Cryogenic Raspberry Pi & Camera CERN schematic

[erpupones]

Hello to all,

We would like to use a raspberry pi 3 model B Vi.2 and the Camera V2.1 in a cryogenic environment (~ -200°C) for an experiment at CERN.

Has anyone already tried to do it?

I'm also searching for a complete schematic of this models with all the internal electrical connections, more detailed than this:

https://www.raspberrypi.org/documentati ... educed.pdf

Thank you.

[W. H. Heydt]

The lowest operating temperature that I know of for a Pi was one monitoring penguins in Antarctica. It ran to at least -42C. I beleive that what stopped it was the batteries failing.

Other processors, since as an Intel Celeron and--if I recall correctly--and AMD chip have been run cooled by liquid Nitrogen. This was done in a effort to set overclocking records, so you might do a search on "extreme overclocking".

Mind you, all of this is WAY out specification for the devices in question. At the temperatures you want to use, I'd be worried about chip packages and the PCB being quite brittle and failing at any shock or impact (but I presume that you are aware of such issues).

[douglas5962]

Two ideas for you.... Use a bore-scope so you can keep the electronics out of the cryogenic temperatures. If you can't move the camera away, you could insulate the camera and Pi so that the heat they produced is use to keep them at slightly warmer temperature.

You might want to check out Cryogel-Z insulation from Aspen Aerogels. I've played with their high temperature products for protecting test equipment and it does a pretty nice job for the thickness.

[jamesh]

Hello to all,

We would like to use a raspberry pi 3 model B Vi.2 and the Camera V2.1 in a cryogenic environment (~ -200°C) for an experiment at CERN.

Has anyone already tried to do it?

I'm also searching for a complete schematic of this models with all the internal electrical connections, more detailed than this:

https://www.raspberrypi.org/documentati ... educed.pdf

Thank you.

I doubt this has ever been done. Quick check round the office resulted in giggling, then "I dunno".

I'm afraid there are no more detailed schematics available. What do you need to know that isn't on them?

[W. H. Heydt]

Well... Let's see.... CERN is a research organization. Seems to me that buying Pi and a camera and trying to run them at cryogenic temperatures would fall into the category "experiment", and darned cheap one at that...considering what the major experimental devices at CERN are...

So my suggestion is: buy a Pi and a camera and try it. You'll only be out about $60 to $80 (plus whatever materials you use that CERN already has on hand. Please report back here with the results so we can inform the next person who asks. (How cold can a Pi be run isn't exactly a *frequent* question, but it comes up surprisingly often.)

[mahjongg]

Hello to all,

We would like to use a raspberry pi 3 model B Vi.2 and the Camera V2.1 in a cryogenic environment (~ -200°C) for an experiment at CERN.

Has anyone already tried to do it?

I'm also searching for a complete schematic of this models with all the internal electrical connections, more detailed than this:

https://www.raspberrypi.org/documentati ... educed.pdf

Thank you.

you are joking right... On both counts...

[ejolson]

you are joking right... On both counts...

While your nose might be right, CERN may also host visiting students and student-guided experiments. Instead of a troll this might be a beginning researcher. Still, it is difficult to know how to answer. If such a question were asked in person, I would next ask whether they have discussed the possibility of using a Pi with the senior scientist on the project.

[mosespi]

Physical cracking of the components will likely result at that low a temperature. Most of the parts on the Pi (and all electronics really).. those Integrated circuits, CPU, memory chips.. are all made out of dissimilar materials; plastic, metal, bonding wires, silicon. All expand and contract at different rates and most generate their own internal heat, -200°C might just be too cold.

Then there is the electrical side of things. Eventually parts fall too far out of spec for proper operation.

Now if I were to test something like this I would =slowly= start lowering the temperature of said Pi while it was in an operating state. Don't shock cool it, that is bad. You may be able to get to -50°C and even -100°C. One thing to note that while it may operate if taken from say 0 to -50°C, it may not power up at that temperature, cold starts may not work. I do doubt -200°C will work.

But if you try it tell us what happens!

Regards,
-Moses

[W. H. Heydt]

Physical cracking of the components will likely result at that low a temperature. Most of the parts on the Pi (and all electronics really).. those Integrated circuits, CPU, memory chips.. are all made out of dissimilar materials; plastic, metal, bonding wires, silicon. All expand and contract at different rates and most generate their own internal heat, -200°C might just be too cold.

