Running peltier thermoelectric cooler at low 'standby' voltage

Thread Starter

wazman

Joined Oct 31, 2016
13
Hello, I have built a small insulated thermoelectric cooler that I am currently bench-testing. The hot side of the peltier assembly is cooled by a heatsink and fan. I am well acquainted with the limitations of peltiers vs compressor fridges - the application for this unit requires it to be very lightweight and simple.

The cooler is intended to be connected to a deep-cycle lithium battery and to run "flat out" during the daytime while the battery can be charged from a solar or other source. It will get as cold as possible, then, at night, run for as long as possible until the lithium deep-cycle battery cuts off - I think that's around 10 volts with the battery I am testing (I am experimenting with various different plate types and voltages to balance temperature drop achieved, power consumption etc). The unit would then rely on its internal "coldth" and insulation to keep the contents reasonably cool until power is restored. It is accepted that the internal temperature will rise over time.

I have noticed that when I disconnect the cooler, the 'cold side' rises in temperature more quickly than I had expected. I attribute this partly to residual heat from the thermoelectric effect 'soaking' back through from the hot side to the cold, once the Peltier effect ceases and the heatsink is no longer being cooled by the fan.

But I can also see that the heatsink and the peltier assembly, when passive, provide a route for ambient heat to be absorbed and transferred to the interior of the cooler.

My question is: as the battery gets low, rather than allowing the unit to switch off completely, should I switch the Peltier plate to the lowest possible voltage, to maintain a net flow of heat from the cold side to the hot, for as long as possible?

It is accepted the unit will still heat up, but the intention would be to prevent inflow/"push back" against ambient heat coming in via the heatsink, peltier plate and spacer/extender block assembly. A bit like maintaining positive pressure in a room to exclude outside air.

Many thanks for any advice.
 

Kjeldgaard

Joined Apr 7, 2016
379
Some years ago, I wrote software for a small peltier cooling / heating system, also with a fan.

At shutdown, it was possible to ramp down the power to the peltier element before the blower was also stopped.

I can't remember the details, but it was something like:
  • Reduce the power to 80% in one step.
  • Ramp down to peltier off in 20 minutes.
  • Let the blower run for another 10 minutes.
 

Thread Starter

wazman

Joined Oct 31, 2016
13
Thanks, I appreciate the idea of running the fan on to remove the residual heat, which is something I had already considered.

However I am not only concerned with removing the residual heat, but, once that residual heat is removed, preventing ambient heat from soaking back through the heatsink, which seems to be a direct route. Or, once the power is off and the residual heat removed, does the Peltier in its non-powered state posses sufficient insulating properties to form a 'thermal break' and prevent ambient heat soaking through?
 

Yaakov

Joined Jan 27, 2019
1,601
Thanks, I appreciate the idea of running the fan on to remove the residual heat, which is something I had already considered.

However I am not only concerned with removing the residual heat, but, once that residual heat is removed, preventing ambient heat from soaking back through the heatsink, which seems to be a direct route. Or, once the power is off and the residual heat removed, does the Peltier in its non-powered state posses sufficient insulating properties to form a 'thermal break' and prevent ambient heat soaking through?
The Peltier device, for its expected operation, should be an excellent path for thermal energy. IF you shut down the full power operation at some specific battery voltage before depletion, you might be able to use an ambient temperature sensor to operate it in such a way as to null out the detected ambient heat extending operation.

TANSTAAFL, however, so this is likely to be just an incremental optimization, and may not even provide that. I think you’d have to model it, or test empirically, or both. However, the intuition is that you could get a slower decay of the cold temperature inside this way.

Alternatively, and I am not sure this is practical at all, you could mechanically disengage the cooling assembly from the compartment and insert an insulator in its place using solenoids, servos, and seem mechanical handwaving which you can imagine from this vague idea.
 

Yaakov

Joined Jan 27, 2019
1,601
Imagining the mechanical system I proposed above, I think it would be far simpler, if done at all, to use the ramp down method of dissipating the heat of the hot side as described above, then mechanically placing an insulating enclosure over the Peltier and fan assembly to thermally isolate it from the ambient environment.

I imagine something like: https://www.amazon.com/Products-Protector-Styrofoam-Plastic-Outside/dp/B008E5C6TG/

Though of course appropriately sized.
 

oz93666

Joined Sep 7, 2010
672
Yes .. that thermal leakage through the turned off peltier is a big issue ....

