Page History

Schedule

Day shift: 9am–9pm /  Night shift : 9pm–9am

...

10:30pm MCC thought they are close to be done with tuning so we started to move Caladium coarse X in reduced x direction and running the Caladium only readout. Before reaching the beam spot, the secondary scatter was ~50 particles/shot but once beam spot came into the frame, that turned to be more like 2000/crossing. Beam spot was also very wide in X ( more than the 2cm MIMOSA aperture). Asked MCC to reduce rate and spot size. a few tries didn't make much impression until a move to close down SL-10 to -0.5. That brought the rate down to ~100 particles/crossing. Beam spot is filling good fraction of aperture ans looks like a good tuning setup to start with tomorrow. The Caladium XYstage X/Ycoarse RBV=115.4/3.8mm (Y is very close to min limit). Run 336 tuning beam spot picture below. 

Image Added

Done for tonight      

...

Activities:  Ben, Timon, Adrien, WilliamPatrick, Sanha, SD

9:45 AM Beam turned on with ~40 hits per beam on MIMOSA

...

• Continued accumulating statistics with the current 2 LBNL RD53a modules at 0 degrees
• Investigate cosmicGui setup and routing beam trigger to HSIO2 for independent Cosmic telescope runs  
• Continue developing the beam vs 40Mhz clock phasing measurement  

Activities:   Ben, Adrien, Elisabetta, Cesar, Sanha, SD 

9:21  Started Run 460 with both RD53a. Same rate as yesterday.

...

14:58 Stopped Run 470.

15:01 Started Run 471.

Data summary:

15:34 Stopped Run 471. Requested access from MCC to fix the trigger for the cosmic telescope. Got cosmicGui running with cyclic triggers after realizing that the key was to adapting a working version of .cosmicDaq.rc. Got clarification from Martin on the HSIO2 cosmic trigger connection scheme: a) 4 LEMO ports on RTM are NIM; b) main board LEMO Ext 1,2 are TTL; c) main board Ext 7,8 take LVTTL (2.5V) are Ored with 1,2; d) TLU RJ45 takes TLU triggers. Found trigger 11 (NIM) used previously by UCSC and connected it to the outermost RTM scintillator trigger port on the HSIO2 board (scintillator0). Also connected LEMO from TLU TTL out Trigger 1 to HSIO2 main board Ext Trg 1. This needs to activate in EUDAQ config to run parasitic to main partition when the is up. Don't have level shifter in hand to adapt the HitOR trigger from cosmic telescope for now.    

16:03 Started Run 472 after access.

16:03 cosmicRun 034 First data in LBNL Telescope via cosmicGui. After coming back upstairs and changing the trigger settings in the cosmicGui, we were able to read out data from the cosmic telescope (see cosmicRun 034). The ESTB beam trigger delay (UCSC LGAD trigger) is set at -104500, and the trigger delay on cosmicGui is set at 228, and telescope/DUT latency set to 10 (reversed def for FEI4, low number means high latency).  

The timing distributions of the hits in cosmicRun 034 peaks at two different bins: a small peak at bin1 and a large peak at bin5. Decreasing the cosmic trigger delay by 4 bins (to 224), we see that both the peaks move by 4 bins, suggesting that the two-peak feature is not due to the detector noise (see cosmicRun 035).  Increasing the cosmicGui trigger deadtime to a large value (100 ticks) to avoid double-triggers, we still see essentially the same behavior in cosmicRun 036. After a longer run (cosmicRun 037) and some analyses & discussions, we infer that the beam is indeed split into two bunches, roughly 100ns apart. We also suspect that certain "bad" pixels might be reading out multiple times after it gets hit by real particles (roughly every 100ps); these pixels have very high hits in only the "second bunch" of the beam (but not the first), and only these pixels read out hits well after the beam is gone (well after the second bunch peak). The behavior seems to happen across all the 6 modules.

16:36 Ended Run 472.

16:37 Started Run 473.

17:10 Ended Run 473.

17:11 Started Run 474.

17:44 Ended Run 474.

17:45 Started Run 475.

18:01 Started cosmicRun 038 with beam trigger delay at -104500 and cosmicGui trigger delay at 224.

18:18 Ended Run 475.

18:19 Started Run 476.

18:40 Data taking stopped abruptly for both the cosmic telescope and the two RD53A modules. Couldn't ssh into computer downstairs that starts the TLU Producer. Shut off beam and requested access to see what happened to computer.

18:45 Su Dong and Sanha accessed the hall to power-cycle the computers downstairs. Suspecting that this could have been a heat issue, we put a large fan blowing at the esadutdaq1 computer.

