How to avoid condensation?

[mecheng66]

I'm tasked with cooling a set of electronics, something I've done many times, but this project has a new twist that I have no experience with: one of the devices has it's own proprietary cooling solution that brings it's temperature below ambient and we can't allow condensation to occur.

The electronics (the conventionally cooled portion) consists of dozens of ICs dissipating a total of about 600W in a space of about 12" x 20" x 24". Normally I'd cool these with about 120 CFM of ambient air delivered by fans, but the cold device is right in the middle of this.

Does anyone have suggestions on how to avoid condensation?

 

[IRstuff]

Two options:
Faster air
Dryer air


But, I'm not sure why this is your problem. Whoever is running sub-ambient would have provisions to prevent condensation, no? After, what spec did he design to?

TTFN

 

[Yorkman]

How cold is the cooling solution? Condensation is only going to happen if the surface temperature gets below the Dew point temperature of the ambient air. Can you regulate the temperature of the coolant without over heating the electroncs. If the surface temperature is indeed below depoint of the ambeint air perhaps dehumidifing the cooling air before it enters the cabinet is the solution. Or perhaps a silica type dessicat that can be change or recharged periodicly What are the conditions of the ambient cooling air?
We use a cooling solution on the heat sinks for the gate drives on our VSD (York Chillers) but we are cooling it with condenser water and that never produces temperatures low enough.

I'm not a real engineer, but I play one on T.V.
A.J. Gest, York Int./JCI

 

[insult2injury]

You might want to investigate the use of heat pipes or possibly some shrouding in the case depending on your geometry.

I2I

 

[Yorkman]

One final thought can the self cooled device be incapsulated so to speak with a foam type insulation/vapor barrier. Since it has it's own cooling system the isululation might not critcally affect it's temperature.

I'm not a real engineer, but I play one on T.V.
A.J. Gest, York Int./JCI

 

[mecheng66]

Thanks for all your suggestions.

IRStuff, how does faster air help?

Yorkman, the surface temperature of the cold device can be at least 20C below ambient.

Any suggestions on the best way to dehumidify 120CFM air?

 

[mecheng66]

Good idea, Yorkman, but it could be difficult: the device being cooled below ambient is mounted to the same circuit board that the other electronics are mounted to.

 

[MintJulep]

Desicant wheel? Suck out the moisture going in, use the exhaust to regenerate.

 

[IRstuff]

Faster air means more heat transfer to the cold surface, thus keeping it staying cold.

Another relevant question is why does it need to stay non-condensing? Can you provide an outlet for the moisture

TTFN

 

[Yorkman]

Mecheng66,
It's 11:00p.m. and I'm still thinking about your question. Something that has occured to me, the cooling fans that you are using to cool the components on your circuit board, are going to actually have a reverse affect on the self cooled device. The ambient air being circulated through the cabinet will likely make the self cooled device run at a higher temperature due to the increased heat transfer caused by the air flow. Now is it enough to cause problems I don't know but it is something to be aware of. Is it possible that if the coolant were "shut off" and the cabinet cooled in your standard method the device would operate safely? Is it possible the designer never considered the air circulation that you are providing and felt the need to supply their own?
Just some thoughts that I had to air so I could get them out of my head so I could sleep. Good Night and Good Luck

I'm not a real engineer, but I play one on T.V.
A.J. Gest, York Int./JCI

 

[mecheng66]

That's right, IRStuff, it doesn't have to be non-condensing. Even as I'm asking how to avoid condensation, I'm also trying to figure out a way to remove moisture, as you said. But it's looking pretty ugly and and I'm hoping a non-condensing solution will be easier... tbd

Yorkman, good point. I'll have to consider the heating effects of ambient air on the cold device. The internal temperature of the device is kept at 0C. No way to do this with ambient.

Someone suggested compressed air. Apparently the labs where these are going have compressed dry air available. Do you think it's feasible?

 

[IRstuff]

Trying to cool a packaged IC from the outside is significantly less than optimal, particularly with a TEC. The proximity of the hot side of the TEC to the actively cooled region is also problematic.

Nonetheless, the most practical solution is probably to foam insulate the device. Since the insulation will take up the delta temperature between the TEC and the air stream, condensation will be minimized. This would be the next best thing to encasing the device in an evacuated enclosure.

TTFN

 

[IRstuff]

Dryer air is of course another option. The dewpoint of the compressed air would need to be monitored to ensure that its sufficiently low enough to make it work. However, the amount of air flow volume for cooling an electronics box is usually a bit much for normal compressed air supply lines.

Is this a one-of? And is this to be a semi-permanent setup within your control?

TTFN

 

[Yorkman]

As to dealing with the condensate can the board be positioned in the cabinet so that the condensate drips/drains off the cold device so as to avoid damaging the other components on the circuit board? For example mounted "upside down" or if the cold device is at the perimeter of the circuit board maybe vertical so that moins off the edge? Just some quick thoughts.

I'm not a real engineer, but I play one on T.V.
A.J. Gest, York Int./JCI

 

[mecheng66]

Thanks for the ideas. IRStuff, the plan is to build about 15 or 20 of these and they'll go into different labs.