HCal Electronics (and - PowerPoint Presentation

Slide1

HCal Electronics (and pulser)

B. QuinnFeb. 23, 2018

1

Slide2

Cabling scheme for HCalOne signal, one HV cable to each tube. (Plus one HV & 6 signal/section for pulser)Cables secured, strain-relieved on gantry behind

HCal.Cables held in place by gantry so they can be quickly re-connected to correct tubes.

(… and gantry moves with

HCal

??)

On

HCal

electronics/cabling platform

(new) patch panel BNC/

Lemo

to short

Lemo

jumpers to amplifiers

X10 amp

Patch panel for output signal to electronics hut (BNC RG58 100 m. long)

Lemo

cable outputs to

UVa

summing module

Discriminator, 16-fold fanout, TTL converter for

pulser

Third patch panel in electronics hut

…. Assuming Nino is used…

Asymmetric passive splitter (board with connection/housing/power for Nino)

Needs to be designed (

JLab

?) noise shielding important

Signal cables run to one of eighteen 16-input BNC sub-panels (S01-S18)

Sequential inputs go to sequential channels of Nino,

fADC

Nino outputs go to VETROC-based TDC or F1 TDC

Slide3

fADC

Nino

Disc.

VETROC

(or TDC)

Sum of 4

Disc

Shielded DAQ hut

Sum of 16

(18 groups)

(10 regions)

18

UVa

Summing mod.

36 in -> 18 in (16 in used)

4 quad linear FO (e.g. Phillips 740 2.5V)

3 quad 4-fold linear FI (

UVa

133)

12

chan

L.E. Disc.(3X4 or 1X16)

Custom

passive

asymmetric

Splitter and

Nino interface

LVDS

To

Ecal

trigger

Lemo

fan

out

100 m

X10 amp

patch panel

PMT

patch panel

Slide4

1-12Cabling pattern.Three 16-channel slotsfor each four rows.

12

8

4

16

1

2

3

4

5

6

7

8

9

10

11

13

14

15

16

1

2

3

4

5

6

7

9

10

11

12

b3

14

15

16

1

2

3

5

6

7

8

9

10

11

12

13

14

15

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1-1

24-1

24-12

{

S16

S17

S18

S01

S02

S03

S04

S05

S06

{

{

{

{

{

S07

S08

S09

S10

S11

S12

S13

S14

S15

Adjacent electronics channels don’t

Fire on same LED pulse and unlikely

to fire in same cluster

Slide5

5Map from electronicschannels to HCal module-Amplifiers-Patch panels-

Ninos-TDCs (or VETROCs)-

fADCs

S01 S02 S03 S04 S05 S06 S07 S08 S09

1 1- 1 2- 2 3- 3 5- 1 6- 2 7- 3 9- 1 10- 2 11- 3

2 1- 4 2- 5 3- 6 5- 4 6- 5 7- 6 9- 4 10- 5 11- 6

3 1- 7 2- 8 3- 9 5- 7 6- 8 7- 9 9- 7 10- 8 11- 9

4 1-10 2-11 3-12 5-10 6-11 7-12 9-10 10-11 11-12

5 1- 2 2- 3 4- 1 5- 2 6- 3 8- 1 9- 2 10- 3 12- 1

6 1- 5 2- 6 4- 4 5- 5 6- 6 8- 4 9- 5 10- 6 12- 4

7 1- 8 2- 9 4- 7 5- 8 6- 9 8- 7 9- 8 10- 9 12- 7

8 1-11 2-12 4-10 5-11 6-12 8-10 9-11 10-12 12-10

9 1- 3 3- 1 4- 2 5- 3 7- 1 8- 2 9- 3 11- 1 12- 210 1- 6 3- 4 4- 5 5- 6 7- 4 8- 5 9- 6 11- 4 12- 5

11 1- 9 3- 7 4- 8 5- 9 7- 7 8- 8 9- 9 11- 7 12- 812 1-12 3-10 4-11 5-12 7-10 8-11 9-12 11-10 12-11

13 2- 1 3- 2 4- 3 6- 1 7- 2 8- 3 10- 1 11- 2 12- 3

14 2- 4 3- 5 4- 6 6- 4 7- 5 8- 6 10- 4 11- 5 12- 615 2- 7 3- 8 4- 9 6- 7 7- 8 8- 9 10- 7 11- 8 12- 9