19:20 ar-eudaq machine was also frozen, so it was reboot.

19:45 Successful data taking resumed for RD53A

20:17 Started cosmicRun 051. LBNL cosmic telescope moved around slightly to better center the beam: current config X stage -3.50mm and Y-stage -83.99mm. cosmicGui trigger delay is also set to 216 now, in order to capture just the first bunch of the beam and avoid the suspected "double-ticking" effects by some bad pixels.

21:01 Ended cosmicRun 051 due to the end of beam time.

Data summary:

Nov 6

Note:  ESTB period is  9:00am–11:00pm  today with extra 2 hours   (Wednesday will be shorter 11am–9pm) 

Plan of the Day:                 

• Switch to tilted data taking with the two LBNL RD53a
• More cosmic telescope data with separate cosmicGui partition 
• DAQ clock phase vs beam tests 
• Depletion HV scan 

Activities   (Adrien, Timon, Ben, Maurice, Charilou, Patrick, Mark, Sanha, SD) 

9:30 Sanha called the beam control room to start the beam. There is a water leakage in the klystrons, and the repair/replacement will take 3-4 hours. Test beam data taking will be postponed until the repair/replacement is complete (probably in the early afternoon).

9:50 cosmicRun 051. Testing the current telescope configuration for any random noises. Cyclic trigger set to 100Hz. Almost no random noise is seen after 100s.

11:00 RD53a DUT stage slight rotated to get slight charge sharing in "phi". Tape measured distances derived an angle of 11.8^o. Reconstructed angle from photo gave 13^o  which are consistent with errors.  

Image Added	Image Added	Image Added
RD53a DUT position before Nov/6 (perpendicular)	RD53a DUT position before Nov/6 (zoom)	Nov/6 DUT after rotation

11:50 Beam is back on (earlier than expected!)

11:52 Started data taking

12:35 Channel 10 beam trigger disappeared while Ch 11 was driving the cosmic telescope at 5 Hz happily. Tried to run ch 10 trigger with a generated 10Hz from the trigger module and that also worked. Call Carsten and Mike Dunning and both had no idea how that could happen. The trigger unit was clearly receiving the accelerator signal correctly to run ch 11 while the down stream part delivering trigger to ESTB was also fine with the generated 10Hz.   

13:05 When in for access to see what's going on. Swapping the trigger channels 10 (originally feeding the EUDET) and 11 (originally feeding the cosmic telescope). The original trigger settings are: width 200ns, delay 110488ns for channel 10 (EUDET) and width 100ns, delay 104500ns for channel 11 (cosmic telescope). This didn't work and quickly found out that Ch 10 is NIM but Ch 11 is TTL so we cannot really swap. So we are back to the original pattern of ch 10 running TLU. The Ch 10 NIM pulse looked a bit marginal with slow ramps barely getting to asym level. Inceasred the width to 300ns and also noticed some ripple after pulsing. Cleanup up the setup so that CH 10 BNC swicthed to LEMO to go in a NIM Fanout unit  input and much more polished square pulse output from the Fanout is sent to the TLU. The fanout unit Thr is tweaked to wipe out the after ripple pulses. This cleanup seemed to have worked, but not clear where things fell exactly before this.       

14:55 Beam back after some struggle at MCC with turning on the A-line magnet. Started Run 497.    

15:29 Noticed the beam was off but trigger was still firing because magnet was still on, not sure for how long but there will be less data in Run 497 because although the number of events kept increasing there were no hits on RD53a bc the beam was off.

15:31 Full-ish(10K but not really) run 497 ended and Run 498 started.

16:05 Ended run 498 with full 10K events and started Run 499.

16:39 Ended run 499 with full 10K events and autostarted Run 500 but got an event mismatch error that killed all of the producers.

16:42 Started Run 501.

17:15 Ended run 501 with full 10K events and started Run 502.

17:34 Ended Run 502 with ~4K events.

17:36 Called MCC to increase rate of electrons from 6 per pulse to 600 per pulse.

17:44 Started Run 505 and will try to run Picoscope during this run. Also since the rate is increased, the runs will only be taken until 2K.

17:51 Stopped Run 505 with 2K hits. Picoscope worked well and will continue to use it. Beam went down momentarily and came back up in a couple of minutes.

17:56 Beam came up and started Picoscope readout before starting data taking for Run 506. Will continue to stop runs at 2K events, Procedure for using picoscope after a run: kill the picoscope program, take down the last 5 digits of the picoscope data file corresponding to the run (ll ~/slactb_scope and take the most recent files last 5 digits) and then start the picoscope program again(build/bin/slactb_scope) and start a new run in RC panel.

