Author Topic: Very high temperature (>150C) electronics  (Read 5774 times)

0 Members and 1 Guest are viewing this topic.

Offline AndyC_772Topic starter

  • Super Contributor
  • ***
  • Posts: 4208
  • Country: gb
  • Professional design engineer
    • Cawte Engineering | Reliable Electronics
Very high temperature (>150C) electronics
« on: October 20, 2017, 05:59:14 pm »
Many of the designs I do these days go in fairly inhospitable places in and around combustion engines. Typically they're subject to shock and vibration, and most important of all, the ambient temperature is very high. The maximum operating temperature is a major selling point for some products; the higher it is, the more applications open up for it.

Normal 'industrial' grade components are rated to 85C, which is no use at all in my market segment. An increasing range of components are available with temperature ratings up to 125C, and these are the parts in my everyday toolbox. It's no major problem; I can get analogue parts, power management, microcontrollers and FPGAs all with 125C ratings off the shelf.

A much, much smaller range of components (op-amps, some PICs) are commercially available rated to 150C. Some discrete semiconductors are available with 175C ratings.

Suppose, though, that I really - and I mean really - need to design an electronic device that can run hotter.

Does anyone have any experience in designing embedded circuits that include significant processing, and which can function at (say) 175C? Or 200C? Or higher?

I'm well aware that I'm into the realms of custom aerospace devices here. I know it'll be expensive, and that's OK.

What I need is to know who can supply the common building blocks (analogue circuits, CPUs, FPGAs etc) with exceptionally high temperature ratings - or, alternatively, if such components simply don't exist, what companies actually do instead.

Really high temperature cables, connectors, PCB substrate materials etc would be handy to know about too, but the first priority is active semiconductors.

All advice, especially from people who have been involved in such projects first hand, gratefully received!

Offline voltsandjolts

  • Supporter
  • ****
  • Posts: 2281
  • Country: gb
Re: Very high temperature (>150C) electronics
« Reply #1 on: October 20, 2017, 06:19:47 pm »
What operating life are you looking for at those temperatures?
If this is for automotive, do the electronics really have to be in such a hot spot?
TI / Honeywell / etc typ. 1000 Hrs at 200/225C but you quickly realise there ain't enough rated stuff for anything but the simplest of systems.
 

Offline JPortici

  • Super Contributor
  • ***
  • Posts: 3452
  • Country: it
Re: Very high temperature (>150C) electronics
« Reply #2 on: October 20, 2017, 06:28:21 pm »
I'd be interested in knowing of connectors, too.
I don't know how much it can help i know of ECUs that are small, built on aluminum substrate. Dice and components placed directly on the substrate and wire bonded, no packages. And of course the enclosure is just a big ass heatsink
 

Offline AndyC_772Topic starter

  • Super Contributor
  • ***
  • Posts: 4208
  • Country: gb
  • Professional design engineer
    • Cawte Engineering | Reliable Electronics
Re: Very high temperature (>150C) electronics
« Reply #3 on: October 20, 2017, 06:30:22 pm »
Depending on the application, anything from 100 hours to 'a few years' lifetime would be acceptable.

Putting the electronics somewhere cooler is the obvious option in many cases, and indeed, that's exactly what I do when possible.

There are, however, applications in which the PCB just has to go where it has to go. Making a board that'll run hotter opens up both technical possibilities and market opportunities.

I know this stuff must exist... down hole equipment for oil & gas, space, high energy physics, combustion engine R&D...

Offline voltsandjolts

  • Supporter
  • ****
  • Posts: 2281
  • Country: gb
Re: Very high temperature (>150C) electronics
« Reply #4 on: October 20, 2017, 06:46:35 pm »
Years at >150C is not feasible with conventional build and difficult with advanced hybrid packaging, IMHO.
Sourcing components is just the first step on a long path.
Simply keeping reliable connections is difficult in the long term, tin whiskers, electomigration, etc... an adventure in metallurgy awaits.
Quartzdyne / Sandia Labs have some interesting papers.
 

