Over-temperature Recovery

taylor2 · ‎2008-04-15

Posted on April 15, 2008 at 12:16

taylor2 · ‎2011-05-17

Posted on May 17, 2011 at 09:51

I am working on pre-release testing of a new product centered around the STR910. This testing includes environmental cycling to make sure the finished assembly performs rationally when running at the temperature margins.

The processor appears to halt most normal operations as ambient temperature approaches 78°C (I have not measured the processor directly, but it is probably getting close to 85°C with self heating). When I reduce the ambient temperature the processor returns to normal operation. These are good results, so far as they go.

Now for my question:

I would very much like to add internal diagnostics to my device that will indicate that an over-temperature shutdown has been encountered. Is there anything on the processor that provides indication of a over-temperature recovery? I would be very well pleased to have access to an interrupt or a register entry that provides recovery notification -- it is inconventient to have unaccounted lapses in my control scheme.

Thanks,

LTR

kais · ‎2011-05-17

Posted on May 17, 2011 at 09:51

Hi,

Sorry to say that the processor can’t detect that an over-temperature shutdown has been encountered.

I guess to have such a feature you should have an analog temperature sensor and comparator embedded in the chip which is not the case with STR9

Eris.

michael · ‎2011-05-17

Posted on May 17, 2011 at 09:51

LTR,

It depends on at least 2 things: whether your processor experiences a voltage drop when it becomes too warm, and if your STR910 (I don't know the data sheet...) has the measures to detect such a drop. If that is the case, you can probably connect a simple sensor to your processor (I2C...) to save the temperature, assuming it was not fried by then :)

michael · ‎2011-05-17

Posted on May 17, 2011 at 09:51

no, no, that won't work...build you circuit, and good luck.

taylor2 · ‎2011-05-17

Posted on May 17, 2011 at 09:51

In the over-temp condition for my particular circuit the processor is cutting out before the power regulation, so there isn't much I can get by monitoring the input voltage.

Once I get to real-world conditions, I'm not really interested in logging the peak temperature -- I just want to generate a diagnostic indicating that the over-temp shutdown was encountered. Without that diagnostic I have the potential of hitting unexplained gaps in my process control (the diagnostic wouldn't prevent the gaps, but it would at least explain them).

The comments I have been getting seem to consistently indicate that there is nothing on the chip that I can use for getting my desired feedback. I would not have complained to be given the lazy man's solution, but it won't be a big deal to implement an external circuit to capture the event.

Thanks for the help.

LTR

michael · ‎2011-05-17

Posted on May 17, 2011 at 09:51

LTR,

Can you maybe use your RTC? maybe you can detect an operational gap like that?