Peltiers - what the heck am I doing wrong!

[LewisM]

OK, a new batch of ex-Hong Kon Peltiers / TEC's arrived just now, and I tested them - good fast cooling, but as usual, the heat wins and transfers through to the cold side.

I had the same issue in a cooler box I made using a peltier purchased at Jaycar. I am using my powertank to supply the juice, RAW, without any inverters, resisters etc. Am I doing this wrong? How do I go about making it so that it stays col on the cold side? (yes, I have used properly thermal-pasted heat sinks and fans, and NO difference - my cooler box becomes a heater very quickly!)

HELP PLEASE, as this is driving me totally insane! I get the same result if I use my mains power supply through a converter.

[mswhin63]

One thing come to mind is the type of heat-sink. Without knowing what you have it is difficult to tell. I have the Jaycar version and found they work well.
There is a few other thing but all require knowing how you have the configuration.

Possible resolution with your current heat-sink is reduce the voltage until you have an equilibrium established. I sound as though your heat-sink does not have good thermal properties for the peltier device.

[LewisM]

Heat sink is an aluminium one bought at Jaycar - finned like a computer one. Heat side has a PC heat sink and fan- I reversed the fan to extract instead of blowing.

If I plug one in direct, without heatsinks etc, I get instant cool, but then the hot overwhelms within 20 secs. I am using 12V DC, at about 3 amps I guess (too much probably). If I use the 1 amp transformer, I get the same. If I drop it to .5 amps, I get nothing

[technofetishism]

Tried blowing instead of sucking with the fan. I don't reckon you'll be getting much airflow over the heatsink otherwise.

[Screwdriverone]

Yep Lewis, I think the problem is the fan direction.

I use a stock CPU cooler which blows ONTO the heatsink surface and the heat flows out through the sides and away from the peltier.

My CPU cooler / fan combo has a copper heatsink plate with an aluminium fin design with the fan PUSHING air onto the fins and plate towards the cooler box.

Also, I am using a Jaycar $14.95 48W peltier and it hasnt missed a beat so far.

Good luck, have you tried stacking two peltiers? Apparently this can work even better than one......apparently, but it would suck a lot of current....

Cheers

Chris

[rmuhlack]

I think that is a significant part of your problem.

In order for your peltier device to cool effectively you need to keep the hot side as cool as possible (as its the temperature differential between the hot side and the cold side that is important). If you let the hot side get really hot, to say 50C, then the cold side isn't going to be much different than the ambient temperature. So this doesnt happen, you need to ideally transfer heat away from the hot side of the peltier to the outside ambient at least as fast as the heat flow across the peltier itself. To achieve this means that your hot side heat sink and fan need to be operating as efficiently as possible.

Now, a centrifugal device like a fan is much more efficient at generating a positive pressure differential (ie increasing the pressure on the discharge side by blowing) than it is a negative pressure (reducing pressure on the inlet side - sucking). It's just like a centrifugal pump. By setting up your fan to suck instead of blow as you have done, it would seem that the air movement is simply not enough to disperse the heat fast enough.

The take home message - get it to blow across your heat sink instead of sucking from it.

Incidently, I have been thinking about building a DIY peltier cooler for my DSLR as well, and have sourced a peltier and power supply. One idea that I have had (to save weight hanging off my focuser) was to use water cooling instead of a heatsink and fan. I was thinking a CPU water cooling block with tubing connected to an aquarium pump and a reservoir of cold/ice water. Haven't tried it out yet but it should in theory be much more efficient than an air cooled setup.

:thumbsup:

[rmuhlack]

one other thing - if you've reversed the direction of air movement by reversing the wiring (as opposed to turning the fan upside down), then that is also working against you, as the fan blades will be shaped to generate a pressure differential and therefore 'move' air in a certain direction (like a wing). A wing travelling in reverse is not going to be very effective

[LewisM]

Nah, I was a pilot for 15 years - I reversed the fan, NOT the polarity :)

[Paul Haese]

What are you cooling? Camera or scope?

I suppose that does not really matter. Do you have a plastic sheet of 5mm perspecs betwen the cold plate and the bottom position of the eat sink? Screws holding everything together should be plastic threaded bolts. Fan direction plays a part but the isolation is more important.

[LewisM]

Yes, I am using nylon/plastic screws and nuts. NONE of the hot side touches or convects/conducts with the cold side, except within the peltier itself.

[Star Catcher]

I have made up a water cooled chamber attached to the back of a large finned heat sink. I did not bother with running any fans across the heatsink, but no doubt I could increase the efficiency if I did so. Using water is more efficient but involves plenty of gear to trip, kick and spill.

I run a simple water pump circulating water through two water inlets to the chamber. The water runs through a couple of metres of copper tubing sitting on the top of the lid of a plastic bucket. This arrnagement dissipates the heat better to the air as opposed to water cooling only in a plastic (insulating) bucket.

I find this arrangement nicely maintains the temperature differential for many hours. The peltier gets super cold and on humid days you can make big ice blocks. :)

Ted

[MrB]

Have you done this to all the Peltiers you are trying to use?
This may not be the cause of your problems, but powering Peltiers with no heatsinking on them is a big no-no, certainly not for 20 seconds.
Atleast, it was a big no-no when I was playing around with them a decade ago. It may not be an issue now, but I would be surprised.

[LewisM]

No, only on a trial one. They cost me $2 ea from China so not a big issue to experiment. The Jaycar one was $19, and not "fiddled with"

[Poita]

For the Peltier to work, you have to get the heat away from the heatsink side as fast as possible.
Turning the fan around will be 90% of your problem.

Water cooling is affordable and tidy now, and is a great solution if done correctly.

[bojan]

Also, there is a maximum junction temperature.. if exceeded the Petlier turns into a block of .. something.
Hot side should never be above 85°C

[solissydney]

Peltier installed back to front

[Poita]

Now that would be hilarious.

(I've done it myself...more than once...)

[pmrid]

O.K. I'll bite. WHich is the hot side? The EBay ones I've bought didn't indicate one way or the other.
Peter

[solissydney]

By testing .
Take the Peltier apart, then turn on the power. Within a few seconds you will know.
I have been there myself.
You could of course also try by cutting the two wires then swapping them.

[MrB]

I suspect most people are buying loose modules(as per this picture (http://miriam-english.org/projects/steg/peltier-module.jpg)) therefore there is nothing to take apart and the wires are already accessible.
As you say, apply power to a loose module and feel which side is warm and which is cool.
Be carefull not to overheat the junction's though, disconnect power the instant you feel heat.

[solissydney]

Loose Peltier???

LewisM has a problem with his Peltier.
I suggest that his problem could be that the Peltier is installed upside down.

[solissydney]

Loose peltier ?

lewisM has a problem with his peltier.
I am suggesting that his peltier is ,installed, upside down.

[MrB]

My apologies, with out a quote I had assumed you were replying to Peter's post immediately above yours.

[solissydney]

I believe the mistake is mine because I kept thinking of an assembled peltier when I suggested taking it apart for testing whether the hot side of the peltier was actually facing the heat sink.

[rcheshire]

Tellurex FAQ for extensive info on TEC's.

Facing the module. Black on left, red on right and you're facing the hot side.

Heatsinks for TEC need to have a rating of approx 2x the power of the TEC. The operation of the TEC produces heat which must also be absorbed. Maybe heat sink is too small.