Offline Jeroen3

  • Super Contributor
  • ***
  • Posts: 4068
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: Very high temperature (>150C) electronics
« Reply #5 on: October 20, 2017, 06:52:29 pm »
You will have to provide active cooling solutions. This is typically how it's done when stuff must be located at the inhospitable places.
 

Offline chris_leyson

  • Super Contributor
  • ***
  • Posts: 1541
  • Country: wales
Re: Very high temperature (>150C) electronics
« Reply #6 on: October 20, 2017, 06:56:17 pm »
Quote
I know this stuff must exist... down hole equipment for oil & gas
The oil and gas guys use bipolar rather than mos for any analog signal conditioning and that's at around 200C. I guess the same would apply to any digital processing electronics and similarly ADCs and DACs.
« Last Edit: October 20, 2017, 07:01:42 pm by chris_leyson »
 

Online mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 13695
  • Country: gb
    • Mike's Electric Stuff
Re: Very high temperature (>150C) electronics
« Reply #7 on: October 20, 2017, 06:59:34 pm »
Digikey search for MCUs filtered by temp range brings up a few 215-220deg.C parts,at £200-600 a pop.
Looking at other categories like opamps and ADCs, there are quite a few >200C parts, mostly from TI or Honeywell, not much under £100 each.

If you're not too concerned about lifetime or maxing out specs it may be worth testing off-the-shelf parts to see how they behave.

Heat pipes might also be worth investigating. For short periods, maybe also phase-change materials.
« Last Edit: October 20, 2017, 07:03:19 pm by mikeselectricstuff »
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 

Offline texaspyro

  • Super Contributor
  • ***
  • Posts: 1407
Re: Very high temperature (>150C) electronics
« Reply #8 on: October 21, 2017, 02:29:09 am »
Back, in the last millennium, when I was designing some down-hole logging equipment, White Electronics was the go-to place for high temp electronics.  I think they were a spin-off of Sandia.

For fun, try building something that works at 200+ degrees C and 5000+ PSI in a corrosive environment.  Also borking a $10,000 multi-layer dewar is always a good way to start your day.

This is some useful reading:
jpt-electronics-components-aug-2013.pdf
 

Offline frog

  • Contributor
  • Posts: 42
  • Country: nz
Re: Very high temperature (>150C) electronics
« Reply #9 on: October 21, 2017, 03:56:02 am »
Possibly interesting but also possibly useless:
Long ago and far away I did an investigation at school (1982 I think) where I heated up a germanium transistor to see what would happen.  I found that the base-emitter drop reduced with temperature, but also the reverse breakdown of the collector-base reduced more or less at the same rate.  At about 150C the device pretty much conducted freely in all directions.  I guessed that the enough electrons were above the work function of the germanium that it was essentially a conductor.  Remarkably this was reversible and the device resumed normal operation as it cooled down.  from what I understand silicon does the same, but at a somewhat higher temperature (175C if you're lucky).  This would be an intrinsic property of silicon and so the only way to get a semiconductor working at a higher temperature would be to use a different material.

Here's a list of semiconductors:
https://en.wikipedia.org/wiki/List_of_semiconductor_materials
Uranium dioxide sounds like it would be a hoot to work with but lead selenide might be better at high temperatures
good luck
 

Offline Niklas

  • Frequent Contributor
  • **
  • Posts: 395
  • Country: se
Re: Very high temperature (>150C) electronics
« Reply #10 on: October 21, 2017, 08:59:39 am »
One of my colleagues works part time as a consultant with downhole electronics for oil and gas. The production series are small, typically between 1 and 5 pcs, so they build and qualify every single unit using a climate chamber. Some parts are used well outside of the specs in the datasheet, but if they pass the test they will be used. The operating temperature is usually high, but stable and without excessive thermo cycling, and the expected lifetime is of course shorter.
 

