How to prevent condensation on Peltier / TEC cooler?

[jbb]

Hi all

First up; this is only a thought experiment here, so don’t hurt yourselves trying to nail down an exact solution.

TL;DR anyone know reasonable ways to keep condensation / icing out of a small refrigerated piece of electronics? Unfortunately water vapour has a way of sneaking into almost anything and condensing.

I’m thinking about electronics for test gear & instrumentation. Many things like crystals or voltage references are placed in constant-temperature ovens to get stable performance.

However, some things get more noisy or leak more current at high temperatures, so I’m wondering about using a Peltier / Thermo Electric Cooler / TEC to make a chilled box. Maybe even getting really fancy with a stacked TEC to get all the way to -40C (for a very very small box and minimal electrical load inside).

Setting aside the design of the TEC refrigeration stage, I’m concerned that water vapour would sneak in (I live in Auckland, a humid city) and condense (bad) and/or freeze (very bad) on things. If cost were no object (hah!) then I could put the whole thing in a metal box with hermetic seals and seal it under vacuum (for thermal insulation).

But cost is an object, so… any suggestions?

[coppercone2]

the reliability is gonna be way lower with this tech then heat because really there is no way to get away from something like hermetic seals.

the answer is always to not put electronics in condensation prone areas

[moffy]

If you are worried about the things inside the cooler maybe a ziploc bag with a slica gel sachet?

[Siwastaja]

Unfortunately water vapour has a way of sneaking into almost anything and condensing.

It ain't magic. The only way is to stop water vapor from entering. It is entirely possible, but you need to pay attention to wire/cable connections etc. passing through your hermetic seal. Then all you need is to manufacture the sealed part in very dry environment (calculate RH at operating temperature). You can add some desiccant inside for some safety margin against very small leaks.

[Psi]

Look up how they protect the cpu/motherboard when doing cry cooling or LN2 / overclocking on PCs.

Fill all the voids with dielectric grease, or kneedable art eraser

Cant condense water if there's no air in the voids.

https://youtu.be/itQLBGQyTX8?t=530

[jbb]

“…the reliability is gonna be way lower with this tech then heat…”

Agreed. I would think very hard before trying to deploy this in an important piece of equipment.

“The only way is to stop water vapor from entering… you need to pay attention to wire/cable connections etc. passing through your hermetic seal.”

Anyone have suggestions for how to do that on a budget?  I was wondering if a re-enterable silicone potting compound might work, but I’m not feeling confident.

I guess a more sensible approach to experimenting at these temperatures would be to make a wee -40C oven and put it into a polystyrene cold box next to some dry ice. Anyone tried that?

[dietert1]

A simple solution is desiccant with a near hermetic enclosure (IP66 metal enclosure or sealed plastic bag). We always get those plastic bags with metal coating when we buy integrated circuits.
If one finds oil proper for the intended temperature that could work. It also helps as heat/cold spreader inside inside the oven.
One could fill/pressurize the oven/freezer with industrial dry gas from a small bottle.
Another way could be a device that sucks out humidity from the oven. Years ago i bought some Rosahl PD4 membrane dehumidifiers but didn't use them until now.

Regards, Dieter

PS: Yesterday i received through hole glas zener diodes (hermetic!) from Farnell and they shipped them in a sealed plastic bag with desiccant and indicator card. Crazy.

[David Hess]

I have seen it done with conformal coating, RTV, and closed cellular foam, but potting by itself is not enough.

We always get those plastic bags with metal coating when we buy integrated circuits.

The common plastics used for bags are permeable to air and water vapor unless metalized.

[coppercone2]

the O2 and H2O molecule are basically the same length in picometers.

IMO its a worse problem then even space, because space is keeping things in not out.

that's why alot of people say underwater is harder then space. under temperature is kinda similar to underwater.

if this was easy we would have a moon base IMO

[CatalinaWOW]

While it is true that water is very effective at getting everywhere, there are two key concepts to keep in mind.  Permeation rate and duration.  Permeation rate has two components-leakage and diffusion.  Leakage is straight gas transfer, and with thought, care, gaskets and seals can be completely eliminated.  Diffusion is much slower, but basically impossible to stop, though proper material selection and physical geometry can reduce the rate dramatically.  Putting desiccant in the protected volume you can keep the humidity below the dew point for some period of time.

Doing this analytically is possible in principal and is done routinely for some equipment configurations, but is difficult and has significant error bars.  It involves estimating the moisture content of items inside the protected volume, all leaks and area, thickness and diffusion rates for all diffusion paths, and moisture capacity of the desiccant all as functions of time, temperature and pressure

It is more practical to take an empirical approach.  Make your chamber as tight as you can, put some desiccant in the chamber and monitor humidity over time (admittedly this part is tough if you are talking about very low temperature).  Be sure to include a pieces of gear similar to your precious to simulate that moisture source.  You can adjust desiccant quantity and sealing effort based on your results.

You can also help the problem by dehumidifying the air around the chamber, or doing a dry nitrogen purge of the chamber.  In many cases a purge is achieved using the boil off from the LN2 used to cool the chamber, but high pressure bottles of dry nitrogen are not horribly expensive.

[CatalinaWOW]

I will add that if you are satisfied with keeping sensible water/frost off your equipment a much simpler approach may work. Condensation will start at the coldest part of your chamber.  If you provide a drain under the cooling heat sink, make sure that the leaks are pretty slow and near that heatsink, and make sure that your initial cooling cycle is slow enough capture and drain the water without freezing and blocking the vanes you may get a good enough answer.

[S57UUU]

Put something colder  (like an unloaded peltier) into the same box?

[mzzj]

Small box from thin copper sheet soldered together on all seams and solderable hermetic feedtroughs for the connections. Purge with dry nitrogen or argon and solder the tube ends.

[Terry Bites]

The TEC is a heat pump. The energy sucked from the hot side must equal the energy shoved out the cold side. 

A larger  heatsink on the cold side will not get as cool for the same throughput while the hot side can sit the set point.
Same energy throughput but with a lower temperature drop.

[AndyBeez]

One question, how are you dumping the heat from the TEC?  Heatsink, heatpipe, radiator?

You might be able to use the waste heat to dry an absorbent filter sponge?

[dietert1]

Since the usual TEC is only 4 to 6 mm thick, if you put it between two flat surfaces, the thermal insulation between those surfaces will be thin and poor. So i started to make packages with copper pieces like shown to increase the space for insulation. In the end one puts the heat into a heatsink. The heat has two components: The heat extracted from the cold side and the dissipation of the TEC itself.

Regards, Dieter

[jbb]

One question, how are you dumping the heat from the TEC?  Heatsink, heatpipe, radiator?

You might be able to use the waste heat to dry an absorbent filter sponge?

This is more of a thought experiment - didn't want to pursue a design concept which was doomed to fail.

Ultimately, if I decide to go this way it'll probably be for short-term experiments, so it wouldn't need to be perfect.