Climate & Nutrient solution controller


Hi guys

I noticed, that my nutrient solution alway stays at about 25°C. I read, that this might be a bit too high for fogponics. I tried to come up with something to adjust it, an maybe even control the temperature (day/night) cycle.
I saw setups/products that were kind of expensive and a bit too big for my system. So I figured to build a simple and inexpensive way to regulate the nutrient-solution. Also I’m curious of what you guys think about my idea.

Rough system setup:

  1. Thermoelectric Peltier Refrigeration Cooling System Kit Cooler Fan TEC-12706 (14$)
  2. GPU Water Cooling Block PC Copper High Performance Liquid Cooler (15$)
  3. Semiconductor BTS7960B Motor Driver 43A H-Bridge Drive PWM (12$)
    4 DS18B20 Waterproof Digital Temperature Temp Sensor Probe 1M (4$/piece)
  4. DC 12V 1A Powerful Micro Brushless Magnetic Amphibious Appliance Water Pump (16$)
  5. 3Pcs Geekcreit™ Doit NodeMcu Lua ESP8266 ESP-12E WIFI Development Board (6$/piece)
  6. Some silicon tubes

My plan is, to replace the peltiers small fan and radiator with the GPU cooler. Attach it to the pump and pump the nutrient solution through it. Using the temperature probes I want to regulate the PWM that powers the thermoelectric element and thus control the liquids temperature. I know that it’s not the most efficient way, but it’s small and cheap to build. I can even heat up the solution, should it get too cold.

So far i got the GPU-cooler mounted on the peltier element (added some thermally conductive paste). And now I have to wait for the pump and tubing :sweat:

It might even be suited to regulate the temperature of a small grow-chamber :slight_smile:

what do you guys think?


Mr. Gadget

New Food Computer Wiki!
Recommendations for Water Chiller?
Where's the cooling aspect of the food computer

Loving this. We are currently hacking a kit from to get it lower cost for the PFC. Would love to see your progress for potential integrations into V2. We def need water temps to be a 10 degree differential from air temps for most crops so water cooling is a BIG need for the community.

V2 Food Computer Kit Build
$300 Food Computer

Thx Caleb! Just got my pump today, but now I need to find the right tubing. The motor is tiny and surprisingly silent.
I will post my progress and share my results. I would feel honored should this concept find it’s way to V2.0 :relaxed:
5000$ for a cooling system is way too much for a small scale hydro/aero system…



Like the idea of using out of the shelve computer cooling bits for this. Some questions though.
The plants like that the soil/nutrition solution is about 10degrees (either celsius or fahrenheit) lower then the grow chamber temp? How important is this? What is the optimal difference?

What kind of cooling power does the Peltier element offer?

Really like the prices on the components you chosen. Great work.


You can check out the following video to actually see the cooling power of a similar setup in action (jump to ~5:20):
Peltier Thermo-Electric Cooler Module+Heatsink Assembly - 12V 5A

As for the effect of the temperature on the plants, I cannot say but I’m sure, that someone here can answer it :slight_smile:


Nice :slight_smile: looks like it is quite powerful and can not be let to run all the time.
Ordered my bits too. Lets see how it works.

Also ordered current sensors to measure what takes how much power :slight_smile:


So here it goes. Finished up the first prototype of my idea. I know it looks kind of hacky, but I had to improvise a bit :smile:

The cooling system is currently controlled by an Arduino UNO.
The ventilator and peltier module are switched on/off using a 2-channel relais. Using the temperature probe (mounted on the hot side) the arduino automatically switches off the peltier, should the temperature rise above a given threshold.
The pump is connected to a motor driver that ouputs a PWM to allow better control over the flow rate. Currently the pump-power can be adjusted manually via a potentiometer connected to one the arduinos analog inputs.
I will test the setup over the weekend and will post the results :slight_smile:


Mr. Gadget


Just a quick update, ran the first tests with the setup. Still needs some tweaking but first results show, that it realy cools the water. Sadly I noticed, that the temperature probes have an offset :confused: So I will run another test hopefully tomorrow.

Test Setup:

  • 8 liter water bucket
  • Ambient temperature 23.9[°C]
  • Inital water temperature 25.5 [°C]
  • Final water temperature 23.5[°C]
    => Runtime: 50 minutes

Measurements were taken inside the intake/outflow pipes.


Just a quick update:
The first version was a bit to unstable (build quality) and the cooling effect was not what I expected.
My thought: The hot side is not getting cooled enough from the radiator in place. I found a working (!!) CPU cooling unit in the electronics disposal bin in our office. Replaced the old radiator with the new one.
Found a solid plastic box and mounted the pump, peltier-element, and the two radiators.

Still need to finish the wiring and controller-software but so far I’m pretty happy with the result:

Hope it will work a bit better that the previous version :grin:


Mr. Gadget

Community Developed Mods
Community Developed Mods

What about evaporative cooling? Coil the nutrient tubing in front of the intake/outtake fan and use the existing internal humidity? Of course if you wanted something to cool more consistently you could have a separate insulated module with a recirculator as well. Not sure how effective this would be through plastic tubing though.