Then there is the electrical side of things. Eventually parts fall too far out of spec for proper operation.

Now if I were to test something like this I would =slowly= start lowering the temperature of said Pi while it was in an operating state. Don't shock cool it, that is bad. You may be able to get to -50°C and even -100°C. One thing to note that while it may operate if taken from say 0 to -50°C, it may not power up at that temperature, cold starts may not work. I do doubt -200°C will work.

But if you try it tell us what happens!

Regards,
-Moses

I can think of three examples of using liquid Nitrogen as a coolant for computer electronics. I don't know that any of them took the temperature down slowly, and one of them was indirect cooling (using a flourochlorocarbon as the coolant and liquid Nitrogen for the cold side heat exchanger). That one immersed an entire PC motherboard. Things went just fine until the coolant started to turn into a gel due to low temperatures. The other incidents I know of used a fairly large (about 2 in. diameter) pipe sealed to a plate that clamped onto the CPU package. The pipe was filled with liquid nitrogen.

So while I agree with you about the thermal stress issues, and the possibility of components going so far out of spec that they won't work at all, the specific examples I've seen written up didn't run into any of those issues.

[Burngate]

Last night while waiting for sleep to come, I tried to do a bit of maths.

With a Pi 3B+ and camera in a box, -200 outside & about 2½W inside (from http://raspi.tv/wp-content/uploads/2018 ... B-plus.png) what sort of temperature can we expect inside, using what materials for the box?

Google is of no use with ones eyes closed, so I came to no conclusions, and leave it as an exercise for the student.

[karrika]

The Pi is not really a good device for cryogenic overclocking. Years ago I tried with a static chip that contained no timers or dynamic delays. You could submerge a Transputer in liquid nitrogen and run it with 10 times higher clock speed. But the Transputer also contained RAM and communication inside the chip. It was a static device that could be overclocked. It needed no external components. The only "problem" was that you also needed to submerge the other Transputers that communicated with it as the link speeds went up by a factor of 10.

The idea was to do 3D FFT for MRI. The cryogenic version was too complicated to make so we settled for arrays of 32 Transputers instead.

Today you have more processing power than that in your Pi

[hippy]

~ -200°C ... Has anyone already tried to do it?

It's not unreasonable to ask - and I don't think laughter, eye-rolling and derision is appropriate - but I think the most likely answer should be obvious.

I'm also searching for a complete schematic of this models with all the internal electrical connections

Being CERN you may be able to obtain the details you need if you approach the Foundation or Trading directly. Mere mortals stand no chance.

[erpupones]

Hello,
I built a system that is able to work at 180°K without problems. The main components used are:

- RPI v3B+
- RPI camera night v2.1
- Silicon heater mats.
- DS18B20 temperature sensors.
- Insulation in stainless steel and polycarbonate.

The system is powered by cable and controlled externally by mouse/keyboard. Tests were made in liquid nitrogen.

Under 170°K, the capacitance of the capacitors in the board drops down too much and the RPI fails.
The camera goes even under 145°K without problems.

PS. Cool down and warm up the components slow to avoid cracks.

[jamesh]

It woudl be most interesing to know the use case for this. If its REALLY interesting, the RPF would probably like to help out.

[W. H. Heydt]

Hello,
I built a system that is able to work at 180°K without problems.

Under 170°K, the capacitance of the capacitors in the board drops down too much and the RPI fails.
The camera goes even under 145°K without problems.

Congrats...and thanks for letting us know.

[erpupones]

It would be most interesting to know the use case for this. If its REALLY interesting, the RPF would probably like to help out.

The project is the development of a mobile robotic system, a kind of submarine, which is able to move inside cryostats in which there is liquid at a very low temperature (-200°C). The experiment at Cern is ProtoDUNE, that's a prototype of the DUNE experiment in Fermilab, Chicago.

https://lbnf.fnal.gov/

https://www.youtube.com/watch?v=nv13DswIKr8

The purpose of this submarine is to allow the vision and even making reparations inside the cryostats.

I've thought to use a RPI as the "brain" of the system. The next step should be:

- modify the RPI to allow it working at LAr temperature (about -200°C)
- the other way is to keep the temperature gradient between the internal and external part of the robot and try to use it to power the device.

The other big problem is finding how to mechanically move the submarine.