Had Anyone Used a Peltier Module to Thermal Cycle Components?

Never heard of it.

Slow response and limited heating and cooling may make them marginal.

Take a look at this:

Peltier - Thermoelectric Cooler Module Calculator » TE Tech

I recall working on a piece of equipment very similar to what you describe. When I opened the card cage to troubleshoot, the change in the cooling airflow caused the chips to warm up and the problem would disappear. It was very frustrating.

The problem with a Peltier device is that you need a heat sink on the hot side to get much cooling on the cold side. The heat you remove plus the heat generated in the Peltier device due to it's inefficiency has to go somewhere. So to cool your component, you will be introducing even more heat into the immediate area.

You might consider a heat pipe solution. A heat pipe will conduct heat efficiently from your component with the far end being cooled by Peltier device or freon spray.

http://www.electronics-cooling.com/1996/09/heat-pipes-for-electronics-cooling-applications/

Just a thought...

Manipulating the cooling air temp and flow that reaches the suspect components looks more practical than peltiers, where among other problems, you must be able to draw like ten times the input heat from its hot side. On low power components, plenty of space and proper temp sensor feedback, well, why not? S.M.

Not I, said the little red hen.

You might be able to get peltier/TEC to work, but as others have stated, you will need either a large heat sink or something like a chill water source.

.

It seems like if you are working with cards that are so finicky, it would be better to introduce as few new currents/circuits/fields in close proximity as possible.

.

What diameter tubing will you typically be able to run in and out of the card case to where the cards are installed/running? Running flexible high temps insulating tubing in to a small 'radiator' of copper tubing shaped appropriately then back out through insulated tubing, could allow you to heat or cool with an attachment to a heat gun or an ice chilled potable compressed air cylinder....add in a non-contact thermometer....and you've probably got an idea that makes you realize I don't really understand the constraints you are probably facing...

Can you heat or cool the entire cage, and reproduce the problem? If so, your best choice may be to diagnose the most likely components which would cause the problem, and replace them. Then heat/cool again, possibly multiple times, to ensure the repair is complete. I used to call this the 'Shotgun Method'. Don't forget that many repairs can be done without power applied. Electrolytic capacitors with low ESR can be quite intermittent, can be tested with an ESR meter, and are cheap and easy to change. Be sure to check for over heating signs, such as a burned or discolored PC Board. Check solder connections with a magnifying glass and good light. For surface mount microprocessors there are solder re-flow techniques which can solve many intermittent headaches. Sometimes you just can't see the flaw, and I have successfully repaired many boards by re-soldering every connection on the board.

Unfortunately no, I cannot really change the temp of the whole rack. It is part of a large multi cabinet installation on a force cooled raised floor.

As far as the shotgun method, many times, that is what we are forced to do. It's just not practical all of the time.

This system is kind of frozen in time as it uses obsolete technology. Many of the components are simply no longer made and we only have a finite supply of replacements. For, example, the other day, I had to replace an 80C186 processor. Not a whole lot of them running around anymore.

As SM has suggested in #3, baffling the air with pieces of cardboard that selectively block airflow to certain cards may isolate the problem to a card. I was able to isolate my problem in #2 to a circuit card. From that point, I cooled down a small bolt with Freon spray and carefully put it on the top of each IC (to avoid shorting anything) until I found the culprit IC. (Cold would make the problem reappear when the cage was open.)

I copied the code from the IC and called the manufacturer. The rep said "Oh yeah, we know about that. I thought we had removed all those chips from production."

You might take a look at other heat sources. A small light bulb or possibly a foil heater might be small enough to place close to a board in a rack.

I'd also be a little concerned with the board design if you are seeing thermally sensitive failures. Was the thermal management done properly? Is the heatsinking sufficient for the parts that are failing? Are there issues with airflow? It might be worthwhile to have a chat with the design engineers.

I may try to find some kind of foil heater or perhaps a resistive element.

As far as being concerned with board design and thermal management, that ship sailed long ago.

These designs were finalized when George Bush the elder was president. Most of the design engineers have long since retired.

The temperature rate of change from a peltier module beside being very slow is also small as compared to the temperature rise or fall for your application. Heat gun or spray freezing is more effective...

I can appreciate your situation as it was really frustrating to troubleshoot intermittent problems even before my time...

The only way I facilitated my troubleshooting intermittent problems during those days was by rigging and using long extended ribbon cables that had enabled the problem card be taken out while electrically alive. By pulling it out of the cage via the extender cable(s) and conducting the test either inside an oven or freezer for extreme temperature testing or during the 24 hour burn-in process.

Good luck!

I worked on an avionics product that was similar in this trouble shooting issue.

Extenders, well, they are tough even if you design module slot specific circuit cards, any microstrip nets will have a crappy ground, as getting a low inductance ground to the extender card is a real tough thing to do, as you probably know.

We specifically designed our production and service center test equipment to be custom for each module type. It could have the covers off the module, and the adapter could be inserted into a temp chamber. In our case we used Thermotron brand, as you could get a small one that was portable for this use.

The only issues that were a PITA were when you had modules that were multicard assemblies. There we would design extenders for the inter board connectors.

I hope your not working for the same company I retired from, as this was not thought through very well. My avionics product was an IPC-6xxx, with module CCMs, DSMs, PEMs, SVMs, and a TCAS module made by brand H.