System Capabilities


Hey everyone! I am awaiting the arrival of all the parts which should arrive in the next few days and I was wondering if I might be able to get a question answered in the meantime. I am wondering what the capabilities of the various sensors and such in a volumetric sense. Essentially how big could the machine be with the current electronics or is there a way I could find this out? Any help would be appreciated!


Are you looking for a Food Computer and half, or Food Computer Squared? MIT has made a Food Server, and they are working on a Food Data Center. However the costs and scale are completely different.

The Fans provide a specific amount of CFM, if you start changing the dimensions of the unit too far, you made need beefier equipment. Grow chamber lights fill the entire box, actually it looks like the box dimensions revolved around the grow lights and the wash basin more than anything else. The height may be subject to change to let the plants grow larger.

EC and PH, those sensors are a work in progress. The ones in the BOM are pretty weak and unreliable (so I have heard) so you need to treat them well and flush the unit head with water/solution before storing between measurements.

Temperature, Humidity, CO2… those could scale further in my opinion.

Heater… that’s the question, the little unit does put out considerable heat and volume… you may get away with more.


Hi @ltchurch

We sourced the actuators assuming the size of the system was about 1m^3

Sensors are size agnostic, but the larger the system, the more difficult it becomes to get to homogeneity - need good wind mixing and water mixing. The pH and EC sensors we have in the V1 are pretty crappy. I would recommend Atlas scientific sensors (pricy but good) - there are good Arduino sketches online that show how to use them (they work over I2C and UART)

The heater is pretty strong for the size but probably can’t do too much larger. Once you get to the size of the Food Server you will need the lower level of industrial for the actuators.

The lights did drive the dimension to some degree, same with the reservoir tub. The height was sized so that a tomato plant could be grown in the FC (though probably the height should be a little taller).


The current software can control the temperature and humidify. CO2 is measured but I do not think it has a control loop assigned. pH and EC do not have a method of control

From my experience the pH and EC sensors are nice but could just as well be replaced by a Hanna combo tool and periodic monitoring. Im working on control system that can monitor and adjust the EC and pH, but you can grow perfectly w/o it.


Hi @dan,
What about CO2 Sensor used in V1 design? Will V2 stick to the same model?


@Rckco I think CO2 is controlled with the vent fan. When CO2 levels go above the setpoint, the fan is turned on to blow outside air in until the levels go down. The problem is that the vent fan also affects air temperature and humidity. Because of that, it could be turned up for other reasons, even if CO2 levels are below setpoint. I can’t prove this, though, because my PFC doesn’t have temperature control yet, and here in Brazil is quite hot, so the vent fan is almost always on to decrease air temperature.


Last I was aware the CO2 sensor, the COZIR will be used in V2.


I plan on putting together recommendations on calibrating the sensors. If anyone has already started this please let me know.

When I last checked my CO2 sensor it was reporting 1300-1400. Which is above the recommended 400-450 in the documentation (if memory serves me)…I moved the FC1.0 up to the dinner room to let it acclimate to fresher air than the basement. I will check the sensor again to see if there are any changes.


Can anyone suggest an alternative to COZIR CO2 sensor? 5K ppm ambient version specified in BOM seems to be out of stock (


@serein 2K ppm should theoretically be fine, since air level is ~400 and for plant growth it should be set up to ~1000-1500 ppm. 10K has a wider range. I don’t know if this lowers precision, but if not, should be OK too.
I used the MG811 sensor instead of COZIR’s, because it was a little less expensive and I didn’t need to import it. But I wouldn’t recommend it, unless it was the only option, for the following reasons:

  1. Every sensor has a different volt x CO2 ppm curve, and this isn’t specified when you buy it. You would have to calibrate it with another CO2 sensor.
  2. It requires a 6V-200mA power source, which you can’t get from the Arduino itself. You would need to power Arduino through an external power supply and have a step up regulator to get 6V (some breakout boards already come with this regulator, but the current requirement still continues).

