High Voltage Testing at the 35t Long Baseline Neutrino Committee - - PowerPoint PPT Presentation

High Voltage Testing at the 35t

Long Baseline Neutrino Committee (LBNC): ProtoDUNE-SP CE & TPC Sarah Lockwitz, FNAL June 23, 2017

ANL (eng), W&M (prof., student, tech), LBNL(scientist, eng), KSU (prof, post doc, student), U.C. Davis (prof., post doc), U of Houston (prof), UCLA (prof), BNL (scientists), FNAL (scientists, techs)

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Motivation

• Design is new
• Resistive cathode
• Aluminum profiles
• Cable-based HV bus
• …more
• And how do these pieces

work together?

Photos from B. Yu

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Motivation

• Predicting HV behavior in LAr is not easy /

designing stable & robust LArTPCs with Vcathode=180 kV is not well-trodden territory

• Breakdown can break components
• it can also set the operating limit of the

detector limiting the physics reach

• Instabilities in the system (glitches, “blips”) are

annoying, and may affect data quality

• Possible light production… HV issues can

cause light affecting our light detection system

Long Bo TPC

This ring is open This is why you should care!

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Introduction to the Test

• The goal of the HV test in the 35t was

to test the ProtoDUNE-SP design with regard to high voltage

• The test was carried out in the 35t in

PC4

• The 35t cannot fit the full size of the

ProtoDUNE-SP field cage.

• We used a partial field cage with

full-scale components (e.g. the profile curvature was the same, etc.)

https://inspirehep.net/record/1397920

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Introduction to the Test

• The goal of the HV test in the 35t was

to test the ProtoDUNE-SP design with regard to high voltage

• The test was carried out in the 35t in

PC4

• The 35t cannot fit the full size of the

ProtoDUNE-SP field cage.

• We used a partial field cage with

full-scale components (e.g. the profile curvature was the same, etc.)

Cathode Resistor to ground Relative Anode Potential (kV)

1.5 m 1.5 m 1.5 m

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

HV Issues We Were Preparing for

• We expected HV problems

to be something like a spark

• Fast discharge/high

current drawn momentarily

• Visible light

Feedthrough went here

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

HV Issues We Were Preparing for

• We expected HV problems

to be something like a spark

• Fast discharge/high

current drawn momentarily

• Visible light

Feedthrough went here

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Monitoring

✴ Power supply: voltage and current monitoring ✴ Current resolution ~10s nA ✴ Two toroids on the HV cable ✴ Sensitive to fast changes in current ✴ Two pick-off points near the end of the resistor

network

✴ Current going down the resistor chain ✴ Three PMTs. No wavelength shifter on them — Just

visible

✴ Cameras — RPi

HV PS

What We Saw Note:

• The supply was a current-limiting supply. When

the output current from the supply reached a set limit, the supply reduces the voltage to stay below the current threshold.

• The voltage listed in the following slides is the

voltage at the power supply unless otherwise

• noted. The voltage at the cathode was 5% less.

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Initial Turn On

• Purge, cool down, and fill went well
• On March 31st, just after we were filled with

not TPC-qualty (7 ppb) LAr we began to ramp.

• We tripped at 160 kV
• It appeared to be a real “fast discharge” trip.
• Let it sit over night at 150 kV, and ramped

without issue to 190 kV the next morning

• It held until we ramped down Monday

morning

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Running with Purification

• The pump started without any issue
• We then ramped back up to 190 kV and noted it

seemed to be getting noisier

• We then saw periods of elevated current; we called

this “streamer mode”

• Current was noisier coming out of the supply & less

current was going to the end of the resistor network

• → Alternate path to ground
• We lowered the voltage and saw this behavior also

at 150 kV

No Circulation Pumps On Current Read Back (and any jumps)

Pick-off points (near end of chain)

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Running with Purification

• The pump started without any issue
• We then ramped back up to 190 kV and noted it

seemed to be getting noisier

• We then saw periods of elevated current; we called

this “streamer mode”

