Berrys Automated Perpetual Grow Journal

Hey everybody, this is the journal for my soon-to-be automated perpetual grow room.  I will be setting up the room and getting the first group of plants ready over the next couple of months.  
 
This will be my first perpetual home grow.  I've done some small grows at home in the past, but they were mainly for just vegging out some plants in a tent and then moving them to a grow, or growing a mother to get some clones for a friend.  I've been working as a grower in the cannabis industry in Colorado for the last few years so I have a good base of knowledge and experience I'm working off of for setting up this grow, plus I have a degree in horticulture.  I've just never had the right space to setup a small to medium sized perpetual grow, but I'm moving into a new house soon, so I'm going to get started on the home grow that I've been dreaming of.
 
As for automation; I've been teaching myself programming for the last couple of years so I'm going to try and automate my grow room as much as possible.  I sometimes have to go out of town for several weeks at a time for work so I've been experimenting with getting a setup working that will monitor and log all the environmental information in my grow room like temperature, humidity, and soil wetness, then the system will automatically power on the lights, ventilation/fans, and irrigation based on the data it gets from the sensors.  I will also be able to access/control the system remotely so I can still keep tabs on things while I'm away.
 
I haven't decided on which strains I'm going to go with, but I will eventually keep a mother plant that I will keep grafting on strains that I want to keep around so that I can clone them out and run through the room at a later point.  For right now I'm starting several seeds I had in storage.  Some of them are bag seeds or seeds I've found in commercial grows from an accidental fertilization from a male that was missed.  The other seeds are some I was given by a friend that are Chemdawg x Island Sweet Skunk.  He grew a few of the seeds and there seemed to be two main phenotypes; tall and short and pretty dense colas that smoked pretty well.  I might replace these plants started from seeds if I get some better clone genetics later, but I'm probably going to run into some bugs in my automation software while getting it going and wanted to use some guinea pig plants that I'm not worried about losing.
 
For now I'm starting the seedlings in a 4x4 grow tent under a T5 veg light.  They are in a potting soil mix that I mixed myself (peat/vermiculite/perlite 60/20/20).  I germinated three seeds so far, using the shot glass method, and then placed them in damp soil.
 
Here's a pic of the first seedling starting to poke through the soil.  
 
 
 

 

All three of the seeds that I planted a few days ago have sprouted.  I'm just going to give each of my plants a letter for a name for now starting with A and proceeding alphabetically as I sprout more plants.
 
Here's some pictures of the seedlings so far.
 
Plant A

Plant B

Plant C

 
I've started working on setting up the automation system.  I'm going to be using a Beaglebone Black to run the automation software and to interface with all of the sensors.  I'm also building a controllable electrical outlet that I can interface with the Beaglebone Black and turn on and off power to things like the lights, pumps, and fans.  I wired up the outlet last night but I must have made a mistake because it still isn't working correctly but I think I know what I did wrong so I should be done with that soon.

 
I've also been testing out an LED light that I built myself.  It has twelve 3 watt leds (6 red & 6 blue).  I'm just going to use it for supplemental lighting for now, but it turned out better than I had expected.

 

Day 3:
 
The three seedlings I've started are all doing great so far.  I've only been feeding them with tap water that has been pH adjusted down to 6.0.  I'm going to wait to start feeding them nutrients until about day 7, or if I notice the color of the leaves start to change into a lighter green or yellow which would indicate they are starting to starve for nitrogen.
 
I haven't decided on which nutrients I'm going to use yet, but I've used Fox Farms nutrients in the past and liked them.  I'm a big fan of coco/perlite for grow medium, so I'll probably end up repotting into that.  I'll be repotting in 2-3 weeks into 1 gallon smart pots, then finishing in 3 or 5 gallon smart pots.
 
I also started working on the automation software that will monitor my grow and control some of the equipment like lights, fans, and pumps.  Eventually I'm going to make it to where I can access my automation system remotely and see information from the sensors, but first I wanted the system to have a user interface that can display relevant information on a monitor inside the grow room.
 
Here is the user interface so far.  On the right side of the window I'm going to add a section that shows whether each plant is wet/dry and needs watering, but I don't have the soil sensors in place yet.  

 

Day 15:
 
My plants ended up getting drought stressed a little when the soil got really dry last weekend.  They are recovering fine now, but they are a little stunted. I kinda expected to run into problems with my software running the equipment, so that's why I started with random bag seeds instead of clones or seeds I've paid for.
 
The seedlings were in some solo cup sized jiffy pots I had laying around and hadn't used yet.  They seem to dry out very quickly, especially because the humidity is so low inside the tent right now.  I'm averaging between 25-30% without a humidity dome over them, and with the humidity dome I can almost get it up to 40%.  I need to go out and get a small humidifier and set it up inside the tent.
 