One solution could be to build a small insulated box around the outside of the hot side of the peltier ... During operation the fan blows outside air down into this box , which swirls around the heatsink and exits through another hole also in the top of the box ....

When power is off the hot side will become cold and cool the air in the box around the heat sink , but this cold air being denser won't be able to escape ... it will envelope the heatsink and greatly reduce thermal leakage.
 

Yaakov

Joined Jan 27, 2019
1,601
Imagining the mechanical system I proposed above, I think it would be far simpler, if done at all, to use the ramp down method of dissipating the heat of the hot side as described above, then mechanically placing an insulating enclosure over the Peltier and fan assembly to thermally isolate it from the ambient environment.

I imagine something like: https://www.amazon.com/Products-Protector-Styrofoam-Plastic-Outside/dp/B008E5C6TG/

Though of course appropriately sized.
Thinking even more, a door in such an insulating assembly could be far more easily operated than moving the whole thing.
 

Thread Starter

wazman

Joined Oct 31, 2016
13
These are excellent thoughts. I did consider mechanical separation - not feasible, there are thermal pastes etc in between the components.

Enclosing the heatsink and keeping it in "still air" when the system is turned off has merit as an idea. It might be possible to accomplish using leaf-shutters on the intake an exhaust. But I am not sure the application is critical enough to warrant going to this extent.

So I suppose really my question comes back to: could the unit 'get through the night' by applying a trickle of current to the peltier, and the fan running, just enough to keep the flow of heat travelling outwards, and prevent ambient heat penetrating inwards? I appreciate too the advice about ramping-down voltage; what I'm wondering is whether a constantly applied low voltage, maintained for as long as the battery can sustain it, would act as a form of electronic "insulation" against ambient heat penetrating the unit. Any specific thoughts on this point?
 

MrChips

Joined Oct 2, 2009
19,383
I don't understand the problem.
The hot side will never be at a lower temperature than the ambient temperature.
Your solution is to keep the fan running as long as the hot side is above ambient temperature.
 

BobaMosfet

Joined Jul 1, 2009
834
@wazman
I think you're missing opportunities here. It's a thermocouple. Which means you don't have to run it all the time. Run it enough to get it to max cool (whatever value you've determined that is for your module), and then turn it off and then watch the output of the peltier module in terms of current and voltage. You can then measure that and use that as an indicator to temp (or rather that temp is where you want it or not). If the value drops too low, then turn the module on again. This way you aren't running all the time (conserves your power).

The instant you turn the module off, it's going to begin transferring energy from the hot-side to the cool. That's what a thermocouple does. But using the above method, you can prevent too much from getting transferred. By running a fan to take heat off the hot side, you slow the whole transfer process back into the cool area down, extending the length of time the module doesn't have to run.

Hopefully, using this methodology, you can simply extend your run time much farther than before, while achieving enough cooling. If you monitor your voltage and current level on your 'power-supply', or using a joule thief even to extend what energy you have to work with, you might exceed your expectations.
 

Thread Starter

wazman

Joined Oct 31, 2016
13
I don't understand the problem.
The hot side will never be at a lower temperature than the ambient temperature.
Your solution is to keep the fan running as long as the hot side is above ambient temperature.
Partly true but not the full picture. The peltier is a heat pump, so when active, heat will always be moved away from the cold side towards the hot.

When the peltier is switched off, yes, you are correct, the hot side can be cooled to ambient temperature by keeping the fan running.

However once the peltier reaches ambient temp, that ambient heat will conduct through the peltier assembly, and reach the colder inside of the cooler unit, thereby warming the interior of the unit and its contents. The peltier assembly is a direct "tunnel" if you like between the interior of the unit and the outside, whereas the rest of the unit is insulated.

To stop this inward flow of ambient heat, I propose continuing to run the plate at a low voltage, just sufficient to keep the heat-pump effect working slightly. Enough that an inward flow of ambient heat via the peltier assembly is not possible. The unit will still warm up eventually of course, but hopefully more slowly than it would if the peltier assembly, in its passive state, were conducting ambient heat inward.
 

Thread Starter

wazman

Joined Oct 31, 2016
13
Thanks for the suggestions and the thought you've put into them. I have several concerns about switching the peltier on/off. When it goes off, heat immediately begins soaking through from the hot side to the cold - although the extender block (example here) would act as a cold thermal mass and absorb some of this heat. The heatsink and fan continue working to draw the heat out as well, of course. But once the assembly levels out to ambient, the fan is effectively pumping ambient heat back into the heatsink and the peltier assembly - warming things up.

