4 ms·

Can Electronics made of COTS components work at cryogenic temperatures? (2020)

31 points by agarttha 5y ago

weinzierl 5y ago
As a young engineer working in aerospace I found it both surprising and remarkable that the satellites we worked with were all designed to operate at and around room temperature [1]. I could understand that you want to avoid temperature changes as much as possible to avoid stress on the components. But why settle on an arbitrary temperature that is comfortable for humans? Of course I learnt quickly that a room temperature design makes a lot of things much easier;-) [1] An interesting tidbit is that in certain fields room temperature is used as a technical term to really mean a well defined temperature, and that there is a difference between North American (25 °C) and the European (20 °C) convention.
- comboy 5y ago
  Clueless questions: 1. Assuming it's easy to keep 20C (I thought it's problematic to keep such a low temp for satellite which has problem radiating heat away), what about transfer, is electronics not yet operating back then or do you use some heating elements specifically for bootstrap? 2. Just to be clear 25C is what is used as room temp in electronics and not something you set your thermostat too at home, right? 3. (easily googlable feel free to skip) - how does heat management look for satelatie electronics? Are peltiers used? Do you just put everything in some heat conductive material? Is there some gas sealed around PCB or is it vacuum? In case of the latter and if everything is not surrounded by some heat conductive substance how do you deal with thermal runaway? Is maybe PCB itself and paths enough to make it less bad than I imagine? Thanks a lot, I'm sorry about so many questions, just got curious.
  - souprock 5y ago
    25C is something you set your thermostat to at home. Well, it would be 77F, which is the same. (note: Florida Man Sets Thermostat)
- madengr 5y ago
  I have designed some satellite microwave PA, and the thermal ranges were -20C to +70C, but yes, not even an industrial temperature range. The biggest issue is getting heat out, though my module was on a cold plate and I doubt it was ever cycled to those ranges. The concern was keeping output power and current within spec over those ranges; getting enough output power (within distortion specs) at hot, and keeping the current under spec at cold (so as not to trip a breaker).
a9h74j 5y ago
The link does not resolve for me. I recall discussions and assumptions that commercial CMOS could be provisionally expected to work to cryogenic temperatures. I have only routinely qualified to -40, and many +85C parts for excursions to 105C.
the-dude 5y ago
Service Unavailable The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later. Additionally, a 503 Service Unavailable error was encountered while trying to use an ErrorDocument to handle the request.
- amelius 5y ago
  It's not easy to run a website at cryogenic temperatures.
onelectrontech 5y ago
Author here. Looks like this page got quite a bit of attention—and unfortunately our server couldn’t handle the load. I’m here to answer any questions while you wait for our website to respond!
- mordechai9000 5y ago
  Dumb question, since I haven't read the article. What constitutes cryogenic temperatures? I worked for a telecom that deployed a lot of equipment at high latitude, in places where temperatures could reach -40 or colder. It wasn't my department, but I had the impression that we weren't really paying too much attention to the lower end of the operating range specified by manufacturers. I recall seeing "ice in the feed horn" as a root cause more than once.
  - detaro 5y ago
    A common definition of cryogenic is "below 120K/-153°C", or more general temperatures where "simple" gasses like oxygen/nitrogen/... become liquids.
    - onelectrontech 5y ago
      That sounds reasonable. I didn't have a specific range in mind, other than "significantly below the minimum specified operating temperature of conventional electronics". The article was mostly focused on space applications, so 120K seems like a good upper bound for extreme cold in these ranges.
- dvh 5y ago
  Is the answer "no"?
  - onelectrontech 5y ago
    It depends--This doesn't feel like a very satisfying answer, but certain devices have been shown to perform better at extreme low temperatures (up to a point). However, there is the ever-present issue of reliability. For instance, certain CMOS devices exhibit improved switching characteristics, but the increased hot carrier mobility can cause degraded reliability, so certain design trade-offs have to be made. Of course, many COTS components specify much higher minimum operating temperatures in their datasheets, so thoroughly characterizing these components is vital for high-reliability operation in cryogenic conditions.
ncmncm 5y ago
Anything works at cryogenic temperatures if you heat it so it isn't. Thus, a big part of the Mars probes', including the Ingenuity micro-helicopter's, power budgets is keeping them warm enough overnight. The Voyager spaceprobes will be shut down soon not because they can't produce enough signal to reach Earth antenna arrays, but because they can't keep the equipment warm enough to work right. It would probably be possible to make stuff that would work cold, but there aren't enough places to use it to support the whole process.
raziel2701 5y ago
I'm getting a 404, what does COTS mean?
- ggreer 5y ago
  Commercial off the shelf. [1] 1. https://en.wikipedia.org/wiki/Commercial_off-the-shelf https://en.wikipedia.org/wiki/Commercial_off-the-shelf
agarttha 5y ago
In our physics research lab we use some COTS components down to 10 mK. Many passive components no longer work and we have to use C0G capacitors and NiCr resistors. Only some discrete SiGe and HEMT transistors will continue to work. But then our thermal budget is 50 uW so design is difficult even if the components remain operational.
agarttha 5y ago
Spacecraft can use https://en.m.wikipedia.org/wiki/Radioisotope_thermoelectric_generator https://en.m.wikipedia.org/wiki/Radioisotope_thermoelectric_... to heat up electronics.

