Hi, a little backstory.

I'm a university student doing a project that requires me to have a server inside a sealed chamber (vacuum).

The thing is the server's power-on button is connected to this little board here: https://imgur.com/z5oeJnS (beside my thumb) and this is the back: https://imgur.com/AfT74nC

It has 13 wires connected to each spot. Anyway, what I'm planning to do is to weld a KJL 15-pin d sub flange: https://imgur.com/OpJO1Vp (front) and https://imgur.com/HWw06rr (back) onto the wall of my chamber.

Then I would cut the 13 wires somewhere in the middle the length, soldering each wire to 13 of the 15 pins in a 1-to-1 positioning, with the little board/button on the external such that I can power on my server from the outside.

So before I jump into this, I'd like to ask for opinion, whether this is a good idea or not. I'm just hoping the 15 pins just work as a conductor from one side to another, and I'm not missing out on anything that may cause any short or open circuit. I'm just a mechanical engineer, and a student at that, so my experience with electronics is rather lackluster.

Thank you.

 

I'm a university student doing a project that requires me to have a server inside a sealed chamber (vacuum).

This is not your main question but I am wondering how you plan on cooling the server. This may be more difficult than extending a cable.

 

As long as the connector is designed and rated for use on a vacuum chamber it should work out fine. The same is true if you plan water lines in and out for any cooling needs.

Ron

 

As long as the connector is designed and rated for use on a vacuum chamber it should work out fine. The same is true if you plan water lines in and out for any cooling needs.

Ron

Heat transfer in a vacuum chamber? That could be a challenge...

 

Heat transfer in a vacuum chamber? That could be a challenge...

We cool KW's of power under vacuum using water coils and a external heat exchange. It shouldn't be a problem with the right heatsink, connections and feedthroughs

https://www.lesker.com/newweb/feedthroughs/instrument_feedthroughs_mpdt_subd.cfm?pgid=cf

 

We cool KW's of power under vacuum using water coils and a external heat exchange. It shouldn't be a problem with the right heatsink, connections and feedthroughs

https://www.lesker.com/newweb/feedthroughs/instrument_feedthroughs_mpdt_subd.cfm?pgid=cf

I know liquid cooling - I was just hoping the OP considered the cooling aspect since there were no questions about liquids -just the wires.

 

I know liquid cooling - I was just hoping the OP considered the cooling aspect since there were no questions about liquids -just the wires.

I would avoid anything but the simplest electronics in a vacuum chamber and can see almost no reason to have a server inside.

 

I would avoid anything but the simplest electronics in a vacuum chamber and can see almost no reason to have a server inside.

Me too. But the story started like this...

I'm a university student doing a project that requires me to have a server inside a sealed chamber (vacuum).

And ends like this...

I'm just a mechanical engineer, and a student at that, so my experience with electronics is rather lackluster.

I don’t imagine a lot of forethought from the professor that “requires me to have a server inside a sealed chamber” and the OP had a good guess as to the type of situation they would be in when selecting s moniker: @toiletmirror

 

Heat transfer in a vacuum chamber? That could be a challenge...

I should point out here that in the wonderful world of vacuum and hot I see two things in my distant past. There were vacuum ovens and vacuum furnaces. In my little world the ovens went to about 400 to 500 F max and furnaces up to 2,000 F. Most ovens ran cooler at temperatures in the 200 to 300 F ranges. I was involved with quite a bit of testing like shake and bake in a vacuum so whatever we were cooking in a vacuum was "instrumentated" so plenty of wires ran in and out of the various chambers. Some of the stuff was pretty cool like shake and bake thermal shock where you have something cooking at 250 to 300 F and turn on refrigeration compressors and dump liquid nitrogen in the chamber, but those were not vacuum chambers. Somewhere around here I have some feed through bulkhead fittings designed for vacuum.

Anyway it was not unusual to instrument things in vacuum furnaces. Mostly with thermocouples. Most just pulled down to atmosphere, not like milli-torr vacuum numbers. The original poster only mentions:

I'm a university student doing a project that requires me to have a server inside a sealed chamber (vacuum).

So I doubt heat is involved.

Ron

 

We cool KW's of power under vacuum using water coils and a external heat exchange. It shouldn't be a problem with the right heatsink, connections and feedthroughs

https://www.lesker.com/newweb/feedthroughs/instrument_feedthroughs_mpdt_subd.cfm?pgid=cf

In theory anyway - there's not just the CPU cooler. There's the onboard SMPSU MOSFETS and at least one BGA chip that has some form of finned heat sink. Regular hard drives need a "normal" thermal environment - no idea about SSDs though.

Might be possible to radiate enough heat away with heat pumps to a substantial black anodised aluminium casing.

Just occurred to me - apart from very recent nitrogen filled types, hard drives aren't sealed. AFAIK: the heads won't float in a vacuum.

