BCS UPDATE. j. welch 2/9/17

Transcription

1 BCS UPDATE j. welch 2/9/17

2 TOPICS RP requirements Shutoff path Beam loss detection scheme Beam loss detectors and FPGAs Current monitors Dumps

3 RP REQUIREMENTS Revised BCS PRD was circulated Tuesday for comments. Still to come: RP requirements for rastering, final protection collimators, auto-bypass for Undulator Complex zone, and bremsstrahlung in the FEE. Selected highlights BCS fault requires shutoff of RF in L0,L1, L2, and L3. MCMs on the laser heater dipoles Shutoff Gun RF if loss near gun exceeds 100 W. More generous overall shutoff times > Choice of continuous LBLM coverage or ACM comparator for S11-S27 BYD magnets are in series with BTH dogleg dipoles

4 BCS SHUTOFF PATH When a (BCS) fault is detected, all beams must be shut off by three independent methods. * 1. Acousto-optic or Electo-optic rep-rate control (pulse picker) device triggers are removed. 2. Mechanical shutter 3. gets complicated * Radiation Safety Systems Technical Basis Document

5 AOM Mech. Shutter Gun L0 L0,L1, L2, L3 up to 1 hr recovery Option 2 inexpensive, reasonable recovery expensive to implement because of number SSAs, reasonable recovery Concerns have been expressed possible gun and or cavity field emission hazards would persist even if the photo beam is turned off successfully if laser heater chicane magnets are off the gun current would reach L1 and be accelerated. RP s BCS PRD under review says turn off Gun, L0, L1, L2, L3 We are going over the details with RP.

6 OPTION 2 Fill time is 10 ms, so beam would go off in ~1 ms when the energy drops below the acceptance. Cryo load change is small and mostly balanced by load resistors that should come on. Recovery time for cavities to ramp up gradient while keeping in resonance is seconds (not minutes)

7 Linac RF Shutoff Mechanisms Control System PPS Fault LLRF 24 V interlock out BCS Fault PPS Breaker 480 VAC 480 VAC Breaker SSA 24 V interlock in 120 VAC SSA shuts down if 480 VAC is out of range

8 OPTION 2 HAZARDS Gun field emission causing radiation in the laser room, or (during gun only running) downstream of the temporary shielding wall. Cavity field emission from L1, L2, L3 causing radiation in the laser room or in the klystron gallery Cavity field emission causing damage to stoppers, protection collimators or dumps.

9 GUN FIELD EMISSION UNDER OPTION 2 SHUTOFF Example of dark current. CM01 is powered in this simulation. (C. Mayes) Gun dark current would mostly get absorbed on the vacuum pipes and in the CM01 cryomodule. Some would reach the laser heater chicane where it would be absorbed because of the ~100x energy mis-match with the laser heater chicane. Power is likely to be very small: 400 na (max) x 750 kv = 0.3 W

10 WORST CASE GUN FIELD EMISSION 10 μa emission: original emitted up to 6 μa. This is possible but unlikely and is less than 5 W. 60 kw emission: theoretical SSA power supply limit (1/2 of 120 kw). Voltage goes to 0.5 MV. This is theoretically possible, but almost certainly will never happen. Working with J. Blaha to define mis-steering case and MCI.

11 CAVITY FIELD EMISSION UNDER OPTION 2 Normal configuration - chicanes trap captured field emission current because of the energy mis-match (forward direction) or wrong sign (backward direction) of the chicane magnets. Field emission currents are expected to be low. For operation 1 na/cm acceptance ( ~ few W) 10 na from worst case CM ( ~ few W) MCI - (maximum credible incident, or, minimal incredible incident?) all chicanes straight, all CM emitting simultaneously 1000 na for a total of 70 kw Not protected by BCS system - it already tripped Shielding was designed for this case. (3 Rem event, 25 R/hr) PPS would likely trip which shuts down all RF. Looks OK to RP at this point, but they are still looking at it.

12 LINAC SAFE STATE C. Adolphson suggested to alternately reverse phase adjacent cavities to put linac in a safe but ready state - no acceleration Fast - a fraction of a second Keeps gradients constant, no ramp up, no cryo change Uses computer control (matlab?) - not hardwired safety systems. Backup measure to prevent the possibility of acceleration. Auto- and operator controlled.

13 BEAM LOSS DETECTION PRD is out: Radiation Detection Requirements for the BCS and MPS, LCLSII-2.4-PR-0868-R0 Selected highlights: Common detectors for MPS and BCS: LBLMs, PBLMs, and IBLMs Complete machine coverage for MPS: gun to FEE Local detectors, PBLMs, to protect safety system items

14 Trip Levels and Response Times

15 PBLMs (diamond detectors, ionization chambers, PMTs)

16 LBLMs (fibers, LIONs) Pairs 100 m long with some overlap

17 CALIBRATION still open issue: More than 200 installed PBLMs and LBPMs that need periodic calibration checking. PRD says use local sources, but hottest portable source at SLAC is a useless 25 mrem/hr. Trip levels are expected to be 1-10 R/hr or higher. Plan B is use the beam to generate pulsed radiation for calibration Need systematic method for making beam get lost in a known way - say every 10 m. Chao thinks it is possible. => ~1000 measurements Need RP modeling of each of those loses to see what radiation reaches the detector

18 DETECTORS (A. FISHER) 100 m fiber and several diamond detectors were installed in Sector 24 in October last year. Beam losses were forced during a machine studies in November and signals studied with good results. ==> Archive data was collected for both LIONs and fibers. It is not yet analyzed. Fibers and diamonds were brought to JLAB and measured radiation during processing of SC cavities. Good signals and much cleaner than JLAB detectors. Next step is to set up approval process with RSC for BCS loss detection, that includes new detectors and FPGAs. (J. Dehong) BLM:LI24:707:QDCRAW BLM fiber 707 () Correlation Plot 16 Nov :29:15 BLM:LI24:707:QDCRAW *5e9*1.6e 19*(c d) () Watts loss calculated from tmits

19 Testing fibers and diamond detectors at JLAB A. Fisher, January 2017

20 FPGA to be used with ACMs, PBLMs, and LBLMs to provide fast digital averaging and diagnostics. match IEC safety criteria redundant FPGAs check each other each line is documented new to SLAC safety systems - expect a vigorous workout to get accepted example:

21 CURRENT MONITORS Average Current Monitors (ACMs) monitor current in various beam lines and cause a BCS trip if the current exceeds a pre-set limit. Average current is defined as the charge in the last 25 μs divided by 25 μs. FPGAs are used to perform the calculation. Digital modeling underway: temperature sensitivity, signal size, loaded Q, rep rate, and bunch charge

22 ACM MODELING EXAMPLE signals at various stages of processing Output of the FPGA averager to be compared with trip level

23 DUMPS Two 120 kw dumps for the undulators, one 250 kw dump in the BSY for idle beam. FDR held December - no report yet guessing more work on intervention plan and shielding, but otherwise ok Next steps are to finish designs and start fabrication Rastering of BSYDUMP is still in conceptual stage.

24 THE END