-- WouterHulsbergen - 09 Apr 2006

Documentation

• Manual for the Hampton Scintillators
• Ortec CFD manual
• TDC manual
• Meantimer manual
• NIM-ECL convertor

Scintillator setup

The three 'Hampton scintillators' have a sensitive volume of 144x40x2.5 cm. Here is a figure made with this macro that shows the position of the scintillators with respect to the barrel. The approximate position of the scintillators with respect to the barrel axis, as seen from the C-side of the detector, is

scintillator	x-position	y-position
HSC1 ('top')	-78.3	200.0
HSC2 ('middle')	60.3	-160.0
HSC3 ('bottom')	120.0	-317.7

The uncertainties in these positions are about 1 cm.

In the manual above the three scintillators are labeled HSC-1, HSC-2 and HSC-3 picture. Each scintillator is read out on both sides by an Hamamatsu R2115 PMT, connected to an Ortec/EG&G type 269 base. In the manual the PMTs are labeled 'A' and 'B'. The following table shows the mapping from the labeling in the manual to the labeling in the cosmic setup:

PMT	PMT name in manual	position	HV cable	readout cable
1	HSC-1B	top A	HV1	PMT1
2	HSC-1A	top C	HV2	PMT2
3	HSC-2B	middle A	HV3	PMT3
4	HSC-2A	middle C	HV4	PMT4
5	HSC-3B	bottom A	HV5	PMT5
6	HSC-3A	bottom C	HV6	PMT6

We are using several scenarios for the HV settings:

PMT	PMT	Reference	HSC Manual HV	Nabil's favourite HV	Final HV
1	HSC-1B	1700	1700	1800	1800
2	HSC-1A	1700	1700	1775	1700
3	HSC-2B	1700	1700	1800	1700
4	HSC-2A	1700	1700	1800	1700
5	HSC-3B	1700	1800	1800	1850
6	HSC-3A	1700	1700	1775	1700

The 'reference' scenario was used for most tests described here. They are a blind copy of those from the test performed in Hampton, except that the last PMT was chosen to be the same as all others. Nabil's numbers were tuned by looking at pulseheights. Wouter HV were tuned such that the 'time-difference' profile of the two PMTs on a single scintillator was flat. (See also below.)

The PMTs are connected via a linear fanin-fanout to an Ortec 935 CFD. The CFD settings are currently:

delay	4ns
threshold	45 mV
signal width	40 ns
walk adjust	0 mV

The delay is the same as used for the test described in the HSC manual. The threshold is the lowest threshold for which the CFD behaves well on all 6 channels that we use. (Below this threshold the output is going crazy, so there is probably some feedback somewhere.) Note that the threshold used for the test described in the manual was 75 mV. The walk adjust has not been tuned.

Signals

Below a plot showing the different signals at the output of HSC1-A (upper), at the CFD-Monitor (middle) and CFD-Output (lower) (time is in Seconds and voltage is in Volts). * Plot

HV Scans

Coincidence rate between different scintillators are typically 1 Hz, which for people as impatient as us, is not sufficient for taking a HV profile based on inter-scintillator coincidences. Instead, we exploit the fact that each scintillator has two PMTs and measure coincidences on a single scintillator. We use '1700 V' as the default HV and measure rate and coincidence rate as function of the HV of one PMT, keeping the other one fixed.

• current versus HV
• raw counting rate versus HV
• 'pseudo-efficiency' versus HV. This is the ratio N12/N2 where N2 is the number of counts in the other PMT and N12 is the number of coincidences.
• 'pseudo-fakerate' versus HV. This is the ratio (N1-N12)/N1 where N1 is the number of counts in this PMT and N12 the number of coincidences.

Timing measurements

How to read the TDC:

• login on lnxpool17 as user sr1daq
• source some setup files
• cd /daqsoft/sct
• source setup_tdaq14_branch_with_tile_and_ltp.sh
• cd /daqsoft/sct/TimingInfo/detectors/Tile/TileVMEboards/i686-slc3-gcc323-opt
• ./v775_test 0x00100000 100000 |tee tdc.events
• the tdc.events file contains the channel id and its DAC value (time of arrival of the signal for this particular channel) delayed with respect to the trigger (common start for all channels).

data[0]=0xf8054435 type=data chan=5 un=0 ov=0 value=1077

data[1]=0xf8064526 type=data chan=6 un=0 ov=0 value=1318

• decode this file/substract/fit/plot...
• how to translate from DAC values to time value??? you've to multiply by 140ns/4096
• the current/default TDC settings corresponds to a time range of 140 ns.
• repeat this at different voltages...

Here are some first results using the TDC. The setup is more or less the following:

• All PMTS were operated at 1700 V. This is probably lower than we want to use.

• The CFD outputs are connected to a NIM-ECL convertor and then to the tdc. The mapping on the tdc is currently as follows:

  	name	tdc channel
  1	HSC1-B	0
  2	HSC1-A	1
  3	HSC2-B	2
  4	HSC2-A	3
  5	HSC3-B	4
  6	HSC3-A	5
  	meantime HSC-1	6
  	meantime HSC-2	7
  	meantime HSC-3	8

• the connection between NIM-ECL and TDC is not great yet: The connectors are actually 34 pins, but since we did not have those, we modified a 16 pins connector. It does the job, but it will fall apart at some point.

• the TDC is operated in common-start mode using PMT1 (HSC1-B) as trigger. The total delay between trigger and tdc signal is about 40 ns. A considerable part of that must be inside the NIM-ECL convertor itself.