19:09 Stopped runs to try and get synchronous front end working.

19:59 Attempts to fix the synchronous front end failed, after returning settings to how it was before trying to fix the front end the chip won't turn on. After many attempts to remotely power cycle and fix the problem, an access was requested to go down and fix it manually.

22:11 After graphing the ToT from runs 522-530, noticed that it was not changing as much as expected, leading us to believe that maybe remotely changing the HV might not be working. To check, we set the voltage to 0 and took data but did not see what we expected. The camera pointed at the equipment rack didn't have the HV PS in the frame so we requested an access to go down and see whether remotely chaning the HV actually works or not. We confirmed that the HV was changing but the script must have an indexing error on its for loop because whenever the HV is set, it always ends up at a value 1 off from what was set. For example if we were at 20V and set it to 15 we would end at 16 and if we were at 10 and set to 15 we would end up at 14. There is also an Errcode flashing n Keithley at each stage. After figuring out that the voltage was changing but not to where we expected, a new webcam was moved to have the HV panel in view so it could be monitored.

22:52 Beam restored and Run 534 started.

23:03 MC called to take away beam after Run 534 was completed.

Nov/7 

ESTB beam **11am–9pm**     

Plan of the Day

• SLAC Rd53a device 0x0A97 (25x100 pixel 150um active thickness) replaces the current channel D 0x0A57 (50x50) early enough in the morning to allow Timon tuning it remotely 
• Setup a cleaned up TTL beam trigger from Ch-11 for cosmic telescope to see if the double timing peaks still exist.
• Main data taking runs will be steep tilted 25x100 sensors. One DUT at a time, first 0x0A54 (100um LBNL) then 0x0A97 (150um SLAC). 

Activities:  Adrien, Mark, Charilou, Cesar, SD 

10:30 Replaced 0x0A58(labelled A57 in the configs) with SLAC RD53A device 0x0A97.

11:12 Beam was up and started taking runs, but no events were being built, called Timon and he said module A54 does not like to be power cycled. So we requested access to go down and fix it. Fortunately on power cycle was sufficient to bring A54 back. 

11:47 Beam was turned back on after module started working again and Run 541 was started with just one Yarr producer in the run control for the LBL RD53A A54 module. Timon is working on tuning the SLAC device to get it up and taking data.

11:55 Lost beam for a couple of minutes but it came back up when Carsten called.

12:52 MC called to say one of the klystrons in the injector is missing triggers so the beam might be coming in and out.

12:57 Stopped run 543 to change HV back to 100V.

13:40 Timon finished remotely tuning the SLAC module A97 but when a run was started there were no hits seen by the module and there were some infrequent readout errors (received data not valid). Timon is attempting to fix the problem but for now only the LBL module A54 is being run.

Access to remove the A97 to leave A54 by itself and changed to large tilt angle ~ -60^o (had to rotate clockwise when looking from above, since mount frame would be in the beam line before sensor if rotated the other way).  Made choice of 60 degrees based on expectation of long cluster of 8x25um pixels for 100um thick sensor. More steep tilt would need to remount RD53a to the opposite side of frame to avoid frame in the beam line.  Couldn't mount 2nd sensor - missing screws to slide 2nd frame into beam line. One device at a time is better anyway to reduce the large multiple scattering at this steep angle. For this tile angle, hits will traverse pixel sideways with a path length of 25um/cos(60deg) ~29um.  Expected signal  peak for such short length is <2000e and really needs pixel threshold to be 1000e or less. Our pixel threshold is tuned for 1000e.        

Image Added	Image Added

The resulting ~4-5 pixel length of "clusters"  might be due to the artificial local XY coordinates are define against the 50x50 chip channel grid to achieve 25 micron pitch with staggered bumps ?    

Run_544 (13 deg)	Image Added	Image Added	Image Added
Run_565 (62 deg)	Image Added	Image Added	Image Added

17:50  Took a short run with cosmic telescope joining the main partition (so that trigger is driven by Ch-10 with NIM Quad discriminator cleanup). The FE-I4 data had no no double timing peak here in comparison to the standalone cosmicGui run just taken with trigger Ch-10 (TTL). This seems to indicate the standalone Ch-11 cosmic triggers was causing the double time peaks. The RTM ports were meant to be NIM so we were triggering on -ve undershoort after ripples of the main TTL trigger pulse ?    

...

21:40. We're done! Called to stop beam for entry. Going to pack up.