Offline max_torque

  • Super Contributor
  • ***
  • Posts: 1272
  • Country: gb
    • bitdynamics
Re: Very high temperature (>150C) electronics
« Reply #11 on: October 21, 2017, 02:25:29 pm »
What volume, what sign off, what warranty?


For low volume, expensive, products you can test every single unit off the end of the line, and use components out of their datasheet ratings.  At this point of course, YOU carry the risk of any failure, so you need to understand very closely what that actually is.  If we are not talking about life rated systems (ie that kill or injury someone when they fail) then you could use non rated, but suitably tested (you'd have to do a statistically valid amount of testing of course!) components.  Anything safety critical, er, nope, you  are going to have to use the suitably rated parts

Components have datasheet temp ratings, within which the function according to the datasheet parameters, ie for example, % error, or ppm drift etc.  Used outside of those ratings they don't generally stop working immediately, but they can (again, statistical analysis / testing required) exceed there maximum datasheet performance windows.


Unless you are forced to run electronics above say 125degC, ie by there being NO OTHER DESIGN OPTION, then i'd say it was always better to simply design to avoid those extremes of temperature (re-location, active cooling, insulation etc)
 

Offline Kalvin

  • Super Contributor
  • ***
  • Posts: 2145
  • Country: fi
  • Embedded SW/HW.
Re: Very high temperature (>150C) electronics
« Reply #12 on: October 21, 2017, 02:53:12 pm »
Why not using active cooling? Combustion engines use active cooling, why not the electronics if temperature control is needed.
 

Online mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 13695
  • Country: gb
    • Mike's Electric Stuff
Re: Very high temperature (>150C) electronics
« Reply #13 on: October 21, 2017, 06:08:55 pm »
Assuming your thing has wires attatched, adding some fluid flow in small tubes alongside the wires could get a lot of cooling for a pretty small tubes.
Heat pipes may be a passive alternative if the distance to somewhere cooler isn't far.
 
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 

Offline joeqsmith

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Very high temperature (>150C) electronics
« Reply #14 on: October 21, 2017, 06:26:03 pm »
You may have seen cooling plates tied into the engines cooling system. 

Offline Neomys Sapiens

  • Super Contributor
  • ***
  • Posts: 3268
  • Country: de
Re: Very high temperature (>150C) electronics
« Reply #15 on: October 23, 2017, 01:08:24 am »
Honeywell does Hi-Temp ICs, including OPAs, VRefs, Gate Arrays. Also Cissoid.
 

Offline AndyC_772Topic starter

  • Super Contributor
  • ***
  • Posts: 4208
  • Country: gb
  • Professional design engineer
    • Cawte Engineering | Reliable Electronics
Re: Very high temperature (>150C) electronics
« Reply #16 on: October 23, 2017, 05:14:36 pm »
Some good leads here, thanks very much everyone.

Looking at the portfolio of available parts, it's clear that:

- yes, it is possible to run active electronics at 225C, or hotter
- their technical performance will be strictly limited, though.

As I suspected, the best option looks to be to put conventional electronics somewhere cooler, if possible. It's the "if possible" caveat that's the real issue here, of course.

Offline max_torque

  • Super Contributor
  • ***
  • Posts: 1272
  • Country: gb
    • bitdynamics
Re: Very high temperature (>150C) electronics
« Reply #17 on: October 23, 2017, 07:43:04 pm »
About 15 years ago there was a push by the car OEMs to mount engine control units directly onto engines.  This turned out to be a bit of a 'mare, and these days, they have moved back to "remote" units, using serial data links to reduce looming complexity, and housing most controllers in a ventilated (often forced fan ventilated to protect against hot soak) "Ebox" usually made of thermoplastic and mounted up under the w'screen scuttle somewhere typically.