16 2-10 3-11 4-12 6-10 7-11 8-12 10-10 11-11 12-12

S10 S11 S12 S13 S14 S15 S16 S17 S18 1 13- 1 14- 2 15- 3 17- 1 18- 2 19- 3 21- 1 22- 2 23- 3

2 13- 4 14- 5 15- 6 17- 4 18- 5 19- 6 21- 4 22- 5 23- 6

3 13- 7 14- 8 15- 9 17- 7 18- 8 19- 9 21- 7 22- 8 23- 9 4 13-10 14-11 15-12 17-10 18-11 19-12 21-10 22-11 23-12

5 13- 2 14- 3 16- 1 17- 2 18- 3 20- 1 21- 2 22- 3 24- 1 6 13- 5 14- 6 16- 4 17- 5 18- 6 20- 4 21- 5 22- 6 24- 4 7 13- 8 14- 9 16- 7 17- 8 18- 9 20- 7 21- 8 22- 9 24- 7

8 13-11 14-12 16-10 17-11 18-12 20-10 21-11 22-12 24-10 9 13- 3 15- 1 16- 2 17- 3 19- 1 20- 2 21- 3 23- 1 24- 210 13- 6 15- 4 16- 5 17- 6 19- 4 20- 5 21- 6 23- 4 24- 5

11 13- 9 15- 7 16- 8 17- 9 19- 7 20- 8 21- 9 23- 7 24- 812 13-12 15-10 16-11 17-12 19-10 20-11 21-12 23-10 24-11

13 14- 1 15- 2 16- 3 18- 1 19- 2 20- 3 22- 1 23- 2 24- 314 14- 4 15- 5 16- 6 18- 4 19- 5 20- 6 22- 4 23- 5 24- 6

15 14- 7 15- 8 16- 9 18- 7 19- 8 20- 9 22- 7 23- 8 24- 916 14-10 15-11 16-12 18-10 19-11 20-12 22-10 23-11 24-12

Slide6

144 BNC cables from

top half of

HCal

(~5m long)

6

9 16-chan

PMT amps

9 16-chan

PMT amps

144 BNC cables from

bottom half of

HCal

(~5m long)

9X16-chan

lemo

/BNC

patch panel

(To electronics

hut)

9X16-chan

lemo

/BNC

patch panel

(To electronics

hut)

144

lemo

cables

144

lemo

cables

G

r

1

G

r

9

}

}

}

S01

S02

S03

S04

S05

S06

S07

S08

S09

144

lemo

cables

144

lemo

cables

G

r

10

G

r

18

9

lemo

cables

9

lemo

cables

9 16-chan

summing

UVa

120

9 16-chan

summing

UVa

120

4 lin.

Fan-

out

3X4-

chan

sum

UVa

-

133

Electronics racks on

HCal

gantry (2 relay racks)

(HV patches and cables not shown)

10

ch

LE-

disc

S01

S01

S09

S09

S10

S16

S16

S10

2x 144

Lemo

Slide7

7

9X16-chanpatch panel/

Nino interface

(BNC from

gantry)

S01

S09

9X16-chan

patch panel/

Nino interface

(BNC from

gantry)

S10

S16

VETROC

or

F1 TDC

S01

S09

S16

S10

VETROC

or

F1 TDC

F250

fADCs

S01

S09

S16

S10

Electronics racks in shielded hut (2 relay racks)

LVDS

(converted

to ECL

for F1?)

Lemo

F250

fADCs

Slide8

8Time (ns) (from impact with front face)

~ 0.2 V signal amplitude=> ~3.0 nVs

=> 60

pC

into X10 amp

after amplifier (3V max)

~2 V amplitude

30

nVs

Slide9

9~2.1 V

After X10 PMT amp

~3000

p.e

!!! wrong!

Slide10

10After X10 PMT ampand 100 m RG58 cable~720 mV

Slide11

11Nino testing at CMU (Status report)Nino puts ~150 mV bias on input signal (for me, boarddidn’t operate with back termination)

Used 370 pF blocking capacitor on input line (wrapped in grounded foil to shield)

LVDS to ECL converter didn’t seem to work with 100 mV

offsets which Nino puts on output. Used two 330 pF caps

to block DC at Nino board.

Nino can sustain very high freq. oscillations, possibly from

Feedback picked up by input lines? Unplugging/plugging

power connector reproducibly initiated oscillation. Cycling

power supply reproducibly stopped it.