The main problem is that I haven't worked out all the little bugs in the automation program's code, so every once in awhile the program will freeze up because of some error I haven't seen before.  When the program freezes all the equipment freezes in its current state.  So if the lights are on they will stay on and worse if a pump is pumping, it will just keep going.  Luckily that didn't happen this time, but what did happen was that the program froze and the pump I had scheduled to come on never did.
 
Here is a picture of my grow tent setup.  I took this the day after the seedlings got really dry and wilted so as you can see it didn't really harm them too much, but as you'll see in the picture of them today they are definitely not as big as they could be at two weeks from sprouting.
 

 
 
Here is what my bag seed seedlings look like at day 15.  Like I said they almost died on day 10 from a fault in the automation software that was controlling the pump that was watering them twice a day.  I have since placed the jiffy pot containers the seedlings are planted in down into some solo cups which is helping to prevent them from drying out too quickly.  The other concern I have about the soil drying out completely is that this can cause the buildup of precipitated salts that can then give the seedlings salt burn.  That's why I think two of the seedlings have some burnt tips on the leaves that were developing right after the soil got completely dry.
 

 
 
Here's a picture of the components inside my automation control box, which is inside the tent for now and connects to the sensors and power relays.
 

 

Day 17:
 
I think I finally got the majority of the bugs out of my automation software, or at least the system will run 24/7 now without crashing.  
 
Right now the system just monitors the air temperature and humidity and turns on and off the lights and fans.  The lights are on a 18/6 schedule and the fans just stay on 24/7.  Next I'm going to add the soil moisture sensors, and have the system water the plants when they are dry.  Until then though I'm going to work on adding some simple features like data logging and email/text message alerts.
 
I'm going to wait on getting a humidifier and see if my low humidity just goes away by itself as the plants get larger, but I'm probably going to have to go out and get a small a/c unit soon to start bringing the temperature down to where it are supposed to be.  Right now it is averaging 75-80 with the lights on and the tent door completely open, and 95+ with the lights on and the tent completely closed up.  I'm glad it got so hot in the tent though because it revealed a bug in the gui that displayed the temperature value incorrectly if the temperature was over 100.

 

I took some new pictures of my plants. I also gave them a good flush and then watered them with some nutrient solution with slightly higher ppm because they were starting to look a little starved for nutrients.
 
Seedling A

 
 
Seedling B

 
 
Seedling C (showing the most nutrient deficiency symptoms, but it is also the tallest and biggest so maybe it has used more energy.)

 
 
All 3 seedlings

 
 
Seedling A under just the led light. 

 

I was looking back through some of my old notes trying to remember when I had gotten these seeds so maybe I could figure out what strains they were a cross of.
 
Two of the seeds were from a commercial warehouse where I know that the pollen came from a TrainWreck strain they had that would always hermie if the grow got too hot.  The seeds were removed from the bud as it was being trimmed.  I think my seedling A is Trainwreck x NYCD and seedling B is Trainwreck x Rare Darkness.
 
The third seed that germinated was from an actual bag of weed I got from Botanico dispensary in Denver.  I'm not certain what the strain was, but I remember it surprising me to see a seed because the weed was so good.  I think it was Rock Lock, but I'm not sure, so this plant is still going to be a mystery.

 

Here are some new pictures.  Plant C is starting to grow faster than the other two, and it also has longer inter-nodal growth making it taller.
 

 

 

Day 35:
 
My plants are doing ok.  They could be bigger, but it isn't bothering me right now because I'm not really trying to go all out yet until I can move my plants and equipment into my new house.  The main reasons I think their growth has been slow are that temperatures, humidity, air-flow, and light intensity are not optimal yet.  I've been having to back off the T5 light so that the temperatures around the plants is in the acceptable range, but the temperature could still be about 10 degrees cooler.  The humidity here in Colorado is always pretty low, and in the grow tent it averages around 25 percent.  So I'm going to get a small evaporative cooler soon that should lower the temperature about 15 and increase the relative humidity to around 45 percent.
 
My soil moisture sensor is in the mail so I'll be hooking that up and adding automated watering to my system soon.  In the mail is also a small webcam that i'm going to connect to my system and point at the plants.  I'll be able to interface the video feed from the camera with my automation program and I'm hoping to add remote viewing capability.  I'll be able to watch my system water the plants in real time from anywhere.  The main reason I want this is because there are visual clues a plant gives, like wilting, when it is in distress that the program can't tell me, but I'll feel more comfortable leaving my plants for weeks if I can verify their health visually.  I'm also going to be adding a push-button to the interface around the same time that will allow me to water the plants at the push of a button is I choose to instead of waiting on the automation software to water them.
 
