Terrarium growing experiments

[Draikan]

Wait until your non-waterproof fan breaks, then replace it with a waterproof one. I have one normal fan that has survived >2 years of humidity and misting. I also have one that died literally on the first day I put it in the tank. If you're already using a 120mm fan, Noctua has a semi-industrial line than would be perfect: Amazon.com

My fan is located in the middle of the tank pointing up. Air flows out of the fan, hits the top of the tank, then goes toward the back wall, then down along the back wall to the bottom of the tank. This way it flows in the same direction as the misters spray, rather than fighting them. I agree it would be better to turn the fan off while misting, I want to do this once I get the irrigation control integrated with the rest of the control system.

I found a grower on youtube who appears to be having a lot of success growing in a controlled environment like we are talking about. He also uses the inkbird controllers. It's worth checking out his methods. He explains his terrarium setup in the video below:

[Draikan]

I finished up the initial firmware for the new sensor. It's now able to control the sensor temperature to anywhere between ambient and ~60C to keep humidity at the sensor from causing drift in the measurement. It can also recognize when it has been sprayed, and instantly heat itself to 60C to remove the liquid water, then resume humidity measurements as soon as the liquid has evaporated. The airflow measurement also appears to be working great, but I don't have a very good source of adjustable airflow setup here so it is hard to test it extensively.

The below plot shows the sensors behavior when it gets wet. I've used a q-tip to place a drop of water right onto the humidity sensing element. This should be similar to spraying it with a mistking. The x-axis on the plot is time, in seconds.

1) I place the water droplet on the sensing element.
2) Air humidity exceeds 100%, indicating the sensor is "under water".
3) The temperature target for the sensor (purple) jumps to 60C (140F), and the heater quickly brings the sensor temperature (red) up to the target. This evaporates the liquid water.
4) The last bit of water evaporates off the sensor and the measured humidity quickly starts to drop back to ambient.
5) By this point the humidity reading has stabilized, and accurate humidity measurement can continue.



Deliberately wetting a humidity sensor like this would normally either break the sensor, or cause it to be inaccurate until the liquid had naturally evaporated. In an environment as humid as a terrarium can be, this natural evaporation could take hours. Here it only took about 2 minutes to evaporate the water drop, and 1 more minute for the reading to return to ambient. This test was done in still air, with gentle flow from a circulation fan the evaporation would be faster. This is all automatic, and the sensor can even indicate when the measurement is invalid due to wetness, preventing the controller from taking action based on bad data.

	Mistking Looking for a misting system? Look no further. Automated misting systems from MistKing are used by multitude of plant enthusiasts and are perfect for Orchids. Systems feature run dry pumps, ZipDrip valve, adjustable black nozzles, per second control! Automatically mist one growing shelf or a greenhouse full of Orchids. See MistKing testimonials

[Metallising]

Neat!
What sensor is that? SHT35 ?
Does the sensor automatically heat itself at a specific humidity level as a non programmable built in function, or do you have to program it?

[fhopper]

Cool stuff Draikan! Are you able to run a calibrating routine each time the sensor reaches 100% humidity in order to verify the accuracy of subsequent readings and negate any potential drift?

[Draikan]

This sensor is my custom circuit board, which uses an SHT35, TMP117, external heater, and an M0+ microcontroller to coordinate and control the measurement:



The SHT35 doesn't do any of this automatically. All the heating and temperature control is additional behavior I have built up to protect the accuracy of the sensor.

I might be able to add an auto calibration routine, but I have not needed to. Ever since I first started heating the sensing elements last year I haven't had any accuracy issues. I'm working on getting a batch of the new custom sensors into the tank for test now. There's no reason to believe these won't be just as stable as the initial prototypes (they should be even better), but I still have to test it to be certain!

[Metallising]

Oh, I liked the idea but now I see it's over my head.

For now I will keep my cheap humidity sensor in a corner where it doesn't get too wet.

[Draikan]

If you have something that works now, don't feel like you have to change it!

Just remember to take baby steps. It's taken me a year to get this far, and I've done it a tiny little bit at a time. I started just by reading data from the SHT35, then by reading the TMP117, then control the SHT35's internal heater, then learn how to combine the data from the two, etc.

I'm pretty reliant on the mistking for automatic watering, and in my tank it gets water everywhere. I have the sensor positioned carefully so it stays as dry as possible, but sometimes when the plants get watered I can see the humidity stuck at >100% for a minute or so, and I know the sensor is getting wet. I want other folks to be able to use these eventually, so it's really important to have something so robust that it appears to "just work". I don't want other users to have to worry about careful positioning or feel like the data isn't 100% reliable. If it's in a terrarium it's probably going to get wet eventually.

	Mistking Looking for a misting system? Look no further. Automated misting systems from MistKing are used by multitude of plant enthusiasts and are perfect for Orchids. Systems feature run dry pumps, ZipDrip valve, adjustable black nozzles, per second control! Automatically mist one growing shelf or a greenhouse full of Orchids. See MistKing testimonials

[Metallising]

My Inkbird humidity sensor just died a few weeks ago after 2 years of abuse. The replacement sensor cost about 15€, I find it reasonable but I would prefer a sensor that would dry itself and eventually last much longer

I have started playing with my Arduino, next week I will get a few 5V relays, a DHT22 sensor and a clock module, lucky me there's a robot parts shop nearby selling this stuff.

The clock module uses a battery to keep the date/time. I will need this to set day / night cycles.
The 5V relays will switch off / on both AC and DC devices based on conditions set in the program.

Have you attempted to control lights and water pump with the microcontroller? Any issues?

[Draikan]

I currently only have the humidity control solenoid valve and circulation fan controllable by the microcontroller.

I'm working on a module to add connections for DC pumps (mistking), and basic input devices (float switches, pressure switches, buttons) right now. This will let my controller monitor the RO system, as well as schedule the 3x weekly irrigation.

I'll do AC devices (lights, heater, etc.) some time after that. Getting the sensor to a shareable state is priority #1 right now.

	Mistking Looking for a misting system? Look no further. Automated misting systems from MistKing are used by multitude of plant enthusiasts and are perfect for Orchids. Systems feature run dry pumps, ZipDrip valve, adjustable black nozzles, per second control! Automatically mist one growing shelf or a greenhouse full of Orchids. See MistKing testimonials

[Draikan]

Tiny update. All is growing well. I've been completely ignoring the plants to work on the next controller. All I do is fill up the water tank on the weekend.





Lepanthes telipogoniflora is finally showing signs of life! It has a single new leaf started. I think it is responding to the spring temperatures, but it would probably still like to be >10° warmer than it is now. My average temperature over the last few weeks is about 70°F.