protoDUNE APA Commissioning Lessons Learned (so far) Andrzej Szelc - - PowerPoint PPT Presentation

protoDUNE APA Commissioning Lessons Learned (so far)

Andrzej Szelc (Manchester) & Serhan Tufanli (Yale)

Introduction

This talk is a summary of things we know about:

• Practicalities of commissioning the APAs in-situ

○ How the measurements work ○ What we can or cannot measure ○ How much we can realistically do in a short time frame

• What happens to the wire tensions after they travelled to CERN and have

been hung in the clean room. Some things we are still in the process of ironing out.

2

“Inventory” Information

So far 2 APA’s have arrived at CERN (both produced at PSL). We have examined both of them:

• APA 1: side B (active) + side A, after transport, side B (after cold box)
• APA 2: side B (active) after transport.

Any measurements or detailed inspection can realistically be done only after the APA is suspended in the clean room (any change observed could be a combination of transport and changing orientation to vertical, more on that later). We have defined commissioning as visual inspection + taking photos of predefined regions (was found to be helpful in MicroBooNE) and measuring tensions of a subset of wires.

3

Visual Inspection

Take photos in predefined regions: All head, foot and side boards. Along combs, and in the central regions - photos taken with region label. “Found” known items - have not found anything new due to travel.

4

APA #1 APA #1 APA #2

Observations/notes on method(s)

• We used the laser tensioning device in two modes:

○ Connected to a guitar amplifier + tuner app on smartphone (consistent with method used at PSL)

○

Connected to laptop with LabView DAQ: calculates FFT and selects peaks, calibrated at Manchester.

• Amplifier + Tuner is really helpful in figuring
• ut the mapping (sound is the key here).

5

Tuner vs LabView cross-calibration

• Plot of tension of same wires between

labview and the tuner.

• LabView setup has been calibrated in

MCR, see difference of O(1Hz) around 60 Hz (in tune with our observations) Estimated error of our measurement is O(1Hz), which means max 3% (depending wire length) systematics in tension of wires.

6

How the measurements actually work

Measure in Region 1 (use scissor lift) Horizontal bar needs to be installed Requires two people (one to operate laser head + one to keep track of mapping). 2 “experienced” operators can do

• (200) wires in ~4 hours.

Region 1

7

Bo

Observations/notes on methods and APA geometry constraints

• Mapping G and X wire plane wires is almost

impossible outside of regions 1 and 5.

• In each region there are 2 possible placements of

tensioning bar.

• In region 5, lowest and middle positions are not

usable due to bolts tethering the APA to ground.

• In region 1 the middle (and top?) position is not

usable due to absence of M1 holes.

• Due to APA geometry, a set of wires in U and V

planes are not possible to test (short wires).

• Level of difficulty grows inward (G, U, V, X), both

because of trying to map what is being measured and reaching the wires to get them to vibrate.

8

Measurement numbers (to give a scale)

• APA #1 Plane A: 112 wires in total ( 43 G-plane, 14 X-plane, 27 V-plane, 28 U plane).

○ After first look at data, did a few extra cross-checks. ○ Measured a few G-plane wires in region 2. ○ *Did not* pay much attention to mapping or selecting outliers - focus on understanding the system and method.

• APA #1 Plane B: 188 Total wires (69 in plane G, 51 in plane U, 37 in plane V, 31 in

plane X)

○ Relatively thorough measurement - measure wires around board boundaries (G-Plane) + selected outliers. ○ Result is a sample of random wires and outliers. Much faster to do adjacent wires than search for a single wire according to mapping.

• APA #2 Plane B: 241 Total Wires (72 G, 46 U, 99 V and 24 in X plane )

○ Procedure as in plane B, APA #1 ○ Add set of wires in zone 5, with relatively low tension - difficult to remeasure tension - inspection. No sag

• bserved.

9

APA #1 Side B - Plane X

• 31 wires were measured

○ Outlier wires: ■ Mean Tension-2.5*RMS > tension of the wire OR MeanTension+2.5*RMS < tension of the wire OR ■ Tension of the wire > 5.3N ○ Random wires from the end and beginning of boards

• Tension is 7% higher than the PSL measurements

with 5% RMS

10

APA #1 Side B - Plane V

• 37 wires were measured

○ Outlier wires:

■ Mean Tension-2.5*RMS > tension of the wire OR Mean Tension+2.5*RMS < tension of the wire OR ■ Tension of the wire <4N OR tension of the wire>5.9N AND ■ Wirelength>600

○ Random wires from the end and beginning of boards

• Tension is 14% less than the PSL measurements

with 2% RMS

11

APA #1 Side B - Plane U

• 45 wires were measured

○ Outlier wires:

