My ICs got a bad cold? What happens?

[TurboPC]

I have a design with regular ICs and FPGAs. They are all rated from 0 to 70 degree Celsius. Although I know what will happen if the temperature is too high, I wonder what is the effect of the cold on them.

What happens if I powerup the board and the temperature is below 0 degrees? Will the chips die or the junctions break?

Is there better chances if the board is powered up at room temperature and then go in a very cold environment?

Thanks for your help.

TurboPC

 

[johnyaya]

I have operated commercial grade (0..70C) devices down to -40 and below without any destruction. Most digital devices tend to have trouble at -40. But none have been destroyed, just stopped working. Don't count on the device meeting specs past the rated temperature though.

 

[Nick C.]

Don't count on the device meeting specs past the rated temperature though.

Especially crystals/resonators.

 

[barrybear]

Try Switching a tv on at -10 degreas and chancs are it will go `bang`.
Another factor is the due point as the devices warm up. The last thing you want is condensation forming all over your board (shorting tracks).
Barrybear

 

[Rayengine]

For common consumer design, it is always working properly under -10 degree. The most problematic may be the drift of crystal and resonator as the previous message. Moreover some electrolytic capacitors also get problem in low temperature.

ICs should be no problem in low temperature!

Rgs

 

[asena]

Chances are that ohmic resistance of tracks will increase, and if you are working with high speed digital systems... you will have losses.

And the previous parasitic capacities that you get from track being close, will work with the resistance.... RC....increase timings.... things will fail.

Dont forget thermal noise of the semiconductors. It may afect your final result. Though we are talking about low temperatures... which may help to reduce noise.


asena 8)

 

[sick_man]

add a 2 watt resistor too waste a few watts

and simply use a hermatic seel too stop any due point problems

this way there is no or very little water

also polystirine is good

as for resonators a 10 k resistor dumping 5 v too gnd beside the xtal should maintain that

i think that at - 20 - 40 deg most chips even military ones will still have problems if used for long periods

silicon is strange stuff at very low tempretures

and so are the caps and resisotors

so some wasted watts are what you need
and hermatic seal

 

[zoovy]

Low temperature Semiconductor issues

This is researched under LTE (low temperature Electronics for deep space applications). Current electorincs for deep space applications use heaters to keep temperature in operating range.

Most semiconductor devices exibit improved performance at lower temperatures because of increased carrier mobilities due to reduced phonon scattering, increased saturation velocities, reduced junction capacitances and line resistances, and increased substrate and metal thermal conductivities

The key here is reduced resistances, reduced capacitances, higher Beta in transistors etc. IF your circuit has a built-in PLL, DLL etc. chances are it will work faster. The lower range of your frequency response will be higher. If your circuit has internal pullup resistors etc. they will have lower resistance. If it has analog circuits internally such as opamps etc. they will work closer to perfect due to increased beta.

The crystals actually get affected from the temperature too.

 

[TurboPC]

Thanks for all these good answers... I appreciate a lot.

I will do some testing on my specific circuit. I'll have a close look at the PLL performances.

Regards, T.