 

So I doubt heat is involved.

I didn’t mean an oven or furnace was involved, but the cpu can quickly reach some amazing temperatures with near perfect insulation (vacuum).

 

I didn’t mean an oven or furnace was involved, but the cpu can quickly reach some amazing temperatures with near perfect insulation (vacuum).

Once I tried firing up a board with no CPU cooler fitted.

My finger was on the CPU when I switched on and the instant burning sensation made me switch off again PDQ.

Some of the more impressive heat sink CPU coolers keep doing their job after the fan conks out - but highly unlikely in a vacuum.

 

It depends on the pressure level you are attempting to achieve.
What level vacuum do you need?

You have two problems. Anything you place inside the vacuum chamber could keep on out-gassing and prevent you from reaching the desired pressure.

Secondly, any feed-through connector has to be designed for that level of vacuum.
There are special epoxy glues that are designed for high vacuum applications. We seal all our connectors with this epoxy.

 

I didn’t mean an oven or furnace was involved, but the cpu can quickly reach some amazing temperatures with near perfect insulation (vacuum).

Oh yeah, pulled a vacuum on impregnation machine and watched the resin boil at room temperature. Like baking in Denver, CO.

Ron

 

This is not your main question but I am wondering how you plan on cooling the server. This may be more difficult than extending a cable.

Heat transfer in a vacuum chamber? That could be a challenge...

We cool KW's of power under vacuum using water coils and a external heat exchange. It shouldn't be a problem with the right heatsink, connections and feedthroughs

https://www.lesker.com/newweb/feedthroughs/instrument_feedthroughs_mpdt_subd.cfm?pgid=cf

In theory anyway - there's not just the CPU cooler. There's the onboard SMPSU MOSFETS and at least one BGA chip that has some form of finned heat sink. Regular hard drives need a "normal" thermal environment - no idea about SSDs though.

Might be possible to radiate enough heat away with heat pumps to a substantial black anodised aluminium casing.

Just occurred to me - apart from very recent nitrogen filled types, hard drives aren't sealed. AFAIK: the heads won't float in a vacuum.

Woah, hi guys. Didn't expect such a discussion overnight (from the other side of the globe).

Yes, it's actually a thermal management experiment, and we have conducted it in other environments (e.g. ambient). So we'd like to see the effects of it in a vacuum. We're using a server because the project is a collaboration with a certain data center. Doubt I'm allowed to reveal more details about this, sorry!

As long as the connector is designed and rated for use on a vacuum chamber it should work out fine. The same is true if you plan water lines in and out for any cooling needs.

It does feel like it SHOULD work but damn I'm having an uneasy feeling about this. Anyway, thank you!

 

I think the idea is nuts for a office environment PC server. Nobody puts generic computer equipment in hard vacuum for operational testing unless they just want to destroy equipment and maybe contaminate their vacuum chamber, plumbing and pumps.

 

The cpu will get incredibly hot. The heat will have no place to dissipate to the environment so it will just heat up the cooling fins until the first failure mode (I’m guessing it will be the gallium/indium-containing solder used to bond the wires to the silicon wafer).

I’m just not sure which piece of silicon will heat up first, the cpu or the mosfet on the power supply. They both have fairly large heat sinks so it might take them a while. Interestingly, the pc might not even boot. The cpu fan may spin so fast with no air resistance that the BIOS check of the back emf to make sure cooling is present may fail.

Who knows. I doubt the thing stays on more than a minute before it automatically shuts down in protection mode.

 

The thing is the server's power-on button is connected to this little board here: https://imgur.com/z5oeJnS (beside my thumb) and this is the back: https://imgur.com/AfT74nC

Hey -- why not just interrupt that power on switch with TWO wires? Grab a piece of 3/4" Delrin in the machine shop, drill the bolt holes, and then a small hole in the center to pass the TWO wires for the switch through. Backfill with Double-Bubble Red epoxy and call it a day.

 

Wow, lotsa jabber. Maybe I should have plowed through it all,,, oh well.

On your actual posed question:

You should use what we always called a "bulkhead fitting". You want to pass those insulated conductors through that wall as simple insulated conductors, and do your connectors separately. If you have signals requiring shielding, you are typically better off to separate them, possibly to a separate fitting, to keep signals clean.

Essentially, such fittings, typically threaded, produce a very tight seal intended to withstand substantial pressure differentials continuously, and chemicals and gasses not known to have deleterious effects on your chosen conductor insulation jackets. They are available with conductors in place and potted, or as ready to assemble pieces which you pass your own conductors through.

 

What kind of signals are on those 13 wires? Some may have trouble going through that connector.

Things to consider minimum and maximum levels are:
Voltages
Currents
Frequencies/rise time/fall time
Impedance