35t TPC FEMB Status 20160302 - 35t - B.Kirby Major TPC FEMB - - PowerPoint PPT Presentation

35t TPC FEMB Status

20160302 - 35t - B.Kirby

Major TPC FEMB Problems

• “High Noise” state - noise levels are unstable

○ Have some probability of going into a “high noise” state that makes data unusable ○ Need to identify best method of recovery, power-cycling works but may be risky

• Noise levels are generally large

○ Raw noise levels generally around 2000-2500e-

• Number of bad channels has increased due to hardware problems

○ FEMB03, FEMB14 producing corrupted data for ASICs 0-3 ○ FEMB09, FEMB10 off due to excessive current draw ○ 7 ASICs turned off after first power cycle following cryostat filling ○ Total # of bad channels = 550

3

“High Noise” State

RMS >100 ADC Pedestal ENC > 20000 e- Oscillatory signal Affects entire detector, never occurs on single FEMB

Feb 26 Start of High-Noise State in Data

4

Run 13159 event 3437 recorded start

• f high noise state

Something simultaneously saturates nearly all channels, high noise starts ~300us after

Current Draw in Normal and High-Noise State

5

LV ASIC current draw in normal state: RCE00 = 1.136A RCE01 = 1.133A RCE02 = 1.12A RCE03 = 0.993A RCE04 = 1.09A RCE05 = 1.03A RCE06 = 1.13A RCE07 = 1.03A RCE08 = 1.03A RCE09 - OFF RCE10 = OFF RCE11 = 1.02A RCE12A = 1.02A RCE13A = 1.02A RCE14A = 1.02A RCE15A = 1.18A LV ASIC current draw in high-noise state: FEMB00 - 1.145A FEMB01 - 1.140A FEMB02 - 1.12A FEMB03 - 0.998A FEMB04 - 1.09A FEMB05 - 1.03A FEMB06 - 1.13A FEMB07 - 1.03A FEMB08 - 1.04A FEMB09 - OFF, FEMB10 - OFF, FEMB11 - 1.03A FEMB12 - 0.900A FEMB13 - 0.898A FEMB14 - 0.913A FEMB15 - 1.04A

Current ~100mA lower on ASICs during high-noise state, turning them off generally stops the noise APA3 FEMBs usually the cause, but not always

High Noise State Recovery

• Need to identify best method of recovering from high-noise state
• Power cycling ASICs with lowered current draw generally fixed problem

○ Sometimes need to power-cycle all ASICs ○ However, power-cycling ASICs may be risking at cryogenic temperature, see FEMB09/10

• Wire-bias trick

○ Likely inducing large transient that saturates amplifiers ○ Does not seem to work consistently

• Resetting / reconfiguring ASICs

○ This does not seem to work ○ Has some probability of causing the high-noise state

• Waiting - the high noise state seems to stop on it’s own on ~2 hour timescale

○ Possibly due to another saturating event?

Noise Measurements

• Noise during “normal” operation is relatively high
• Tried increasing ASIC gain and shaping time on Feb 26 to improve S/N
• Currently running with 14mV/fC gain, 3us shaping time
• Run 13851 recorded Feb 29, ~4300e- ENC raw, ~2500e- ENC after 11kHz

subtracted

Run State Collection Noise (ENC) Induction Noise (ENC) Collection Noise 11kHz Subtracted Induction Noise 11kHz Subtracted 13851 Normal running 3200 e- 4300 e- 1600 e- 2500 e- Raw Noise Vs Channel #

Feb 17 “Turn Everything Off” Test - Noise Increases

Run # ELOG System State Collection Noise (e- ENC) Induction Noise (e- ENC) 11481 5927 “Everything off” 2310 4070 11484 5929 Turned on purity monitor, RTD, PC4 2340 4120 11487 SSPs turned on 2630 4430 11488 CCU and PTB turned on 2640 4490 11492 Wire-bias ramped, “high noise” state 11493 ASICs power-cycled, “high noise” removed 2690 4620 11494 Drift field bias ramped up 2870 4810 5936 Cameras back on 11504 5940 “Everything on” 2940 4780

8

Feb 22 Single FEMB Runs

9 FEMB 1us 2us 3us FEMB00 12402 12404 12403 FEMB01 12408 12410 12411 FEMB02 12414 12413 12412 FEMB03 12415 12416 12417 FEMB04 12420 12419 12418 FEMB05 12421 12422 12423 FEMB06 12426 12425 12424 FEMB07 12427 12428 12429 Run ENC @ 1us ENC @ 2us ENC @ 3us FEMB04 1020e- / 1880e- 1070e- / 1700e- 1160e- / 1740e- FEMB05 1000e- / 1850e- 1050e- / 1830e- 1080e- / 1710e- FEMB06 980e- / 1650e- 1000e- / 1570e- 1020e- / 1670e- FEMB07 1020e- / 1840e- 1080e- / 1780e- 1100e- / 1850e-

Noise Measurment in ENC Collection / Induction ENC After 11kHz Subtraction

• Single FEMB study done with wire bias +

drift field off , other systems on (?)

• Single FEMB noise ~1000 to 1100e- ENC
• n collection planes, 1600e- to 1900e- ENC
• n induction planes
• Can this be lowered?

Noise Measurements - APA0 sees “Ringing”

“Ringing” packet: ~28kHz This is a sign that large noise seen on channel inputs is partly due to external pickup

Noise and Cosmic Track S/N

Collection plane signals clearly visible Induction plane signals lost in the noise Example Induction Channel Waveform with Track Signals Example Collection Channel Waveform with Track Signals Online Channel # Waveform Sample # Waveform Vs Channel After Correlated Noise Subtraction

Bad Channel Status

• Number of bad channels have increased due to three problems:

○ Corrupted data on FEMB03 ASICs 0-3, FEMB14 ASICs 0-3 (might be recoverable) ○ FEMB09, FEMB10 power draw off ○ 7 bad ASICs following cryogenic power cycle (try running at 900mV baseline)

• Total number of bad channels = 550

○ Includes individual bad channels, does NOT include broken wires

Summary

• Understanding FEMB instability and large noise levels is the top priority for

cold electronics

• Significant improvement in noise levels due to choice of ASIC settings

○ However noise levels are still high, evidence of pickup from non-TPC related systems

• Propose another “everything off” test to pin down noise sources, 1 day

○ Turn everything non-TPC related off, take data with single FEMB and identify best possible noise level (pumps probably can stay on) ■ Vary FE-ASIC settings to see if improvement is possible ○ Turn on FEMB ASICs on one APA + wire-bias + drift field ○ Take ~4 hours of cosmic data ■ Determine if signal to noise is “acceptable”, improved over normal running mode ■ Measure how often “high-noise” state occurs ○ Add additional system components and measure changes in noise levels

• Reducing noise levels is valuable as it will improve proposed detector studies,

and connection between noise levels and system instability is not clear