Hi all,
Well, I was changing the power supply to the Zyxel after realising I was using another 12V one with it and not the original (had the Fritz!Box one connected to it). I noticed this while I was packing and saw the Fritz!Box didn't have a power supply in the box. Anyway, after switching them over I noticed my trickling CRC errors have returned.
However... what I did notice was whenever I slightly move the cable in the master socket (test socket) my CRC errors spike. At one point the SNR even dropped a fair bit until I wiggled the connector again.
My question is whether this is normal, for an ever so slight wiggle of the connector plugged in the master socket to cause CRC errors to spike? I don't imagine it is. This would likely explain why the SFI engineers found no issues when they bypassed the master socket/test socket and directly connected to the wires, at least from what I call. From what I recall the final visit the SFI engineer noticed the same problem when he plugged in his JDSU via the master socket and not directly to the wires behind it.
I'm currently trying to stop the trickling CRC errors with the theory that it is indeed perhaps a defective master socket.
EDIT: P.S. I'm pretty sure it's not the cable itself as I previously tried other cables when I had this issue in the past. Moving house in just over a week so hopefully this issue will no longer haunt me
.
EDIT 2: I've re-synced with a more aggressive cap/banding in order to try and preserve fastpath as I prefer lowest latency over highest speed. So far it looks like that's helped. This is a very strange problem, but as I say I hope at the new house I won't see such issues and therefore won't have to apply such an aggressive cap/banding.