13
\$\begingroup\$

My company is working on developing a product that will go into commercial freezers, so my boss asked me to provide operating temperature specifications for the product. I can find "operating range" temperatures listed for everything but the PCB itself, which is just plain old FR-4.

Wikipedia helpfully lists "Temperature Index" (whatever that means) as 140 C, but there are no indications of a minimum temperature.

I'm not really worried, as I'm sure that the other components on the board will be the limiting factors, but for the sake of completeness, I would like to have it listed.

Does anyone known the minimum operating temperature of FR-4? (And what would the failure mode be?)

\$\endgroup\$
  • 2
    \$\begingroup\$ It's not what you're asking, but...In addition to looking at the operating temperature range of each part, remember to look at how they interact. Different materials have different coefficients of thermal expansion (CTE), and parts with CTE much different from that of FR4 tend to be the ones that fail (either break or pop off the board) due to temperature swings. Large ceramic parts (like some crystal oscillators) are the most common source of problems in my experience. \$\endgroup\$ – The Photon May 17 '13 at 2:30
  • \$\begingroup\$ I'm assuming you're working with (or considering) SMT parts here. \$\endgroup\$ – The Photon May 17 '13 at 2:31
  • \$\begingroup\$ Thanks, these are good things to keep in mind. The board is SMT, and has a SAW resonator on it for RF communication. \$\endgroup\$ – TimH May 17 '13 at 16:56
  • 1
    \$\begingroup\$ TimH, one thing is, if your regulatory environment allows it, use tin-lead solder for this board instead of lead-free. The lead-free types are stiffer and more brittle so more likely to crack the solder joints or crack the parts when thermally stressed. \$\endgroup\$ – The Photon May 17 '13 at 17:06
  • \$\begingroup\$ Duly noted. I am aware of the issues with the RoHS directive, and prefer to use tin-lead solder wherever possible, just as a general rule. This case would be even more important. \$\endgroup\$ – TimH May 17 '13 at 17:12
13
\$\begingroup\$

FR4 PCB is glass-reinforced epoxy laminate. Several research studies have been published of the effect of low temperatures on such material.


These are two key assumptions regarding the device touched upon in the question:

  • The freezer temperature will not reach as low as those used in these studies
  • There is limited risk of vibration or impact upon the actual circuit board, hence brittle cracks are unlikely but not ruled out

If either of these assumptions is invalid, the material in question needs to be reconsidered. There do exist special-purpose industrial ceramics / alumina PCB substrates designed specifically for extremely low temperatures, typically used for device deployment in space or in cryogenic equipment. Those materials might be more suitable in this case.

The point to note in such environments is the possible fracture of packages, casings and solder junctions for the electronic components on the board, not just the PCB itself.

A commonly recommended procedure for deploying circuit boards in extreme temperature conditions, is to bring the device to the desired temperature in slow stages, thus avoiding any rapid shrinkage or thermal shock to board or parts.

\$\endgroup\$
  • 1
    \$\begingroup\$ This is precisely the type of answer for which I was hoping. Thank you for your thorough, carefully thought-out response. \$\endgroup\$ – TimH May 17 '13 at 17:07
  • \$\begingroup\$ @TimH You're welcome. I've been dealing with the vagaries of very similar materials being used as thermal insulation in a cryogenic chemical reactor design I'm helping with. \$\endgroup\$ – Anindo Ghosh May 17 '13 at 17:15
4
\$\begingroup\$

FR-4 easily survives cryogenic temperatures. It has been used in various hobbyist satellites, tested in liquid nitrogen and my place of employment uses some FR-4 adapter pcb's in a cryogenic vacuum chamber. If FR-4 successfully functions at -60*C and -90*C, then it should survive the freezer. Due to difference in temperature coefficients we use copper laminate pcb's for temperatures under 40* Kelvin.

\$\endgroup\$
3
\$\begingroup\$

When it comes to PCB laminates, there is 'FR-4' and there is 'FR-4' and other types.

I am not repeating myself: all that FR-4 really means is it meets a flame retardant standard (usually 94V0).

There are a number of laminate manufacturers out there, such as Isola who make a wide range of laminates each with different properties for different markets.

Hitachi Chemicals also make a wide variety of base laminates. Although the answer above is excellent, sometimes the best people to ask are the laminate manufacturers.

As you appear to be in a bit of a niche area, I think the laminate manufacturers may be best placed to help you from this point.

\$\endgroup\$
-2
\$\begingroup\$

FR-4 has no "operating" temperature. It has a melting point or a glass transition point which is about 120-140 C. If you freeze it, it just gets colder that's all.

You might get failures as a result of repeated sudden changes in temperature, i.e. thermal shock. There are tests you can do to see how much thermal shock your PCB can handle before vias and tracks crack and stop being usable but that's a different issue.

\$\endgroup\$
  • \$\begingroup\$ Thanks for the answer! The PCBs in question are going to pretty much stay in the freezer, so repeated thermal shock should be minimal. \$\endgroup\$ – TimH May 16 '13 at 23:09
  • \$\begingroup\$ "a melting point of 120-140 C"?!? I don't think so! It can easily withstand soldering temperatures, which are 180-200 C minimum. \$\endgroup\$ – Dave Tweed May 17 '13 at 11:14
  • \$\begingroup\$ You're right @Dave Tweed, the glass transition temperature is just the temperature at which the glass begins to rapidly expand. It is not the melting point. \$\endgroup\$ – TimH May 17 '13 at 16:59
  • \$\begingroup\$ Even still, the glass transition temperature for silica-based glass is on the order of 550-600 C. Please correct your answer. I think the temperature at which the epoxy binder starts to break down defines the actual operating limit of FR-4 anyway. \$\endgroup\$ – Dave Tweed May 17 '13 at 18:27
  • 1
    \$\begingroup\$ For the sake of clarifying the confusion here, I note the following: epoxy is not a thermoplastic; it has no melting point (it chemically degrades rather than melting). Its glass transition temperature, which will indeed be something like that quoted in the above answer, is the temperature above which it can plastically deform. \$\endgroup\$ – Oleksandr R. Jun 11 '16 at 23:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.