filterless Peltier PWM

[star882]

Is it OK to directly PWM a Peltier at 20kHz or higher? It's for a portable refrigerator/freezer and I can greatly simplify the circuit if I eliminate the filter in the buck converter.

[BWRX]

What kind of a load does a peltier junction (thermoelectric cooler) look like to a power supply? Mostly resistive? It should be able to handle a PWM signal but you might not get the same level of cooling as you would with a DC voltage.

[jcx]

no - in fact the filtering requirement may need to be much larger than you would like to get good heat pumping efficiency

any ac content will cause loss/heating in the Peltier element severely degrading its cooling performance, they really want very low ripple DC, don't skip the filter

[djQUAN]

I have read somewhere that the PWM for a peltier should not be more than 1kHz.

and if you go too low, there will be too much repeating thermal stress on the device which could cause failure.

[Hartono]

IIRC, the point in using PWM to control Peltier device is that, it doesn't like low value DC.

[luka]

Hi

But peltiers run on DC votage how will be destrayed if you go to low with freq.???

[Hartono]

Well I just remember reading it somewhere before,I do a search and found this :

It seems that not the DC itself that cause damage, but the cycling. and this only happen in temperature controlled operation, when the device must be turned ON / OFF

The problem in using DC to control peltier temperature is the thermal cycling, which lower the life expectancy of the device. DC control will do ON and OFF cycle causing thermal cycling.

PWM is also pulses of DC but the ON/ OFF cycle timing is controlled to minimize thermal cycling. the key is in the timing

http://www.tetech.com/techinfo/


Hartono.

[djQUAN]

Originally posted by luka
Hi

But peltiers run on DC votage how will be destrayed if you go to low with freq.???

turning it on and off causes thermal stress. repeated on and off causes thermal stress again and again so that would eventually break the junctions.

[luka]

Hi

Yes I agree with you on that part, didn't get it before

[star882]

Originally posted by jcx
no - in fact the filtering requirement may need to be much larger than you would like to get good heat pumping efficiency

any ac content will cause loss/heating in the Peltier element severely degrading its cooling performance, they really want very low ripple DC, don't skip the filter

http://www.maxim-ic.com/appnotes.cfm/an_pk/1757
The filter in the circuit is very simple.

I think I'll try powering a Peltier from a 3.3v power supply and from a 12v power supply PWMed at high frequency to have an effective voltage of 3.3v and compare. Then I'll add a simple filter and see what happens.