21569 SOM Fault Light Suddenly Always On

Hi,

The SYS_FAULT LED glows, when exceptions or error occurred or try to boot the processor when the .ldr file is not flashed into external SPI flash.

Please provide your comments on below points to assist you better on this issue.

1. Could please confirm that this is the first time you are facing this issue after using the board for a few weeks?
2. Did you perform any booting in this board? If so, please provide more information on this (like application, boot format, ldr generation)
3. At present are you able to connect the board in debug mode?
4. If you are able to connect with board, we recommend to flash simple LED application with command -erase all

Regards,
Divya.P

It wasn't even connecting to CCES through the ICE1000 debugger.

We've actually managed to repair the SOM.

The LED is connected to the SYS_FAULT pin, and according to the 569 HW ref manual, this is asserted when the DSP is not detecting an input clock (there may be other reasons too but I couldn't find any in the manual)..

So we probed on a scope the input clock pin and indeed there wasn't a clock anymore. After removing the zero ohm resistor on the oscillator CLK0 output there was no output from the oscillator.

The oscillator on the SOM luckily has a second 25MHz clock output CLK1 which was still working fine, and it's only connected to one of the SOM headers.

We managed to connect this clock onto the pad of the corrupt clock to feed it into the DSP.

It's now working fine again.

I'm not sure why one clock would die in this way, other than perhaps a loading issue presented by the DSP clock input, so am hoping this won't happen again with the second clock.

Hi,

To assist you better, please share us the probe image of input clock (SYS_CLKIN0) and share us the silicon revision and BOM revision of the board.

Regards,
Divya.P

6102.SOM photos.zip

Attached is a zip file with photos of SOM and its box showing serial numbers and revisions.

The CLK0 output, which goes to SYS_CLKIN0, measured on pin 22 of the SI5356A had dropped to about 400mV pk-pk both with and without the R23 zero ohm resister removed in between CLK0 and SYS_CLKIN0.

The other 25MHz clock was still at around 2.8V pk-pk, which is what was routed to SYS_CLKIN,

I will send a scope image soon,

Thanks

Hi,

Can you please share the scope images.  This would be helpful for us to check your query with our Internal team.

Best Regards,

Santhakumari.K