Electronics in an Arctic Environment?

[wosser]

Hi guys, wondered if you'd be able to impart some advice on a rather interesting task I might be embarking on soon...

At work we have a product that has up to now only been used in temperate or hot climates.  I would like to do some research into what it would take to redesign this product so that it can cope with cold climates instead (as a separate variant of the product).

The product is usually classed in the computer and telecoms equipment category, it will need to run off a battery (sealed lead acid type).  I'm expecting that the gear will be expected to operate at -10° C for extended periods of time, and with repeated thermal cycles perhaps once a month when the gear is brought into a warm area for reconfiguration by personnel.  It probably won't be subjected to any harsh impacts or vibration (hopefully!).

The existing kit is simply just a few PCBs mounted on standoffs inside an HIPS plastic (IP20 rated) enclosure with an exhaust fan on the back.

I'd welcome some general tips about what type of environmental phenomena I should be thinking about and what can be done to defend against them.  What I've come up with so far...

• Thermal cycling of components and battery (frequently moving from warm to cold conditions over several years)
• Humidity changes (bringing cold gear into a warm building...) condensation?
• Thermal expansion / contraction stresses... could cause mechanical failure?
• Changes in air pressure /density (due to moving over mountainous terrain / helicopter transport etc)

Initially my chief concerns are preventing moisture building up inside the unit and thermal degradation of the main battery.

This kind of design is rather outside my normal engineering experience and so I'd appreciate some hints about how I can begin looking into solving this interesting set of problems.

I suspect a good all-round solution will be a lot more involved than simply cramming the enclosure full of insulating material and duct tape.

I'd really appreciate any ideas from you guys.

Cheers.

Wosser.

[HackedFridgeMagnet]

Hi
I have no direct experience of electronics at this temp, but a bit in other harsh environments.

You need to know what temperature is your absolute minimum operating temp. Get someone to put a figure on it or get it from the weather bureau.
Also any other abnormal environmental conditions that you consider important and email it to the bosses so they know what you are designing to.
It covers your arse and give the others a chance to check to see if you are missing something.

That said the only one I can think of is the battery. You will need a temperature compensated battery charger, if it charges at -10C. Also I think your battery energy will be down at this temp, to maybe 80%.
www.batteryweb.com/manuals/techman.pdf

If you can I would seal it to IP65 or whatever the equivalent NEMA is,  if possible. With a greased ORing Seal and/or potted. If you can control the moisture inside the unit most your problem is gone. IP65 connectors are quite expensive though.
With changes in temp the box my warp a little and you can lose your seal. Double check this doesn't happen.

-10C is cold for us but I think most of the electronics will love it.

Is it outside? Wind and Ice are big problems if it is. Especially for antennas.

Do you have comms to the unit to get status? Especially battery voltage.
can you do in field firmware updates?Hard to get I know.
getting a tech to remote sites is very costly so it's best if it is not necessary.

Any chance of doing thermal/humidity stress testing before you go, in an environmental chamber?

Do a test report before deployment, and include every test you do and send it to the interested parties so that
a. you can remember what your unit is capable of.
b. others can say something if they think an important test is missing or done wrongly.
c. you have shown due diligence if something does go wrong. Things often go wrong first time in the field.

Good Luck

[wosser]

Hi HackedFridgeMagnet,

I think we'd have to use greased o-rings for sealing the case as we wouldn't be able to pot it because it requires physical reconfiguration internally from time to time.  I suppose the humidity could be kept under control by putting dessicant granule packet inside the case before closing it and replacing them every time the unit is opened.

Yes, would definitely do a proper environmental test at a certified laboratory, I believe we did so a few years back when this product first went to prototype.  That was only for hot areas though.

Good point about the battery charger temperature compensation, hadn't thought of that.

The unit won't be directly exposed to outside elements, it'll always be in a container of some sort even if it's just a wooden hut in the middle of nowhere.  A penguin might roost on it if it's lucky.

[notsob]

minus 10C doesn't sound very cold, I've been in Canada in -25C on a sunny day and -35C at night, also battery performance in the cold will be an issue.

