1
\$\begingroup\$

What exactly happens to an IC that is (for example) specified at maximum 85 degrees C? Also, being made from I assume similar materials, how is it possible for another chip to have a higher rating like 105 or 125 degrees C?

Edit: my question title was originally "What happens when an IC goes over it's maximum operating temperature specification?" but I have changed it to reflect the text of my question

\$\endgroup\$
  • 1
    \$\begingroup\$ possible duplicate of electronics.stackexchange.com/questions/13873/… \$\endgroup\$ – MarkU Jan 8 '16 at 7:10
  • \$\begingroup\$ Hello and welcome to Electronics.StackExchange! Usually an IC will fail if overheated due to the previous link. Depending on what the die inside is constructed of exactly, the electronic function of the IC, the case material (plastic or ceramic), even assembly methods, all contribute to how much heat an IC can tolerate. Note that many IC's will also change behavior or cease to function normally if too cold as well. \$\endgroup\$ – rdtsc Jan 8 '16 at 7:21
  • 1
    \$\begingroup\$ The first part of your question is answered in MarkU's link. The second part is a good question. Please consider editing this to just that to prevent it from being closed. \$\endgroup\$ – Passerby Jan 8 '16 at 7:31
  • \$\begingroup\$ Thanks, I didn't find that question when I searched, but I was more wondering the part about why there are 125 degree rated parts, so I suppose you are correct I should update my question. \$\endgroup\$ – bearded_badger Jan 8 '16 at 8:40
4
\$\begingroup\$

A maximum temperature listed in a datasheet is the maximum temperature at which the manufacturer guarantees that the IC will be working. Depending on the design this temperature can vary, 85, 105, 125 degrees are common junction temperatures (the temperature at the silicon die).

All the components on the IC's die are temperature dependent, their behavior changes over temperature. This is a physical process intrinsic to how semiconductors work, it cannot be avoided. This causes voltages and currents to change over temperature making the circuit operate differently at different temperatures. The design of the circuit and the required properties of the circuit are what determine the temperature range of that circuit.

It is not so that an IC specified to work up to 85 degrees will stop working above that temperature ! In many cases it will still work but it might have a lower gain for example.

Nearly all ICs can withstand a junction temperature of at least 180 degrees or so without being physically damaged. This includes the ICs specced for 85 degrees. Think about it, it these would be damaged above 85 degrees, how can you solder them on a PCB ? You could not so the 85 degrees is not the point where damage occurs, the circuit just stops working as expected/specified.

A circuit in an IC specified to work up to 125 degrees could be the same as a circuit specified to work up to 85 degrees but the specifications might be more relaxed for the 125 degrees specification. Or for the 125 degrees ICs the manufacturer selects the "best" ICs. This does not mean they will measure them each at 125 degrees ! This is too time consuming. What is possible is that they select the ICs with the lowest power dissipation (at room temperature) and rate these for 125 degrees. Then the "lesser brothers" with an increased power dissipation might be rated for only 85 degrees and sold at a lower price.

| improve this answer | |
\$\endgroup\$
0
\$\begingroup\$

I'm not sure this is addressed enough in the other linked answer, but transient failure (ie it works fine again once cooled down) is about timing.

Very simply, we can model the delay in a logic stage as an RC timing circuit, where the resistor is the connecting wire (plus output FET channel) and the capacitor is the FET gate being driven. Increasing the temperature increases the resistance, and therefore the delay. When a signal is delayed enough so that it doesn't arrive at the next D-flip-flop input in time, an error occurs, which may leave the chip's control logic in a jammed state.

When designing a chip, the designer selects a "worst case" temperature to use while running simulations. This determines the design temperature limit.

The effective temperature limit differs not only between designs but between individual parts, due to manufacturing variation. People overclocking PCs have made use of this a lot.

| improve this answer | |
\$\endgroup\$

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.