller reports, not only your click. Fourth, and separate from all of these, is the Watering switch in the top bar. Arming never turns Watering on and never starts a pump. Next, learn how to hand control back. Understand the return handover A shift handover happens at a set time, not halfway through a task. Disarm plan asks the Controller to go back to manual targets at the next lights-on. Until then, the plan’s targets stay in force. Wait for the status to read draft before you edit. From then on, the plan is yours to change again. A plan hold is a different thing. A hold happens when a zone’s day in the plan is missing, finished or invalid. The Controller then stops steering that zone by the plan, rather than invent targets. Safety watering still runs during a hold, such as the overnight rescue shot below the Rescue floor. So a hold pauses the plan, and the zone can still get water. If water flows when it should not, use the physical shutdown at the equipment first. Disarm plan and a hold are software states, and they come after that. Next, practise a second block on your own. Your turn: another dated block A second try without the worked example shows what you really learned. Choose another demo zone and plan days eight to fourteen. Say its Zone start date aloud, then work out the calendar date of day eight. Find every block that uses your chosen profile. Check the days either side for gaps or overlaps. Use Review and save on the demo draft, and leave Arm plan alone. Then open the Recipes tab. A recipe is a saved copy of a plan, kept in this browser for this site, this room and this mode, demo or live. The button Load into local draft copies a recipe into your draft. It keeps each zone’s current start dates. Saving a recipe changes nothing in Home Assistant. Loading one changes only your draft. A recipe is a notebook, not the plan the Controller follows. Next, recall the whole sequence. Recall what each action does Close the screen and teach the timetable idea to someone else. What does a profile contain? Two endpoints, one at each end of the slider. When is day one for a zone? Its own Zone start date. Now separate the four actions. Review and save stores a checked draft. Arm plan asks for activation at the next local lights-on. Activation is when the Controller takes over the dated targets. Watering is the room’s own switch, and arming never touches it. Then explain the way back. Disarm plan returns manual targets at the next lights-on, and you edit once the status reads draft. Finish with the hold: plan steering pauses, and safety watering continues. Tomorrow, repeat the exercise with a different start date. Your practice record should show the block, its preview and every block that shares its profile. Next lesson, you compare what happened with what was planned. INDEPENDENT PRACTICE Create a demo block for another zone and explain save-to-activation timing. Compare your reasoning: The block is valid; your explanation covers activation at the next local lights-on, who owns the targets, and what Disarm plan and a hold do. SOURCES Planning guide: endpoints, holds and scheduled handover: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/GROW_PLANS.md User guide: Calendar and Recipes: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Current source: save, arm and disarm review: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/frontend/src/pages/grow-planner.tsx D4 — Compare reality with the plan Read History like a bank statement and targets like a budget: compare matched periods, keep estimate labels and date every reference. A budget and a statement answer different questions A budget says what you planned to spend. A bank statement shows what was actually spent. The app has both kinds of information. Targets are the budget. History is the statement. Open History in the isolated demo. It has four main tabs. Timeline is one list of the Controller’s records, such as shots, phase changes and target changes, newest first. Water use shows the litres each zone was given today, this grow week and since the grow started. A run is one grow in one room, from start to finish. Compare runs puts the run in progress beside an earlier run, week by week. Grows holds the run records: the name, start date and end date of each run. Before you open any tab, write down your question, the room, the zone and the time period. For example: did Zone two in Room A get more water on day ten than last run? A clear question tells you which tab to open. Next, make sure two runs are fair to compare. Match the comparison before calculating Comparing one child’s height at age five with another’s at age eight tells you little. You compare at the same age. Grow age is the number of days since a run started. Compare day ten with day ten. Next, match the zone. Check it had the same plant count and the same setup in both runs. Then match the units and the type of EC. Pore-water EC and feed EC are different measurements, as you learned in A five. Last, check coverage. Coverage means how much of the period actually has recorded data. In Compare runs, the section called Data quality and history coverage shows each day’s recorded range as a bar. Example: one run has twenty four hours of records for day ten, and the other has six hours. That is not a fair match. A day with no records stays empty, and empty is different from zero. Next, try a fair comparison. Worked example: day ten, 12 L versus 15 L Now a fair comparison. Two fictional runs have complete day ten records for the same zone, with the same plant count. Run A shows twelve litres of estimated delivery. Run B shows fifteen. Subtract twelve from fifteen. That is three. So Run B received three litres more estimated delivery on day ten. Remember from B four: delivery is water supplied through the drippers, runoff is water that drains out, and uptake is water that enters the plant. Water use counts delivery, worked out from each shot’s runtime and the configured dripper flow. Keep the word estimated in your sentence. Three litres more delivery does not show the plants took up three litres more, or that a target change caused it. Climate, plant size and other changes can all play a part. A careful sentence is: Run B