PIC Low temperature performance?

[MikeMl]

I am building a RDF transmitter for a high altitude balloon. I plan to use a PIC to control when it transmits (timing), to generate a MCW ID, to switch power levels, etc. During the early phase of the flight, the transmitter will not be active, but the PIC has to keep track of time in order to turn on the transmitter about two hours after the launch.

I'm worried about the PIC hanging up/resetting at the extreme cold of near space. The temperature gets to -60degC, which is 20degC below the -40degC temp range specified by Microchip. While extremely cold, all the PIC needs to do is keep the timing going. As the balloon descends, things warm up above 0degC, so only the timing function has to keep running.

What are the pitfalls? A separate temperature-based watchdog (i.e. reset the PIC when the temperature rises)?

 

[Pommie]

Does it matter if it resets? As long as you have the WDT running it should start up again once warm. If you need the time to be accurate then I assume you have GPS on board and you can sync with that.

Another alternative would be a heating resistor. To save power, you could power it from 2 pins and only turn it on when it's getting too cold. You would of course have to insulate it.

Actually, I just had a thought, would the actual program survive those temperatures? Tales of resetting car radios by freezing come to mind.

Mike.

 

[BeerBelly]

It sounds like a fun adventure.
Why did you decide not to use an insulated box?

 

[Mr RB]

I would stay away from mechanical oscillators like xtal or resonator and go for the internal osc or external RC osc. And like Pommie said, use the WDT and the newer PICs have clock failsafe detect etc.

The big problem at such low temps might be your batteries!

Do you have a flight recording from another balloon? How long is the thing dwelling at the high altitudes? If it only for some minutes you can probably just insulate the whole electronics compartment, and/or add some small heating resistors on the PIC, GPS and battery.

 

[ronsimpson]

I have worked outside at -40C all day and -50C for hours.
I noticed problems with capacitors at -40C. The ele. capacitors change value. You need to look at all your capacitors. The voltage regulators may oscillate because the capacitors became too small. Resistors are not much of a problem. I don't know why silicon would fail. The internal RC oscillator in the PIC will be off frequency a little. Look at the temperature curves on the PIC data sheet. See if the curves turn up/down at the low end. If the curve starts up from -30 to -40 then you can predict where the curve will go from -40 to -60. Look for internal clock frequency and max clock frequency. Just because PIC has not tested the parts at -60 does not mean they won't work there.

 

[languer]

I have worked outside at -40C all day and -50C for hours.

I thought you were going to follow this with: "and I am still kicking and have not reset", or some smart comment like that.

Why did you decide not to use an insulated box?

This is really an excellent question. Most of the projects I've seen to weather balloons, or "near-space" projects use some sort of insulation for this very purpose. Similar to what has been said, I've never seen silicon go bad at very cold temperatures (capacitors yes), but I have seen bad things happen when uC or uP get very cold and code operation becomes erratic and all sort of coold things can happen; thing the code never intended to - but because the uC/uP operation is unknown in these states, all bets are off. I mean I've seen WDTs go out, yet some code still running, in unpredictible ways. In very rare cases, to prevent that (in some products) we disable (i.e. remove power) from "intelligent" circuits if the temperature drops below our test ranges (somehow you can always trust those darn OPAMPs and transistors).

 

[languer]

Forgot to add that this is one of my favorite columns from Nuts&Volts: Near Space by Paul Verhage

 

[MikeMl]

The entire RDF beacon (batteries, controller, transmitter, antenna) will be packaged into a "hula hoop" (about 1.5m dia) made out of 1" PEX tubing. This hoop is what the shrouds of the parachute are tied to. It is used to hold the parachute canopy open so that the parachute catches air as soon as the balloon falls into sufficiently dense air to provide drag. The hoop goes along for the ride on every launch. This way the RDF beacon doesn't occupy a "spot" that would normally be used for student-built payloads. The hoop is visible in **broken link removed**. The wall of the tubing is only about 1mm thick, so it is not going to provide much insulation.

My goal is to make this beacon as robust and simple as possible. The other "trackers" are too complex, they are dependent on a line-of-sight radio path to an APRS digi-peater, GPS lock, etc.

I want to make mine dirt simple, no GPS, no external dependencies on LOS to a digi-peater, etc, simple primary batteries. One idea is to have the PIC execute code with counting loops for the timing so that the chip self heats, rather than shutting down then cpu and using the super low power timers. I could put a little bit of foam insulation around the PIC.

Which types of capacitors do best a extremely low temps?

 

[languer]

That is a pretty cool idea. Could you wrap the hoolahoop thingy on some material which helps with a little insulation? I'm thinking wool, but that's not too static friendly.

Ceramic caps are quite reliable over quite a range of temperature extremes. Electrolytics do not have good performance at extremes (either because of venting, or freezing).

 

[nsaspook]

A limiting factor for PIC operation at below -40 is the flash (Electrically Re-programmable) operation. Some people have used the OTP version of the PIC16C7X at -65 in some milspec products without any problems. YMMV

 

[BeerBelly]

Coloring the hoop black may warm it some. Spray in foam like they use around windows and doors, Maybe a surface mount resistor on the surface of the PIC. The timer in the PIC could be used to supply power to the resistor when you expect it will be needed

 

[languer]

This peaked my interest, so I looked at some of the old articles I had on the BallonSat thing and visited Paul Verhage's webpage. The lowest outside temperature he recorded on the last flights was around -55C; this is really not that far off from the temperature rating of the PICs. So even though I would thread carefully with the capacitor selection, you could still provide some reasonable insulation to the PIC and keep the PIC working. You could even have the "heater" resistor just in case. You could (perhaps) place the circuit inside a ziplock bag and then fill the area around it with some insulation material (think roof or window/door). I would do some experiments inside the freezer. For example, cut a piece of the hula-hoop large enough to fit the circuit, insert the circuit, add whatever insulation material you come-up with, close the ends, and measure temperature differential in the circuit and outside of the hula-hoop (i.e. freezer).

 

[large_ghostman]

we did some cold temp experiments (not micro's) we were fault findinding,dad borrowed a co2 cylinder from the pub, if you turn it upside down and open the valve with a couple layers cotton cloth over the nozzle you get dry ice (-78c) maybe you could use that for testing?

hi from logan