260 likes | 275 Views
HBD Status. TK Hemmick for the HBD Crew. Brief Overview. HBD West ~1 Month of stability tests completed (looks good) 100% functional Cosmic Rays and Scintillation seen in all GEMstacks. Presently under Nitrogen Flow. Required for delivery:
E N D
HBD Status TK Hemmick for the HBD Crew Detector Council Meeting
Brief Overview • HBD West • ~1 Month of stability tests completed (looks good) • 100% functional • Cosmic Rays and Scintillation seen in all GEMstacks. • Presently under Nitrogen Flow. • Required for delivery: • Re-install air-cooling system (plumbing presently off) • Final noise chasing (Cu tape & mylar shields) • New Survey (Really Necessary???) • PHENIX Gas System Ready. • HBD East • 4/10 GEMstacks installed (fast method) • Need more GEMs (10 more on the way from WIS) • Just waiting around (tent is pretty boring right now). • Initial Tests on new OFFLINE algorithm begun. • Affects choices for items stored in new Compact CNTs. • Trigger on HBD • Possible but needs dedicated team & serious work Detector Council Meeting
Cosmic Measurement System Scintillator Setup Coincidence CAMAC Detector Council Meeting
Result from the 1st Test • Cosmic Ray Trigger paddles cover 14 pads. • In “Forward Bias” one expects both ionization and scintillation signals from cosmic rays. • Source of exponential as scintillation confirmed in two ways: • When recording a nearby pad that it not under the paddles, the MIP disappears, but the pedestal and scintillation remain. • When recording data in reverse bias, MIP disappears, scintillation & pedestal remain. Pedestal MIP Response Scintillation Mean Scint = 20 ch Landau MPV = 300 ch Landau MPV ~15 Detector Council Meeting
WN4 Spectra Detector Council Meeting
Some GEM Reminder Numbers • GEMstack = 3 GEMs. • GEM has 28 strips. • Strip has 20 MW • Each stack runs from a triple chain. • If one strip is shorted, the GEM is 20 MW instead of infinity. • Current Draw in the chain increases by 1.5%. • Old Model: Trip @ standing current + 1.5% • New Model: Trips are more complicated. Detector Council Meeting
Cascading trips. • When operating under Gain Conditions, a discharge will propagate and avalanche. • Avalanche will produce large current surge & may trip subsequent GEM strips. • Tripping the strip in a neighbor GEM produces a MUCH larger current surge. • 1.5% trip settings have been too low. Detector Council Meeting
Who’s Afraid of the Big Bad Trip? • Run 7: • Trip settings were 1.5%. • Mesh-GEM spark phenomenon & LeCroy 1471 FPGA bug combined to deliver MASSIVE discharge from the mesh if the GEMstack was disconnected. • Tripping the LeCroy to “save the GEMs” saved nothing and initiated a damage-inducing trip. • COMPASS: • NO TRIP CIRCUITS!! • Occasional discharges do not damage GEMS. • PHENIX Future: • Compromise… • Run trip settings low in software and high in hardware. Detector Council Meeting
Trip Rate reasonable • Hardware trip = one strip in 2 Gems. • Software trip = 1.5% • Above result is what happens with all stacks in the HBD on at the same time. • Additional benefits of new scheme: • Trips are no longer “massive” (every stack). • Trip rate is smooth function of threshold (how low rate do you want?) • BIG NEWS: • Lots of trips early in the testing…NO DAMAGE! Detector Council Meeting
Our trouble maker for big trips was stack South 4. With new resistor chain change voltage to dV=460V (all other were at 480 V): Works stable at this dV, with large signal !!! Detector Council Meeting
Update on the HBD gas system • The main goal is to increase flow to the two detector as much as possible without compromising gas quality. • Three places changes can be made • The Monitoring Hut • The rack in the mixing house • The small bypass panel in the IR (this probably will not do much) The Hut • Modify Flow path to gas analyzers. (DONE) • Reduce restrictions by replacing flowmeters and valves (DONE) • Install gauge and new valve to make switching from vacuum in the monitor to gas a little easier. (DONE) • Replace H20 sensors with ones that work. Detector Council Meeting
