PIC temperature, when is it too hot

[Demon]

16F18877 specifically

There's a bunch of charts in the datasheet with fancy curves, as well as the 85C limit for extended operation (section 37.2).

1. At what point is it recommended to start glueing a heatsink on these puppies?

2. Is there a rule of thumb? Or is it arbitrary?

3. And if you do cool PICs, what adhesive do you use? What heatsinks?


9817

[tumbleweed]

In all the years I've been using PICs I've NEVER had one get even slightly warm.
They usually only pull a few 10's of mA's max, but I suppose if you had all the IO pins trying to drive too much current it might increase the temp a bit.

Is yours getting hot?

[pedja089]

There is max current per port, and Vdd pin. That current is much lower than sum of all 20mA per pin. So you should not try to pull 20mA from each pin...
Pic should not get warm at all. Power disipation should be in mW.

[Demon]

In all the years I've been using PICs I've NEVER had one get even slightly warm.
They usually only pull a few 10's of mA's max, but I suppose if you had all the IO pins trying to drive too much current it might increase the temp a bit.

Is yours getting hot?

2/3 the pins are sinking inputs, the rest are unused for now.

Thermal cam says 36.4C.

I didn't have a thermal cam before, so I have no clue what "normal" is. It is a busy PIC though, 2 ADC and 2 PWMs.

[Demon]

There is max current per port, and Vdd pin. That current is much lower than sum of all 20mA per pin. So you should not try to pull 20mA from each pin...
Pic should not get warm at all. Power disipation should be in mW.

Good idea, I'll get a current reading on both power pins.

[Ioannis]

According to the 16F18877 data sheet, the Absolute Total Power dissipation is 800mW, which is quite a lot!

Absolute Max current on any standard pin +/-50mA, but this does not mean you can source or sink on all pins at the same time that current. You have to stay lower than 800mW total.

In any case these are impressive numbers but for good reliability better stay lower.

Ioannis

[HenrikOlsson]

37°C is nothing. At twice that I'd start thinking if there's something wrong.

[Demon]

I tested current draw at the VDD and VSS pins (like pins combined):

VSS 10mA
VDD 120mA

[Ioannis]

This means that your PIC is sourcing (from HIGH output pins) 110mA to some loads.

FOLLOWING SENTENCE IS NOT CORRECT. PLEASE DISREGARD: Total power dissipation is about 600mW close to the limit at the room temperature. Better do something to reduce that or make sure that PIC will not be in a hot environment.

Ioannis

[HenrikOlsson]

How do you figure the power dissipation is 600mW?
It would be if all the of the 120mA would go out the Vss pin, thus 5VÄ*120mA. But now, most of the current is going out the I/O-pins with FAR less voltage drop than 5V.

Let's be conservative and say that the voltage drop across the high side switch is 1V, in other words the pin sits at 4V when high.
Then the power dissipation of the PIC core is 5V*10mA=50mW and of I/O is 1V*110mA =110mW for a total of 160mW which is a whole lot less than the 800mW max, In reality I'd think (but do measure) the voltage drop across the high side switch is far less than 1V.

/Henrik.

[richard]

and of I/O is 1V*110mA =110mW

provided that the current is spread over a number of i/o pins and that no individual pin is overloaded

[Demon]

provided that the current is spread over a number of i/o pins and that no individual pin is overloaded

I rechecked my work to be sure, this is my busiest 40-pin PIC that I mentioned previously:

- 16 rotary encoders pins pulled HIGH,
- 4 momentary switches pulled HIGH,
- 2 ADC inputs,
- 2 PWM:
__- pin for LCD backlight, 3mA (2N2907)
__- pin for LED strip, 0.0mA (IRFZ44N) - so low it doesn't even register on my ordinary multimeter.
- USART TX and RX pins pulled HIGH,
- remaining unused pins are output pulled LOW

(waiting for new fuses to retest using my TES multimeter)

[HenrikOlsson]

provided that the current is spread over a number of i/o pins and that no individual pin is overloaded
Yes, the voltage drop across the switch increases as the current goes up. That's why I was being conservative with the 1V voltage drop.
But really, as long as the voltage drop isn't MORE than 1V it wouldn't matter (from a power dissipation standpoint) if the 110mA exits the device thru one pin or 25. Of course, you'll likely destroy the pin but it's still 110mW dissipated.

Point is, 600mW isn't being dissipated in the PIC. Keep the current to within specification and you'll be fine. No heatsinks, fans or watercooling needed. If you think you need that (chip is burning hot) something is wrong.

BTW: Am I the only one not being able to edit my posts? All I get is a blank "edit box" as if I was making a new post.
EDIT: Hmmm, not this time, on this machine. Weird.

[Ioannis]

What I was thinking... Henrik is right as all that current is flowing out of the PIC and not to ground.

So indeed, the only power loss is the voltage drop on the output buffer times the specific pin current.

Sorry about that, guess it was too late at night to post anything...!

Ioannis

[tumbleweed]

- 16 rotary encoders pins pulled HIGH,
- 4 momentary switches pulled HIGH,
- 2 ADC inputs,
- 2 PWM:
__- pin for LCD backlight, 3mA (2N2907)
__- pin for LED strip, 0.0mA (IRFZ44N) - so low it doesn't even register on my ordinary multimeter.
- USART TX and RX pins pulled HIGH,
- remaining unused pins are output pulled LOW

It's hard to see how that ends up sourcing 110mA out of the IO pins... I would have guessed less than half that.

[richard]

BTW: Am I the only one not being able to edit my posts? All I get is a blank "edit box" as if I was making a new post.
EDIT: Hmmm, not this time, on this machine. Weird.


not anymore

i cannot edit my posts either

[Ioannis]

Maybe clearing cache, or a different browser resolves the problem?

Ioannis

[richard]

Maybe clearing cache, or a different browser resolves the problem?
tried all that. problem still comes and goes, seemingly at random

[Ioannis]

Then only boss can check that...

Ioannis

[Ioannis]

It got me too... I really cannot understand what is going on.

It showed me an empty text box. Was it the same with you too?

Ioannis

[richard]

That's it, an empty box

[richard]

That's it, an empty box

But not this time, but reason for edit box has gone missing


Third try reason box is back

[Ioannis]

At the same time, same browser, one message that had a degree character won't let me edit, other message does.

Do not know if that degree symbol (°) does anything cause there was no other strange thing there.

Ioannis

[Ioannis]

Yes, I confirm it does...! The degree blocks me for editing!

I don't know about yours Richard...

Ioannis