# My machine uses 4 peltiers 12V 10 A so how many volts my source should use to get machine max power

#### Long292

Joined Nov 18, 2018
16
This is the first time I post on forum so I hope my matter can be solved by all of you. Thanks

#### AlbertHall

Joined Jun 4, 2014
12,019
It depends on how the peltier elements are connected.
They could be connected in parallel in which case you would need 12V at 40A
They could be connected in series in which case you would need 48V at 10A
They could be connected in a combination of series and parallel in which case you would need 24V at 20A.

#### Long292

Joined Nov 18, 2018
16
I have read on the internet if the peltiers are use in one time we must connect series but I don't know is it true??? So I still don't find out the result for this source

#### ArakelTheDragon

Joined Nov 18, 2016
1,353
I have read on the internet if the peltiers are use in one time we must connect series but I don't know is it true??? So I still don't find out the result for this source
Have you already connected them or are you connecting them now? Can you give us pictures of the connections?

#### dl324

Joined Mar 30, 2015
14,457
I have read on the internet if the peltiers are use in one time we must connect series but I don't know is it true??? So I still don't find out the result for this source
Can you cite a source?

Peltier devices allow a range of input voltages. One would think you'd get better performance if you connected them in parallel.

Last edited:

#### Long292

Joined Nov 18, 2018
16
Can you cite a source?

Peltier devices allow a range of input voltages. I'd think you'd get better performance if you connected them in parallel.
I need it uses electric and makes water cold or hot so if I connect it in parallel it won't active as I want . Is this right???

#### Long292

Joined Nov 18, 2018
16
Have you already connected them or are you connecting them now? Can you give us pictures of the connections?
I still not but I intend to connect it series

#### recklessrog

Joined May 23, 2013
985
https://tetech.com/faqs/
Have a look at their FAQs and other companies who manufacture Peltier devices. There is a lot more to know than just how much current to pass through them.
I was involved in making a cooler for a ccd imaging sensor for an astronomical camera. We ended up with a stack of three that were thermally in series but each one had it's own temperature monitor and its own supply. We were able to cool the sensor to -100 deg C to reduce the dark thermal image and obtain very long exposures without having to always subtract a dark frame. As condensation was a big problem, everything was hermetically sealed and the chamber was filled with anhydrous Nitrogen.

#### ArakelTheDragon

Joined Nov 18, 2016
1,353
I still not but I intend to connect it series
Please post some pictures and more details.

#### wayneh

Joined Sep 9, 2010
17,168
I need it uses electric and makes water cold or hot so if I connect it in parallel it won't active as I want . Is this right???
If current flows through a Peltier, one side gets hot and the other cold. If the heat on the hot side is not dissipated, both sides will become hot and the peltier can be easily destroyed. This is much more likely if you operate the device near its specified maximum. In other words if you supply 12V to a nominal 12V peltier, it will draw a high current and will operate very inefficiently. Typically, they will consume 10 units of energy for every one unit moved from cold side to hot side, with all eleven units appearing on the hot side, requiring good dissipation to avoid destruction.

See post #2. Those are all valid arrangements. You could also run them at reduced voltage and reduced current. For instance, 2 parallel strings of 2 in series on a 12V supply. They’ll be more efficient but will have a lower capacity. But they’ll be less likely to fail.

#### crutschow

Joined Mar 14, 2008
29,768
If you connect them in series, make certain they are electrically isolated from each other.

#### ebp

Joined Feb 8, 2018
2,332
I really don't understand your question or requirements.

If what you are talking about is Peltier modules intended for cooling and you are trying to use them as generators, you will find that the maximum power you can get at the maximum allowable temperature will be a tiny fraction of the power you can put into them for cooling. 12 V at 10 A from a Peltier module is going to require a large, expensive module that must be operated with a very high temperature differential across it. It will be made specifically for power generation. There is NO practical hope of getting 120 W from modules intended for cooling without building a very large array.

#### MisterBill2

Joined Jan 23, 2018
11,875
Please post some pictures and more details.
What we need in order to provide useful advice is a better description of what you want to achieve. pictures at this point are less useful, but if there are data sheets for the devices that you have, the recommended current and forward voltage information would be useful, as well as any information about the claimed efficiency. AND, any description of the recommended heat sink capabilities would be useful as well. The problem is that while the devices transport heat, they also generate a bit of heat, and that is part of the heat sink consideration. So we also need to know the temperature that you intend to cool things down to.

