High-Temp Electronics For Venus Exploration, recent advances Options

[dtolman]

(MOD NOTE: Started a new topic for this discussion to continue. Please remember the 'no sci-fi engineering' provision of rule 1.9. Have fun!)


Also, since I'm thinking about surface operations on Venus, the state-of-the-art in high temperature electronics has advanced quite far in the past decade.
Its now possible to buy off the shelf chips from vendors designed to operate at the 250-300 C range.

Meanwhile basic functionality has been tested at and beyond the temperatures needed for long-term surface operations on Venus:
http://www.grc.nasa.gov/WWW/SiC/
http://www.gizmag.com/extreme-silicon-carb...ctronics/16410/
http://www.grc.nasa.gov/WWW/SiC/publicatio...Contact2010.pdf

Another decade or so and a long-term Venus lander could be possible with (practically) off the shelf electronics!

[tasp]

Having electronics operate at a higher temperature works 2 ways to make for a longer operating time on the surface.

Firstly, it takes longer to heat them up more, and also, the rate of heat flow into the probe decreases as the temperature difference between inside and outside decreases.

Additionally, having a wider temperature operating range might allow the use of different/better phase change heat absorbents.

We could be seeing quite a jump in operating time that's possible.

[Robotbeat]

Of course, if you can operate at 500C, you don't need to have phase/change...

Resistors, Capacitors, inductors, all those sorts of things can work at 500C, though not the usual components. A resistor is fundamentally a simple device, as are capacitors and inductors... They can technically be constructed entirely with just some conductor and some insulator material.

Now that we have silicon carbide ICs of tens of gates operating at 500C, lots of stuff is possible. A turing-complete computer (provided you have some sort of memory, could be coil or capacitor based) can be made with just a couple hundred gates, though you have to be satisfied with very low performance (maybe just a 4-bit or 8-bit computer, but it's Turing Complete so can simulate higher). Whether it's worth it...

[hendric]

Now that we have silicon carbide ICs of tens of gates operating at 500C, lots of stuff is possible.

http://spectrum.ieee.org/semiconductors/de...ng-temperatures

has some interesting details. Problem now isn't the semiconductor per se, but everything around it such as interconnect between the transistors (metal layers) and packaging etc.

[stevesliva]

In addition to high-temp industrial applications, I wonder whether technologies like this could eventually avoid the need for ridiculous active cooling of server farms. I think the ideal would be servers that reliably operate around 100C, so that you could simply cogenerate (if that is the right term) by circulating water and creating steam. Right now 40C is the typical limit.

[nprev]

Silly question: How are they doing surface-mount boards? My first guess is silver solder, which is what is used on aircraft engine thermocouples, but curious to see if a better technology has arisen.

[siravan]

I wonder about sources of power. Is there any high temperature solar panel? Also, there is not much light on the venus surface. I remember Venera description of "same as noon in a winter day in Moscow", which is probably not a whole lot. RTG is worse. To generate any electricity using a thermocouple, one needs a temperature gradient, which would be extremely hard to maintain on venus.

[ElkGroveDan]

Silly question: How are they doing surface-mount boards?

Tungsten solder, Nick. I have a patent pending.

[nprev]

I'm officially your distant relative now, Dan, and I got a few bills...

Seriously, this is not a trivial problem. A Venus lander of any sort either requires a VERY well-controlled temperature environment for its electronics (which IIRC was the Venera strategy) or not only electronic components that can survive extreme temp swings during the cruise phase but also the same for all required electrical connections.

The latter are obviously quite vulnerable to materiel expansion/contraction cycles, which ultimately loosen connections over time and introduce either spurious or high-resistance interface points which can do all sorts of nasty things to signal flow. (For example, MIL-STD-1553 data buses really don't like impedance changes; tends to turn perfectly good data words into gibberish.)

How I see this playing out is a loosening of environmental control constraints over time for the silicon itself but capability bounded by a hard constraint on connection methods unless suitable (and workable, as Dan pointed out) connection/wiring alloys of some sort can be developed that are extraordinarily resistant to thermal cycling. That's a rather daunting challenge in metallurgy.

[mcaplinger]

The latter are obviously quite vulnerable to materiel expansion/contraction cycles...

The cruise environment is fairly benign and once you get to Venus it's just hot all the time, there isn't a lot of diurnal variation. Exactly what the interconnect is made of is another question. That said, the SiC components are at a very early level of development so I'd say we are quite a ways away from having real and capable systems that would work at Venus ambient conditions.

RTG-powered refrigeration systems are feasible:
http://web.archive.org/web/20010106124800/...p/vgnp.txt.html

but for the foreseeable future I think short-lived landers are all we are likely to see, unfortunately.

[Explorer1]

If one is willing to constrain the landing site by altitude the engineering challenge can probably be lessened as well. Maxwell Montes is almost 100 C cooler that the mean; any lander would last a bit longer there and at the other high points.

[ElkGroveDan]

Seriously, this is not a trivial problem. A Venus lander of any sort either requires a VERY well-controlled temperature environment for its electronics (which IIRC was the Venera strategy) or not only electronic components that can survive extreme temp swings during the cruise phase but also the same for all required electrical connections.

It's a real problem but I'm confident that existent materials technology can solve it -- especially the problem of conductive connections. I would imagine a mechanically secure solution such as clamping or spiral threading which is secondarily secured by a high temperature polymer. (If I were building it in my garage I'd be using RTV silicon or two part-epoxy which are reliable to 600F. I'm sure the JPL toolbox has even better stuff in it).

The electronic components are another story but frankly I'm pretty certain there are already temperature rated ICs and such in use in military and industrial applications which offer a starting point for this kind of high-temperature electronic circuit design -- probably servos, relays and optical applications as well.

[mcaplinger]

(If I were building it in my garage I'd be using RTV silicon or two part-epoxy which are reliable to 600F.... frankly I'm pretty certain there are already temperature rated ICs and such in use in military and industrial applications which offer a starting point for this kind of high-temperature electronic circuit design...

Venus ambient is around 450C or almost 900F. There are no ICs I'm aware of that get anywhere close. Most mil-spec parts go to 150C or maybe 200C in a few cases (junction temps, not ambient.)

[siravan]

One small thing in favor of high temperature electronics for Venus is that, while very high, the temperature is relatively stable. You don't need components that works at both 20 C and 450 C. They only need to work at 450+/-50 C, which is probably somewhat easier task. The lander can have a separate set of electronics for the cruise and native high temperature ones after landing.

[mcaplinger]

You don't need components that works at both 20 C and 450 C.

That would enormously simplify testing, though.

For our Venus proposals we spent a fair amount of time worrying about how we could even simulate the environment in any practical way. How do you put a scope probe on something that has to be at 450C to work?