• Current was higher coming out of the supply &

less current was going to the end of the resistor network

• Alternate path to ground
• We lowered the voltage and saw this behavior

also at 150 kV

Pick-off Points — voltage /current near the end of the resistor network Power Supply Read back (current = green)

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Continuing with the Run: Contamination

• We then wanted to better check the

effect of purity on the current features

• We bypassed the filters and returned

the condensed liquid to the cryostat

• Still had current features, but their

magnitude was smaller

• Running at 175 kV in ~350 μs LAr

Current

Pick-offs (end of chain)

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Purity Effects

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

What Happens at 100 kV?

• Let’s re-purify and answer this
• Ran at 100 kV in 3 ppb O2 LAr for

a baseline and restarted purification

• Ran at a steady ~4 ms lifetime
• Still observed streamer-like events

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Running in Stable ~4 ms LAr

• Lowered the current limit of the supply to 24% over

the expected current draw; considered this a “tight”-er limit

• Still “jumps/spikes”, 1-2/hour, no streamers
• Changed to 90, 70 120, 140, and 160 kV (with

corresponding current limit),— still spikes, similar rate

• (Note the order)
• After running at 70 kV, the current limit was raised from

21 μA to 40 μA, and a two minute streamer occurred

• Tighter current limits quench the long-streamer

behavior

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

How do the features behave in time? Are there charge-up effects?

• We turned the voltage to 0 for ~67 hours 


and then ran at 150 kV

• We observed charge-effects. No activity for 


4.5 hours (previously 1-2 features/hour).

• More followed. Appeared to get busier in time.
• Saw this for other voltages too
• There is a hysteresis — one can go to a higher 


voltage and have a stable period, but going down did not lead to a stable period.

• This was not simply a current limit effect
• We record spikes even in the tight current limit running
• We relaxed the current limit and did not observe streamers after being off

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Where are the Sparks? A Quick Note on What We Were Seeing

• We were expecting HV issues to be like sparks we’d seen in other test setups
• Light produced & relatively fast discharge (resistive cathode may have slowed this)

producing a signal on the toroid

• While ramping up the first day, we experienced a trip of the PS. This appeared to be like

something we expected — there was a signal on the toroid

• We had something similar while ramping the day after the pumps were turned on (4/4)
• The PMTs were instrumented shortly after we purified
• We did not see light or toroid signals for any of the streamer or current activity in the

month after April 4th

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

A Note on Diagnostics on These HV Events — Dark Sparks

• We have 3 PMTs in the cryostat (no TPB on them)
• The PMTs were running after the first purification
• We do not see light associated with these events (H.Y. Liao’s work in our weekly

meeting) — they do see noise

• The PMTs were working — they see the purity monitors and Cherenkov light from

showers

• Toroids do not see activity for the streamer events (slow developing).
• See two types of signals on the toroids 


— small signal (10s mV)

• Large (multiple Vs) signals we expect 


from breakdowns — only a few of these 
 (1st day, 200 kV event, one while 
 ramping…)

• It seemed to be hard to get a fast-discharge-like


event in the pure[r] states

Evidence the PMTs are working

Cherenkov light from shower

Photo: G. Horton-Smith

~95 μA current jump at 180 kV on 4/4 25 V 200 μA current jump at 170 kV on 4/25 30 mV

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

What’s Going On? Going Out with a Bang

• Toward the end of the run, after not seeing traditional 


sparking behavior, we went to 200 kV — the maximum 


• f the supply
• Current limit was loose (200 μA limit; ~48 μA exp.)
• Ran for 8 hours — streamers/activity
• Then, we suffered a major E&M event (30 kV swings)
• Probed issues that could occur
• Instrumentation tripped off (Keithleys, scopes, PMT HV, cameras) — some didn’t recover
• Our LAr pump was turned off
• Lesson: Look at conditions for cryo-system — make it resistant to sparks
• The field cage, resistor chain, and HV feedthrough + filters were not damaged.