Great work :slight_smile: I knew not throwing away my old CPU coolers would pay off one day :slight_smile:

Is this possible to connect to the FC’s Arduino/RPi setup or does it require its own brain to function?


@silversson I designed the setup, so that it does not need a lot of expensive hardware. I wanted the cooling
part to be kind of self regulating (simple code with safeguards in place should something go wrong :wink: )
I’m looking into the idea, to use a RGB-LED light strip amplifier, to control the whole setup.
This would make it a lot easier to build the cooler and much more cost effective.

As already mentioned (or not) cooling with a thermoelectric cooler is not very energy efficient and generates a lot of heat. My thinking was, to use the generated head to our advantage :smiley: if needed said generated heat could be used to increase the grow chamber temperature. Should it be already warm enough, the hot air could be diverted to the exhaust vent. It just needs some clever engineering :wink:

I’m sorry for the slow progress… :sweat: I only get to stay at home during the weekends as I currently have to attend a military refresher course and at the same time prepare myself for a robotics competition I’m joining next week. :heart_eyes:

As I did not yet built my PFC, I still have some questions:

  • How much current do all 12V components need (worst case)
  • How much relais are available (not yet used)?
  • What kind of interfaces are available that might me used to interface with another module? (I2C,CAN-bus,serial)

Kind regards and have a nice weekend

Mr. Gadget


Have you tried just using ice to cool the nutrient solution down?


Yes I did. well, not ice but those cooling pads. Cools down the solution down very fast, but it’s just a temporary solution… And one would need to change the cooling pads again and again.


  • Cheap and reusable
  • Cools down solution fast


  • Cannot be really automated
  • Increased chance of contamination of solution
  • Increased effort for users
  • No fine tuning possible

Not sure:

  • Are the plants affected by the sudden temperature differences?


@jake check this out.


The 12V power supply in original BOM has maximum output of 30A. This 12V PS is needed for 90W grow LEDs and approx. 10W (2 meters) for white LEDs. Maybe Arduino can be powered from this source (let’s say another 5W). Both ventilation fans should be 8W. So 113W in total. It means there is 360-113=247W left. Your Peltier element will take another 60W, so there is still 187W remaining.

As for relays: I m not sure either but maybe one is not used (?). But be careful, the electromechanical relays are not designed for huge DC currents (also, they are not designed for fast PWM). The better choice is to use MOSFET transistor, for example IRLB8743PBF (I used them for 10A switching, PWM, no heatsink, without any problem).

Interfaces: There are three interfaces which are commonly used in Arduino world – serial (a.k.a UART), I2C and SPI. The Arduino have several serial (UART) interfaces so this can be used. I2C supports several slaves on one bus and right now, only one slave is used.

I2C was designed for short “inter-integrated circuit” communication, thus small distances – few centimeters. In my last project, I used Modbus over serial line. This is very simple, bidirectional master-slave protocol with CRC checking. It supports up to 248 slaves on one bus and if you use RS485 (differential signals) you can use it on long distances too. Without RS485, you can use it on plain UART. I think Modbus is good choice, because it is easy to make design more modular.

If you are interested, I can help you with implementing Modbus. I have Modbus bootloader and simple client code for ATMega88/168/328 AVR family.


Not everything needs to be completely automated, When the cost of automation outweighs it’s usefulness it’s hardly useful.

If you actually use ice, there is no risk of contaminating your nutrient solution. If you freeze pH adjusted water it wont even affect the pH. If your plants aren’t drinking that much water yet, you can freeze the nutrient solution and it wont even affect pH, but it wont dilute the nutrient concentration either. Some people even freeze water in plastic water bottles, essentially making their own non toxic ice packs/cooling pads.

Granted using Ice isn’t ideal in a lot of situations like growing outside in summer heat, in large systems needing a lot of water volume, or if you simply don’t have room in the freezer. But when your growing inside in already climate controlled rooms and only need to change the temps a few degrees then it’s a viable option. I wrote this article “Nutrient Solution Temperature is Important” that also explains other nutrient solution cooling options.

If you have to cool down the water temp a lot (15+ degrees), there could be some initial temporary shock to the plants, but that’s far less damaging than the heat stress from high root zone temperatures. Slight changes (5-7 degrees) to root zone temps are natural occurrences in nature from differences in daily temperatures, as well as day/night changes.


Cool input … sense…ible options , level parameters … successful growth


@MrGadget @silversson @Peperoni @JoshSinykin and all those who finished the built of the PFC: my water temp is between 28 and 31 celsius, and I read that other nerdfarmers have the same issue. Did any of you have adopted a water cooler? I am really tempted to test one of this.

@caleb you mentioned that your team was working on a cooling mechanism for the 2.0. Would you suggest us to wait until you release the 2.0?


H i Pau, I have the same problem and I ordered a Peltier cooler. It has not arrived yet so I cannot comment on the effectiveness yet…
Meanwhile the outside temperature has lowered a bit here in NL (we had extremely warm weather the last weeks) and the water temp has now come down to 26-27 ° C. I think it is still a bit too warm although I read various recommendations on the ideal water temperature. Anyone here lik to comment on this ?