I noticed last week that the IPC receive errors at all end station SUS and HEPI for the LSC signals is non-zero. Following the discovery of LLO LSC transmit errors to the end stations due to CPU overrun on the LSC model, I verified that our errors are also due to the LSC model running longer than its average of 40uS. The attached plot shows the error going to ONE at the time the LSC cpu increased to 44uS. The error latches, which is why we don't see any indication of the subsequent 45 and 44uS excursion.
I am running a script which clears the diagnostics on H1SUS-ETMX every minute. The next time an IPC error is seen it will reset after a minute ready for the next event.
All the error rates I've seen so far are only for ONE transmission of the 16384 sent that second. So the error rate is about 3 in an hour or 5.e-6 %