@JoshSinykin Did you mean the recommended ppm for plant growth was 400-450? I have read some books/papers, and they say it is around 1000-1500 for optimal growth. There is a really nice guide to growing hydroponic lettuce from Cornell with a recipe that I’m following on my first batch. Here’s the link:


This sounds exciting. I would be interested to hear how you get on in this area.


the 400-450 was not for plant growth, just what you should typically see from the COZIR in regular air measurements.


I’ve been modifying the code to allow for a calibration mode. All of the sensors do nothing when they are written to. I have changed this. For example, I’ve just added a water level sensor. It can be read by looking at ATWL 1 in the data stream (returns the % of full). To calibrate the sensor, the tank is first filled, then a write to ATWL 1 1 automatically reads the level sensor and sets it as 100%, then scales the remaining readings.

Of course, this is only one small step in the complexity of the system. The controller code must read/write the sensor to the django api.

Then, the UI must allow for a button of some sort to activate this mode.

I plan to implement EC and pH calibration in a similar way. For example, the pH sensor is inserted in a standard, a SWPH 1 1 takes the reading and uses it as some type of calibration point. If a two point calibration is desired, a SWPH 1 2 can be sent after the probe is inserted into the second standard.


@dan, from recent commits on github I see that there are changes in sensors for V1.1. I see rgb from atlas, am2315 for air temp and humidity. Is this just widening the choice for sensors or preferred solution for v1.1?
Can someone from the team comment on those choices? I


As I understand now, the EV and PH sensors cannot be used continuously. So I ask: why do we need these sensors to be connected to the Arduino ? I think it is a lot cheaper to buy handheld meters (I see them for sale for 20 Euro online) and type the values is in the software, instead of measuring it with the DFRobot sensors and than still have to do all the manual labor…
Maybe it is even possible to hack these handheld meters and get them connected to the Arduino somehow… ?
digital PH meter: (19,95 euro)
digital EV meter: (19,95 euro)


I think this is because in the long run human time is precious and often error-prone, while computers and sensors are cheap (or getting cheaper fast) and pretty solid and stable.

Whatever can be automated, needs to be automated.

And AFAIK v1.1 design will have it measured continuously and with feedback loop.


Found an interesting post on HackADay where someone build a 3 dollar/euro EC Meter for Arduino based on an American or European socket. He is using it for his aquaponics projects.


Hi Mateus.
Did you buy a sensor module based on the MG811 sensor or just the sensor and build the module (incl amplifier) yourself ? I ask because I found a module by Sandbox Electronics which is ready to use on Arduino and still a lot cheaper than the Cozir.


@Peperoni I bought a module containing the MG811 sensor. It’s not Sandbox’s , but a similar one. It is less expensive then the Cozir sensor, but I wouldn’t recommend it, unless it was your only option.

One of the problems is that it needs a heater for the sensor to work properly, and it requires 6V and ~200 mA. This is a lot for the Arduino when it is working with the USB 5V power supply, which has ~500 mA total capacity for the board and the sensors and actuators. So I had to use a 7.5V external power supply, which has 1 A capacity. I also had to use a step-down voltage regulator because my module had to be supplied with 6V (maybe yours can be supplied with 5V and has a built in step-up regulator , but the current limitation still applies, keeping the need of a 5V external power supply).
With some extra effort, this can be done.

The other problem is that every sensor has a different CO2 concentration x voltage output response curve, and they don’t come calibrated. Unless you could compare your sensor with another one that you know is calibrated, you would have to go with a mix of theoretical curve from the datasheet and your readings at CO2 ambient air concentration.

I only bought it because it was the only one I could find (and I didn’t know all of this beforehand since information about it is scarse).

Community Developed Mods

You are right… I took another look at the datasheet ( which convinced me to stick with Cozir for now. Hopefully they ship to Europe without problems.
Thanks !