Slide12

12Fixed-size LED pulses to PMTTrying to determine effective threshold:

Nino threshold set to max by on-board pot. Reproducibly 600 mV signal through 40 dB just fires Nino-> threshold ~5-6mV

Timing resolution (best case)

LED pulse (600 mV) to LE disc. with 120 mV threshold

and to Nino through 26 dB (20 X lower amplitude,

20 X lower threshold)

Nino -> LVDS/ECL -> ECL/NIM -> TDC

RMS resolution: Nino 182

ps

LE 132

ps

Slide13

13Fixed-size LED pulses to PMT, 100 m cable to NinoTiming resolution (still fixed signal size & shape)

LED pulse (600 mV) to LE disc. with 180 mV thresholdand to Nino through 20 dB after 3X amplitude loss in cable

(30 X lower amplitude, 30 X lower threshold)

RMS resolution: Nino 151

ps

(LE still ~132

ps

)

Slide14

14Cosmic (vertical)Hcal pulses through100 m cable to Nino

Cosmics ~110 mV to LE disc. with 30 mV threshold

and to X10 amp (~1.1 V) to Nino through 26 dB after

3X amplitude loss in cable

(6 X lower amplitude, 5 X lower threshold)

Trigger counters to TDC

Hard to identify peak in Nino signal…. ~ns width

Suspect change in delay of 100m cable over days of cosmic

run

Slide15

15Cosmic (vertical) Hcal pulses all timing signals through100 m cable

Discriminate trigger counters (and make coincidence) before cable, re-discriminated

after cable then TDC

Cosmics

~110 mV

X10 amp to cable (/3) to splitter (82% !!)

to LE disc. with 30 mV threshold

(11 mV effective)

and X10 amp to cable (/3) to 26 dB (/20)

(36 mV effective)

Slide16

16Analysis of 10k cosmic events(85 ps

trigger width subtracted in quadrature)No walk correction (ADC cut,

ToF

cut, ignore tail)

Nino:

s

=438

ps

Leading edge:

s

=322

ps

With walk correction (thanks Juan Carlos) from ADCNino:

s

=333 ps

(corrected with NINO width)

Leading edge: s

=263 ps

With (black box) neural network (thanks again) correction

based on: Nino start time, time above threshold and ADC

Nino: s=272

ps

NINO width and start time drifted over run?? (ADC didn’t)

Slide17

17Expected number of photo-electrons (

npe)

Cosmic muon through vertical

HCal

deposits ~80 MeV in

scint

. ~380

p.e.

observed:

~5

p.e.

/ MeV in

scint

.

(Normalized to that,

Geant

predicts

~6 p.e.

/with Cherenkov turned on ??)

3X3 cluster

‘Max’ 250-700 MeV

single (highest E)

module

‘Max’ 175-500 MeV

875-2500

p.e.

Slide18

18Drive PMT with variable-brightness LED pulser

Digitally-selected brightness 0 & 1-63 0-11,500

pe

added optical filter between LEDs and fibers

0-2300

pe

(c.f. ~400 for

cosmics

)

Can simulate pulse height variation of

cosmics

to try to

understand timing behavior.

Take data at 100 Hz instead of 1 every 15 minutes!

Slide19

19

ADC spectra for bit patterns 0, 7, 15, 23, 31, 39,47, 55, 63

Pulser

& fibers are very stable so number of

p.e.

can be

found from ((Mean – pedestal)/width)

2

Slide20

20Because of fiber-to-fiber variations (and rough LED calibration)number of p.e. may not be a monotonic function of pattern

Can be calibrated in hours

pulser

run (for all 288 PMTs) and shows great stability.

Allows setting of proper HV and gain matching in advance

ADC (mean)

s

with HP filter

s

(

ped

noise sub.)

calculated

npe

Slide21

21

A Surprise

Timing of the pulses shows

jumps depending on bit pattern.

Cause not understood.

Pulser

is primarily for setting

gain, not timing.

Timing information needed

for bench test of NINO (and

handy for use with

HCal

).

Calibrate timing offsets with

constant-fraction discriminator

(or scope).

Slide22

22

Leading Edge

disc. before/after

Calibration removes jumps

allows walk to be studied.

(Another set of calibration

constants, one for each of

16

pulser

boards)

Slide23

23

?

?

