hot weather, "problem uploading to board"

I have an arduino nano v.3 on a windowsill in the sun.
Some afternoons, the IDE on the linux PC won't upload the program and reports
"problem uploading to board"
Next is random between "programmer is not responding" or a wrong response
Throughout this, the nano continues doing its existing program for example blinking an LED, and is listed in

 ls /dev

as

 /dev/ttyUSB0

Repeating the upload from the IDE sometimes got a different "incorrect response" failure and sometimes worked normally.
I have noticed that it has been more reliable in the morning and late evening.
Is the uploader expected to be unreliable at elevated temperature?
Do I need a cooling fan as well as the "sun hat" shade which improved things somewhat?
Is there anything else which experienced users recommend?

you mean it's directly exposed to the sun ?

Absolute Maximum Ratings for ATmega 168/328

Operating Temperature.................................. -55 C to +125

Storage Temperature ..................................... -65°C to +150°C

Is the uploader expected to be unreliable at elevated temperature?

It is not the uploader it is the whole system. Do you know what temperature it actually is?
It is likely that anything over 40C will start to give problems.

I have an arduino nano v.3 on a windowsill in the sun.

Then move it out of the sun.

It is definitely nowhere near the "absolute maximum ratings" for operating, but that end of the room does feel pretty warm today. 40C ambient is possible. I've been in hoter ambient than that in Australia in the shade. The nano is so small that it might be getting a bit warmer than ambient in normal use. By normal, I mean cpu busy for 20% of the clock cycles, about three digital out pins connected through 10k resistors .. not overclocked and cooked ..

What do other nano users find?

ad2049q:
It is definitely nowhere near the "absolute maximum ratings" for operating, but that end of the room does feel pretty warm today. 40C ambient is possible.

In my opinion if it's directly exposed to the sun then yes 100 degrees of Celsius could be reached on the ATmega after a period of time..
But anyway it's pretty warm and probably that's why you face those problems...

Try the internal temp sense sketch, and see what it reads out.

I'll try uploading that tonight if it magically clears up its problem. Once that is in, I'll report what that shows on the next sunny afternoon.

zaxarias:
you mean it's directly exposed to the sun ?

Absolute Maximum Ratings for ATmega 168/328
Operating Temperature.................................. -55 C to +125
Storage Temperature ..................................... -65°C to +150°C

The chip's Absolute Maximum Ratings is not the same as the systems Absolute Maximum Ratings .

Grumpy_Mike:

zaxarias:
you mean it's directly exposed to the sun ?

Absolute Maximum Ratings for ATmega 168/328
Operating Temperature.................................. -55 C to +125
Storage Temperature ..................................... -65°C to +150°C

The chip's Absolute Maximum Ratings is not the same as the systems Absolute Maximum Ratings .

i never said it is.. i just mean that the chip might be overheated causing those problems ..

Well, now hat the sun has gone behind the hedge, it is showing (uncalibrated) 16C rising to 18C in the first ten seconds.
Raw wADC values are about 362 to 366 at present. I think I might try that again tomorrow.

Does anyone know reason to object to making a "torture chamber" to get it up to 60C ambient in a cardboard box with a fan-heater? Apart from the flammability of cardboard, that is.

No problem, we used to do it all the time when I was at work. It is a standard technique.
We had a small heat chamber called "Bernie", it was a passive chamber, just an insulated box.
On the other hand "tostie" had a small heater in it.

Well, now that I've got it showing a steady wADC=366 on an LCD screen, I can test it beyond the range at which the PC won't talk to it. K type thermocouple shows about 26C ambient with no sun yet.