Peltier calc help

[Mosaic]

Hi all.
I am trying to size a Peltier unit to chill 27cu inches of air in a 1/4" thick plexi box with 1" thick inlaid styrofoam insulation.
Compartment temperature target is 0°C - 10°C. Ambient up to 35C.
The actual usable compartment would be 3" x 3" x 3" .
I can do a pulsed duty cycle to manage the temp.

Any advice?

 

[ADWSystems]

Will there be air movement in the box? Changing the temperature of stagnant air is difficult, even a small circulation has a big effect (think ceiling fans).

Will the PJ be mounted on the top or the bottom of the box? Top would be best, cold air sinks. No sense in cooling cold air.

How fast do you need to cool worst case (from 35C to 0C)?

Peltier units work by creating a temperature difference between to the two sides. The cold side will be the box side, the warm side will be mounted to a BIG ol' heatsink. The heatsink will act as an 'anchor' for the temperature difference. If the heatsink temperature changes, the cold side will move (almost) in direct correlation.

You are aware that peltier junctions are power hogs. A quick search found a 40mm X 40mm 72W unit that requires 6A at 12V.

 

[Mosaic]

Will there be air movement in the box? Changing the temperature of stagnant air is difficult, even a small circulation has a big effect (think ceiling fans). PJ is at top, relying on convection in the 3x3x3" space.

Will the PJ be mounted on the top or the bottom of the box? Top would be best, cold air sinks. No sense in cooling cold air. --- YES

How fast do you need to cool worst case (from 35C to 0C)? -- No rush, 2-4 hours

Peltier units work by creating a temperature difference between to the two sides. The cold side will be the box side, the warm side will be mounted to a BIG ol' heatsink. The heatsink will act as an 'anchor' for the temperature difference. If the heatsink temperature changes, the cold side will move (almost) in direct correlation.

You are aware that peltier junctions are power hogs. A quick search found a 40mm X 40mm 72W unit that requires 6A at 12V ...That's ok, compact size and convenience matter more, here .

So 72W of heat plus whatever is in the 27 cu in. volume, assuming worst ambient is 35C.
Given perhaps a 40C Delta across the Peltier, we need to hold the hot side to no more than 50°C to hold the 10°C ceiling temp.

 

[ADWSystems]

12V 10A power supply (adding some for margin) plus a temp controller, isn't exactly small.

It's all about the heatsink. Getting the cold side cold is easy, keeping the hot side within the junctions Tdelta of the cold side is the trick. You are already thinking in that direction.

My notes on sizing a PJ were for direct contact application. The project was to cool and item in direct contact with the cool side of the PJ. Your application is slightly different. Sizing the heatsink for direct contact was based on the degC per W of the heat sink combined with the power consumed by the PJ plus the target temperature difference. Using the 72W PJ (as an example), and (if we could find) a 1/2degC per W heatsink; the heat sink would raise in temperature by 36degC due to the PJ plus the worst case difference (35C to 0C) for a total of 71degC. So as the PJ starts to cool off the temperature will start to drop and then rise as it follows the rising temperature of the heatsink. Selecting the heatsink and keep it's temperature somewhere within the Tdelta of the PJ is the trick. I glanced at Mouser and Tdelta of 60-65degC is quite common. Additionally at ambient of 35degC (which is quite warm by the way, 25degC is more typical) the ambient air will try to cool the heatsink rather than warm it.

What are you using for a temperature sensor and heater controller?

 

[Mosaic]

Was planning to use a dual thermistor and a microcontroller to do the sampling.
This forced air Hsink calc seems interesting.
https://www.farnell.com/datasheets/54432.pdf

This calc seems to do the job.
https://tetech.com/cooling-assembly-and-heat-load-calculator/
But their TE units with fan and heatsink are amazingly expensive. Their smallest, the AC-027 claims to be able to do the job down to 7°C

 

[chemelec]

You can use a Thermistor and a meter to Monitor the actual INSIDE Temperature.

Most Important is to Keep the Heatsink from getting Too Hot.
So as far as the Controller is concerned, it only needs one thermistor to monitor the Heatsink Temperature and control an On/Off Duty Cycle for the Peltier device..
This is because as the Heatsink gets Hotter, so will the inside temperature.

This Circuit worked very good for me.

 

[ADWSystems]

My notes on sizing a PJ were for direct contact application. The project was to cool and item in direct contact with the cool side of the PJ. Your application is slightly different. Sizing the heatsink for direct contact was based on the degC per W of the heat sink combined with the power consumed by the PJ plus the target temperature difference. Using the 72W PJ (as an example), and (if we could find) a 1/2degC per W heatsink; the heat sink would raise in temperature by 36degC due to the PJ plus the worst case difference (35C to 0C) for a total of 71degC. So as the PJ starts to cool off the temperature will start to drop and then rise as it follows the rising temperature of the heatsink. Selecting the heatsink and keep it's temperature somewhere within the Tdelta of the PJ is the trick. I glanced at Mouser and Tdelta of 60-65degC is quite common. Additionally at ambient of 35degC (which is quite warm by the way, 25degC is more typical) the ambient air will try to cool the heatsink rather than warm it.

I made a small mistake/omission. The heatsink won't get to 70degC. The temperature will make a few ups and down as it stabilizes. First the box will start to get cold. But the heatsink (the master temperature reference) will start to rise in temperature as the PJ heats and the heat is shifted from the box to the heatsink. As the box "runs out of heat", the temperature will again begin to drop, back to the target.

 

[Mosaic]

In terms of how much 'heat' watts a PJ moves.....if a PJ consumes 72W, what is the rate of energy (Watts) can it pump from the cold to the hot side?
EDIT:
Item 5 seems to answer this, 30% to 70%
https://tetech.com/faqs/#1

 

[ADWSystems]

That article also mentions the nest efficiency is achieved around 2/3 of Vmax and 2/3 of Imax. I did not know that. I had been targeted a convenient number less than Vmax and Imaxm ie., if Vmax=14.4 and Imax=4.6 I would target something like 12V and 4A.

 

[Mosaic]

Using this:
http://www.engineeringtoolbox.com/air-properties-d_156.html
I calc the energy required to remove 30°C from ambient air (27 cu inches) is 16J.
The TE calc for the insulation @ 1" thick indicated a <3W leakage of heat. So the energy input into the chilled volume is about 3 J /s.
I have this TEC on hand:
http://www.hebeiltd.com.cn/peltier.datasheet/TEC1-12706.pdf

The perf. graph shows I need about 4.5A @ 13.5V to get a 50C delta with a 50C hot side.
That's about 61W of power.
At a low 30% eff. the TEC is pumping about 19 W to net 80W of heat on the hotside.
19 W of cooling less 3 W of leakage leaves 16W to apply to the 16J.
That mean if air were 100% thermally conductive it would drop to 0°C in 1 second! If these calcs are right then I should have no problem cooling this small volume rapidly. It might make sense to place the TEC cold side on an alum plate to increase cooling efficiency in the insulated space.
Clearly I can duty cycle the current to perhaps a 20% square wave to permit convection to work and not freeze the peltier with condensation. That should average the heat dissipation down to 12.2W +4W of pumped heat for a net 16.2W to manage.With a max 35°C ambient (tropics) I'd need about a 1°C/W heatsink to target about a 50°-55° C hot side.


Sounds ok?