Thermal / Short-circuit Protection?

Arduino chips (atmega) are fairly tolerant to short circuits if its within its recommend operational voltages ~5V or less.

The I/O pins almost certainly have short circuit protection, shorting across other pins however carries far more risk, if your lucky you just strain the silicon and its thermal protection kicks in before any serious damage happens.

I've seen true short circuits on IC's, when first semester students connect 20V to a 5V 74XX chip the wrong way. You can actually see balls of molten silicon embedded in the plastic,

No they aren't! The IO pins fail all the time. If the output is commanded HIGH and then shorted LOW, (and vice versa), it's gonna fail, either really quickly for a true short, or maybe take a touch longer if the short is above 40mA. There is no short circtuit protection, there is no thermal protection. The VCC pins can support 200mA each, and if you short the output, the 40mA output drive transistors will give it all they have until they go.

randomname:
The I/O pins almost certainly have short circuit protection,

No they don't...

Firstly without a doubt the chip has thermal detection, its stated in the spec sheet as part of the hardware. Considering we have no access to this thermal detection (probably just a calibrated resistor), its part of some hardwired logic, aka thermal shutdown is only logical choice. Thermal protection is so easy to implement, it would be very odd that any micro controller designed in last 20 years doesn't have it.

As far as short circuit projection, I never said it had current limiting. But I've never lost a pin to a 5V short, if you lost a pin you probably had more then 5V or are just really unlucky. The internal resistance of the logic in a I/O pin is clearly enough to prevent current runaway, any slight defects in the silicon or copper during manufacturing is the more likely cause of pin failure, if it failed when you shorted it was probably going to happen eventually down the road even under normal load. If you leave the short for an extended period of time, 60 seconds+ then ya that might be enough to cause thermal degregation to the point of failure. I'm well aware that you wouldn't do it on purpose and that your putting a lot of strain on the pin, but even now as a test I've been shorting out my I/O pins and its been peachy fine, had I fired the pins I would have had to live with it.

I make no claim this holds true for other manufactures, my experience with atmel is that they are very well designed and to put into terms, very forgiving.

I'm not saying it can't kill the pin/port, I'm saying its unlikely. Don't bother screaming at me cause you barely passed high school electronics (do they even still offer that?), I would never recommend you make a short to ground on purpose.

randomname:
Firstly without a doubt the chip has thermal detection, its stated in the spec sheet as part of the hardware.

Which chip? And did you really mean thermal protection?

its stated in the spec sheet as part of the hardware.

Where?

Don’t bother screaming at me cause you barely passed high school electronics

Can I scream at you because I have over 40 years in electronics and you are talking tosh?

I went to the cinema today with my wife and we saw an add where someone answered a facebook question with a link to a song. She challenged me to answer a question on the forum with a link to a song.

So @randomname listen to the chorus - this one’s for you:-

Where?

atmega32U4 (lenardo)
Listed under "Peripheral Features"
Page 2, 5th line down "On-chip Temperature Sensor"

atmega328
page 1, makes reference to it
confirmed on pg.261, section 23.8

Don't you just hate how they bury this stuff where know-one can find it? If you ask nicely I'll tell you how to find it easy next time.

Can I scream at you because I have over 40 years in electronics and you are talking tosh?

40 years of doing it the hippy way doesn't hold a lot of weight.

1 Like

Yes, one of the ADC channels can measure the chip temperature. That's all it does.
Your sketch can read that and do something. Or not. The chip will not shutdown, will not protect itself, and may operate erratically, or not at all, if it gets too hot.

Thermal protection is so easy to implement, it would be very odd that any micro controller designed in last 20 years doesn't have it.

And yet it doesn't. That's up to you as the designer.

CrossRoads:
Yes, one of the ADC channels can measure the chip temperature. That’s all it does.
Your sketch can read that and do something. Or not. The chip will not shutdown, will not protect itself, and may operate erratically, or not at all, if it gets too hot.