I suspect this can all be balanced, in the way that you have suggested, but it might prove more complicated than just having the unit ticking over to at least prevent the heatsink and peltier assembly acting as a thermal bridge, until power supply is restored and the unit can go back up to full power. Think I need to do some more testing to see how low I can run my peltier plate and still get a net movement of heat in the 'outward' direction.

Thanks for your thoughts



@wazman
I think you're missing opportunities here. It's a thermocouple. Which means you don't have to run it all the time. Run it enough to get it to max cool (whatever value you've determined that is for your module), and then turn it off and then watch the output of the peltier module in terms of current and voltage. You can then measure that and use that as an indicator to temp (or rather that temp is where you want it or not). If the value drops too low, then turn the module on again. This way you aren't running all the time (conserves your power).

The instant you turn the module off, it's going to begin transferring energy from the hot-side to the cool. That's what a thermocouple does. But using the above method, you can prevent too much from getting transferred. By running a fan to take heat off the hot side, you slow the whole transfer process back into the cool area down, extending the length of time the module doesn't have to run.

Hopefully, using this methodology, you can simply extend your run time much farther than before, while achieving enough cooling. If you monitor your voltage and current level on your 'power-supply', or using a joule thief even to extend what energy you have to work with, you might exceed your expectations.
 

MrChips

Joined Oct 2, 2009
19,383
Same difference. I still don't see what is the problem.

Run the peltier with x-current to bring the cold side to y-degrees below ambient temperature while the cooling fan is running.
All you have to do is choose x and y that makes you happy.

Why don't you run an experiment and plot y vs x and perhaps you will find your sweet spot?
 

Thread Starter

wazman

Joined Oct 31, 2016
13
Yes, I think we are in agreement after all, and experimentation is the next step. I can run the plate in situ at different voltages, see how much current it is drawing, and use my non-contact thermometer to measure the temperature drop below ambient.

I also have a little Inkbird temperature logging unit that samples temp data at preset intervals, logs it, and and transfers it to my phone via Bluetooth. I can try a hard switch-off vs continuing to run the peltier at low voltage/current, and produce graphs to show whether keeping the peltier running produces a worthwhile delay in the interior of the unit warming up.

Thanks very much

Same difference. I still don't see what is the problem.

Run the peltier with x-current to bring the cold side to y-degrees below ambient temperature while the cooling fan is running.
All you have to do is choose x and y that makes you happy.

Why don't you run an experiment and plot y vs x and perhaps you will find your sweet spot?
 

oz93666

Joined Sep 7, 2010
672
Same difference. I still don't see what is the problem.
TS plans to feed no electricity at all to this unit for most of the night ....Is relying on good thermal insulation to keep things cool inside ... The peltier will give an easy path for all the cold to escape.
 

MrChips

Joined Oct 2, 2009
19,383
TS plans to feed no electricity at all to this unit for most of the night ....Is relying on good thermal insulation to keep things cool inside ... The peltier will give an easy path for all the cold to escape.
Cold cannot escape.
You need to stop the heat from getting in.
Solution: pack some ice on the hot side,
or cool it down with another peltier.:rolleyes::D
 

wayneh

Joined Sep 9, 2010
16,114
Yes, I think we are in agreement after all, and experimentation is the next step. I can run the plate in situ at different voltages, see how much current it is drawing...
You shouldn’t have to reinvent the wheel. The data sheet for the TEC should give you the current versus voltage and the delta-T versus current information. For the current, a TEC resembles a resistor as long as you’re above about 1.5V.

The number I’d be looking for is the heat transfer coefficient of your box. Measuring that will allow you to relate the cooling load to your battery capacity.
 

Yaakov

Joined Jan 27, 2019
1,601
@MrChips comment got me thinking. Do you have enough capacity to freeze water during the day when you have maximum current available? This would increase efficiency considerably.
 

Thread Starter

wazman

Joined Oct 31, 2016
13
Not really, and ice just adds dead weight, and takes up space, unless you end up drinking it of course

@MrChips comment got me thinking. Do you have enough capacity to freeze water during the day when you have maximum current available? This would increase efficiency considerably.
 
Top