Peltier Thermal Cycler

Forum 2005-2010 (read only) Hardware Interfacing
1

Hi all,
I'm planning to make a Thermal Cycler using a Peltier unit or two, an LM35 temperature sensor, and an arduino. Because I can't link on my first post, a Thermal Cycler is just something that heats and cools a lab sample to a set of very specific temperatures in a predetermined pattern.

Having fought with a thermistor for ages and realising there were easier ICs to do the job (LM35), my next problem is the peltier unit, which requires more than the 5V, xmA output that the arduino provides.

I've ordered some Solid State Relays, but just found out that AC relays cannot necessarily work with DC. Damn. As you can see, I'm relatively new to Electronics, and I'm trying to keep this project simple and idiot-proof.

So my question is, any suggestions on how to drive a peltier unit (~50w) with arduino, simply? Relays seem to be the best option, but a lot of the high current/voltage options require too much switching voltage for the arduino..

Thanks!

2

A FET solid state relay will handle DC so you could use that.
A mechanical relay can't be driven direct from the arduino you need a transistor to drive it.
http://www.thebox.myzen.co.uk/Workshop/Motors_1.html

A peltier device is just like a motor from the point of view of driving it, current one way to cool the other way to heat. So you can use a motor H bridge driver, many of which you can buy ready built.
Anyway that gives you some words to search for.

3

Thanks for the quick answer!
I'd been considering a motor controller chip, but I'm not familiar enough to know I wouldn't burn it out right away with a too-high current or whatever. In addition, I gather PWM doesn't suit peltier units. Do you know any chips that suit these criteria well enough?

4

Do you know any chips that suit these criteria well enough?

Any decent motor controller should cope with DC current as well as PWM.

In the end, it all comes down to PWM anyway; it's just a matter of frequency :wink:

5

I gather PWM doesn't suit peltier units.

I haven't herd that, I can't see straight off why that should be.

I wouldn't burn it out right away with a too-high current

Just make sure that the motor driver can handle the current of your peltier device and add a bit extra for margin, say 10 to 20%.

6

Alrighty, back to the Maplin Catalogue!
Thanks for the feedback. Might have more questions when I see the datasheets..

7

Oh, quick question regarding PWM:
If indeed peltiers don't like PWM, is it possible to smooth out the signal with a capacitor to make it more closely resemble set-voltage DC? Like an AC->DC circuit does?

Any recommendations on what sort of cap to use?

8

Yes it is possible to do this BUT if you need to apply this to a large current then it is difficult to control it because the driving device ends up dissipating all the unwanted power as heat. That means it has to be able to handle a very large amount of power.

9

I know this thread is really old, but I'm planning on doing the same project for PCR and was hoping you could post something about how the build went :slight_smile: