(4 mm Al mask was put upstream of AIDA near 20:20 on 7 June)

01.15 Run R1219 started on neutron rich setting.

            Temps ok. Leakage currents ok.

             Plate with holes in front of AIDA , after F11 plastic. F11 plastic count rate ~700cps.

             Keeps getting 'Attempting resychronisation' message in merger terminal window.

             driver 0 10205 consuming almost 100% of CPU. Merger ~20% and data links ~1%

              See attachments 1-4

02.15 Run R1220 n-rich setting (BigRIPS run 1218, EURICA off)

          EURICA team retires for the night and cannot get it up and running

          Biases and leak shown as attachments 5-6

          Merger info is shown in attachments 7-9

          CPU Load in attachment 10

          Timestamps against the ADCs and also SYNCS, respectively, attachments 11 and 12

          3:31: Run stopped for the operator to tune the beam

          F11 plastic rate was 700 cps (~heavy ion injection rate)

3:43 Run R1221 n-rich setting (BigRIPS 1219, no EURICA)

          4 mm Al mask was removed.

          Pause calls vs. Timestamp is attachment 13

          Merger screenshots are shown as attachments 14-16

          F11 plastic rate is 850 or 900 cps after retune

          ADC v Timestamp is shown as attachment 17.  It can be seen the rate was roughly constant before the operators stopped (corresponding to a long gap of no hits, but now  we go up and down quite a lot in a rather ugly way

          CPU stats (driver to write to HDD in green, merger in red) as attachment 18

          At the time of ~90 in attachment 18, the linkers start to consume all available CPU resources, shown as attachment 19

04:35 Run R1222 n-rich setting (BigRIPS 1219 / 1220)

          Ask the operators to reduce the beam intensity by "1/2" -> F11 plastic rate goes from 850 cps to 170 cps, so more like 1/4 effectively

          Merger stats are shown as attachments 20, 21, 22. 

          nnaida1 and nnaida24 seemed to often show pauses (among other ones)

          trying to understand what's going on with the display in Midas, because it has looped around to time zero and is putting new data on to the old data...?

         It seems this condition is much better than before, though

         05:33 stop the run

05:38 R R1223 n-rich setting (BigRIPS 1221)

           Ask the operators to resume the previous beam intensity.  (Now like 600 to 800 cps at the F11 plastic)

           attachment 23 shows the beam coming on near time zero.  within <100 seconds, the driver has to fight with many link64 instances and cannot operate correctly

           we understood how to clear the midas histograms so that we can make sense of things.  attachment 24 is ADC, and attachment 25 is pause. 

           you can see some data come smoothly until the system locks down

06.11 R1223 stopped.

06.30 Started run R1224.

            Moved Pb bricks downstream of F11 plastic inwards, from a separation of 15cm, to 4cm in an effort to reduce rate in AIDA.

            F11 plastic rate ~900cps.

            No change to vis scalar rates and data still being dropped.

06.45 Stopped run R1224.

            MIDAS plot of ADC vs Timestamp (attachment 26) shows characteristic behaviour.

            Upon writing to disk, for the first minute or so link64 processes consume minimal CPU power. But then very quickly they ramp up consumption and ultimately choke the merger (it seems).

            This is reflected in the ADC v TS plot. Continuous ADC data for first ~1min.

06.54 Moved Pb bricks to separation of 2cm. Started run R1225.

            AIDA vis scalar rates reduced slightly, but no change to CPU usage and usual happens - data lost, link64 CPU usage spikes and chokes merger (seemingly).

            F11 plastic rate ~900cps.

07.00 Run R1225 stopped.

           SYNCs stopped being produced. Lost contact with nnaida1.

07.25 Pb bricks moved to 1cm apart. Run R1226 started.

            No change to CPU usage (link64 still hogging everything). MIDAS online monitor lost in power cycle so cant view in real-time, but anticipate same outcome.

            Sopped producing SYNCs 07.39. Run 1226 stopped.

08.20 Some trouble restarting DAQ and merger.

            As nnaida19 has lost the 'H' for histogramming on the run control page, nnaida21 has now lost the 'X' for data transfer (see attachment 27).

           Removed nnaida21 from merger and normal state of play resumed.

           4mm holey plate + the one attached to it we installed dowstream of F11 plastic.

08.50 Run R1227 started.

           CPU usage by link64 ~1% and driver ~70%, much more reasonable. When reloading merger web page, channels flash between bright green and olive. Good.

09.09 Changed to reference mass setting. Run R1227 stopped.

09.12 Run R1228 started on reference setting with holey plate still in place.

           R1228 stopped 10.32 (for B2F/F11 entry)