pulser B Quinn Feb 23 2018 1 Cabling scheme for HCal One signal one HV cable to each tube Plus one HV amp 6 signalsection for pulser Cables secured strainrelieved on gantry behind ID: 815722
Download The PPT/PDF document "HCal Electronics (and" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
HCal Electronics (and pulser)
B. QuinnFeb. 23, 2018
1
Slide2Cabling 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
Slide3fADC
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
Slide41-12Cabling 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
Slide55Map 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
Slide6144 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
Slide77
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
Slide88Time (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
Slide99~2.1 V
After X10 PMT amp
~3000
p.e
!!! wrong!
Slide1010After X10 PMT ampand 100 m RG58 cable~720 mV
Slide1111Nino 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.
Slide1212Fixed-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
Slide1313Fixed-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
)
Slide1414Cosmic (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
Slide1515Cosmic (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)
Slide1616Analysis 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)
Slide1717Expected 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.
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!
Slide1919
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
Slide2020Because 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
Slide2121
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).
Slide2222
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)
Slide2323
?
?
NINO ‘corrected’
mean time (for a later run)
Leading edge discriminator
‘corrected’ mean time
Slide2424StatusPulser 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)
Slide2525P
Slide2626
Slide2727StatusProgrammable
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.
Slide28GEp 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
Slide2929
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)
Slide3030Sum X -0.393Sum X -0.697 X 2
integral/4.7 nsdecay time=230 ns
X -1
6.5 ohm
Slide31Plan for LED pulsing.
Don’t worry about firing adjacent
PMTs simultaneously but avoid having
adjacent electronics channels for same
LED pulse.
Slide32Simple 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
Slide3333Considering possible routes for cables (assuming no local hut)
Slide3434Choose 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
Slide3535~240 mV
~4400
p.e
Expect 6200-7200
p,e
PMT @ 1325 V
Low Q.E. ?? Need more testing
Slide3636
{
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?)
Slide3737Remaining 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)
Slide3838For 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?