Small temperature chamber design

Hello!

I´m quite new to Arduino but have had some success building a data-logger. Now I want to build a temperature control chamber to perform entomological experiments inside. Humidity control is not important as the insect chamber inside the temperature control chamber will have separate air circulation. So I will not call it a climate chamber. It should have temperature logging ability and maybe an lcd screen to display set-point temperature and actual chamber temperature. Set-point temperature adjustment button or knob would be really cool.

The chamber inside dimensions should be about 20 x 20 x 12 cm. The the temperature control range should be +5 *C to +40 *C if lower temperature can be achieved then it`s a bonus. +/- 1 *C accuracy over 1 minute and longer period. Ripple could be a bit higher but no more than 3 *C peak-to-peak. Avoiding ripple could be a challenge, as the chamber is small.

I try to keep it simple and low cost. My first idea is to use old computer PSU to provide low voltage DC; an TEC1-12706 thermoelectric element (for low cost and low noise), which should have about 60W of cooling power; an old AMD CPU active copper core aluminium cooler for outside heat exchange; copper heat plate, tubing and aluminium radiators for inside heat exchange; a smaller fan to mix air inside the chamber; plywood and styrofoam as chamber walls and insulation; Arduino UNO as cnotroller; relay module with two 10A relays to switch on and reverse polarity of TEC1-12706; SD card and RTC modules to log temperatures with timestamp; DALLAS DS18B20 sensors for temperature monitoring (one for ambient, and two for chamber); 16x2 LCD screen.

As high power current control electronics seems expensive, I plan just to use the 12V rail on PSU for TEC1-12706 and vary the ON and OFF times of the element by switching the relays. The DALLAS DS18B20 has accuracy of 0.5 *C so I might need to average values of two or more sensors to get a more accurate result. I have 5 of them to use.

I ask for design advice and useful comments. Maybe someone has experience with similar builds or climate chambers?

The ambient temperature is going to be 30 *C maximum so 25 to 30 *C max. temperature differential should be manageable.?

How to solve the temperature set-point input? Button or knob adjustement would be much more convenient than having to hook UNO to my PC in order to adjust it.

Also I have no clue about the temperature control algorithm. The easiest way, I guess, would be to run some sort of if statement to compare set-point temperature and chamber temperature and switch the TEC1-12706 on or off when difference exceeds some value. But my worry is that this creates too much ripple, as heat transfer from the element to the chamber takes time. Also the TEC1-12706 performance could be influenced by ambient temperature as the cooling by the CPU heatsink gets better when ambient temperature is lower. But this is just a theory at the moment and needs to be tested first.

If anybody had thoughts or suggestions then I would appreciate it.

Thanks!

How to solve the temperature set-point input? Button or knob adjustement would be much more convenient than having to hook UNO to my PC in order to adjust it.

That's the simple part. Connect a potentiometer to 5V and ground, with the center tap going to an analog pin. Read the value. Map the value read from the range 0 to 1023 to the range minSetTempDesired to maxSetTempDesired, resulting in a set temp.

That sounds simple indeed. Is there a limit to the potentiometer value? Low resistance would draw too much power. The higher the resistance the better?

10K pots tend to be the best choice. With higher resistance you can start running into problems with being unable to get stable analogRead() values because the ADC in the Arduino likes to see enough current in order to run properly.

Put a 10K resistor between 5V and the potentiometer and you're never using more than .5ma.

You probably should prepare to have also a humidity sensor as active heating and cooling can disrupt humidity which may have unwanted side effects.

Could you elaborate on the:

robtillaart: "unwanted side effects"

? The actual insect chamber inside the temperature control chamber will have separate air circulation as it is connected to a infrared gas analyzer (I measure CO2 produced by insects).

Would the temperature chamber moisture somehow influence the temperature control and should it be somehow included in the control algorithm?

If its about condensation build up then I cant see problems at the moment. I can just cover the inside of chamber walls and floor with thin plastic sheet.

temperature and humidity are separate and independent as far as temperature control is concerned. higher humidity means more mass and water vapor is larger than air so it will take more time to heat and cool. but generally speaking it is close to insignificant.

understand that the ONLY POSSIBLE place for normal temperature loss is at the skin or envelope of the structure. the animals inside, if they are warm blooded, will radiate heat, if they are not, then there is minimal heat gain inside the structure.

if you create one chamber, such as the lab room, and keep that within a normal 4 degrees that building temperature control offers, then your chamber skin will only change by that same amount. add insulation and you will have even less of an effect.

as for the power needed, the major condition is to maintain temperature. will almost no loss to the room, or a constant loss, all you need to provide is enough to compensate for that loss.

a human residence operates on unbelievably sloppy control. it first over heats, then allows the space to under heat and the sea-saws back and forth around the set-point often a range of */-2 degrees or a 4 degree spread. the reason is that for residential control the heating or cooling source is not variable.

in your case, you could have multiple heaters that are all undersized and bring one on at a time. with an elevated space temp, and normal heat losses, a space temperature drop of 1 degree for 30 minutes could be negated to 1 degree every 60 minutes with one heater. the second could bring that up so that it only has to come on at times to keep the temperature within a 1 degree space temp.

air flow or ventilation air would need to be tightly controlled as that is the variable. if your incoming air were 2 degrees colder, it would cool off the space quickly.

in selecting a heater, consider thermal mass. an exposed wire coil has little or no thermal mass. encapsulate that in some concrete and you have created a thermal mass that would radiate heat with no applied power. your sensor would need to be on the skin or embedded in that and the heater would be able to run to keep that mass near a constant temperature. the mass will have very long temperature changes. look at a pot on the stove. the water takes forever to get hot while the gas flame is 1,900° !

one wonderful opportunity you have is to measure the input watts of your heater. instead of only relying on the feedback of space temp, you could also monitor input watts. knowing you need more when the temperature if far from the desired level and less when it is near that setting, you can deliver watts in a very controlled manner.

