Another oops to amuse you.....

[Andrew Gabriel]

(poor cooling design by Inno3D). Radeons don't break like that. They run at

100C for hours without breaking, they just start making graphical errors. Why can't they make them with a thermal throttle like the Intel processors?

My nVidia GPU has a thermal throttle at 125°C. The value doesn't appear to be configurable (at least, not via the GUI).

It normally ran at 47°C, but the fan is now knackered and it climbs to 65°C every few months, until I pull the fan apart and clean out the bearing again, and then it's back down to 47°C until next time.

I've bought a replacement card, but the driver doesn't like it. Haven't got around to updating the driver yet.

[SteveW]

Modern, surface mount components can actually be glued in place with a solder paste and then the whole board heated to melt it. You can buy special ovens for doing it or convert an ordinary worktop oven with external control to ramp and hold the temperature at controlled rates. It would not therefore be suprising that a fix might work by baking the board in an ordinary oven, however the poor control may well damage components unless you are lucky.

SteveW

[Lieutenant Scott]

n in use (poor cooling design by Inno3D). Radeons don't break like that= . They run at 100C for hours without breaking, they just start making g= raphical errors. Why can't they make them with a thermal throttle like = the Intel processors?

At 125C it would probably break first. That is an absolutely stupid thr= ottle level.

My Intel processors have always throttled at 95-100C. My AMD GPU makes = mistakes at 100C. My Nvidia GPU broke at 95C. There is no way 125C is = good for your GPU.

[geoff]

In message , Lieutenant Scott writes

I think you should refrain from going near anything other than wax crayons

hang on ... 200C?

Perhaps those as well

[geoff]

In message , "Dave Plowman (News)" writes

Clue ...

See what temperature your electrolytic capacitors are spec'd to

[Lieutenant Scott]

[geoff]

[Lieutenant Scott]

>> Must be true, its on the internet

If it was false there would be contradictory articles.

[John Rumm]

>>>>>>> Must be true, its on the internet

Its not false, but it is poor practice.

Using a pre-heated oven is also likely to exceed the component temperature maximum slew rates. Most components will not only have a maximum temperature, but also a max temperature rate of change specified (e.g. 4 deg C per sec). This applies bit on heating and cooling. Exceeding that can stress them and lead to premature failures. That's without the problems of exceeding the maximum temperature limits for the larger discrete components.

[Lieutenant Scott]

>>>>>>>

It's probably the best that can be done with domestic appliances though.

[geoff]

You could write one, couldn't you

[Lieutenant Scott]

Yes, when trying to fix a fault likely to be completely unrelated to solder joints, in an old Indesit oven, this method goes wrong.