Peltier device used as a heater, trying to build temp controller. Any Suggestions?

Discussion in 'The Projects Forum' started by Redline870, Nov 19, 2015.

  1. Redline870

    Thread Starter New Member

    Nov 13, 2015
    22
    3
    I am working on building a microscope using 3D printed parts and inexpensive optics. The microscope itself is complete, looks and functions great.

    However I need a heated stage for viewing micro organism growth over extended periods of time. I decided on a peltier device sandwiched between 2 pieces of aluminum. I went with a peltier device as it has a uniform heating surface, is very compact, and because there are models with a hole in the middle. It MUST have a hole in the middle so that light can pass through the slide. The temperature on the stage MUST be between 32-34 degrees Celsius. I am working on building a small temperature controller for it which I am kind of struggling with.

    For the temperature controller I am using 2 linear regulators in series, one with fixed resistor values to supply a consistent input voltage to the second regulator. The second regulator is connected to a thermistor that is mounted on one of the pieces of aluminum. I tried building it with only one regulator but couldn't get a consistent voltage out, The voltage out changed depending on input voltage. The peltier only needs 1.2V at .39A to reach 33 degrees celsius. The regulator output must be above 1.25V or it shuts off so I decided to build it with a 2.0V output and then use a 2.2ohm 5 watt resistor to step down down the power. This is looking more and more to not be the most ideal setup.

    My question is: Does anyone have a better idea? I'm getting annoyed continuously playing with this thing and not making any head room. Has anyone else built a hot side peltier temp controller or have any ideas how to?

    Help would be VERY appreciated. Thanks
     
  2. Dodgydave

    Distinguished Member

    Jun 22, 2012
    4,988
    745
  3. wayneh

    Expert

    Sep 9, 2010
    12,126
    3,048
    You might also benefit from this project, which was about making a cooler but making a heater would be similar. It uses simple thermostat (on/off) control.
     
    Last edited: Nov 19, 2015
  4. Redline870

    Thread Starter New Member

    Nov 13, 2015
    22
    3
    Wow quick responses, thank you.

    wayneh - Thanks for the suggestion. However I don't really want it turning on and off continuously. I would like it to vary the voltage based on temperature as it has to remain precisely at 33 C. It can't fluctuate any more than 1 degree + or -. How precise of control did you have with your project?

    Dodgydave - I have very limited knowledge in electronics, I am by no means an electrical engineer. I am working with one to try and get this built but it doesn't seem to be working properly. This is the wiring diagram to what he has come up with as of now:

    [​IMG]

    Should I abandon this all together and go with what you suggested? Also what would be used to monitor the temperature, would a thermistor work in this case? Sorry for what might be stupid questions but I have very little experience with electronics. I won't be able to meet with the engineer until tomorrow and I was hoping to arm myself with a little more knowledge before I present the idea to him and tell him we are abandoning his idea.

    Lastly do you any example schematics of this being used in a heater or peltier application? I did a quick search but they seemed to be mostly motor speed controllers.

    Thank you very much to you both for your help.
     
  5. Dodgydave

    Distinguished Member

    Jun 22, 2012
    4,988
    745
    Its the same principle, motor or heater, low voltage higher current is best to go with a pwm controller for the peltier less power wasting,

    a voltage regulator like a Lm317 will give you 1.25v minimum but will waste power by getting hot,

    To monitor the temperature you're going to need a thermistor or LM35 sense circuit with an op amp, to feed into the pwm.
     
    Last edited: Nov 19, 2015
  6. #12

    Expert

    Nov 30, 2010
    16,298
    6,810
    It seems to me like you are trying to use, "dead simple" to achieve, "dead accurate".
    Here's one I did decades ago with a 10K thermistor, and it's deadly accurate (compared to a degree C). If you understand how it works, you just change one resistor to change the temperature range or to match whichever thermistor you have handy. You can use it to drive the ground pin of an LM317 in order to have sufficient current capabilities without loading the regulator chip.

    http://forum.allaboutcircuits.com/blog/lm723-as-a-thermostat.532/
     
  7. wayneh

    Expert

    Sep 9, 2010
    12,126
    3,048
    As I recall, I got control to about ±0.1°C.
    There's no way you'll drive enough current thru that 2K resistor. Or is that your 2.2Ω, 5W?
    I don't care for that circuit. You want one controller, not two in series. I'd be tempted to use a constant-current configuration, since the output of a TEC is proportional to current. So if you want to run your TEC at 390mA, design a constant-current regulator to do that.
    I prefer the LM35 like I used in that cooler project. It gives a proportional output that you don't need to calibrate except for very sensitive work. A thermistor arrangement can certainly work, and be calibrated to your target temperature.
    These are not stupid questions. There are a lot of different ways to approach this and no hard and fast rules for choosing between techniques. One thing that's important is to know what you want, in terms of numerical specifications, and to separate that knowledge from the set of techniques you might use to meet those specifications.

