21/04/15

1325: Recabled the MACB distribution tree as per PC-S instruction yesterday.

          Now have nnaida5+6, 13+14 in MACB3 (nnaida5 in output 1) and nnaida3+4, 11+12 in MACB10 (following the numbering system used in yesterday's ELog entry).

          Complete system now happily resyncs on initial reset after power cycle.

          nnaida7+15 still give a temperature reading of 0^oC and nnaida8 is still running hot, but stable, at 65^oC.

          Visual inspection of cooling crate didn't reveal and air gaps between cooling plates and the FEE cards, or anything else to suggest there would be lower than expected cooling performance.

1340: After complete system start up, nnaida3+4, 11+12 now show absolutely no data (empty histograms, no good events/SYNC pulses in stats). (Attachments 1-4, example good spectrum/stats attachments 5-6)

          nnaida5 shows nonsensical data in histograms. (Attachments 7-8)

1500: Swapped HDMI cables from MACB NIM mdules cabling for FEE modules nnaida1+2, 9+10 for nnaida3+4, 11+12 - problem appears to move implying MACB is the problem

          Swap MACB #13 for MACB #4 - problem resolved (still in position #10 as per CG numbering system used previously).

1600: nnaida5 does not produce sensible data - typically ~3-4 active channels with ADC and fast disc data with rates ~kHz/channel

           Swap nnaida5 & 6 for (currently not used and not powered) nnaida27 & 28. Update /etc/dhcp/dhcpd.conf, restart dhcpd.

1800: Sync problem resolved, stable operation all afternoon including several system resets - no ReSync issues noticed. SYNC pulses provided to all FEEs consistently.

          Connected detector BB18 2977-20 (same detector as used during Feb 2015 tests, with same kapton config as per 1520 on Feb 19th) to system with +ve 100V bias provided to n-side.

          Updated firmware in new nnaida5+6 (formerly nnaida27+28) to be in line with all those currently in use.

          Used saved ASIC settings from Feb 2015 visit and see similar FWHM in most spectra to those found during February after connecting heavy duty copper braids in loop between FEE modules in use.

          nnaida5 still does not produce sensible data with almost identical spectra/stats to those seen before the module was replaced.

          To be investigate tomorrow.