I didn't mention earlier, that the 25MHz clock for the LAN9514 is generated by the CM4S.
The clock is configured in the dt-blob.bin with this code:Exactly the same configuration is used on the CM3.
Now we performed some measurements on both systems. The generated clock on the CM3 looks like this:
For comparison, the generated clock on the CM4S looks like this:
Both are screenshots with infinite persistence so we can verify the jitter. The only apparent difference between the 2 clocks is that the clock signal generated by the CM3 has a full swing from 0V to 3.3V, while the one generated by the CM4S oscillates between 0.5V and 3V.
Could this be causing the Ethernet issue we see with the CM4S ?
The clock is configured in the dt-blob.bin with this code:
Code:
clock_setup { clock@PWM { freq = <2400000>; }; clock@GPCLK0 { freq = <12288000>; }; clock@GPCLK1 { freq = <25000000>; }; };Now we performed some measurements on both systems. The generated clock on the CM3 looks like this:
- Image may be NSFW.
Clik here to view.
For comparison, the generated clock on the CM4S looks like this:
- Image may be NSFW.
Clik here to view.
Both are screenshots with infinite persistence so we can verify the jitter. The only apparent difference between the 2 clocks is that the clock signal generated by the CM3 has a full swing from 0V to 3.3V, while the one generated by the CM4S oscillates between 0.5V and 3V.
Could this be causing the Ethernet issue we see with the CM4S ?
Statistics: Posted by ondwis — Thu Nov 21, 2024 11:08 am
