ProtoDUNE-SP 35T HV Testing Sarah Lockwitz, FNAL DUNE LBNC Meeting, - PowerPoint PPT Presentation

ProtoDUNE-SP 35T HV Testing Sarah Lockwitz, FNAL DUNE LBNC Meeting, October 26, 2017

Introduction and Motivation • HV performance in noble liquids is not well-understood 1.5 m • Wanted to test the performance of the ProtoDUNE-SP design with respect to high voltage 1.5 m • We’ve been doing a partial-sized but full scale (e.g. 1.5 m curvature of the profiles is the same, etc.) test in the 35T at FNAL Cathode • The test is a 1.5 m x 1.5 m x1.5 m cube • HV is applied to the cathode and correctly graded Resistor to ground down 10 profiles of the field cage Potential • Due to space limitations, we then have a resistor Relative (kV) Anode chain to ground. LBNC: 35T HV Test Sarah Lockwitz October 26, 2017 2

Preliminaries: Instrumentation • Reminder, this is just an HV test… there are no anode read out wires ✴ Power supply: voltage and current monitoring ✴ Current resolution ~10s nA ✴ Two toroids on the HV cable. Can distinguish between PS issues and something in the cryostat. ✴ Sensitive to fast changes in current ✴ Pick-off points near the end of the resistor network. Two in phase I; three in phase II (one on the beam plug). Monitor the voltage across 1 M Ω to ground ✴ Current going down the resistor chain ✴ Three PMTs. Phase I — just visible. Phase II, one had a wavelength shifter on it. ✴ Cameras — RPi (phase I only) HV PS ✴ Ground planes: 1 M Ω to ground monitoring stray currents (phase II only) LBNC: 35T HV Test Sarah Lockwitz October 26, 2017 3

Preliminaries: Instrumentation • Reminder, this is just an HV test… there are no anode read out wires ✴ Power supply: voltage and current monitoring ✴ Current resolution ~10s nA ✴ Two toroids on the HV cable. Can distinguish between PS issues and something in the cryostat. ✴ Sensitive to fast changes in current ✴ Pick-off points near the end of the resistor network. Two in phase I; three in phase II (one on the beam plug). Monitor the voltage across 1 M Ω to ground ✴ Current going down the resistor chain Slide from H.Y. Liao ✴ Three PMTs. Phase I — just visible. Phase II, one had a wavelength shifter on it. ✴ Cameras — RPi (phase I only) HV PS ✴ Ground planes: 1 M Ω to ground monitoring stray currents (phase II only) LBNC: 35T HV Test Sarah Lockwitz October 26, 2017 3

Preliminaries: Footnotes • We have a “current-limiting” power supply • When the supply hits a user-defined current limit, it reduces the voltage… to stay at or below that current limit. • There is a time needed for the supply to limit. Fast discharges can momentarily go above this limit. • In between the PS and the HV feedthrough, there are two 100 M Ω resistors as planned in the ProtoDUNE-SP setup • This makes the effective voltage at the cathode ~5% less than the PS reading • Note: this implies that if there is an 1 μ A change in current, there is a 200 V change on the cathode LBNC: 35T HV Test Sarah Lockwitz October 26, 2017 4

Preliminaries: Definitions • Streamer: Sustained, elevated, noisy, current. Does not match the expected current draw for the system implying an alternate path to ground. • Current-limit event: The supply hits the user-defined current limit and reduces the voltage (the event is slow enough to be controlled by the CL loop). In this run, the CL was set ~25% over the expected draw. • Discharge: Fast change in current as seen by the toroid. May or may not trip the PS. • Trip: The unit went beyond the set current limit so quickly, it did not regulate. If the jump is beyond a defined trip threshold (typically defined to be 3-4x the current limit), the supply turns off. This turning off is done in software. • Glitch/Spike/Blip: A temporary current draw that does not current limit or trip the supply. LBNC: 35T HV Test Sarah Lockwitz October 26, 2017 5