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

A Note on These Current Features

• Between the PS and the field cage, there are a series of

resistors totaling 200 MΩ

• A 1 μA change in power supply current is a 200 V change on

the cathode

Cathode HV Impurities Current Read Back Current Set Limit

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

A Note on Diagnostics on These HV Events

• Looking at the pick-off points and

current measurements, one can make a model for where the leakage current to ground is occurring (assuming one discharge point per event).

• Three independent models

pointed to something near the cathode, but not the cathode

• One attractive candidate are the

cathode Field Shaping Strips (design since changed in PD-SP)

V

Ist

Filters Good side of FC Current to Gnd Anode to Gnd 3 non-streamer walls of FC Photo: G. Horton-Smith

Possible culprit?

Pick-offs + PS Current = ?

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Stage II Test

• There was always a plan to have a second run

including the beam plug once it was finished.

• That is being installed now
• The cathode field shaping strips have been modified

(smaller, will fit inside the envelope of the field cage)

• The ground planes above, below, and behind the

field cage are being instrumented to be read out

• Better identify where the source of the extra

current to ground is.

• A PMT with TPB is being installed
• Scheduled to begin purging in July

Beam plug in CA Beam plug @ FNAL Monday

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Closing Thoughts

• The 35t cryostat attained a 5-6 ms lifetime after identifying and addressing a stratification issue
• We do not immediately spark at any voltage tested up to ~190 kV on the cathode
• We see high current draws that seem to come from a feature near the cathode
• Long (in time) current streamers appear to be controlled by “tight” current limits
• Visible light is not associated with these
• We don’t know the effect of these features on a real TPC (electronics, scale). I note a 1 μA drop at the supply is a 200 V drop at the

cathode.

• Time at voltage is a dependent variable. Charge-up effects?
• Purity effects: Greater magnitude current draws in purer argon (but current spikes/jumps in the entire range of argon studied),

“difficulty” getting spark like behavior in very pure argon (streamers)

• The field cage, resistor network, HV feedthrough, HV filters & cables all survived (likely the largest LAr HV event on site).
• One should carefully consider fault conditions/interlocks for the cryogenics system. Interlocks need to be robust against E&M events.

Back-up Slides Follow

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Streamers?

• Reminded me of something from a

few years ago

• While looking for breakdown with a

1.3 mm ball, sparks were seen for gaps (between the sphere and ground) < 20 mm.

• At 20 mm, we just noted an

increasing current draw with the voltage from the PS

• No light, no sound — just a

steady current draw.

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Purification Success! Attained a τ ~ 5-6 ms

PrMs

O2 Analyzer

O2 analyzer switched to vapor space

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Purification Success! The Before: A Stratification Issue

• The returned filtered liquid

was assumed to be warmer than the bulk in the cryostat (flowing faster)

• But before returning, the

liquid passes through a phase separator that puts the liquid in the same state as the liquid-gas interface

• f the bulk liquid — cooler

Credit E. Voirin

• A. Hahn: 10.1109/NSSMIC.2014.7431158

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Purification Success! The Before: A Stratification Issue

• The return and the intake were located at the bottom of

the cryostat

• Temperature stratification

Credit E. Voirin

Input Return

Credit: E. Voirin Credit: A. Hahn

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Purification Success! The Before: A Stratification Issue

• The return and the intake were located at

the bottom of the cryostat

• τ ~ 3 ms

Credit E. Voirin

Input Return

Credit: E. Voirin

Longer PrMs

Credit: A. Hahn

LBNC: HV Test in the 35t Sarah Lockwitz June 23, 2017

Purification Success! The After

• The liquid return was modified to be about 2/3
• f the height of the cryostat. Plots below

show measured and predicted lifetimes vs. height in LAr

Credit: E. Voirin

PRM4% PRM3% PRM2% PRM1%

Old arrangement New arrangement