I re-potted my plants into 1 gallon pots with coco coir and perlite. I've also upped the ppm of the nutrient mix I'm feeding them to around 1200.
 
Here's some pictures of my plants today.  Plant C is the one on the right and is doing much better than the other two.  Plant A and B suffered from a nutrient deficiency, probably Magnesium or Calcium.  Plant B, the middle plant, had it the worst and also had the poorest root development when I was re-potting them.
 

 
Plant A

 
Plant B

 
Plant C (I really hope this one is female)  I topped this one a couple of days ago.

 

I finally got around to adding data logging and email/text alerts to my automation program.  Right now I only have it set up to email me once an hour if the temperature inside my grow tent is above 100 degrees.  I didn't want to set up the text message alerts until I know I've worked out the bugs, otherwise I risk sending myself hundreds of text messages in a few minutes accidentally.
 
Right now I have the system updating a google doc spreadsheet with the temperature and humidity every 5 minutes.
 
Here is what the graph from last night looks like. Blue is temperature and red is humidity.
If you click the chart it will bring you to an interactive version.

 
I've been having to close up the tent completely at night because the light bothers my SO and I can't keep the room's door shut for very long before my cat opens it.  So, I closed up the tent and had my program shut off the light when the temperature reached 95 and turn it back on when it had dropped to 85 (in retrospect I probably could have set this to 80).  The ambient temperature is usually around 70-75 in the room that the tent is in, and around 75 inside the tent with the lights off while zipped up.  If I didn't have it shut off at 95 the temperature would probably rise up to 105-110 quickly.  On the chart above the lights are only on when the line is rising and are off while the line is on a downward trend.  That means that the light was only on about half the time last night, but that's better than not being on at all which is what I was trying to achieve.  Hopefully once I get the swamp cooler installed I'll be able to maintain a much lower temperature, higher humidity, and keep the lights on all night long with the tent closed up.
 
Here's a picture of my Plant C.  This one is looking the healthiest so far.

 

Thanks.
 
I'm eventually going to try and devise a system of peristalic dosing pumps that will allow me to maybe mix nutrients from concentrates and also raise and lower the pH of my reservoir.  I'm also going to get a small RO unit that will be controlled by the program keep a water tank full.  Then I'll have another tank that pumps in water, nutrients, and pH balancing chemicals to make the proper mix.  I have a pH meter hooked up to my system already but because I am still hand watering them I haven't had need for it to be running, but once I set up the auto-watering then I'll put the pH probe in the nutrient mix reservoir to ensure the mix is the right pH before the system will feed the plants with it.  I'll have to get an EC chip and probe once I start putting together the auto mixing part of the system.
 
I only know python, but luckily it is a very diverse language that you can learn quickly and do pretty much anything with.  There are also tons of resources available for python on the internet.  I pretty much google "how to <insert problem here> in python" and the right answer will come up 95% of the time.
 
I wouldn't mind collaborating on this at all.  I actually have plans to install a copy of my system in another home grow here in Colorado as soon as it works reliably enough.  A couple of friends have a basement grow, but they can't be in it or check on things for most of the day, so once I get the system where it will water the plants automatically and monitor the environment it will be perfect for them, plus its cheaper, and more customizable than the commercially available grow conroller they were thinking of buying.
 
I don't want to share my code yet though, it is just too ugly and uncommented to do any good.  I'm improving it every day right now, but once I get all the features in and sensors added that I want, then I'm going to spend some time rewriting the code to make it better and more reliable.  After that I might be willing to upload my code to github or something for further improvements from others.
 
I can tell you exactly what hardware and give code examples of how I did things if you want to build one and need help.

 

My plants are doing well.  Plant A is coming out of its nutrient deficiency and the other two are doing great.  I still haven't been able to keep the temperatures down so they could probably be doing better, but I should hopefully be fixing that problem in the next couple of weeks.
 
I'm going to be taking my first round of clones as soon as Plant C has grown back enough after I topped it last week.  I want to determine the sex of my plants soon so I'm not wasting time on any males, but I'm not sure If I want to switch my plants to flower yet, so I might end up putting the clones into a separate area that I'll put on 12/12 from the start and just look for flowers or pollen sacks.  If any of the clones turn out to be male then I'll just get rid of the plant they came from.  I'm going to get some better genetics later when I start filling out my new grow room, but I'm still curious about how these plants might turn out so I'll still run these plants all the way through flower if they are female.
 