NINO ‘corrected’

mean time (for a later run)

Leading edge discriminator

‘corrected’ mean time

Slide24

24StatusPulser almost ready as test bed for simulating cosmicor nucleon events (to aid in testing of NINO vs. standard leading-edge disc.) Is NINO comparable to commercial leading-edge disc

.?? Are LE disc. available if not?

All bugs in

pulser

system must be uncovered and resolved by summer so

pulsers

can be installed.

Still to do: QE vs. gain of PMT…. do bases need to be modified?? (XP 2262/GHC base and XP 2282/

JLab

base)

Slide25

25P

Slide26

26

Slide27

27StatusProgrammable

pulser moved from test stand (light-tight box)to same

HCal

module used for cosmic tests. Fibers inserted.

Control cables run, loading of patterns works,

debugging pulse command to board

Data with Nino card soon.

Slide28

GEp trigger (or possible HCal trigger)HCal is 12X24 array of modules.

Form overlapping regions

by taking

all possible 2X2 sets of (4X4 module)

groups

. Each group is a member of four

regions (or less).

Total of 10 regions to be summed to give

total energy in region. Each sum can be

compared to threshold.

Ten logic signals to send to

ECal

to look forenergy in region of HCal expected to

correspond to ECal

hit.

28

Slide29

29

1

2

1

2

2

2

2

2

2

2

4

4

4

4

2

1

1

2

For

GEp

trigger

18 (4X4 module) groups require:

18 (16+

chan

) summing modules

Have

17

UVa

120s (more available?)

(Could turn some modules into 2X18

chan

sums with fairly simple mod.)

17 modified and tested (2 repaired)

14 linear fan-out channels (4 modules)

10 fan-in

chanels

(3 modules)

Have

2

UVa

quad 133s (more

available?)

(Built-in disc. BUT 100 ns delay !!

and ECL out)

Slide30

30Sum X -0.393Sum X -0.697 X 2

integral/4.7 nsdecay time=230 ns

X -1

6.5 ohm

Slide31

Plan for LED pulsing.

Don’t worry about firing adjacent

PMTs simultaneously but avoid having

adjacent electronics channels for same

LED pulse.

Slide32

Simple cabling pattern.Adjacent electronics channels don’t fireon same LED pulse and unlikely to fire in same cluster.Electronics module problems show up as many bad HCal

signals in two rows, easy to spot.

Somewhat easier to track down single-

module problems.

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

Slide33

33Considering possible routes for cables (assuming no local hut)

Slide34

34Choose PMT gain so expected signal sums to ~60 nVs after amp giving ~1V output. Discriminate at

eg. ~0.25 V.

Summer output vs. pulse integral

Slide35

35~240 mV

~4400

p.e

Expect 6200-7200

p,e

PMT @ 1325 V

Low Q.E. ?? Need more testing

Slide36

36

{

Zeners

on last 3 stages

total 500 V

{

@ 1325 V

only 185 V K-D1

probably need a

Zener

(“Easy” fix on 200 CMU

XP 2262 bases…

~100

JLab

XP 2282 bases are SMD

but lower-gain 8 stage?)

Slide37

37Remaining tests:(This summer)Double-check QE at low HVCompare with Zener

on K-D1Test XP 2282 at low gain for QE drop

Test timing with Nino boards with/without 100 m cables

Require voltage-divider boards to give mV signals

with isolation capacitors!

Also test with

cosmics

(less smooth signal)

Make final decision on use of Nino boards

Decide whether 100 m cables OK

…if so, decide cable route (and rack positions in hut)

Slide38

38For all experiments:288 ~5m BNC/lemo cables24 12-chan amps &2 NIM bins288

lemo/BNC jumper cables288

chan

BNC-BNC patch panel

288 long BNC RG 58 cables (from

ECal

)

3 long cables +1 SHV for

pulser

Asymmetric splitter panel in Electronics hut

with connection/housing for Nino cards

18 Nino cards (with pin connectors)18 sixteen channel LVDS cables18 VETROCs (or F1 TDCs)

288 lemo cables

18 fADCs

4VXI crates 4 ROCs & TIs

288

chan HV (2 1458 crates

24 1461N)12 24-conductor HV cables

24 24-chan. SHV boxes288 1.5 m SHV

288 5 m SHV

For GEp

:288+ lemo cables 18

UVa 120 Summing modules (modified)

4 quad linear FI/FO (e.g. Phillips 740 2.5V)3

UVa 133 quad 4-chan summing modules

12 channels L.E. Disc.3 NIM bins

}

Acquisition

in done?