Peltier Cascading - Does It Work??

In this video, we will test whether or not cascading 2 TEC or peltier modules together increases cooling capacity and whether it is worth doing or not. Cascading is a concept where the hot side of one peltier is cooled by the cold side of another peltier to hopefully make the first peltier perform better. Confused? Watch the video and I'll explain.

[charlesgalea]

I started my interest in peltiers only yesterday and i learned that stacking works properly using different voltages on the stack. Thank you for unknowingly proving that .

[andrewbrooks]

This is not a really good model of cascading. Generally, when you cascade, the situation involves a small yet very hot chip, & the appropriate size Peltier cannot provide enough cooling. So, the first small Peltier with a modest heatsink is applied, then a larger Peltier and heatsink attached on top of the first heatsink. Sometimes third & even 4th layers are applied. This allows greater cooling, with less risk of condensation, because any bottom area of the Peltier not in direct contact with the hot chip is ineffective, and risks condensation. Also as a way to cool when the target is crowded by other components which would prevent an oversized Peltier, even if you were able to prevent condensation.

[brightknight]

Thank you for running this experiment! Hearing that you are interested in efficiency, are you aware that we receive power from the peltier when there is a temperature difference between the 2 sides? What about keeping the cascade setup that you have now, but instead of putting in current into the top peltier, we see how much power we can extract from the top peltier? For the top peltier, the temperature of the top side is ambient temperature and the temperature of the bottom side is the top side of the bottom peltier, which should be hot since you are running current through it to cool off the heat sink. You should be able to increase the overall efficiency of the entire system. It would be interesting to know if this would work or not. I would greatly appreciate it if you could run that experiment. Thanks.

[haikiri]

The major problem I noticed in these videos is the management of the heat itself.
Heat sink and insulation are the key words.
You are basically running 72 Watts through the first TEC so if you get a radiator from a car or an Air Conditioner unit from a scrap yard you could add 20 watt from a table fan and water pump and get a lower temp with way more efficiency instead of the (9.5 x 12) 114 Watts by stacking 2 TEC.

[bertzerker]

I'd be keen to space the cascade with a sandwich plate, 'semi' heatsink (as the cascade device is not directly loaded and each share an added value of safety from a serious breach that can occur during operation. I hear that these things can stall on start up if the opposition of dissipation is too high). The values of different peltier operation characteristics can better sympathise as one configuration might be premium to lower current operation or to the needs of higher demand such as on group for body operation and another group for a freezer unit.
Then use 'the semi-stack' to improve the heat dissipation from a small group of primary stage sinks.
Basically ramp up the the top deck current while boosting the primary array, overall it could be a more linier change with the compensations made by a programmed triggering network of PWM controllers.

It will save on space and the extra cost of heatsink material. It can even eliminate the need for water cooling which could be a payoff for the ecconomics of a project for years to come. Then again the water thermo distribution could work as a payoff in itself.
I just don't want to buy a conventional fridge or putty up an ugly old styrofoam box.

You got it neat and exceptionally organised, nice to see the effort in effect!
Grats on the video ✌

[Bodragon]

(0:36) - The hot side touching the cold side is not *_"back to back"_*
If anything, *_that_* would be *_back to front_*
_(or front to back)_
>

[russellking]

I have tested a similar setup, except using water cooling - results below:

Maximum water temperature in the LQS was 35C at max load. Due to the heat distribution in the water, hot-side thermal conditions were therefore considered to be fairly static. this was monitored during the experiment.

I used water as the thermal conductor between the two peltier stages - this allowed me to know if the junction between the two where above or below freezing (the two units would be frozen together if below freezing).

Single peltier cooling allowed -18C to -22C at 12 volts.

Cascade cooling, with both units operating at 12volts in open air gave -22C to -26C at the cold side.
note the junction between the stages was above freezing

cascade cooling with the 1st stage at 12 Volts and 2nd stage at 5 Volts gave -33C to -35C at the cold side
note the junction between the stages was below freezing

The issue here is the heat dissipation at the cold side of the junction in the cascade setup. This did not allow for delta temperatures to be obtained. My next experiment will involve two 1st stage peltiers with a copper conductor to aid heat dissipation for the 2nd stage to the 1st stage cooling. This should increase thermal dissipation at the cascade junction by up to 80% (expected deviation due to unueven heat dissipation and thermal conductivity at the copper junction).

[silvenshadow]

Great video, would be great to see how it goes generating power instead of heating and cooling.

[kcconnor]

I'm also playing with Peltiers, with a fermentation refrigerator.

I've got two different cooling plates: One is a quad-set of 12706 chips arranged on 80x80mm water cooling plates, and the other is a single 12715 chip on a pair of 40x40mm plates. Both perform almost identically.

My next experiment is to cascade/sandwich a 12715 and a 12706 together. the 706 will cool the cold water circuit that chills the fridge, and the 715 will take all the heat from the 706 and pipe it to a large radiator array.

My thoughts are, this will lower the ambient starting point for the temp differential that the 706 experiences, since the "hot" side of the 706 will be slightly below ambient. Right now, my "hot" circuit is around 90 degrees F and my cold circuit is cooling the box to around 54 F, with ambient at 78 degrees.

Reason: The 706 moves about 50 watts of heat, but produces an additional 70 watts for a total of 120 watts. Stacking multiple 706's will eventually cook your stack of them. The 715 moves about 140 watts of heat, but produces another 180 watts of heat. So, cooling a 706 with a 715 will give a thermal advantage to the hot side of the 706 and lower the starting point of the deltaT spread.

[fissb]

Great that you tested the concept, the manufacturer in China Had done that.

[fuzzywzhe]

Have you tried limiting the current to each junction? I would EXPECT, slowing down the cooling rate but keeping the power consumption constant should increase efficiency.

[quantumlab]

I have gotten a single peltier module down to about -40 degrees Celsius. I used a heat sink submerged in a water bath full of ice. The water bath was around 5-9 degrees Celsius. Also I got a cascade peltier system like yours down to about -15 Celsius. Pretty cool (pun intended).

[ChrisDIYerOklahoma]

Very cool...learned a lot. See you are using the Uni-T clamp meter...thanks again for the review. Mine is awesome and use it quite a bit!

[bransen]

Great content

Batterie Temperature Management

[markgray]

Cascade/series yeah. It takes two modules to carry the heat of one. Place two on large hotside sink, cut 1/4 inch or better plate of aluminium to match size of the two, place one on top of that, place 1/2 inch or better riser cut to size of one, place cold sink on riser, use only two screws/clamps to sqeeze together, insulate with foam. Run peltier devices electricaly in series. Nice 50 watt cooling using 12v 150w devices. Fans Yeah.

[abrahamf]

If you want to cascade the first stage should be more tecs covering a bigger area. That should make the cascading effect more noticeable

[nilko]

Górny peltier powinien być zasilany 2-3X mniejszą mocą niż dolny (dolny na radiatorze).

[coltoncarew]

Is it best to wire the Peltier units in series? Parallel? Or Combination?

[EricPham-grpg]

I believe if you separate two cell into two different temperature region in the house like the roof and floor then connect them together like a loop with positive goes into positive of load like fan or light or air-conditioner then go to the other side of cell on cool foundation then when sun raise it drive referigerator and charge battery but it need more cells in parale and series together to make it power refrrigerator and microwave and stove

[jayc]

You're saying it proper!
Pel-Tee-Yay