For mass production, the extra cost of parts, especially long life high temp caps, mean't it just was sensible to design for above 125degC, despite the extra unavoidable costs from the remote location of such devices from their sensors and actuators etc
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 14080
  • Country: de
Re: Very high temperature (>150C) electronics
« Reply #18 on: October 23, 2017, 08:26:42 pm »
If it has to be really high temperature higher bandgap semiconductors are the way to go.

https://sic.grc.nasa.gov/files/IEEEProceedings6-02.pdf

Nothing to buy at the normal distributors though.
 

Offline cdev

  • Super Contributor
  • ***
  • !
  • Posts: 7350
  • Country: 00
Re: Very high temperature (>150C) electronics
« Reply #19 on: October 23, 2017, 08:42:41 pm »
Electronics seem to fare much better than people do in high temperatures!
"What the large print giveth, the small print taketh away."
 
The following users thanked this post: NiHaoMike

Offline m98

  • Frequent Contributor
  • **
  • Posts: 613
  • Country: de
Re: Very high temperature (>150C) electronics
« Reply #20 on: October 23, 2017, 09:56:48 pm »
Besides active liquid cooling, I'd suggest using aerogel insulation.
For real expert consultancy, testing and parts procurement, you could try getting in touch with TESAT Spacecom.
 

Offline Kalvin

  • Super Contributor
  • ***
  • Posts: 2145
  • Country: fi
  • Embedded SW/HW.
Re: Very high temperature (>150C) electronics
« Reply #21 on: February 23, 2018, 03:09:57 pm »
Old topic, but this new article "Tackling high-temperature data acquisition and processing platform development" covers some aspects of high-temperature embedded design up to 200 Celcius.

https://www.embedded.com/electronics-news/4460345/Tackling-high-temperature-data-acquisition-and-processing-platform-development

- The VA10800 is an ARM© Cortex©-M0 based microcontroller that operates at 50MHz.
- Polyimide PCB board.

More details in the article.
 

Offline AndyC_772Topic starter

  • Super Contributor
  • ***
  • Posts: 4208
  • Country: gb
  • Professional design engineer
    • Cawte Engineering | Reliable Electronics
Re: Very high temperature (>150C) electronics
« Reply #22 on: February 23, 2018, 05:16:07 pm »
Old, but still absolutely relevant, and thanks for the link.

Offline coppice

  • Super Contributor
  • ***
  • Posts: 8605
  • Country: gb
Re: Very high temperature (>150C) electronics
« Reply #23 on: February 23, 2018, 05:38:37 pm »
Old topic, but this new article "Tackling high-temperature data acquisition and processing platform development" covers some aspects of high-temperature embedded design up to 200 Celcius.

https://www.embedded.com/electronics-news/4460345/Tackling-high-temperature-data-acquisition-and-processing-platform-development

- The VA10800 is an ARM© Cortex©-M0 based microcontroller that operates at 50MHz.
- Polyimide PCB board.

More details in the article.
For those with deep pockets there have been 200C MCUs around for a while, but 200C flash has been a problem. It looks like there are 200C rated  SPI flash chips now, but I couldn't download the datasheet to see what their downsides might be.
 

Offline GreggD

  • Regular Contributor
  • *
  • Posts: 136
  • Country: us
Re: Very high temperature (>150C) electronics
« Reply #24 on: February 23, 2018, 06:43:57 pm »
While investigating a large ceramic 0.1uF cap on the Rode NT-1A microphone I speculated it could be NP0 (COG).
Not for the prevention of piezoelectric noise as the diaphragm will pick that up but so that it will still be 0.1uF with 70-90V on it.
This seams ok as some in a 1812 package can be had for $0.40 at 2000.

At the bottom of my Digikey search for these showed 399-5748-1-ND, "CAP CER 0.1UF 100V C0G/NP0 1210 " -55°C ~ 200°C Down-hole rated. For the bargain price of $19 each or $10.45 at Qu 2000.

Interesting data sheet.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf