1
\$\begingroup\$

I am thinking about using peltier for both heating and cooling to maintain desired temperature profile in small area (2ft square at max)

Can continuously changing polarity damage peltier module and more importantly, will it be as efficient in reverse polarity?

Note that it will only be turning on once a minute for a few seconds to maintain 38°C

What side should I put/face inside the enclosure, forward or reverse?

Ambient weather swings alot from 30 to 43 so heating and cooling cycles can both be considered 50% of the time because enclosure will have the ability to hold the temperature as it will be sealed.

\$\endgroup\$
1
  • \$\begingroup\$ No problem. I've used peltiers for years to temperature control stages. I Drive them with an H bridge at 10-20kHz, so they can heat and cool. It is important to LC filter this and not use PWM directly as mentioned by Jonathan S. \$\endgroup\$
    – tobalt
    May 28 '21 at 17:10
2
\$\begingroup\$

Can continuously changing polarity damage peltier module and more importantly, will it be as efficient in reverse polarity?

Yes it can, I've burned a few out this way. You can kind of think of a peliter as a car, you need to slow it down or stop it before you go into reverse.

The problem is you've built up a thermal gradient when you pump current into the peltier, which in turn builds up a voltage gradient. If you throw the peltier into 'reverse' then there will be a large voltage gradient which can source a lot of current, when the circuit is reverse this can lead to a large amount of current being sourced until the thermal gradient across the peltier starts to reverse (a large thermal sink can take some time for everything to fully reverse).

This large amount of current has two effects:

  1. if the driving circuit is an h-bridge it can exceed the current rating for the mosfets (relays not so much but this leads to problem 2)
  2. the sudden change in thermal gradient can cause mechanical failures of the peliter

As far as efficiency goes, cooling will always be less efficient than heating because of the ohmic internal heating of the peliter. The material of each peliter also has resistance, this resistance causes heating related to the amount of current through the peliter (P=I^2*R). This internal heating means that heating something will always be more efficient than cooling it.

The cold and hot side markings (or in the datasheet) indicate whether it will cool or heat relative to the voltage placed on the terminals (kind of like polarity marks on a battery) as long as you keep track of the polarity, orientation does not matter for most applications.

What can be done about this?

If your using a control circuit (like a PID) a slew rate limiter can be used to protect very fast transitions. (I've ran this successfully in a number of designs and it protects against fully positive (or very high) on to fully negative, usually with 10's of seconds on the slew rate)

You can introduce some dead time logic to give some time for the switch from positive to negative.

\$\endgroup\$
6
  • \$\begingroup\$ Thankyou sir, As my application only requires maintaining temperature, if ambient temp is below 38 (<=37) it will only be heating occasionally and cooling if >=39 and i will be using relays for simplicity but yes temp sensors (have multiple) can go crazy sometimes \$\endgroup\$
    – asim
    May 28 '21 at 17:40
  • \$\begingroup\$ If you are only occasionally cooling then just introduce some dead time or lower the current before reversing polarity \$\endgroup\$
    – Voltage Spike
    May 28 '21 at 17:58
  • \$\begingroup\$ Anything i can do to limit current using arduink? I know i can read voltage across a current sense resistor but how can i limit current without a building a proper buck regalator? Will simply switching n channel mosfet via pwm and using capacitor do the trick? \$\endgroup\$
    – asim
    May 28 '21 at 19:29
  • \$\begingroup\$ Thats a tough one, a buck would work or some kind of switcher that supports current limiting works. You can't use PWM to limit current directly because PWM switches voltage and your either at the rail or your not. You could monitor the current with a sense resistor and use that to change the PWM. It really depends on the current and power you need, lower than an amp and a constant current circuit with a DAC would work. \$\endgroup\$
    – Voltage Spike
    May 28 '21 at 19:33
  • \$\begingroup\$ Like using the method i used here: electronics.stackexchange.com/questions/551663/… @VoltageSpike \$\endgroup\$
    – asim
    May 28 '21 at 19:38
1
\$\begingroup\$

You can do this, the Peltier will be just fine and the orientation doesn't really matter (efficiency will be the same). Just keep in mind that Peltier elements don't like being switched on and off a lot, but as long as you only switch it once per minute and don't use PWM to control it, it should be fine. Changing polarity once a minute won't hurt the Peltier either. Also, don't underestimate the amount of heatsinking your Peltier device will need.

\$\endgroup\$
2
  • \$\begingroup\$ electronics.stackexchange.com/questions/124211/… says using it in reverse will yield bad efficiency , are you it will be the same? Datasheet doesn't mention it \$\endgroup\$
    – asim
    May 28 '21 at 15:46
  • \$\begingroup\$ That answer says that using it flipped (but still for cooling) will yield worse efficiency which is not the same thing. Efficiency for heating and cooling should be fairly similar. \$\endgroup\$ May 28 '21 at 16:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.