PIC stress testing...

[ELCouz]

Hi,

I have some general use boards (using PIC24FJ128GA010) designed for me.

What kind of stress test (using CCS C compiler of course ;) ) can you do to see if the PIC is alright.

I heard some pic are counterfeit and / or recycled in the mainland china.

The PCB engineer and factory lines are in China.

What kind of endurance testing can you do with a PIC?


Just asking to be sure everything is legit!

Thanks!
_________________
Regards,
Laurent

-----------
Here's my first visual theme for the CCS C Compiler. Enjoy!

[temtronic]

hmm maybe the 'into the freezer, then the fire' test ?

See if the boards run and BOTH ends of the thermometer.
Also maybe a 'test' program that cycles the I/O lines at near max load/frequency. Best to have a contract clause saying IF they fail, mfr has to 'make it right'.

I know it's a lot cheaper to have them made 'overseas' but once out of the country you really have no control. I used to have a local PCB mfr make mine and I supplied the PICs from Toronto..sigh that was 15+ years ago....

Jay

[asmboy]

A counterfeit PIC, especially of the pic24 family seems VERY unlikely to me.

I'd test run actual programs on it, make sure i could read the DEVID and DEVREV registers and successfully program the flash configuration words.

There is way too much complexity in the part for anyone to copy it without the chip masks and actual process information at the fab level.
Making chips is a far cry from running a photocopier.......

The fact that it runs your 'tested' program would give me all the confidence i need. Were it a fake chip ( which i doubt is reasonably possible) i think it would not work at all. For a "genuine fake" chip - endurance and thermal stress would be the LEAST of your problems...........

I suspect you are worrying about a very improbable thing.

[Ttelmah]

A lot depends on what your end product is rated to handle?.

The standard is to temperature/humidity stress cycle sample boards for a number of hours. You may also use altitude tests stresses etc..
The normal code is what is used, with the 'caveat' that this should be written to record/report unexpected things. So code for this will do things like have a checksum on calibration & variable storage.

The PIC's, involved are actually genuine parts, marketed by Microchip as 'special needs' parts. These don't have the normal line testing and as a result are very much cheaper than standard parts, and are sold to customers who are prepared to accept their status in exchange for the low price. These have been reported to have been re-badged, and sold as 'real' parts by some suppliers in the Far East. There have also been completely bogus parts (often not even a chip from the same family), but these don't function at all (and won't program)....

[newguy]

We had a few test boards made in China. Once. Never again.

First and foremost, we asked that the boards be conformal coated. Gave them a list of connectors to mask and the conformal coating we wanted (Dow 1-2577). Boards we got were dipped in conformal coating. No masking. Conformal coating they used was indestructible (certainly not Dow 1-2577). No chemical I tried would even soften that stuff. Their advice to remedy the situation? "Get a small pointy tool and pick the coating out of the connector holes." High density approx. 0.010" square holes. Yeah, no.

Noticed that some of the components weren't what I specified. Wider capacitance and resistance tolerance than what I wanted. Some of the ICs were commercial temperature range, not industrial like I specified.

We demanded a refund and actually got it.

We now use a board manufacturer in California and I couldn't be happier with the service and quality of the boards they deliver. They never substitute parts on me either. Oh, and they don't dip our boards in adamantium-based conformal coating either.

[ELCouz]

[ELCouz]

[ELCouz]

[newguy]