HBD Gas Rack: In Progress Solenoid valve to the vent had to be changed to accommodate much higher flow rate than designed (10x higher). Custom valve ordered and in hand Pressure transmitter may need to be replaced. This was at upper reading limit in Run6. With the new flowmeters, the higher pressures in the rack may not be needed Old flowmeter with built-in valves needs to be replaced with low impedance flow indicator and separate valve. This valve controls how much gas passes through the purifiers. Right now all gas pass though it. This flowmeter was already replaced once mid-Run6 to increase flow to the HBD Old flowmeter without built-in valves needs to be replaced with low impedance flow indicator. This indicates how much total gas is going to the system. This flowmeter was already replaced once mid-Run6 to increase flow to the HBD. It need to be replace again for the higher flowrate. Replace Purifier to handle the higher gas flow--DONE Detector Council Meeting
Algorithm Ideas • Old: Clustering • Clusters could grow to any size or shape • Fails with scintillation • New: Hub and Spoke (HnS) • - Reconstructed electron location is used as center for search region • Hub: multiple-pad sum at center • Spoke: multiple-pad sum at some R and phi away from center Detector Council Meeting
Two Possible Pad Sizes: 3 vs 7 - Two options that are presently coded Detector Council Meeting
Detector Council Meeting Singles Doubles
HBD Trigger (aka WTF??) • Physics: • Charged pion signal spin asymmetry is MORE sensitive than neutral pion spin asymmetry. • We don’t record this since we don’t have a high momentum pion trigger in pp collisions. • When operated in ++ field, the HBD only sees blobs for high momentum pions (req’s RICH coincidence) • Hardware • The new RPC trigger module is 100% HBD-compatable. • Adding these to the HBD crate will allow HBD signals to reach the trigger system. • What it takes: • Extra RPC trigger boards…$$$. • Extra code to have HBD data flow on backplane… Chi? Chuck? • Extra code (modules?) in the trigger system to use HBD data. Lajoie • NEW CIRCUIT BOARD to put an optical copy of RICH signals onto new Fiber media Chuck. • Additional dedicated people putting in significant effort.(Stony Brook cannot provide item #5). Detector Council Meeting
Extras Detector Council Meeting
HBD System Trigger • The ADC boards has 2 pairs of cables output per board, 4 LVDS signals. • It cover 48 channels • If you output 48 bits per beam clock, one bits per channel. We only need one pair cable. • The RPC trigger module could handle 10 ADC boards • The maximum data output is 384 bits. • This is less than 480 bits for 10 ADC boards. • The optics we used are the chip sets 8b/10b encoding. Except for the upgrades detector system, all other system use GLINK chip set • If HBD trigger need to merge with the existing detectors to form a trigger. We have issues with L1 system. Detector Council Meeting
RPC trigger module • RPC TDC board has 64 channel per board • For every beam clock, we output 64 Y/N bits by 2 differential LVDS cable in the backplane. • 320 bits per cable. • We need to move 3 TDC’s data by 1 optical cable • 64*3 bits per beam clock. • The optics will run 2.8 gbits/sec. • One trigger module will handle 6 TDC modules. • 2 optical cables • The input cable can handle maximum 10 pairs cable for the L1 trigger primitives. • We could just connect all the traces to the FPGA. Detector Council Meeting
HBD West Trip statistics Nominal Resistor chain current = 146uA Mean and Peak Trip limit = 170 uA Detector Council Meeting HV output number
Mean and Peak Trip limit = 175 uA 3 South 4 South 0, North, South Detector Council Meeting
More statistics: • At 170uA there were 37 trips over 93 hours. 93/37=2.5 hours • At 175uA there were 32 trips over 90 hours. • 90/32=2.8 hours • Trip pattern (size, stacks) changed, but frequency was not • There are two outliers: South 3 and 4. May be 0 North/South too? Detector Council Meeting
Measurement with track trigger only across few pads, few others have no PIM track at the same time (no overlap with trigger counters): Pad with no track across it. Noise + scint light. All track crosses N=66226 Pure light signal between chan. 30 and 300. N_scint=433 This is first estimate: Probability to get electron per track=433/66226=0.0065 Detector Council Meeting
2nd test: Another pad with scint. Light andMIP track Assume all signals with >300 correspond to MIP across this pad. N_mip=303 events Look at other 3 pads (like ch10) where for sure was no any MIP at all – scint. counters were far away For 3 pads with no tracks, we find only 14 scint. signals. Probability to get light per pad will be (14/303)/3 pads= 0.0066 Probability to get electron from scintillation from track close to the pad is 0.006-0.007. This is the largest light signal. Other tracks will give smaller probability Detector Council Meeting
New resistor chain. Zero voltage bias Bias = +100 V (but low statistics) Detector Council Meeting