picaxe stability at low temperatures

[alistairsam]

Hi,

I've completed a few projects where the circuit board is housed in a plastic box and needs to operate outdoors at close to 0 deg C ambient with medium to high humidity and have found that they behave erronously as the temperature drops.
I had a read of the microchip datasheet and the max ambient range is -40C to +125C, so it is well within tolerance.
Is this normal or expected and if so would I need to maybe insulate the box?
came across this thread as well http://www.picaxeforum.co.uk/showthread.php?12932-PICAXE-stability-vs-operating-temperature

I'm using 28x2's, 20x2's and 20m2's and they've all operated perfectly indoors on my work bench but when I use it outdoors, it starts behaving erratically after an hour or so. eg, a 28x2 as a simple pulse generator for a stepper controller, works perfectly inside, but outside, starts pulsing erratically or stops after an hour.
another 20m2 as a dew controller driving FET's, same thing, stopped pulsing (not using pwmout), but started after a power reset and then stopped after a while.
I'm using them with telescopes, so they're mounted in a jiffy box that's not airtight, but doesn't get damp inside even if there's heavy dew.
any suggestions?
What should I test?

power supply is via a 5v voltage regulator. I've used a 9v DC adaptor as well as a 9v battery.

 

[Buzby]

Hi Alistair,

A few years ago I did some low temperature tests on a 20X2, in a fridge and a freezer.
The internal resonator frequency drifted, but not so much that the chip stopped working.

As you say, 0'C is well within the spec, both for the PIC and just about every other electronic component you are likely to be using.

It is far more likely that the problem lies elswhere.

Damp, although you you can't see it, could have a significant effect, especially if you have an external resonator or crystal.

Battery performance drops off dramatically as the temperature falls, and any high current loads could cause significant voltage droop.

Even mechanical contraction can affect poorly fitting plugs and sockets.

Of course, all this assumes you have built the boards to a good standard. ( Especially decoupling and bulk capacitors, unused inputs tied somewhere, and the PICAXE download circuit. )

That's about all I can think of.

EDIT : Looking at your earlier posts, are you certain it's not noise pickup in the long cables ?
It's probably best if you tell us a bit more about the different environments 'on the bench' and 'in the field'.

Cheers,

Buzby

 

[MFB]

I also think the problem lies elswhere. Most probably in the battery because its internal resistance increases when the temperature drops. As a quick bit of diagnostics, try increasing the amount of decoupling on the PICAXE power supply lines (but be aware that the characteristics of electrolytic capacitors also change with temperature).

 

[bfgstew]

Just add my two pennies worth.

Any electronic cabinets we have at work due to the enviroment (between +12 and -20 deg C) all have small heaters in to stop condensation building up and giving the components a controlled atmosphere to operate in, just a thought and am sure easy enough to pick up off the net.

Stewart

 

[Eclectica]

Hi alistairsam,

IMHO if you want to eliminate the moisture issues:

I have had various 08x2 based devices working quite happily in a non-heated glasshouse for years. The boxes were rated at IP67, cables entered via IP68 glands (rubber inserts of the right size for the cables!), and had sachets of dry silica gel in the box too.

There was a single failure in one of the 6 boxes after 2 years, and this was due to the seal breaking down and allowing the box to 'breath' moist air. This resulted in moisture getting into the box, saturating the silica gel, condensing on the inside of box/circuitry, and causing erroneous erratic behaviour (like working during the day and not at night, then finally stopping altogether).

I now use breathing cable glands on all my new designs which incorporate a gore-tex material (http://www.wiska.de/eng/301,ventgland.html) and have had no problems since. So this might be worth a try - if you can get them that is ;0

Also, on anti-condensation heaters, a few resistors on the board dedicated to dissipating 0.5w or less may well be enough in a small enclosure - make sure you are using an adaptor of course. (You could also switch them on under PICaxe control using a temperature sensor in the box, or even the PICAxe own sensor (readinternaltemp).

Hope it works out for you.

Regards
Eclectica

 

[hippy]

Beyond guessing what may have gone wrong and proposing ideas to avoid such things occurring the only way to really get to the bottom of things is to analyse what it is actually doing and determining why.

This can often mean debugging within the environment where the failure is observed, comparing what was expected with what you get, determining what does work and what doesn't like normal debugging.

 

[fernando_g]

Agree with Hippy, you need to start troubleshooting.

A way could be to create a sample program to exercise the Picaxe's functions, and report back via debug.

But from your explanation, I also believe it could be humidity condensation. Someone has already proposed turning on some power resistors when the temparature drops and/or the RH increases above a certain point. Picaxe controlled, of course. This routine could be running sporadically, once every couple hundred of your main program loops.

Now, it will increase your current consumption, whic could be an issue with batteries.

Another solution would be to conformal coat your board. There are plenty of supplies available at hobby levels, such as this one:

http://www.google.com/#q=conformal+coating+spray&hl=en&prmd=imvns&source=univ&tbm=shop&tbo=u&sa=X&ei=ZJ71T8n8LITCswail53PAg&ved=0CHAQrQQ&bav=on.2,or.r_gc.r_pw.r_qf.,cf.osb&fp=3e2c8727f5c690de&biw=1131&bih=1034

If you follow this route, ensure that the board has been completely cleaned and dried before you apply the coat. Otherwise fluxes and contaminants will have no other way to go but to the electronic components, causing a failure.

 

[Jeremy Harris]

I can confirm that there's no problem with running Picaxe chips at sub-zero temperatures. My environmental logger has had a wireless linked outdoor sensor, with soil and air temperature sensors, a solar panel to keep the NiMH battery charged up and also measure sunlight, and a humidity sensor. The 08M2 in it has been running continuously for several months now, including several nights where the temperature dipped below zero. It's not glitched once, so far. It's run at 100% RH a fair bit recently, too, with no adverse effects.

 

[HertzHog]

I designed a maximum and minimum recording thermometer maininly for an incubator. But I have tested it in a deep freeze and even with dry ice and had no problems well below zero. I knew it might not appreciate humidity so I kept it dry. It was fine outside as a frost monitor too. My chip was an 18m2. HertzHog.