What We Saw in Stage I • Stage I was dominated by streamers Elevated current • Could be somewhat limited by a tight current limit Power Supply Read • They were observed as low as 70 kV back (current = green) • In pure argon, the streamers seemed to prevent the development of a tripping event • In contaminated argon, we could trip • We were able to get to 190 kV (180 kV) in “dirty” Noisy/low current argon (~7 ppb or 500 μ s) • We were able to hold 160 kV in pure/TPC-quality Pick-o ff Points — argon (~70 ppt or 4 ms) with streamer activity voltage /current near the end of the resistor network LBNC: 35T HV Test Sarah Lockwitz October 26, 2017 6

• Between the PS and the field cage, there are a series of resistors totaling 200 M Ω • A 1 μ A change in power supply current is a 200 V change on What We Saw in Stage I the cathode Current Current Set Limit Read Cathode HV Back What We Saw in Stage I Impurities LBNC: 35T HV Test Sarah Lockwitz October 26, 2017 7

Photo: G. Horton-Smith What We Saw in Stage I Possible culprit? • We believe at least some of the streamers were caused by a feature 
 on the cathode frame (field shaping strips) • At the end of the run, after not seeing trips in pure argon, we raised 
 the voltage to 200 kV, and put a loose current limit on it (200 μ A limit; 
 ~48 μ A exp.) • Ran for 8 hours then had a “major EM event” (30 kV swings on the cathode followed by a large breakdown). • A lot of instrumentation tripped/turned off (PMTs, cameras, an argon pump, scope, multimeters) • The field cage, resistor chain, and HV feedthrough + filters were not damaged LBNC: 35T HV Test Sarah Lockwitz October 26, 2017 8

Photo: G. Horton-Smith On to Stage II Possible culprit? • Optimistically, it currently takes between 2.5-3 months to turn around the 35T (time & $$$) • With this very much in our minds, we decided to make two major changes: • Modify the cathode field shaping strips • Install the beam plug LBNC: 35T HV Test Sarah Lockwitz October 26, 2017 9

What We Have Seen… What can we hold? • Have not been able to go above 160 kV because we will current limit then trip Current Limit Trip • Not a gentle thing to do • Tried to answer “What voltage can we run for a day?” • Couldn’t reliably hold 130 kV • The answer seemed to be 120 kV; even held for 100+ hrs days before being turned off. Current+Current Spikes • Had blips every few hours • Eventually did trip at 120 kV 110 kV 115 kV • (An aside, also held 145 kV for 41 hours) • Never tripped at 110 or 115 kV. One could argue 115 kV had more activity. LBNC: 35T HV Test Sarah Lockwitz October 26, 2017 10

What We Have Seen… • The streamers have vanished…. with the exception of one minor streamer (<1 μ A) observed at 145 kV for a few minutes. • With regard to purity — Running in 300-500 μ s argon does not enable us to run at 140 kV. • Ground plane data — The ground plane data have had activity in different locations (not all activity is localized to one area). • We have seen a number of events near the beam plug end wall. • The beam plug is normally filled with nitrogen. If there were breakdowns through the nitrogen, the performance could be improved by pulling a vacuum. • We pulled a vacuum and saw no major change in performance LBNC: 35T HV Test Sarah Lockwitz October 26, 2017 11

Status & Plans • We are still analyzing our data • Our team consists of people who have worked on liquid argon since the ICARUS R&D days, people who have first reported that the dielectric strength is an order of magnitude less than presumed, and experienced general hardware types. We are in communication with the Dark Matter community (and more) on HV issues. • We plan to warm up the cryostat over the next month and introduce breathable air • Remove the beam plug and test it in a small cryostat alone before sending it off to CERN (December) • Install a normal end wall where the beam plug was and retest (installation ~2 weeks, cool down/fill ~month; late January 2018). • We are open to ideas from the collaboration. LBNC: 35T HV Test Sarah Lockwitz October 26, 2017 12