31 points by agarttha 5y ago

As a young engineer working in aerospace I found it both surprising and remarkable that the satellites we worked with were all designed to operate at and around room temperature [1]. I could understand that you want to avoid temperature changes as much as possible to avoid stress on the components. But why settle on an arbitrary temperature that is comfortable for humans? Of course I learnt quickly that a room temperature design makes a lot of things much easier;-) [1] An interesting tidbit is that in certain fields room temperature is used as a technical term to really mean a well defined temperature, and that there is a difference between North American (25 °C) and the European (20 °C) convention.

comboy 5y ago
Clueless questions: 1. Assuming it's easy to keep 20C (I thought it's problematic to keep such a low temp for satellite which has problem radiating heat away), what about transfer, is electronics not yet operating back then or do you use some heating elements specifically for bootstrap? 2. Just to be clear 25C is what is used as room temp in electronics and not something you set your thermostat too at home, right? 3. (easily googlable feel free to skip) - how does heat management look for satelatie electronics? Are peltiers used? Do you just put everything in some heat conductive material? Is there some gas sealed around PCB or is it vacuum? In case of the latter and if everything is not surrounded by some heat conductive substance how do you deal with thermal runaway? Is maybe PCB itself and paths enough to make it less bad than I imagine? Thanks a lot, I'm sorry about so many questions, just got curious.
- souprock 5y ago
  25C is something you set your thermostat to at home. Well, it would be 77F, which is the same. (note: Florida Man Sets Thermostat)
madengr 5y ago
I have designed some satellite microwave PA, and the thermal ranges were -20C to +70C, but yes, not even an industrial temperature range. The biggest issue is getting heat out, though my module was on a cold plate and I doubt it was ever cycled to those ranges. The concern was keeping output power and current within spec over those ranges; getting enough output power (within distortion specs) at hot, and keeping the current under spec at cold (so as not to trip a breaker).

a9h74j 5y ago

The link does not resolve for me. I recall discussions and assumptions that commercial CMOS could be provisionally expected to work to cryogenic temperatures. I have only routinely qualified to -40, and many +85C parts for excursions to 105C.

the-dude 5y ago

Service Unavailable The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later. Additionally, a 503 Service Unavailable error was encountered while trying to use an ErrorDocument to handle the request.

amelius 5y ago
It's not easy to run a website at cryogenic temperatures.

onelectrontech 5y ago

Author here. Looks like this page got quite a bit of attention—and unfortunately our server couldn’t handle the load. I’m here to answer any questions while you wait for our website to respond!

mordechai9000 5y ago
Dumb question, since I haven't read the article. What constitutes cryogenic temperatures? I worked for a telecom that deployed a lot of equipment at high latitude, in places where temperatures could reach -40 or colder. It wasn't my department, but I had the impression that we weren't really paying too much attention to the lower end of the operating range specified by manufacturers. I recall seeing "ice in the feed horn" as a root cause more than once.
- detaro 5y ago
  A common definition of cryogenic is "below 120K/-153°C", or more general temperatures where "simple" gasses like oxygen/nitrogen/... become liquids.
  - onelectrontech 5y ago
    That sounds reasonable. I didn't have a specific range in mind, other than "significantly below the minimum specified operating temperature of conventional electronics". The article was mostly focused on space applications, so 120K seems like a good upper bound for extreme cold in these ranges.
dvh 5y ago
Is the answer "no"?
- onelectrontech 5y ago
  It depends--This doesn't feel like a very satisfying answer, but certain devices have been shown to perform better at extreme low temperatures (up to a point). However, there is the ever-present issue of reliability. For instance, certain CMOS devices exhibit improved switching characteristics, but the increased hot carrier mobility can cause degraded reliability, so certain design trade-offs have to be made. Of course, many COTS components specify much higher minimum operating temperatures in their datasheets, so thoroughly characterizing these components is vital for high-reliability operation in cryogenic conditions.

ncmncm 5y ago

Anything works at cryogenic temperatures if you heat it so it isn't. Thus, a big part of the Mars probes', including the Ingenuity micro-helicopter's, power budgets is keeping them warm enough overnight. The Voyager spaceprobes will be shut down soon not because they can't produce enough signal to reach Earth antenna arrays, but because they can't keep the equipment warm enough to work right. It would probably be possible to make stuff that would work cold, but there aren't enough places to use it to support the whole process.

raziel2701 5y ago

I'm getting a 404, what does COTS mean?

ggreer 5y ago
Commercial off the shelf. [1] 1. https://en.wikipedia.org/wiki/Commercial_off-the-shelf https://en.wikipedia.org/wiki/Commercial_off-the-shelf

agarttha 5y ago

In our physics research lab we use some COTS components down to 10 mK. Many passive components no longer work and we have to use C0G capacitors and NiCr resistors. Only some discrete SiGe and HEMT transistors will continue to work. But then our thermal budget is 50 uW so design is difficult even if the components remain operational.

agarttha 5y ago

Spacecraft can use https://en.m.wikipedia.org/wiki/Radioisotope_thermoelectric_generator https://en.m.wikipedia.org/wiki/Radioisotope_thermoelectric_... to heat up electronics.