# How to regulate temperature of an IC/PCB -in a vacuum

I have to regulate the temp of an IC. (70deg. C)

The part will be in a vacuum chamber.

Electrical connections can pass in-out of the vacuum chamber through a port (plenty of connections available.)

I have done this before in air using TECs to regulate the temperature (both heating and cooling) using a control loop with a thermistor for sense feedback.

I am predicting this won't work in a vacuum.

TECs only create a temperature differential. You need a constant temperature reference to go up and down from (like a chunk of metal with fans.)

If that chunk of metal cannot thermally conduct to air it will continue to heat-up. TECs are not 100% efficient; a portion of the input power goes directly into heating the system, regardless of what direction it is being driven.

What I need to do is take thermal energy out of the system.

I don't think it is impossible.

Imagine a cylinder filled with air compressed by a piston, sitting a vacuum.

It takes a constant force for the piston to maintain position.

• Compressing the piston further would heat up the cylinder (energy-IN)
• Decompressing the piston would cool the cylinder (energy-out)

But even this example wouldn't work forever because of the friction losses from the piston actuating.

You need to thermally conduct in-out of the vacuum chamber.

But can you do this with electronics? Would it be a TEC across the vacuum chamber barrier? If so, I don't have that.

• you can still radiate. If it is ovenized just for temp stability that could be enough thermal load to control against (but veeery slow response time of control) Commented Jan 20, 2021 at 17:25
• Convecting the heat to a solid mass? Commented Jan 22, 2021 at 20:20

In my experience you can conduct a great deal of heat away from your device simply by conduction through its electrical connections. If that is not sufficient, use a heat pipe (as is done in laptops, for example) to conduct the heat away from your device to the wall of the chamber or to a heavy cable that leaves the chamber.

• This worked BTW, heating the ground plane @ the part... Heat conducts away through cable shield(s).
– Tony
Commented Aug 9, 2022 at 14:15

You can conduct heat out of a vacuum chamber through the wall. A glass wall if necessary, but a metal wall is better. You could either have a cold wall, with the TEC inside, or a hot wall, with a TEC or even passive cooling outside.

If your vacuum system will permit it, you can pass two pipes into it and use liquid cooling to get heat out.

Isolated objects in the vacuum will still radiate heat to the walls, mainly infrared. As radiation goes as the fourth power of temperature, this mechanism will work a lot better at higher temperatures.

Radiation isn't great at 70°C. The exchanged heat is not a function of the temperature difference in radiation, but a function of the absolute temperature of the DUT in the chamber to the power of four minus the absolute temperature to the power of four of the wall.

Attach a heat pipe to the DUT in the chamber.

As @Neil_UK says, you can run tubes through a heat exchanger in the chamber or try to conduct the heat out through the walls. Either way you will need to get a lot of heat out if you expect to cool the chip effectively as TECs have miserably low efficiency.

In fact, you might prefer to eschew the TECs and run temperature-controlled fluid through a heat exchanger. Oil or water 'temperature controller' devices are available which pump liquid through an external system (such as a plastic mold- which is, in many respects, a heat exchanger). They are commercially available in units that can easily handle the 70°C temperature or you can fabricate something with commercial components.

In any case, to actually conduct the heat to (and from) the chip you should consider the conductivity between the die and mounting.