/* ramble mode off */

naftaliin:
Could you elaborate on the:

robtillaart:
“unwanted side effects”

?

dying insects

naftaliin: Could you elaborate on the:

robtillaart: "unwanted side effects"

Would the temperature chamber moisture somehow influence the temperature control and should it be somehow included in the control algorithm?

humidity is greatly effected by temperature. if you have a near constant make-up air temperature, you could humidify that air stream. a water bowl in the space might serve to maintain an acceptable range of humidity. larger bowl, more surface area, more evaporation.

humidity would need to be controlled as a separate function with separate control devices. water level in a funnel or 'beach' in the space could offer variable surface area. there are lots of ways to control humidity in a duct.

It seems that I have not been clear enough before.
Insects will NOT come in contact with the air in temperature chamber. They will be in a seperate smaller chamber inside the temperature chamber where humidity and air flow is being controlled allready (see drawing). I just need a device that can fit the smaller insect chamber and allow to perform experiments in different temperatures.

chamber.jpg

Has been a while, but I have an update on my temperature control chamber. I decided to use two cpu heat-sinks (one on either side of cooling module). I think it is going to be quite close in terms of available pin count on my UNO. Running an LCD, SDcard, two relays, RTC module, 4 temp sensors and one potentiometer uses up a lot of pins. I also wanted to add a push-button to store the mapped temperature that is being chosen with the 10k potentiometer. This would be handy because the potentiometer output value might not be stable over many hours or somebody might touch it accidentally.
Only have pins 0 and 1 left and would not want to sacrifice serial communication ability.

Would it be possible to somehow send the store temperature command through the same analog pin that the potentiometer is connected to? Maybe having the push-button short the analog pin to ground and using some logical test to recognize this action differently from normal potentiometer operation?

Hi,

Can you please post a copy of your circuit, in CAD or a picture of a hand drawn circuit in jpg, png or pdf? I can't tell from here , but are you using all the analog pins, as they can be configured as digital inputs if need.

Tom.......... :)

I can tell you for a fact, you will destroy the TEC1-12706 thermoelectric element if you switch immediately from cool to hot and back again with out about a 5 minute pause between switching. Been there and done that to see if was possible. It was not possible. With a 5 minute pause between the change direction, the unit has been running fine for almost a month, now.

Paul

TomGeorge:
Hi,

Can you please post a copy of your circuit, in CAD or a picture of a hand drawn circuit in jpg, png or pdf?
I can’t tell from here , but are you using all the analog pins, as they can be configured as digital inputs if need.

Tom… :slight_smile:

Well I have a table (attachment) with the pins used. I use analog pins for relays and possibly for button input too. Just seem to be one pin short. Would like to have one button to change display information.

Paul_KD7HB:
I can tell you for a fact, you will destroy the TEC1-12706 thermoelectric element if you switch immediately from cool to hot and back again with out about a 5 minute pause between switching. Been there and done that to see if was possible. It was not possible. With a 5 minute pause between the change direction, the unit has been running fine for almost a month, now.

Paul

Hm… Why would it destroy the element?
The only thing I can imagine that would destroy it would be excessive heat, that would melt the solder inside. I would have thought that if the thermal masses of both of my heat-sinks are similar, then the total heat-energy would not change much when switched as heat from the hot side would be rapidly sinked by the cold side. But I guess it is possible that there might be just too much heat for the TEC to pump through when one heat-sink is really hot.

I will have both heat-sinks equipped with DS18B20 sensors that measure up to 125 *C and it should not be hard to switch the TEC automatically off if certain temperature gets exceeded by either sensor.

Might actually use the NANO board cause it has 2 extra pins.

" But I guess it is possible that there might be just too much heat for the TEC to pump through when one heat-sink is really hot."

The heat problem is in the actual metallic parts inside the device. It takes only a single one of the 127 points to overheat and open up to kill the device. They are all in series.

Google "Peltier device" and look for the various ways they can fail.

The other way, and I found it also, is for the device to not be completely sealed so air cannot get inside. One of mine was that way. One morning I found a small puddle of water under the device. I thought it got so cold it made ice(supposed to be impossible). Then it quit working. The water formed on the actual bimetalic components inside and destroyed one. These were cheap Chinese devices I bought on Ebay.

The latest bunch of devices got an extra covering of RTV around the edges and the wire leads. They are working just fine.

I developed my project with a UNO and then changed to a NANO without any pins because I needed a very small controller. I hot screwed the relay board with two relays to one side of a perf board and hot glued the NANO to the other side of the perf board. I soldered wires for the relay board and three LEDs directly to the NANO. Green LED = cold, yellow LED= pause, red LED = hot.

Paul

I too use those cheap eBay modules. In fact, the first 12706 modules I ordered turned out to be fakes. They were marked as 12706 but would not draw more than 3.2A @ 15V. So I suspect they were actually 12703 modules with fake markings.

Then I ordered 12715 version and this time it looks like the right one. Draws about 10A @ 12V. My 12V rail on ATX supply drops to about 11.6V when TEC 12715 is running so I think the 10A relay should just be able to handle it.

Thanks for the warning Paul! I will put some extra superglue or something around the wire leads of the element just in case, to stop water vapor from getting in. Looks like one lead is not properly covered by silicone and I can see the edge of the wire insulation. The silicone around the edges seems intact.