[Kremmen]

OK, you said that the device will be in a container of some kind so i don't know if this applies. But if a penguin can get there (unless it was a joke) then so can driven snow. Now that is a bitch, especially if you have any dissipated heat from the device warming the casing. Then you are guaranteed to get ice buildup. A case in point: my house is partially heated with a common air-air exchanger of the type you see all over the place, but normally used for cooling. In freezing conditions the outside unit is periodically cycled with hot gas to clear the radiator grille from built up frost. This in turn creates an iceberg under the unit, easily over ½ meter thick in one winter. In a similar case your device could be entombed in an ice block. What to do about it is of course another story...
Also, all external connectors to the device housing should be fully hermetical, preferably o-seal types. Anything else and failure is just a matter of time, and not too long either. Avoid PVC isolated wires since they get hard as stone and brittle. Rubber, silicone, stuff like that stays supple. Even if the device does not need to be moved per se, just handling the frozen solid cables is a PITA.

minus 10C doesn't sound very cold, I've been in Canada in -25C on a sunny day and -35C at night, also battery performance in the cold will be an issue.

. True, even here hundreds of km south of the Arctic Circle i recall a January where every day was -30 or below. Up there it is routine.

[wosser]

Hmm, food for thought indeed.

Come to think of it, we had -17°C here in the UK a couple of winters back - in a "temperate" country totally unprepared for that kind of cold it was pretty unpleasant.

[SeanB]

I will add that you will have to use an insulated case, with an IP68 rating, and a separate sealed and vented enclosure inside for the battery that vents to a non freezing ceramic micropore filter so that gas generated by the battery does not get released into the case. the battery will probably be happiest if you have the electronics oaround it inside warming it up, it will not like being frozen at all. I would recommend rather using a few large LiPo packs as a power source, they will survive better cold temperatures ( no water inside at all) and they have a much higher energy density. If you use the same mass of LiPO packs you will be able to only discharge them down partially, and will have a longer life on them. You will not then need a battery vent, just good insulation around the battery pack and electronics.

Use MIL circular connectors for all external connections, as they can be capped if not in use, and are available in IP68 ratings with assorted contact styles and contact numbers. Design it as if it is going to be tied to a rope and dropped over the side of a ship into ice water, it must survive the drop and the cold water as well, even if it hits an ice floe on the way down.

[Achilles]

I think the biggest issue will be temperature changes. I saw the same with our stuff that is operated in the tropics. You just have a few Kelvin fluctuation over the day, but when there is a high humidity it's starting to get interesting.
As I disassembled an enclosure (IP67) last year, I opened it and a massive amount of water was greeting me when I opened the lid. It hasn't been maintained for over two years and has a built in Battery with solar charger.
Interestingly, I tested the logger which was assembled inside and it still works great. Just the relays and solar charger are broken.

So, if the stuff is too tight, you may have problems as well. A good insulation and stable Temperatures are the way to go I guess. You should also think about Wind and the Windchill. Most arctic areas are wide open and the stuff is pretty exposed there. So you'll have a lot of wind which could press humidity inside the enclosure.

[Neilm]

I have designed equipment that works at -10C or less with no real problem. The biggest issues we have are due to the battery reduced capacity or lower terminal voltage and the LCDs on the instruments. These LCDs will freeze and stop working. There are some available that will work down to -30 and below, but these are expensive. Plastics can get brittle when cold but -10C should be OK.

If you have an exhaust fan on the unit, I am assuming that it generates heat. If so, is the -10C the outside temperature? Depending on how much heat it does generate you might find that the internals do not get much colder than freezing.

The worst case conditions from what you have said will be moving it back indoors. The temperature jump when it is moved inside will cause condensation inside the unit and you will have to let it recover before using it.


Neil

[MartinX]

-10°C is not a problem for the electronic components at all, the problem is moisture especially moving the unit into a relatively moist warm environment will cause condensation as Neilm points out, and dealing with it is a matter of correct procedure when the unit is brought inside. Conformal coating of the boards will be a good idea.

How exposed will this thing be? If situated on a mast or something similar dual enclosures is a good solution, the outer enclosure will then take the  grunt of rain, wind, snow and ice.

Lead acid batteries are not a good idea at all, when discharged the freezing point of the electrolyte will go up and the battery can burst if it is discharged too deeply and it freezes.

[wosser]

This is good advice guys, thank you very much for your time, all of you.

I'll look into LiPo cells for starters, then IP68 cases.

Your help is appreciated.

Cheers.

W.

[qno]

Check out NASA.
Apollo 11 used silver-zinc batteries

[wosser]

 

[Christe4nM]

Hi Wosser,

Unfortunately I've got no answer directly to your question, but I was triggered by this words "It probably won't be subjected to any harsh impacts or vibration (hopefully!)."

