SPD commissioning Ricci, Míriam, Daniel, Edu, Hugo et al. from 1.

SPD commissioning

Ricci, Míriam, Daniel, Edu, Hugo et al. from

1

Outline

• SPD calibration with cosmics• Vref calibration

• LED intensity: limits and current status• SPD hardware• LED features

2

Calibration with cosmics

• Have to set threshold @ 0.5 – 0.7 MIP• MIP computed by multiplying calibration of

components– nPhes from Evgeni: cosmics with the detector “on

a table”

• Check MIP computation with cosmics:1) Make sure that SPD occupancy looks uniform on

ECAL&&HCAL cosmics2) Try to compare efficiency with expectation

3

R. Vázquez

Uniformity• Cosmics allowed to solve a bug in HV settings

in a few PMTs• Reasonable uniformity after correction:

4

Threshold @ 0.5 MIP1 M events

Threshold @ 1.5 MIP0.5 M events

R. Vázquez

Global SPD efficiency

• Tried to define a clean sample:– Track extrap. >2 away from acceptance limits– No multiple hits in SPD– Track extrap. x, y < 150 mm– Energy in PS > 40 ADCs– PS cell and track extrap. in 3 & only 1 cell– Look in any BX

• Still, efficiency lower than expected– Reason: time of arrival

5

R. Vázquez

Efficiency vs arrival time

6

R. Vázquez

0.5 MIP 1.5 MIP

100% 50%

Arrival time (ns) Arrival time (ns)

With perpendicular cosmics

• Selecting cosmics with x/x < 1.2

7

0.5 MIP 1.5 MIP100% 50%

Arrival time (ns) Arrival time (ns)

Threshold voltage vs Vref

Vth = 0.65 · Vref ·(n/63), n internal DAC reg.

• Physics setting: in each VFE, minimum Vref that can provide 1 MIP in all channels– To minimize step size, hence resol on threshold

• But will need 2 MIPs for threshold scans• How stable is calibration under changes in Vref?

• Checked by measuring noise offset by performing threshold scans with different Vref

8

Threshold dependence of Vref

• Checked with Vref = 555 mV vs Vref = 1560 mV

• Dispersion lower than 1 step (5-15 mV)• Slight bias due to choice of step (will be optimized)

9

Míriam

Voff (V)

LED system

• Light intensity being adjusted by Evgeni• Limit on LED light?

– Occupancy wrt 40 MHz @ 2·1032: (0.001,0.03) (40K, 1.2M) MIP·s-1

– LED system: 40 MHz / (3564 * 5 ) = 2.2 KHz

• A LED flashing @ 20 MIP is contributing between 3% and 100% of the light from physics

10

Measuring LED intensity

11

• Idea: change SPD threshold to a very high value and look for signal

• Electronics allows for a maximum of 6 MIPs if we want to have uniform thresholds in the whole SPD

Max SPD threshold in MIPs

LED results: intensity

• Complete results from runs from 25th Jan at: http://lhcb.ecm.ub.es/~lhcb/2009/jan/27/

• All in the form of LED-timing scans: LED trigger delay from 0 to 25 ns with steps of 1ns

• Look at many BXs: can study duration of signal• Edu, Evgeni will explain details, here just flash

some results

12

Míriam/Daniel

VFE 35 / Trigger 40 / Vled(8) = 1750

0.5 MIPs

0.5 MIPsOld HV settingsVled = 1700

2 MIPs

6 MIPs

13

Míriam/Daniel

C side

Too much light!

VFE 21 / Trigger 43 / Vled(7) = 1750

0.5 MIPs

0.5 MIPsOld HV settingsVled = 1700

2 MIPs

6 MIPs

14

Míriam/Daniel

C side

Far too much light!Remember subtractor 20% of previous BXHow many MIPs?

VFE 40 / Trigger 40 / Vled(8) = 1750

0.5 MIPs

0.5 MIPsOld HV settingsVled = 1800

15

Míriam/Daniel

C side

Too little light

VFE 91 / Trigger 41 / Vled(8) = 1700

0.5 MIPs

0.5 MIPsOld HV settingsVled = 1700

16

Míriam/Daniel

A side

Far too little light

2/Feb/2009 SPD shutdown activities and LED checks

17

SPD electronics status

• Checks on 1 VFE which showed problems:– VFE number 2 (crate PRS0, CB5, channel2) on C-side

– No communication between VFE and CB before Christmas. Now OK.

– Possible causes:• VFE problem• Control/clock cable problem (as 2 other from same bundle)• Metal dust inside EE box

– Actions done:• As VFE worked last week, just clean all dust.• Wait until all detectors are closed on C-side (if the cable is the problem) or the

problem apears again

– Solution if the cable is found broken:• Use another box control/clock spare cable and extension.• It is not feasible to pass a few new control/clock cables to the problematic zone

through the chain.

2/Feb/2009 SPD shutdown activities and LED checks

18

SPD shutdown activities: tests

• VFE tests:– LED measurements

• Uniformity study and adjustment (Evgeny)• Run types:

– Simple measurement (just apply a trigger pattern)

– Time alignment scan (change delay on the trigger of the LEDs)

– Threshold scan (modify threshold values on each step)

– Offset measurements• Threshold scan with and without HV• Repeat for various Vref and to check stability

– Mapping test• Check the DAQ vs. VFE channel mapping• To do it once

A-side V(2)=2800V(7)=2450V(8)=1700V(9)=2400

V(10)=2500

C-side V(2)=2100V(7)=1750V(8)=1750V(9)=2100V(10)=1750

19

Bias applied

2/Feb/2009 SPD shutdown activities and LED checks

20

VFE 33 / Trigger 42 / Vled(8) = 1750

0.5 MIPs

0.5 MIPsOld HV settingsVled = 1700

2 MIPs

Problems with LEDTSB, VFE configuration?

2/Feb/2009 SPD shutdown activities and LED checks

21

VFE 38 / Trigger 41 / Vled(8) = 1750

0.5 MIPs

0.5 MIPsOld HV settingsVled = 1700

2 MIPs

6 MIPs

Trigger signal reflections

BACK-UP

22

Occupancieslumi2 lumi2 lumi2

lumi5 lumi5 lumi5

23

Conclusions on LEDs

• C-side:– Most VFE have signal with threshold @ 6 MIPs

• VLED(2, 7, 9, 10) too low; V(8)~OK?

• A-side:– Some VFE signal too low (@ 0.5 MIP)

• VLED(2,8) too high; V(7, 9, 10) ~OK?

– Problems with triggers 31, 32, 38, 39 (VLED not controlled) and 42, 56 (not triggered)

– VFE 65(?), 66 no HV

24

‘Teor’Delays from data

Some discrepancies for inner, ~ 3ns. (Check teor fiber delays.)Wait for fine time alignment with first collisions. 25

Time alignmentMíriam

With 2.6 M cosimic events analyzed:

Material traversed by cosmics

26

R. Vázquez

(-5<t<5) (15<t<25)

Offset difference in Volts

V V

VV

27

RUN 36187 (Vref = 555 mV) vs RUN 36204 (Vref = 1560 mV)Míriam