We have a BBB-based design currently undergoing FCC testing, and they’ve found a bit of noise at 625 MHz. Has anyone else encountered this, or know what it might be, or how to mitigate it?
In order to mitigate it you need to figure out where it is coming from. A probe test will help in that area. Have you run one? Using ferrites on all cables generally helps in these types of issues.
It seems to come from right around the processor, but we couldn’t pinpoint it any more precisely than that, or even to one side of the board or other.
Sounds like you may have grounding issues or unterminated pin . Using a probe should let you be able to determine which side of the chip it is.
It’s not something unique to our board–a stock BBB has the same problem.
OK. Understood. Different test equipment will yield different results. And it can be effected by the SW you are running.
625M clock noise maybe from 125M clocksource， or 25M
This might be from ethernet
Do you use GIGA ethernet ?
Try to unplug the RJ45 cable
In a similar problem it helped for me to replace the 0.1R that connectes Ethernet shield to the frame of the Beaglebone by a simple solder bridge.
This resistor is located on the solder side of the Beaglebone, just below the network connector.
I have a spectrum analyzer and a 3GHz active probe and was astonished about the amount of infuence.
Sounds crazy but prooven many times meanwhile.