#### Janis59

Joined Aug 21, 2017
1,383
Peltier voltage may change vs temperature in some 2 or 3 fold region, thus there are no better algorythm as CC-PS. And then, regulating the CC amount You maximize the dT for the bottom cascade, then for 2 lower, and then for top one. Never the current (even in equi-sized constructions and far below in pyramidal constructions) will be identical for all platelets in tower. Upper will demand the less power and bottom will love more power. Thus, if the maximizing the dT is the priority, there ought be separate adjustable CC-PS.
But when one is using 130 C capable platelet being happy with 40 C then all may have an identic currents. However, in the situation of last, ultimatively the best solution may be to choose the one-platelet modell.

#### MisterBill2

Joined Jan 23, 2018
11,875
Are you choosing to connect them thermally in cascade?? (thermally in series) or electrically in series but thermally in parallel? Your connection arrangement you choose will depend on the voltage and current ratings of the power source that you have available. And if you have already decided to connect them in series, why ask us??? IN any instance the voltage that you need is that voltage that will cause the desired current to flow. For that you will need to be able to measure current with reasonable accuracy.

#### Janis59

Joined Aug 21, 2017
1,383
Mister Bill2: Exactly, thats was my point in more fluent English.

#### recklessrog

Joined May 23, 2013
985
Used as generators, there may well be different requirements but thermally, the same basics should apply as to cooling.
When you want a high temperature difference between the cold and hot side, you need to put Peltier coolers in MECHANICAL THERMAL series, you create a stack that will look like a pyramid so read on to see why. This is because the first (coldest point of the stack) not only transfers heat from the cold side to it's hot side, but in doing so, produces heat itself. They are not very efficient even if carefully designed and operated.
Now, next in the stack you need a copper or aluminium spreader the same size as the next cooler which will have to be physically larger and be able to take the TOTAL heat now produced by the first one. The second cooler will now need a similar arrangement with another spreader to thermally connect it to the 3rd cooler which must be able to cope with all the heat from the initial point you are trying to cool PLUS all the self generated heat of the cooler stack.
The spreaders Must be very flat and in good mechanical contact as this can be where a lot of efficiency is lost if not done correctly.. Relying on a dab of heatsink compound is not good enough, they need to be lapped perfectly flat with no nicks, burrs or scratches. Same goes for whatever final heatsink you use.
Finally you need to be able to transfer that total heat into either the air with a large heat sink and fan, or better still, a water cooled heat sink.
Because each of the coolers will be most efficient at a different voltage and most likely different current, (you can get some that are specifically designed to work together in stacks at the same current) you will need to power each one from a separate source.
Just wiring them ELECTRICALLY in series will over power the smallest and under power the largest ones leading to destruction of one and horrendous inefficiency.
You really need to contact a manufacturer, tell them your thermal requirements, and they will be able to calculate the correct coolers for you to achieve what you want to do.
Simply putting 4 the same in ELECTRICAL series or parallel, is ok if you only want one stage of cooling. Thermally you will only have a temperature difference between hot and cold that the devices are rated at depending on thermal transfer in and out. Usually done when you need to shift more heat at a low temperature difference than one or more coolers can provide. The thermodynamics involved can become very complicated.
Another fly in the ointment is Thermal gain, basically it's thermal feedback where heat generated in the stack gets back to the cold points. Thermal insulation from hot to cold points can become tricky.

Last edited:

#### MisterBill2

Joined Jan 23, 2018
11,875
Going back over the thread, it appears that your intention is to have hot water and cold water, possibly at the same time. That is certainly possible with these devices. The way to do it is to have two containers that have thermally conductive surfaces on the portion that contacts the peltier devices, so that heat energy can be transferred from the cold water container to the hot water container. That should work quite well. The arrangement of devices, cascade or parallel, will depend on how cold and how hot you want the water and the maximum obtainable temperature difference that your devices can obtain. The limitation is that you do not want ice forming in the cold water container, because a freeze would stop the flow and also cause damage because ice expands as it forms.
As an example, if the supply water temperature is 30 degrees C and the maximum obtainable difference is 20 degrees C then you could get cool water at 10 degrees and hot water at 50 degrees, with all four devices in parallel. The electrical connections would still be dependent on your power source. The advantage of having all 4 in parallel electrically is that you could switch on more or fewer devices depending on how fast you wanted to produce the cold or hot water. In addition, 12 volts at 40 amps is easily produced by an automotive alternator, if you need to generate your own power. But if local (mains) power is available then probably a series connection is better because 48 volts DC at 10 amps can be conveniently produced by a more common transformer and less expensive rectifier diodes.

