2

I'm relatively (ok, very) new to electronics. I'm trying to find where to buy parts (or even just more information on what sort of materials I should be looking for) to build a circuit board with a micro controller and RF transmitter that can withstand being exposed to a hot, dry environment. The upper end of the temperature range would be 200C (~392F), and would be need to be able to operate under constant exposure to this temperature for 8-12 hours. Ultimately, it would need to be fairly small (we're hoping to keep it smaller than the size of a couple 9V batteries stacked together)

Not being a traditionally educated EE or material scientist, I'm very much out of my element here - referrals to other sources of information, web sites, etc., so that I can at least get an idea of direction would be tremendously helpful.

I would also be interested to hear more information about thermal protection methods - would there be a way that I could build an enclosure to protect the internal electronics from the same environment. I would think that no matter the sort of "housing" I could build, without an active cooling system (which is all but prohibited given the target form factor), then the electronics themselves would need to be able to tolerate the temperature - but I would certainly love it if someone could correct my naive assumptions here.

Thanks for any advice/help.

Blup1980
  • 6,120
  • 3
  • 26
  • 47
Brandon
  • 123
  • 4
  • 2
    200C is *at least* space grade gear. – Ignacio Vazquez-Abrams Dec 13 '13 at 05:20
  • 4
    Your average semiconductor will melt at around 150C and many components are spec'd up to some 80C, so you have to find a way to keep thing cool, well below 80C. – jippie Dec 13 '13 at 06:19
  • Can you provide more details on what you are trying to do, most, if not all, semiconductors will not survive 200C. Perhaps most of the circuit can be kept cooler and only a small sensor needs to be exposed to this temperature? If not you are going to need some form of active cooling. – Warren Hill Dec 13 '13 at 07:04
  • 1
    You may find this site of some use: http://www.extremetemperatureelectronics.com/ – Warren Hill Dec 13 '13 at 07:16
  • 2
    As others say, most regular parts are specced to 80C, Automotive/Military grade gets you to 125C, after that you're into "if you have to ask" spec. However, it's very unlikely you've hit upon something that hasn't been done before, so I'd look at any other devices operating in similar conditions and see what they're doing. – John U Dec 13 '13 at 09:08
  • What are you trying to do? – Lior Bilia Dec 13 '13 at 09:31
  • What is your power budget? Do you have an amp or two to spare at 12V? Or is there a source of lower temperatures nearby - e.g. this is in a furnace? –  Dec 13 '13 at 11:02
  • 2
    I've managed to get 120degC out of a small module incorporating a PIC and a low power FM transmitter - transmission was starting to get flaky though (above 110C). – Andy aka Dec 13 '13 at 12:13
  • 1
    @jippie You may have been using hyperbole, but I just wanted to clarify that semiconductors most certainly will not _melt_ at 150C. Electronic systems used in deep boreholes operate at temperatures up to 200C. If you meant that average electronics will not function reliably at 150C then I agree. – Joe Hass Dec 13 '13 at 13:02
  • To put things into perspective, your standard 63/37 _solder_ will melt at 183 C. – Kevin Chen Dec 13 '13 at 17:44
  • Thanks all for the information. I can't really discuss what we're trying to accomplish, but you've all really clarified to me exactly how far "out of my element" I am getting. Time to start talking to specialist design firms if we're to continue much further down this path - as pjc50 suggests. Thanks for all the comments and information. – Brandon Dec 13 '13 at 19:51
  • You mention batteries for a size comparison; but you aren't actually going to have 2 9V batteries in that environment, right? – Kaz Dec 13 '13 at 21:19
  • separate the sensor from the mcu – geometrikal Dec 14 '13 at 11:55

3 Answers3

6

This is well into the territory of "hire a specialist design firm". It's going to be expensive, so you might as well do it properly and hire people who know where all the pitfalls are and where to get all the parts.

TI have some high temperature microcontrollers, and a design guide linked from that page. The SM320F28335GBS is good up to 210 degrees (which is cutting it a bit close). While you can buy from Digikey in single quantities, they cost $300.

The other approach you could take if the operating time is limited to a few hours in that environment and the box can be sealed with no wires in or out, is to insulate it very heavily. Add some thermal mass, preferably with a phase change. Do not exceed the stated cooking time.

You'll end up with a something like a meter cube of fiberglass or foam insulation containing a sealed shoebox of methanol with the electronics floating in it, which will stay below 70C until it boils dry, hopefully for long enough for whatever your heating cycle is. At which point you need to cool it down again equally slowly. I think that's the only viable strategy for normal temperature electronics which doesn't involve active cooling.

Edit: normal batteries do not like high temperatures either, although there are special high-temperature molten-salt ones used in some missiles.

pjc50
  • 46,540
  • 4
  • 64
  • 126
3

Most commercial ICs and components have heat ratings, especially resistors and capacitors. You might have to look at the data sheets to find them, though. As for an enclosure, anything that is heavy and metal will act as a heatsink to dissipate excess heat, but if the circuit is to be used in a hot environment this won't help much.

None of this will really help with up to 200˚ though. With that, you have 2 options - custom parts (very expensive and slow) or protecting the entire setup from the heat (bulky). With protection, the idea is to add shielding on the outside (foil, or higher-tech materials) and inside that, as much insulation as possible. This only works if the circuit is only to be subject to heat for a limited time. Also, this setup keeps heat in as well as out, so design your circuit to run as cool as possible.

felixphew
  • 445
  • 7
  • 16
1

Here is a paper by Honeywell on IC development for -55 to 250 but it outlines some of the challenges you will definitely face design for that operating range. The silicon on insulator technology will give operating range of -55 to 250C. However I'm not aware of how broad the is the availability of devices built with this technology. Also at your maximum of 200C just regular lead solder or unleaded solder will probably start to melt especially after 8-12 hrs of operation so you will definitely be looking at different soldering methods

Kvegaoro
  • 3,141
  • 17
  • 14
  • I had come across this paper before, but being the neophyte that I am in this area, wasn't sure I understood everything I was reading. Thanks for posting it - definitely appreciated. – Brandon Dec 13 '13 at 19:50