shows three litres more estimated delivery on matched day ten. Next, date the reference you compare against. Date the reference as carefully as the reading A budget is only useful if you know which month it was written for. A target reference is the same. Here, a reference is the set of targets you compare the readings against. In Compare runs, open the full-resolution chart and select a Target reference. There are three choices: Current configured daily plan, Saved run daily reference, and Current phase reference. A snapshot is a copy saved at one moment, with a time stamp. When you save a run record in Grows, the app saves a snapshot of the targets at that moment. Example: today is the twentieth of October, and you record last month’s run for the first time. Its snapshot holds today’s targets, not last month’s. Targets nobody saved cannot be rebuilt. Read the note beside the reference. It says when the reference was captured and where it came from. Comparing old readings with the wrong month’s targets leads to false conclusions. Next, label the delivery figure correctly. Your turn: name what the estimate means A taxi meter estimates the fare from time and distance. Delivery figures work the same way. The app multiplies each shot’s runtime by the configured dripper flow. That gives estimated delivery. Measured delivery would need a water meter or a catch test. Complete the sentence in your head: runtime times flow estimates delivery, not plant what? Pause. The answer is uptake. A per-plant figure adds one more step. It divides the zone total by the plant count from setup. Example: twelve litres across six plants is twelve divided by six, which is two litres per plant on average. Some plants may have had more and some less, if a dripper was blocked. When you copy a number into a report, keep its label beside it: measured, estimated or unknown. A label lost in a report becomes a false fact for the next reader. Next, try a messy comparison. Try an incomplete comparison Comparing one day’s shopping with a whole week’s would make the week look expensive. Here, one record covers day six only. The other covers days one to six. One run is also missing its pore-water EC readings. Are their totals comparable? Decide first. They are not, because the periods differ. Choose a matched period with good coverage, such as day six in both. Why might data be missing? Home Assistant stores past readings in a part called the Recorder. The Recorder keeps readings for a limited time, and that time is called retention. Older readings are removed. Water use can also read Home Assistant’s long-term statistics, which are kept with no time limit. A longer chart cannot bring back readings the Recorder removed. Write one thing you can support, and one question you cannot answer yet. Next, recall the whole method. Recall an observation before proposing a cause Explain the budget and statement idea without the chart. A recorded reading is the statement: what the probe actually measured. A saved reference is a budget written at a known moment. A current reference is today’s budget, which may differ from last month’s. Now repeat the day ten result in one sentence: Run B shows three litres more estimated delivery on matched day ten. Then list other things that could explain it: climate, plant count, flow settings and other changes that week. Finish your practice record with five labels. The scope, meaning the room and zone. The matched period. The units. The source of each number. And the coverage. Those labels let the next person check your comparison instead of trusting it. Next lesson, you use the same care when something goes wrong. INDEPENDENT PRACTICE Write one defensible observation and one unsupported conclusion for another demo comparison. Compare your reasoning: The observation names scope, interval, units and source; the limit avoids inventing past targets, uptake or causation. SOURCES User guide: History and Grows: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Planning guide: comparison, retention and delivery estimates: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/GROW_PLANS.md Current source: Compare runs: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/frontend/src/pages/comparison.tsx Current source: grow and run recording: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/frontend/src/pages/grows.tsx D5 — Troubleshoot before you change a recipe When something looks wrong, make it safe, keep the evidence, find the real cause and prove the repair, before touching any target. Repair the evidence before turning the dial If a fridge thermometer’s battery dies, turning the thermostat colder does not fix the thermometer. A fault is anything that stops the system working as it should. Examples are a dead probe, a valve that will not close, or a wrong reading. A fault needs a repair, not a new target. Use the same four steps every time, in this order. One, make safe. If water keeps flowing when it should not, use the physical shutdown first. Remember from C one: the physical shutdown is the hand-operated way to cut water or power at the equipment itself. Two, preserve. Write down what you saw, with times, before anything changes. Three, investigate. Find the real cause with competent help, meaning someone trained on your site’s equipment. Four, verify. Prove the fix worked, with fresh readings and a look at the equipment. Example: a reading drops to zero at two o’clock. Changing a target would leave that reading broken. Next, work through an old reading. Worked example: a four-hour-old reading A note on the fridge saying the milk is fine means little if it was written last week. In this fictional example, a zone is holding. Holding means the Controller is giving it no shots for now. The zone’s moisture reading was last reported at six ten. It is now ten ten. From six ten to ten ten is four hours, so the reading is four hours old. Lowering a threshold would leave the old reading in place. Instead, follow the four steps. Make safe: nothing is flowing, so the zone is already safe. Preserve: record the room, the zone, the time and the reason shown in Why slash next action. Investigate: open Equipment, Probes. Read the probe’s health, such as Silent