    One issue I see is the temperature of what needs to be 33°C? The temperature of the cells in the field of view may not be at the same temp as whatever probe you are using. I'd be tempted to use an aluminum (or even copper) block for thermal mass and then use the TEC or any of another heat sources to provide the small amount of heat you need to maintain the temperature of the block. This is easier than a "fast" controller without a time delay. It would also be prudent to estimate how much heat is needed. A 1W heater is different than a 10W heater.
     
    Last edited: Nov 19, 2015
  8. Redline870

    Thread Starter New Member

    Nov 13, 2015
    22
    3
    Thanks for all the responses. Accuracy to 0.1C is way more than enough for this project. If I can get .5C accuracy out of this I would be ecstatic. I will run the options by the engineer and see if he can make one of these work. I have access to a circuit board mill and many 3D printers as this project is taking place at a university so I'm hoping we can make this into a nice little compact unit that is fairly simple and easy to assemble.

    The only reason I am trying to keep it simple is because the end goal for this project is to put together kits to give to high schools that can't afford microscopes, and give them a sort of DIY microscope for them to assemble. They will have limited soldering and assembly skills so we don't want to overwhelm them. We are hoping it will spark interest in the high school students and will make them want to continue their education after high school. Whether it be using the microscopes for biology or chemistry, or if their more interested in the assembly/design then they can look in to engineering programs.

    Either way I need to leave the decision to the engineer I'm working with as he is much more knowledgeable than me. I will let you know what we decide on and if anyone is interested, post some pictures when it's complete.

    Thanks again for everyone's help!
     
  9. KeepItSimpleStupid

    Well-Known Member

    Mar 4, 2014
    1,146
    204
    You would be much better off using PWM for control and a constant temperature sink.
    so, you water cool one side and just set the value to the pelter device. that worked fine for me.
    A flowmeter is used for water flow. Bad part - you dump water down he drain.

    It's usually pretty good day to day.
     
  10. Redline870

    Thread Starter New Member

    Nov 13, 2015
    22
    3
    Dodgydave - Thanks for the post and looking for one for me. In the last sentence the writer states:

    "Set the temperature adjust pot for what ever temp you want and then start adjusting the Peltier voltage pot until you reach about 7 to 7-1/2 volts. Do this in reasonably small steps and let the temp stabilize a little before increasing the Peltier voltage up."

    Does this mean you need a multimeter every time you turn it on?

    KeepItSImpleStupid - The microscope is made of 3D printed parts and only stands about a foot or so high. I don't think water cooling is really an option for something this size. Also is water cooling really necessary for only 33 degrees Celsius? I want to use the hot side of the peltier to put slides and petri dishes on. Even if this was my only option that would require building some sort of reservoir to place between the aluminum plate attached to the peltier, and the sample to be viewed through the scope. I don't think this would allow me enough room to focus the optics.
     
  11. Redline870

    Thread Starter New Member

    Nov 13, 2015
    22
    3
    This is the stage I have constructed:

    [​IMG]

    The one on the right is not fully assembled and was just used to show the underside of the warming plate and where the thermistor is mounted. I would like to build a controller that could be wired to that.
     
  12. KeepItSimpleStupid

    Well-Known Member

    Mar 4, 2014
    1,146
    204
    Without a constant sink on the other side, you might have to implement PID control.

    We wanted to control at 25 C with light (heat source) shining on it. The "hole in the center" TEC's cost too much. We ended up settling for 28 C and used a fan.

    The stereoscope with the peltier cooler was pretty much commercial with a vacuum hold down.

    Cooling loops can be made with a dovetail cutter, The copper pipe is stuffed into the dovetails and hammered into place.

    I'm just saying that the method works. I did other stuff too.
     
  13. wayneh

    Expert

    Sep 9, 2010
    12,126
    3,048
    Fwiw, many pieces of lab equipment that require a steady temperature are design to be attached to a water bath. All the temperature control is handled by that commercial device and the rest is just plumbing. I believe you can buy a flow-through microscope stage.
     
  14. KeepItSimpleStupid

    Well-Known Member

    Mar 4, 2014
    1,146
    204
  15. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,290
    1,255
    Just thinking out loud. I think something like #12 talked about in post #6 might work just fine.
    If you don't want to build something you can buy it cheaper.:D
    http://www.ebay.com/itm/50-110-C-W1...523020?hash=item4ae2417ecc:g:CfUAAOSw0vBUl-WQ
    Hopefully you have a 12 volt supply. So just use the first regulator and set it to where in the coolest room the hot plate is just higher than your set temperature. Maybe 1.4 volts. 35 degrees.
    Then attach the themistor and set the temperature where you want it.
    While this is a bang bang system it should hold a few tenths of a degree.
    I think I would add a little foam between the hot & cold side.
     
  16. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,290
    1,255
  17. #12

    Expert

    Nov 30, 2010
    16,298
    6,810
    I did some modifications of the LM723 thermostat.
     
  18. Redline870

    Thread Starter New Member

    Nov 13, 2015
    22
    3
  19. Dodgydave

    Distinguished Member

    Jun 22, 2012
    4,988
    745
    Yes, try it set to constant current mode, on an heatsink, use a 12v psu, the most current you can get will be 1.5 Amp, still better with a pwm.
     
  20. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,290
    1,255
Loading...