My soil moisture sensor should be arriving in the mail today, so I'll be able to start working on the auto-watering part of my system in the next few days.  I only ordered one of the sensors so far, mainly because I've ordered other sensors that ended up not working for what I needed, so I'm going to test this one out before I buy more of them.  I started off using a $50 soil moisture sensor that ended up having some serious design flaws (the sensor itself is fine, but they used cheap unshielded wires for the cord so the data fluctuates wildly because of interference in the wire I think), so I'm going to try the cheap $5 electrical resistance hydrometer type sensor and see how accurate those can be.  For right now I'll just put the moisture sensor in one of the three pots and just water the other two when that one gets dry, they're all drying out about the same rate right now anyways.
 
Here's some pictures from today.

 
Plant C after being topped.

 

My soil moisture sensor and webcam came in the mail today.  
 
I've already connected the moisture sensor to my system and it seems to be working.  I just watered my plants last night and the soil still feels wet to the touch, and the sensor is reading about 89% wet.  The sensor sends back an analog signal on a scale of 0(dry) to 950(in water) so I just programmed it so that it takes the value from the sensor divides by 950 and multiply by 100 to give me a percentage.  I'm going to set up the system to log this value over the next few days so I can see how much it changes based on how dry the pots get, that way I can calibrate the system to water at the correct level of soil moisture.  Long term I'm going to be watching to make sure the system is consistently getting the correct soil wetness measurement and there isn't any drift or degradation over time.  I'm also going to have to watch and make sure that the sensor itself doesn't get corroded from being in humid soil all the time.  
 
The webcam is going to take me a little more time to get hooked up.  The main obstacle in my way is the fact that the BeagleBone Black only has one usb port and I'm currently using it to boot my OS from an external hard drive.  I've found this to the most reliable way to run a long-term project on either a BBB or Raspberry Pi, there are too many issues with booting from an SD card all the time.  I'm going to have to get a usb hub to add more ports, but none of the ones I currently have work with the BBB.  I researched a little and found one at MicroCenter that will work, but I'll have to go buy that later.
 
I also discovered a few bugs in my system as I was trying to get the soil moisture sensor working.  I noticed that if my air temp/humidity sensor becomes disconnected then the program won't start at all, which I suspect might be a problem with my code in the arduino uno (microcontroller) part of my system.

 

If anybody plans on trying this build then be sure to get a different camera.  I got the logitech c310 but it doesn't seem to play well with my program.  The python module that I'm using for most of my gui stuff is pygame which is intended to be used to design video games, but it works well for an information display too.  Pygame can import the picture from a webcam, but the webcam has to be compatible with linux and v4l2.  Unfortunately these must not be the only requirements for because when I try to use my camera with pygame it doesn't work like it is supposed to, because from what I can tell from researching the camera exports the wrong file format for the program to import.
 
For some reason I can't use the webcam like normal, but I did figure out a way to get python to get the webcam to take a single picture and save it.  Then I can just call up that picture with any program.  It won't be a streaming 30 fps video feed, but I will be able to see the plants and the picture should update every second.
 
The soil moisture sensor is doing great.  The only thing I don't like about it so far is that it fluctuates a lot and isn't very precise.  This morning the readings have been fluctuating between 85% and 90%, so I'm thinking of writing in something that will average out this value over time.  Maybe I can store the values from the moisture sensor for 1-5 minutes, then return the average.

 

I figured out a way of using my webcam so that I can now see it in my program.
 

 
I couldn't connect the webcam to the beaglebone black yet because I'm using up the only usb port, but I have a linux laptop on a shelf right outside the grow tent that I usually have the grow automation program running on through ssh'ing into the beaglebone inside the tent, but it displays on the laptop screen through x11 forwarding. 
 
I have the webcam plugged into the laptop, and then the laptop takes a picture with the webcam every second and saves it to a webserver on the laptop. Then my program ( on beaglebone black) downloads the image from the webserver ( on laptop)  and use it in the display.  
 
If I were to set up the necessary port forwarding in my modem/router I could see this same display from anywhere with internet access.

 

My automated grow is still going well, but I did get rid of plant B. 
 
Plant B just wasn't growing as fast as the other two, and it was taking a very long time for it's pot to dry out.  I'm not sure what was wrong with it, but the rootball was definitely on the smaller side and had barely started to grow into the coco after being repotted.  The other two have roots coming out of the bottom holes in the pot already, so plant B was definitely stunted.  Above the soil it was the same slow growth and the plant was a third of the size of the other two.  It was a bagseed, so it could possibly have been just bad genetics.
 
I was worried that the soil moisture sensor would get corroded or covered in rust because it was sitting in moist soil all the time, and I was right.
 

 
So for now I'm going to stop using this soil sensor, and I'm either going to have to devise a system that will insert this same type of soil sensor and remove it after taking a reading, or I'm going to have to find a different design of sensor that can stay in the soil for prolonged periods of time without rusting.