Last year at university (undergrad. EE student here) we've spend quite some time doing requirements engineering for electronics in a design project for an actual customer. From that background I'd suggest that you try to find the people you intend to sell your product to. I mean the actual users. Talk with them and try to find out how they use similar equipment. What difficulties they experience, what annoys them, or makes them really happy. Don't ask them straightforward what they want in product 'x' but, try to get every bit of information about how they use it in their daily job. That should give you all information you need to create a product that ideally hits the 'perfect fit' mark with the first prototype. It's not just about the technical performance, but in a large way about how you implement it in such a way the user feels your product is better, more intuïtive, fits their current range of equipment etc. It could be as simple as button size so that users wearing arctic gloves can still operate the equipment. Whether this applies to your product doesn't matter, just know that there probably are hidden user requirements you cannot come up with yourself.

Finally a trick I know a military radar designer uses: they do sometimes have to use commercial off the shelve components that do not comply to for example millspec temperature ranges. They accommodate by controlling the enclosure's inside environment, so that temperature and humidity are always within the component's spec, no matter the environment outside the enclosure.

Hope this helps. Goodluck with the redesign process.

[Paul Moir]

If you're looking for a relatively inexpensive IP68 connector, I've had very good luck with the Souriau UTS series in a food processing environment (washdown with chlorinated cleaners).  Digikey carries them.  One nice feature I've noticed is that when overstressed, the locking pins on the receptacle shear.  This leaves the connector and wiring still operable but without locking, and replacement of the shell is pretty easy with a pin removal tool.

Operation is pretty nice too:  it's much quicker than a circular Mil connector.

[SeanB]

Exactly the ones I had in mind. Note that they do actually mate with the metal cased military versions in certain pin configurations.

[SgtRock]

Dear Wooser:

--The case insulation should be  Goldilocks specified to the extent possible. Silicone gaskets to prevent liquid intrusion from the top and sides. Drain and vents at bottom. A duct with a fan running from the top of the space a bottom vent to vent heat. A motor and linkage to open and close vents, or possibly passive vents could be air operated (best). Thermostat with hi low capability (if possible to avoid microprocessor). With regard to condensation and thermal stress, when moving, seal case and to not open until equilibrium is established, as with musical instruments. Do not uses sockets, and use lock tabs on all push on plugs, or use solder connections. This will avoid thermal creep. If you must use socketed relays, strap them down. Establish worst case scenario for heat and cold and build for that first. It would be nice if you have a year long temperature graph for the approximate location(s) the instruments will be going to, just to give you an idea of the extremes

"Don't eat yellow snow."
Gator Dundee 1948 -

Best Regards
Clear Ether

[Neilm]

Lead acid batteries are not a good idea at all, when discharged the freezing point of the electrolyte will go up and the battery can burst if it is discharged too deeply and it freezes.

There are some Lead acid batteries that do work at -10C, the Yuasa NP series can operate to -20C. I have used these in this application. They can also be charged at negative temperatures which is an advantage over Lithium as this can't be charged at less than freezing unless you want to pay silly money for them.

One problem you might have is that charging batteries can produce explosive gasses so you will need to vent during charge to stop them building up.

Neil

[Mcfly]

You can probably just keep a couple chickens and feed inside the whole assembly.  The heat should keep all of the components in working order...  Seriously try it out.     
it worked for brits:
http://en.wikipedia.org/wiki/Blue_Peacock

[MartinX]

Lead acid batteries are not a good idea at all, when discharged the freezing point of the electrolyte will go up and the battery can burst if it is discharged too deeply and it freezes.

There are some Lead acid batteries that do work at -10C, the Yuasa NP series can operate to -20C. I have used these in this application. They can also be charged at negative temperatures which is an advantage over Lithium as this can't be charged at less than freezing unless you want to pay silly money for them.

One problem you might have is that charging batteries can produce explosive gasses so you will need to vent during charge to stop them building up.

Neil

Yes they work in low temperatures but if they are discharged deep enough the electrolyte will still freeze, this may or may not damage the battery, this is not mentioned in the data sheet. It is possible to incorporate some protection circuitry that disconnects the battery at some point to avoid deep discharge, in that case you could use an ordinary car battery they work quite well in low temperatures, my car started just fine the other day and it was -15°C, and they are cheap, maybe a bit big and heavy, snow mobile battery maybe?

If you want a device that can be left in a cold climate unattended for any period of time, if that is the requirement here, then lead acid batteries is not the best choice in my opinion.

[PuterGeek]

You haven't described the equipment very much so we are all kind of shooting in the dark.

In your subject you say arctic. If you mean that literally the conditions will be beyond brutal. Increase your safety margins significantly more than you normally would. If you typically have a 20 to 50% safety margin I would at least double that.

You probably need to use military grade components (-55 to 125C) or industrial  (-40 to 85C) with a heater. Both Antarctica and the arctic easily see temperatures below -55 so you may need a heater even with Milspec components. LCDs and any type of lead chemistry battery don't like to be frozen so a heater during storage and transport would be required. LED displays with a power disable and lithium batteries would be better.

I agree on using an IP68 (NEMA 6) enclosure inside an insulated enclosure. Ideally the insulated enclosure would be at least IP 65 also. Temperature and barometric pressure cycling will cause even an IP68 enclosure to 'breathe' so you will have atmospheric exchange. The good news the cold makes the humidity near zero but that makes static electricity problems much worse. I'm not sure if the aurora Borealis will cause problems but you might want to find out.

A metal enclosure, ideally stainless steel or aluminum, would be better than any plastics both for ruggedness and heat transfer. I would probably use powder coating if it must be painted and insure the paint will survive the extremes.

Avoid intentional venting if at all possible or use an IP 68 vent (hard to come by). Every enclosure penetration is a leak waiting to happen so the fewer the better. Avoid control and connectors penetrating the enclosure  if at all possible. If they can be kept inside they can't be a leak source.

You definitely need written procedures on reconfiguration and returning to the field. Baking out any moisture and using desiccant packs would be a good plan. I would silicone conformal coat the boards if possible so any moisture that gets trapped inside can't cause corrosion.

Assume it will see vibration if only during transport so anticipate rough handling and transport at least. Silicone everything with any mass (electrolytic caps, large resistors, LEDs, etc.). This includes silicone on the bottoms of radial lead electrolytic caps (except the smallest ones) where they mate with the board. You definitely want locking connectors and if you must use ribbon cables, tie wrap or silicone their locks.

If I'm correct about the vibration, any mass along with solid wires (cold makes them more brittle) will eventually break. If I'm wrong, consider the wasted silicone as insurance. To paraphrase my father, 'it won't break if you put it together too well.'

Hope that is helpful.

[Kremmen]

[...] I'm not sure if the aurora Borealis will cause problems but you might want to find out.
[...]

No it won't.
Solar particle streams visible as aurora will cause electron avalanche phenomena in the upper atmosphere, detectable by incoherent scattering at some radio frequencies but otherwise won't affect ground based electronics. There are several atmospheric science efforts confirming this:
- Ionosondes in Northern Finland above the Arctic Circle typically find echo returns at E, F1 and F2 layers of the ionosphere btw 100 to 600 km above ground.
- EISCAT UHF radars in Norway, Sweden and Finland map electron densities at similar or little lower altitudes 77 to 340 km.
- The STARE aurora radars in Norway and Finland measure electron densities and motion velocities.
- Additionaly there are several RIOmeter stations (Relative Ionospheric Opacity) that can detect the presence of free electrons in the atmosphere, specifically in the lower ionospere in layer D.
- The IMAGE magnetometer grid extending from Southern Finland all the way to Longyearbyen in Svalbard islands, the northernmost town in the world. This detects the aurora induced electric currents in the plasma sheet.

All systems confirm that auroral effects are limited to the extreme upper atmosphere. So, no trouble from aurora down here. Solar flares are another thing but let's not go there since the question was about aurora.

[wosser]

You haven't described the equipment very much so we are all kind of shooting in the dark.

Yeah, I wanted to keep it a bit vague because the nature of the gear means it might need to be reconfigured considerably in order to perform other purposes.

Imagine a sandwich of PCBs of mainly surface mount devices, microcontrollers, sram, PLDs.  The device and a suitable LiPo battery would just about fit inside a shoebox.  Four cables (of various types) are required to exit the device and be exposed to the inside of the vehicle or building it's contained in.  At least 1 of those cables makes it out to the fresh air outside.

[peter.mitchell]

They can also be charged at negative temperatures which is an advantage over Lithium as this can't be charged at less than freezing unless you want to pay silly money for them.

The A123 Systems Lithium Iron Phosphate batteries can be charged/discharged in very low temperatures, losing around 10-30% of their rated nominal capacity at -20oC. These cells are both physically and electrically incredibly strong, plus, they have very very low self discharge which I get the feeling would be important.

[Neilm]

The A123 Systems Lithium Iron Phosphate batteries can be charged/discharged in very low temperatures, losing around 10-30% of their rated nominal capacity at -20oC. These cells are both physically and electrically incredibly strong, plus, they have very very low self discharge which I get the feeling would be important.

When I was looking at them they were quite expensive (although that might have been due to the pack manufacturer). We stopped looking at them as they filed for bankruptcy.

Neil