or Not reporting. Then, with competent help, check its cable, its power and its mapping, the link between the name on screen and the real probe. Verify: wait for fresh readings and a normal status before you think about targets again. Next, two readings that disagree. Conflicting tank evidence needs a real check If a car’s fuel gauge shows half full but the low fuel light is on, one of them is wrong. A tank level sensor measures how full the feed tank is. Some tanks also have a low float: a floating switch that reports low when the water drops past it. Here, the tank display says eighty percent, but the low float says low. That is conflicting evidence: two sources that cannot both be right. Averaging them would invent a level, and choosing eighty because it reassures you would hide the fault. Keep the protections switched on. When a tank level is mapped, the Controller holds shots at eight percent or less, or when the low float reads low. That hold stops the pump running a dry tank. Look at the tank itself, safely. Then check each sensor’s mapping, calibration and last report time. Record which check settled the disagreement. Next, a saved setup the Controller has not accepted. A stored setup needs Controller acceptance Posting a letter, the letter arriving, and the reader acting on it are three different events. Setup changes have three similar stages. If you skipped the setup lessons, here they are in plain words. Equipment, Setup is where an administrator links each zone to its real valves and probes. That link is the mapping. Each saved version of the setup gets a number, called a revision. Stage one: Save configuration stores the new revision in Home Assistant. Stage two: acknowledgement. The Controller reports which revision it has accepted, like a signed receipt. Check that it matches, and that the heartbeat, the Controller’s regular report that it is running, is recent. Until the Controller accepts the revision, keep the room’s Watering switch off. Stage three: commissioning. Commissioning is a supervised test with real water. It proves each valve feeds the right drippers and then stops. Next, the most urgent case. Your turn: OFF displayed, water flowing A light switch in the off position with the bulb still glowing tells you the switch is not the whole story. The screen says off, but water is visibly flowing. Finish the first response aloud: physical shutdown. Close the water or cut the power at the equipment, using your site’s procedure. Then investigate with competent help. There are four common causes. A sticking valve is a valve that stays partly open after it is told to close. A siphon, which you met in C one, is water that keeps running downhill through the pipe on its own. A wrong mapping means the screen controls a different valve. An old timer or another system may still switch the same valve. Remember from C one: each output needs one owner. Stop all, Disarm plan and a plan hold are all software requests, so they come after the physical shutdown, not instead of it. Next, write a fault report. Try another fault report independently A good doctor’s note says what happened and when, before guessing why. In this case, a shot ran for Zone two, but Zone two’s moisture did not rise. Zone three, next door, rose instead. Write the observation first: room, zone, event time, what you expected and what actually happened. Example: Room A, Zone two, shot at ten fifteen, expected a rise of about two points, saw none. Zone three rose two points at ten seventeen. Then list possible causes without choosing one. The valves may be mapped the wrong way round. The probes may be in the wrong pots. Delivery may have failed. Check the zone names, the times and the commissioning records. Arrange checks with competent help, rather than switching valves one after another to see what happens. Leave out passwords and access tokens, the digital keys that let software sign in. A clear report lets someone else find the cause quickly. Next, recall the sequence. Recall the repair sequence Teach the four steps to someone else without notes: make safe, preserve, investigate, verify. Then explain how repairing a sensor differs from changing a target. A repair fixes the information. A target change tells the Controller to aim somewhere else, using whatever information it has. Verifying means fresh readings and a normal status. It also means a look at the equipment in its agreed final state, such as valves closed. After a valve or mapping repair, a supervised commissioning test may be needed again. Ask your site’s responsible person what you do not yet know about the physical shutdown. Where is it? What does it switch off? Who may use it? Finish your record with the fault, the repair evidence and the name of the person doing the next check. Next lesson, you build a short daily routine. INDEPENDENT PRACTICE For the four-hour-old moisture reading, the tank conflict and the continuing-flow case, state the first response and evidence needed. Then write a report for the neighbouring-zone example. Compare your reasoning: Each answer includes a safe first action, relevant source checks and verification without a protection bypass. SOURCES User guide: emergency warning, setup and troubleshooting: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Planning guide: hold and disarm limits: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/GROW_PLANS.md Workbook: three failure scenarios: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/WORKBOOK.md D6 — A five-minute daily check and a useful handover Run a short preflight, keep a five-part diary, make one reasoned decision, treat the AI assistant as advice to check, and leave a handover the next person can use. A short preflight, then a brief diary Before take-off, a pilot walks round the plane with a checklist. That short check is called a preflight. After landing, the pilot writes a log. Use the same rhythm each day. First, a preflight: a quick look at the screens and the room before you decide anything. Then a daily diary: a short written record of what you found and what you decided. The diary has five parts. One, identity and freshness: which room and zone, and how old each reading is. Two, state: the phase and