We measure 10000 triggers. The results are summarized in these two plots:

• times measured in HSC1. Note that there is a considerable spread in channel 6, which is the trigger itself. This is quite surprising. Was is also intersting is the distribution of the difference: The fact that it is not flat suggest that the efficiency is not flat!
• times measured in HSC2.
• times measured in HSC3.
• differences in the meantimes. This plot is the best advertisement: it seems that we can obtain a timing resolution of about 0.5ns.

Studying the time resolution

If the efficiency of the scintillator is more or less constant over its length, the 'time-difference' distribution follows a 'block-pulse' convoluted with a gaus. The results from a fit of this function to the time difference distributions can be found here. The trigger uses a coincidence on the meantimes of top and middle scintillator. This is important because once we apply coincidences, the 'illumination' of the scintillators is no longer entirely constant over their surface. An example of this effect is seen in this plot, which shows the time difference distribution in the middle scintillator for events with a coincidence in top and middle and for events which in addition require a signal in the bottom scintillator.

Ignoring this effect for the moment we can derive some interesting information regarding the resolution and signal velocity. The 'sigma' derived from this fits is the squared sum of the individual PMTS. We therefore have

scintillator	average PMT time resolution [ns]	signal velocity [m/ns]
1	0.430 +/- 0.014	0.1611 +/- 0.0004
2	0.361 +/- 0.013	0.1585 +/- 0.0003
3	0.371 +/- 0.019	0.1607 +/- 0.0005

The errors are statistically only and should be taken with the usual grain of salt. However, the fact that the derived 'velocity' is reasonably equal in the three modules is a convincing sign that things are reasonably well understood.

Using the triple concincidences we can also test the 'z' resolution: This figure shows the z-position in the middle scintillator following two calculations, namely from the time difference in this scintillator and extrapolated from the time differences in the other two scintillators. For the velocity of light we used 0.160 m/ns. The correlation is remarkably good, though not excellent: It is as if either of the two calculations is biased on the middle of the scintillator.

Charge measurements

How to read the ADC:

• login on lnxpool17 as user sr1daq
• source some setup files
• cd /daqsoft/sct
• source setup_tdaq14_branch_with_tile_and_ltp.sh
• cd /daqsoft/sct/TimingInfo/detectors/Tile/TileVMEboards/i686-slc3-gcc323-opt
• ./v792_test 0x00200000 100000 |tee adc.events
• the adc.events file contains the channel id and the "charge".

data[0]=0x36 type=data chan=0 un=0 ov=0 value=54

data[1]=0x62 type=data chan=16 un=0 ov=0 value=98

• decode this file/fit/plot...

The mapping...

	name	adc channel
1	HSC1-B	7
2	HSC1-A	6
3	HSC2-B	5
4	HSC2-A	4
5	HSC3-B	3
6	HSC3-A	2

as an illustrative plot: the charge seen in the different channels (one of them is not connected...????) charge collected in the ADC

Charge and Time measurements

How to read the ADC and the TDC at the same time within the TDAQ framework ?:

• login on srcsctdaq1 as user sr1daq
• soiurce setup.SCT.sh
• start (???)
• data are stored in /data/sct
• their format is binary
• either work within ATHENA (need help from Luca Fiorini)
• either use eventStreamDecoder to decode the file like this
• eventStreamDecoder daq_SCTEB__0002456_file01.data > run_2456.data
• it contains the same data structure as for the TDC/ADC/ standalone programs...
• it contains: the L1Id: L1ID = 0x4 BCID = 0x0 TType = 0 det type = 0
• a TDC fragment is: SubFragmentID = 1 Length = 10 Found TDC fragment. Decoding...
• a TDC data stream is as before:

data[0]=0x147 type=data chan=0 un=0 ov=0 value=327
data[1]=0x220 type=data chan=1 un=0 ov=0 value=544
data[2]=0x3f6 type=data chan=2 un=0 ov=0 value=1014
data[3]=0x32f type=data chan=3 un=0 ov=0 value=815
data[4]=0x479 type=data chan=6 un=0 ov=0 value=1145

• an ADC fragment is:

SubFragmentID = 2 Length = 38 Found ADC fragment. Decoding...
data[0]=0xf8004035 type=data chan=0 un=0 ov=0 value=53
data[1]=0xf8104056 type=data chan=16 un=0 ov=0 value=86
data[2]=0xf801401e type=data chan=1 un=0 ov=0 value=30
data[3]=0xf8114050 type=data chan=17 un=0 ov=0 value=80

• decode, fit, plot

here as an illustrative example, the time resolution achieved with the TDC with and without the meantimer.

* without the mean-timer (we simply use the time information from the different PMs...)
figure

* with the mean-timer
figure

Time Of Flight (TOF)/Momentum measurement.

An ongoing discussion with Wouter on whether the momentum that we can extract from the TOF makes much sens... is the TOF that we get from the two meantimers, the real TOF, or is it simply the TOF, plus some systematics due to the electronics + some aspects of the setup that we clearly do not understand. Anyhow, one can extract from the TOF, a momentum. Whether it makes sens or not, we need to investigate this in more details. Here is a plot showing the TOF from the Momentum, using HSC-1 and HSC2 only.
figure
The plot shows the TOF limit of 13 ns for muons with momenta above 0.5 GeV, in agreement close to the TOF that we measured. The plot includes also the resolution on the TOF (~600 ps ). This needs further investigations...