■ Mean Tension-2.5*RMS > tension of the wire OR Mean Tension+2.5*RMS < tension of the wire OR ■ Tension of the wire <4N OR tension of the wire>6N AND ■ Wirelength>600

○ Random wires from the end and beginning of boards

• Tension is 5% more than the PSL measurements

with ~5% RMS

12

APA #1 Side B - Plane G

• 69 wires were measured

○ Outlier wires: ■ Mean Tension-2.5*RMS > tension of the wire OR MeanTension+2.5*RMS < tension of the wire OR ■ Tension of the wire > 5.5N ○ Random wires from the end and beginning of boards

• Tension is 15% higher than the PSL

measurements with 2% RMS

13

Calibration measurements on APA#1 Side A (plane G, Region 2)

• Region 2 seems to be less over-tensioned than Region 1

○ (small sample size, though)

14

APA #2 Side B - Plane X

• 31 wires were measured

○ Outlier wires: ■ Mean Tension-2.5*RMS > tension of the wire OR MeanTension+2.5*RMS < tension of the wire OR ■ Tension of the wire > 5.3N ○ Random wires from the end and beginning of boards

• Tension is 14% higher than the PSL

measurements with 1% RMS (previously 7%)

15

APA #2 Side B - Plane V

• 37 wires were measured

○ Outlier wires:

■ Mean Tension-2.5*RMS > tension of the wire OR Mean Tension+2.5*RMS < tension of the wire OR ■ Tension of the wire <4N OR tension of the wire>5.9N AND ■ Wirelength>600

○ Random wires from the end and beginning of boards

• Tension is 6% less than the PSL measurements

(previously 14% less)

16

APA #2 Side B - Plane U

• 45 wires were measured

○ Outlier wires:

■ Mean Tension-2.5*RMS > tension of the wire OR Mean Tension+2.5*RMS < tension of the wire OR ■ Tension of the wire <4N OR tension of the wire>6N AND ■ Wirelength>600

○ Random wires from the end and beginning of boards

• Tension is 1% more than the PSL measurements

(previously 5% but with large spread)

17

APA #2 Side B - Plane G

• 72 wires were measured

○ Outlier wires: ■ Mean Tension-2.5*RMS > tension of the wire OR MeanTension+2.5*RMS < tension of the wire OR ■ Tension of the wire > 5.5N ○ Random wires from the end and beginning of boards

• Tension is 26% higher than the PSL

measurements with 2% RMS (previously 15%)

18

Observations

Wires in vertical planes (G and X) go up in tension between PSL and CERN. And 16% and 26%(!) (G) 7% and 14% (X). U plane sees a slight rise in tension (but essentially close to zero). V plane sees a decrease in tension (14% and 7%). Not a very large data sample yet, but it might make sense to compensate for this effect during DUNE production.

19

Possible reasons for the measured difference

• Different APA orientation during the tension

measurements

○ Horizontal at PSL ○ Vertical at CERN

• Systematics due to hardware and measurement method

○ Hardware: Different amplification hardware between PSL and CERN ○ Method: Constraining wires with clips on combs

20

Other Considerations

Traveller documents provided are extremely useful - defines the subset of wires to be measured. We are in the process of devising a traveller document - lite. Easier to streamline into ntuples and measurements. Easier to keep consistent columns. Would be useful to have a small application for data entry (a relatively small project).

21

Conclusions

• We have re-measured the tension of o(500) wires at CERN and

commissioned 2 APAs.

• One side of APA takes about a day. And can only reasonably measure a

fraction of wires in a part of the APA.

• Having a more automated method (e.g. electric) would allow a much more

thorough examination.

• Observed

○ ~12% increase in tension for X layer wires ○ 10% decrease in tension of the V layer wires ○ 2-3% increase in the tension of the U layer wires ○ 20% increase in the tension of G layer wires

• Probably need a detailed FEA analysis and engineering if we would like to

understand all the details.

22

Back up

23

APA #1First Measurements on Side A (plane X)

• All wires measured were in the higher half of spectrum.
• “Wire-by-wire” comparison indicates a systematic shift of 10% (note that

mapping might not be exactly precise)

24

First Measurements on Side A (plane V)

• Tension looks visibly shifted down. “Wire-by-wire” comparison suggests a

drop of 10-15% in tension.* *Note that here mapping here is not precise for all wires. Although, in these cases, the wires are same length, so general

tension comments apply in region.

25

First Measurements on Side A (plane U)

• U-Plane wires seem to be in the right ball park.
• Wire-by-wire comparison seems to show a small rise in tension (0-10%).

○ Mapping not perfect, so side B should be more precise.

26

First Measurements on Side A (plane G)

• G-Plane tension seems systematically higher (15-20%).

27