the reason. Three, water: how much was delivered. Four, feed and context: the tank, the EC and the room conditions. Five, the decision. About five minutes is a good routine on a normal day. A fault takes as long as it needs. If a source has no current reading, write unknown, as you learned in D one. Next, fill in the first four parts. Fill the first four parts with evidence A weather report gives the place, the time, the conditions and the outlook, in that order. The first four diary parts work the same way. Identity and freshness: write the room, the zone, the time, the Controller’s report age from the top bar, and each probe’s report age. State: the phase, the reason from Why slash next action, the last shot, and any equipment state the screen confirms. Water: the delivery total, and if useful the per-plant average, each labelled measured, estimated or unknown, as in D four. Feed and context: the tank level, which type of EC you are reading, and anything unusual in the room, such as a hot afternoon. Example: Room A, Zone one, eight fifteen. Controller one minute, probe two minutes. P one. Eleven litres estimated. These four parts are the evidence. The fifth part, the decision, rests on them. Next, a decision to change nothing. Worked example: a reasoned no change A health check-up that finds nothing wrong is still a useful check-up. In this fictional example, Training North, Zone three, has recent reports, the expected P two phase and a clear waiting reason. Today’s estimated delivery is fourteen litres. The last five days ranged from thirteen to fifteen litres, so fourteen sits inside the usual range. The tank level and the low float agree. The decision is no change. Write that decision down, with the reason and when you will check again. A decision record is the part of the diary that says what you decided, why, and when you will look again. For example: No change. Delivery within the usual thirteen to fifteen litres. Next check at two in the afternoon. Changing a target every day only to have done something adds noise and hides real problems. Next, record a change when one is needed. Your turn: finish the change record A note on a recipe is only useful if it says what you changed and how it turned out. When a change is justified, the decision record needs more. Write the scope, meaning the room and zone. Write the before and after values. Write the reason, and what you expect to see next. Then make the change with the path from D two: draft, Review changes, Apply, then readback. Finish with a follow-up: a time when you will check the result. Example: Room A, Zone two. Rescue floor thirty to thirty one percent, as agreed with the head grower. Expect the Rescue floor line one point higher on the chart. Readback thirty one. Check tomorrow at seven. If you change two things on the same day, record both. A better day after two changes cannot tell you which one helped. Next, the AI assistant. Use the AI assistant as advice to check A sat nav can suggest a route, but you still watch the road. The app has an AI assistant. AI stands for artificial intelligence: software that writes answers in everyday language. Find it under Settings and help, AI assistant. Its decisions also appear as marks on the zone page chart. A filled mark is an action it took. An outline mark is advice, no action or a refusal. The AI assistant can organise a question or explain a proposal. It does not measure the crop. Check what its suggestion rests on: the readings, their report times, the room, the zone, the units and its assumptions. Example: it suggests a lower re-water threshold, but cites a probe reading that is three hours old. That suggestion rests on old evidence. Any change you accept still goes through review and readback. Keep passwords and access tokens out of your questions and screenshots. Next, prepare a handover. Try another day and prepare the handover A good shift handover means the next person can carry on without phoning you. First, practise another day. Choose a different demo zone, and suppose one source has no current reading. Complete the five parts, then decide: keep the settings, investigate, or make one justified demo change. Say why. Then write a handover note: a short note that gives the next responsible person what they need. Include who is responsible now, the software versions, where the backup is, how to recover, and the agreed final state of the equipment. A backup is a saved copy of the whole setup that can be restored after a failure. A backup counts once someone has restored it successfully in a test. An update installs a newer software version. Rollback means going back to the previous version if an update causes trouble. Follow your site’s update and rollback guide. Next, teach it back. Teach it back, then recall it later You remember best what you teach, and what you recall later. Teach another learner to find a zone’s full address, check report ages and explain what Watering and Stop all act on. Then show them the change path in the demo: draft, Review changes, Apply, readback and restore. Some skills cannot be learned from a screen. Where the physical shutdown is, and what commissioning proved at your site, need a competent local person to show you. Tomorrow, recall without notes the difference between a draft, Apply and Arm plan. In about a week, solve an old-reading case with new numbers. Spacing practice out like this makes it stick. Keep your diary short, specific and honest about anything still unresolved. That completes the daily operation lessons. INDEPENDENT PRACTICE Complete a daily record for another demo zone, including one unknown source. Compare your reasoning: Another operator can identify scope, evidence age, uncertainty, your decision and exactly how and when it will be checked. SOURCES Workbook: daily record, handover and delayed recall: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/95237d5a7f4971d98ca12104c586b1e7d277fbcf/docs/training/crop-steering-from-zero/WORKBOOK.md User guide: control scope, AI actions and readback: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/USER_GUIDE.md Install, upgrade and rollback guide: https://github.com/JakeTheRabbit/HA-Irrigation-Strategy/blob/main/docs/INSTALL.md