So the decisions are based on your situation and your requirements, which I did not see much details about them.

#### Janis59

Joined Aug 21, 2017
1,383
Re: Reckless
""next in the stack you need a copper or aluminium spreader""
Yes, and no. Just peltier even the largest and how much a smaller has ENORMOUS Watts per square centimeter. The flux density figure. Thus, if the surfaces will not be polished and glade in range of micrometer or better, the all generated cold will be lost in there.
Shall illustrate with an example. One client was willing to provide a certain experiments in the hot African desert with our reference spectra source, thus the minus 40C was the demand at roughly 10W. So, the best case for Peltiers, but the special demand was no ANY water if the water must be transported on camels maybe thousand km afar. But air cooling for Peltiers is sth very very uncommon.
So I took in the trash one Pentium core radiator with fan, mounted on the best Peltier what was possible to buy for a money at that time and got the cold end at first minute minus 80, next minute zero, next minute plus 80, next minute plus 160. So - what were happening? The 200W radiator was inable to sink the 10W!!! Why? Because it was calculated to get those 200W at about 2x2 inch why my device was 1/4x1/4 inch and the same power. Then I solved this by using a best play-obsessed people videocard 8x heatpipe radiator, what helped a halfway - with a hard but minus 50C it gave when in air is plus 50C for element designed for 137C difference.
By the way, in thermally serial stack, any next platelet MUST be about half x half size or even less the towers one level lower platelet, so making a pyramide, because efficiency of Peltiers is rarely over the 8...10%. And very rarely therefore someone uses the 4-level towers, whilst three level towers are more widespread... 10% x 10% x 10% =0,1%. Push in 100W for to get out so small as 0,1W of cold.

#### recklessrog

Joined May 23, 2013
985
Re: Reckless
""next in the stack you need a copper or aluminium spreader""
Yes, and no. Just peltier even the largest and how much a smaller has ENORMOUS Watts per square centimeter. The flux density figure. Thus, if the surfaces will not be polished and glade in range of micrometer or better, the all generated cold will be lost in there.
Shall illustrate with an example. One client was willing to provide a certain experiments in the hot African desert with our reference spectra source, thus the minus 40C was the demand at roughly 10W. So, the best case for Peltiers, but the special demand was no ANY water if the water must be transported on camels maybe thousand km afar. But air cooling for Peltiers is sth very very uncommon.
So I took in the trash one Pentium core radiator with fan, mounted on the best Peltier what was possible to buy for a money at that time and got the cold end at first minute minus 80, next minute zero, next minute plus 80, next minute plus 160. So - what were happening? The 200W radiator was inable to sink the 10W!!! Why? Because it was calculated to get those 200W at about 2x2 inch why my device was 1/4x1/4 inch and the same power. Then I solved this by using a best play-obsessed people videocard 8x heatpipe radiator, what helped a halfway - with a hard but minus 50C it gave when in air is plus 50C for element designed for 137C difference.
By the way, in thermally serial stack, any next platelet MUST be about half x half size or even less the towers one level lower platelet, so making a pyramide, because efficiency of Peltiers is rarely over the 8...10%. And very rarely therefore someone uses the 4-level towers, whilst three level towers are more widespread... 10% x 10% x 10% =0,1%. Push in 100W for to get out so small as 0,1W of cold.
As I mentioned in a previous post, we needed to get down to -100 deg C and tried every combination imaginable. Without the thermal spreaders we had no chance. The purpose of the spreaders is often miss-understood. It is there to ensure that the whole surface area of the larger cooler can absorb heat from the footprint of the smaller one. The ceramic type of material on the surfaces of the coolers is not particularly good at transferring heat laterally across it's surface, hence the need for the spreader. That in itself introduces the problem that you now have a surface that will happily radiate some of the heat into the surrounding area as thermal feedback.Insulating the exposed surface can prove problematic unless careful thought is given to the thermal path.
We, and a peltier cooler manufacturer, both tried many combinations and arrangements before we finally succeeded. Annoyingly, within 6 months of finalising the astronomical camera, a new improved version of the imaging ccd became available that worked at much lower current, produced less heat and had 3 times more pixels! A single stage cooler was all that was required for that one.
That's development for you

Last edited: