LED driver at -40degc ambient

[ocset]

Hello,
We have an 40W offline LED driver which uses no electrolytic capacitors...only ceramic ones.
Also, it uses only sequence switched linear regulators, -no SMPS.
It is only surface mount components.
Optocouplers, resistors, caps, diodes, FETs,  SFH5711 light sensor, MCP9701 temp sensor, MCP6001 opamp, BZX84C-XX zeners.
Do you think we will have problems at -40degC?
The PCB sits on a rubbery thermal pad which is on a heatsink

We are getting it in the oven at -40degc, but wonder if there are any long term effects.
The product is totally enclosed..no vent holes....a rubbery gasket makes the seal between the optic and the heatsink

[ArthurDent]

The question is will it work at -40 degf   

[ArthurDent]

Seriously I would wonder if a through hole design might be better. I'd wonder if parts and circuit board would expand and contract at different rates and perhaps break ceramic caps or component solder bonds with the board. I would think that leaded components would allow some give that you might not get with surface mount. Might be worth looking into for long life.

One place I worked in winter (Mount Washington, New Hampshire) would get temps approaching the limits you're using. When the rebar reinforced thick concrete walls approached these temps we would hear load noises like gun shots as the rebar and concrete contracted at different rates and 'readjusted'.

[mdszy]

Seriously I would wonder if a through hole design might be better. I'd wonder if parts and circuit board would expand and contract at different rates and perhaps break ceramic caps or component solder bonds with the board. I would think that leaded components would allow some give that you might not get with surface mount. Might be worth looking into for long life.

One place I worked in winter (Mount Washington, New Hampshire) would get temps approaching the limits you're using. When the rebar reinforced thick concrete walls approached these temps we would hear load noises like gun shots as the rebar and concrete contracted at different rates and 'readjusted'.

I feel like you'd also have to worry about the solder joints cracking at that point.

[Ian.M]

Yep, ROHS Sn based solders can turn into non-conductive grey powder at very low temperatures.  See https://en.wikipedia.org/wiki/Tin_pest

[Zero999]

Don't you need a freezer, not an oven to test at -40^oC?

If it's purely linear it might work, once it's heated up.

[ocset]

Yep, ROHS Sn based solders can turn into non-conductive grey powder at very low temperatures.  See https://en.wikipedia.org/wiki/Tin_pest

Thankyou Ian.M    ...wow!......thats Brilliant info!..........
I've worked a lot of places and nobody knew about that.

I hope the EU   have a clause in their regulations so that we can get our boards  made up with leaded solder.........this is the problem when , as a country, you completely give away all the rights to make laws to a foreign power...........sometimes i wonder what "rewards" our British MEPs received in order to sell off our country (UK) down the river to the EU.
Mind you, we havent yet been hit as badly as Poor Greece. Greece went bankrupt after its leaders decided to take Greece into the Euro currency......Greece would never have been able to borrow enough money to go bankrupt if it had not joined the Euro......The Euro was always going to help some and hurt others.....different countries , with totally different economies, cannot all survive under one currency....for a start, the country then cant change its own interest rate....its basic economics. Sorry to moan.

[Ian.M]

You may find this page helpful: http://www.dbicorporation.com/tinpest.htm

[Zero999]

Yep, ROHS Sn based solders can turn into non-conductive grey powder at very low temperatures.  See https://en.wikipedia.org/wiki/Tin_pest

Thankyou Ian.M    ...wow!......thats Brilliant info!..........
I've worked a lot of places and nobody knew about that.

I hope the EU   have a clause in their regulations so that we can get our boards  made up with leaded solder.........this is the problem when , as a country, you completely give away all the rights to make laws to a foreign power...........sometimes i wonder what "rewards" our British MEPs received in order to sell off our country (UK) down the river to the EU.
Mind you, we havent yet been hit as badly as Poor Greece. Greece went bankrupt after its leaders decided to take Greece into the Euro currency......Greece would never have been able to borrow enough money to go bankrupt if it had not joined the Euro......The Euro was always going to help some and hurt others.....different countries , with totally different economies, cannot all survive under one currency....for a start, the country then cant change its own interest rate....its basic economics. Sorry to moan.

Come on, you should know that political rants like that are against the rules!

If you're using leaded solder, then I don't see the problem. I remember working on a design for a defence company awhile ago, which had to function down to -40^oC, but a part was only rated down to -25^oC, so we tested it down to -40^oC in an environmental test chamber and it worked perfectly.

[mzzj]

I hope the EU   have a clause in their regulations so that we can get our boards  made up with leaded solder...rant on

Nope, your led drivers are not on  exception list. 

Seem like Tin pest on lead-free solder joints is mostly academical problem, not too many real-world cases. I'd worry more about mechanical stresses from thermal cycling. 

[mzzj]

You may find this page helpful: http://www.dbicorporation.com/tinpest.htm

I'm afraid that guy is borderline nutcase:
http://www.dbicorporation.com/rohsbib.htm

[mdszy]

You may find this page helpful: http://www.dbicorporation.com/tinpest.htm

I'm afraid that guy is borderline nutcase:
http://www.dbicorporation.com/rohsbib.htm

For myself and my companies, in early 2005 I adopted the following version of the "Precautionary Principle", which so far has worked very well for me:

Don't buy-- or fly in-- anything manufactured in Europe.
Don't buy new electronic/electrical stuff, if you can find an older item (manufactured before 2006) that will do the job.
If you buy a new, expensive electronic/electrical item, get the longest warranty that you can for it.
If you buy a new, inexpensive electronic/electrical item, buy at least three, to give yourself a fighting chance that one of them will still work when you need it.
If you are replacing an older, working unit, HANG ON TO THE OLD ONE!

Holy shit, I'm glad I don't work for this whacko...

[Ian.M]

You may find this page helpful: http://www.dbicorporation.com/tinpest.htm

I'm afraid that guy is borderline nutcase:
http://www.dbicorporation.com/rohsbib.htm

Yes, he comes across as fairly obsessive, but he *has* collected a comprehensive list of technical references, including ones to books and papers that are not readily available online.

[mdszy]

You may find this page helpful: http://www.dbicorporation.com/tinpest.htm

I'm afraid that guy is borderline nutcase:
http://www.dbicorporation.com/rohsbib.htm

Yes, he comes across as fairly obsessive, but he *has* collected a comprehensive list of technical references, including ones to books and papers that are not readily available online.

He seems to make it sound like lead-free solder shall be the downfall of humanity, though.

[Ian.M]

I hope the EU   have a clause in their regulations so that we can get our boards  made up with leaded solder...rant on

Nope, your led drivers are not on  exception list. 

Seem like Tin pest on lead-free solder joints is mostly academical problem, not too many real-world cases. I'd worry more about mechanical stresses from thermal cycling.

The problem is: Grey tin auto-catalyses the low-temperature transition from white (metallic) tin to grey tin.   As long as your equipment regularly cycles all tin in it to above +13.2°C, for long enough for the fairly rapid reverse transition to occur, you wont have any tin pest problems unless it is significantly sub-zero for a continuous extended period.  However, if some tin plated part of your device doesn't get above +13.2°C, (e.g. a shield or housing) it doesn't stop the clock on the tin pest, and at sufficiently low temperatures, a single grain of grey tin dust forced into intimate contact with the solder surface (e.g by differential expansion/contraction of parts in mechanical contact or near contact due to thermal cycling) can cause Pb free high Sn percentage alloys  to develop tin pest far more rapidly than would normally be the case.   

In Treez's application, that would mean that if the lights have to withstand Arctic winter conditions, there is a risk of tin pest related failure if they aren't used long enough to fully warm up on a daily basis.

Combined with your point about fatigue due to thermal cycling, choosing a ROHS compliant solder alloy for extended low temperature high reliability applications is *NOT* at all easy.   Even NASA don't have any good answers except:

3.2.6 TIN PEST (TIN DISEASE) MITIGATION
Lead-free Tin (Sn) and Lead-free Tin (Sn) technology shall not be used in applications with exposure to
temperatures at or below -30°C -22°F), unless combined with one of the following alloys by weight
percentage:
a. Not less than 5 percent lead (Pb)
b. Not less than 0.3 percent bismuth (Bi)
c. Not less than 0.5 percent antimony (Sb)
d. Not less than 3.5 percent silver (Ag)

https://standards.nasa.gov/file/283/download?token=rMus0Zx4

I don't have the data to asses the risk in relation to the risks of other possible low temperature and thermal cycling related failures.   Maybe if you plow through the references from DBI corp's tinpest page, you'll have enough data, but its going to be highly dependent on the exact solder alloy used, and the the purity of tin plating on other parts of the device.

[mzzj]

I don't have the data to asses the risk in relation to the risks of other possible low temperature and thermal cycling related failures.   Maybe if you plow through the references from DBI corp's tinpest page, you'll have enough data, but its going to be highly dependent on the exact solder alloy used, and the the purity of tin plating on other parts of the device.

I don't have any probability numbers to throw either but given that we have had lead-free electronics for over 15 years without hearing of any case of tin pest affecting electronics in Lapland I'd say that it's not on top of the risk list.
I'm living in the coldest EU country, temperatures get below -40  in northern parts of Finland every year, yet I haven't heard of single case of tin pest actually affecting electronics.
 

[Ian.M]

It can take a total of a couple of years below -10°C to initially trigger it, so summer temperatures are a key factor as if they reliably get above about 15° C, it will 'reset the clock' on it.   That means its got a vanishingly small probability of occurring except in outdoor fixed installations in the high Arctic or above the snow-line, or installation in walk-in freezers etc.  Any portable electronics or vehicle is likely to be brought into a shirt-sleeve environment at least annually.   

If Treez is selling fixed exterior LED lamps to the Alaskan oil industry he probably should worry about it.

[tooki]

The question is will it work at -40 degf   

It's treez we're talking about. Make sure to suggest testing it at -40˚K, too!

[tooki]

It can take a total of a couple of years below -10°C to initially trigger it, so summer temperatures are a key factor as if they reliably get above about 15° C, it will 'reset the clock' on it.   That means its got a vanishingly small probability of occurring except in outdoor fixed installations in the high Arctic or above the snow-line, or installation in walk-in freezers etc.  Any portable electronics or vehicle is likely to be brought into a shirt-sleeve environment at least annually.   

If Treez is selling fixed exterior LED lamps to the Alaskan oil industry he probably should worry about it.

Wouldn't one solution be to simply install small thermostatically controlled heaters into the units to always keep them above the critical temperature?

[Ian.M]

Yep.  However,  running a heater 24/7 isn't going to do you any favours for energy efficiency, especially if your refrigeration plant has to get rid of the waste heat from the small heaters keeping the lights nice and toasty.

Also, this reasonably recent paper is scary:
https://www.emeraldinsight.com/doi/full/10.1108/SSMT-10-2013-0027
If their results are reproducible, it looks like the combination of low temperatures, the wrong solder alloy and mechanical stress can induce tin pest *WITHOUT* inoculation by grey tin dust in under a fortnight, which means a massive logistics problem maintaining heated storage so it isn't already compromised when you install it.

[Zero999]

Wouldn't one solution be to simply install small thermostatically controlled heaters into the units to always keep them above the critical temperature?

Wouldn't that defeat the main benefit of LED lamps: efficiency?

I suppose the heaters could be put on a timer, so they only have to turn on for an hour or so a week, to get the temperature above 15°C.

Painting the case black may also help absorb sunlight in the summer to raise its temperature above 15°C, even if the air is cooler.

It can take a total of a couple of years below -10°C to initially trigger it, so summer temperatures are a key factor as if they reliably get above about 15° C, it will 'reset the clock' on it.   That means its got a vanishingly small probability of occurring except in outdoor fixed installations in the high Arctic or above the snow-line, or installation in walk-in freezers etc.  Any portable electronics or vehicle is likely to be brought into a shirt-sleeve environment at least annually.   

If Treez is selling fixed exterior LED lamps to the Alaskan oil industry he probably should worry about it.

What's worse, extreme cold or lack of warmth in summer?

His product might not be used in such extreme places as Alaska. There are a few places in the UK which barely get above 15°C in summer, but they don't get much below -10°C in winter. One example is Cairn Gorm, which usually gets below -10°C in winter, but doesn't normally get much above 15°C in summer.
https://en.wikipedia.org/wiki/Cairn_Gorm#Climate

[tooki]

Wouldn't one solution be to simply install small thermostatically controlled heaters into the units to always keep them above the critical temperature?

Wouldn't that defeat the main benefit of LED lamps: efficiency?

Absolutely!! It really does beg the question of whether LED is optimal for all situations.

I suppose the heaters could be put on a timer, so they only have to turn on for an hour or so a week, to get the temperature above 15°C.

Painting the case black may also help absorb sunlight in the summer to raise its temperature above 15°C, even if the air is cooler.

Aye.

Yep.  However,  running a heater 24/7 isn't going to do you any favours for energy efficiency, especially if your refrigeration plant has to get rid of the waste heat from the small heaters keeping the lights nice and toasty.

Well treez seems to work for a company making street lights, so that's the intended application, I assume. But you'd be right, of course. (I sometimes chuckle when I go past the hot food bar in a US grocery store, throwing off tons of heat into a space that's being refrigerated down aggressively…)

[Ian.M]

What's worse, extreme cold or lack of warmth in summer?

I don't know - if I did, (and had the data to back it up) I could make good money as a low temperature reliability consultant!

I also don't have data on how long and how much it needs to be above the transition temperature to reliably convert all grey tin back to white tin.

If anyone wants to experiment, put a ziploc bag with a small coil of >99% Sn SnCu solder deep in the freezer, and when its well chilled bend it a few times without letting it warm above zero, then leave it in the freezer and check it weekly.  If the paper I linked is reproducible, there's a significant chance you'll have grey tin to play with in a few weeks time.

[Whales]

Wouldn't one solution be to simply install small thermostatically controlled heaters into the units to always keep them above the critical temperature?

Wouldn't that defeat the main benefit of LED lamps: efficiency?

Absolutely!! It really does beg the question of whether LED is optimal for all situations.

Not entirely.   LEDs + dedicated heaters can outlast filament bulbs, incans and the likes by large factors (if designed correctly).  This might be very important in cold locations if changing the bulbs/units is hard, or it might not be important at all.

[mzzj]

Yep.  However,  running a heater 24/7 isn't going to do you any favours for energy efficiency, especially if your refrigeration plant has to get rid of the waste heat from the small heaters keeping the lights nice and toasty.

Also, this reasonably recent paper is scary:
https://www.emeraldinsight.com/doi/full/10.1108/SSMT-10-2013-0027
If their results are reproducible, it looks like the combination of low temperatures, the wrong solder alloy and mechanical stress can induce tin pest *WITHOUT* inoculation by grey tin dust in under a fortnight, which means a massive logistics problem maintaining heated storage so it isn't already compromised when you install it.

Why everyone is fixated about heaters when there is tens of watts of heat coming from the leds itself in the same package? 
Unless this is a lighting for darkened cryo storage room 

I found the same paper about tin rot earlier but it was bit too much accelerated testing again vs real-world probabilities.
One point  from that paper is that it might be bad idea to tighten any mounting screws over tinned areas of pcb.
And I'm not saying that you should ignore tin pest in every case, life supporting equipment on McMurdo station for example..

[David Hess]

I have run boards down to -55C without problems and we destructively tested them down to −196C (liquid nitrogen) temperatures without issues.  The only way you will really find out is to do destructive vibration testing inside of an environment chamber to see what breaks first.

You may need to qualify components for operation outside of their specified temperature range.

[mmagin]

The product is totally enclosed..no vent holes....a rubbery gasket makes the seal between the optic and the heatsink

I would also consider whether the daily / seasonal temperature variation will lead to air (and moisture) gradually 'pumping' past the seal.
Also, how is the device going to perform at the maximum expected temperatures?  Will it be in direct sunlight?