Correct, nor is there any indication whether this temperature sensor is even suitable for thermal protection use. It could be in a totally inappropriate region of the die for that.

The very idea of thermal protection in an MCU is a bit odd. It’s a low power device with modest current source and sink limits. It’s not a darned voltage regulator or something. It’s supposed to be operated within limits. Why waste power running thermal protection circuitry?

For people who doubted its very existence, and couldn't seem to read a spec sheet, you seem to know quiet a lot about the chip and what it does or does not do.

If they designed everything on the assumption that it would operate within its limits, the very concept of current limiting and thermal protection wouldn't exist.

Low power is relative. Temperature is temperature, your trying to protect the silicon. Any time your sinking or sourcing a load of any size you have the potential to over heat. Saying its small and drives small loads so it can't get hot is silly, you might as well say regulators are large and drive large loads to it can't get hot.

If you've been following CPU's in the last 2 years you might recall the issue they are having. Even though they are consuming less and less power, the surface area is decreasing faster, resulting in hotter chips, the heat dissipation is getting worse.

As to wether they have thermal protection on this chip.... Personal observation suggests they do, I've prototyped tons of 32u4 and a few have had there share of cooking without any damage after a cool down.

Read the datasheet - "thermal" is not shown at all.
"Temperature" is, in terms of calibration being off at high temperatures, performance charts over temperature, etc. "Protection" is also used, but in any way related to temperature.
So you may have overstressed some parts and reduced their longevity with high temperatures, but the chip did not do anything to protect itself while you were doing that.

my experience with atmel is that they are very well designed and to put into terms, very forgiving.

I agree.

randomname:
...and couldn't seem to read a spec sheet...

Speaking of "couldn't read a spec sheet", you have yet to provide a datasheet reference to "thermal shutdown".

As to wether they have thermal protection on this chip.... Personal observation suggests they do, I've prototyped tons of 32u4 and a few have had there share of cooking without any damage after a cool down.

Well, in my personal observation, people who go the the movies on a Tuesday afternoon are idiots. You see, once, when I went to a movie on a Tuesday afternoon, there was this one guy who was clearly an idiot.

What's the phrase ... oh yes ... "specious reasoning". The fact that you have not noticeably damaged a processor does not mean the processor is not actually damaged. The fact that you have not noticeably damaged a processor does not mean the processor includes circuitry to perform a controlled thermal shutdown.

Don't bother screaming at me cause you barely passed high school electronics
40 years of doing it the hippy way doesn't hold a lot of weight.

@randomname, no more personal insults.

randomname:

Where?

atmega32U4 (lenardo)
Listed under "Peripheral Features"
Page 2, 5th line down "On-chip Temperature Sensor"

atmega328
page 1, makes reference to it
confirmed on pg.261, section 23.8

Don't you just hate how they bury this stuff where know-one can find it? If you ask nicely I'll tell you how to find it easy next time.

Can I scream at you because I have over 40 years in electronics and you are talking tosh?

40 years of doing it the hippy way doesn't hold a lot of weight.

You have not proved your case in the slightest.
You are making it up,
The data sheet says nothing about thermal protection. It is easy to have a temperature measurement channel, it is just a diode, protection is altogether another matter. And that is all the data sheet says.

Can you please indicate what experience you have with electronics over what period of time at what level. You mentioned students, are you a teacher? It seems you have a lot to learn.

randomname:
Page 2, 5th line down "On-chip Temperature Sensor"

atmega328
page 1, makes reference to it
confirmed on pg.261, section 23.8

Don't you just hate how they bury this stuff where know-one can find it? If you ask nicely I'll tell you how to find it easy next time.

LOL, pure comedy.

randomname:

Can I scream at you because I have over 40 years in electronics and you are talking tosh?

40 years of doing it the hippy way doesn't hold a lot of weight.

OK, now you've outed yourself as a troll.

I'm only giving you 1/10 - you got a few bites but it's not really subtle enough.