DIRC R&D Software User's Guide
20020817:
MISSINGFILES flat inter note results
2002-10-10 14:49:04 INFO mayankj I repeated some of linescans at y = -4,
17 and 39. There is a uniform shift of 2.6 mm or so in all these three
scans.
So, good news : no gradual shift !! but lack of knowledge about what
happened !
2002-10-10 15:05:00 FEEDBACK mayankj hi! When I checked the limits of
motion controllers, I found out that the shift of 2.6 mm or so is due to
motion controller at address 8.
So, my hypothesis is that it blew up when it was at 2.6 and when reset,
it showed the position as 0. But I dnon't think anyone reset it. When
power is switched off and then back on, I don't know if it knows it's
settings.
I will experiment with it after talking to Thomas.
20021012:
flat inter note results
2002-10-16 11:28:27 INFO hadig Hi,
I just updated the CVS sources to have a faster 7186 readout. My program
reading both 3377 and the 7186 now needs 4s for 40000 events, consistent
with the 10kHz rate.
During the studies, I also found (one of) the reason why we could not
read faster than 10kHz : The ADC we use as a trigger has a conversion
time of 60mus, i.e. can tolerate rates up to 16kHz only. As a
comparison, the 7186 has less than 7.2 mus conversion and processing time
* window size and the 3377 has 1.8 mus + 100 ns per hit + 320 ns
window.
The 7186 window is still set to 200ns or 400ns but I will set it today
to 100ns to get the 25ps resolution. In total the 7186 will be slowest
with 7.4 mus which would allow 135kHz operation.
During my check I found that there is quite some waiting time for the
next signal if I leave the ADC out of the run, so with the following
setup, we could try higher rates again:
* 7186 set to 100ns range
* dummy loop connected, so that we surely get a hit in the 7186
* remove the ADC from the (readout) system
* use the 7186 as LAM provider
* readout both 3377 and the 7186 all the time
2002-10-16 14:30:56 INFO hadig Jerry and I just tried with the new
configuration.
* With 20kHz, the system could take the full rate.
* With 50kHz, the system took 25kHz rate
So my software is able to take between 25kHz and 50 kHz ...
We left the setting at 20 kHz.
I also changed the setting of the 7186 to 100ns range/25 ps resolution,
it now works. However, there is something really strange going on with
the conversion. For common start, the Philips needs a conversion delay.
This delay has to be at least 125ns for the 200ns range option and can
be 0 for the 100ns range option. Before, it was plugged into 2000ns
delay. With the new setting, it did not work anymore. I tried different
cable lengths, set the delay to 0 etc and nothing worked. At the end, I
got it to work, if I put the delay cable into a higher number channel
(as those get converted later ...). I increased the delay and then I
could move the cable into lower number channels again. The lowest delay
I found to get channel 0 working was 4000ns ... I checked on the scope
that the delay between the common start and the signal is 30ns. The TDC
value is 410 counts (10ns) and with maximum scale of 4095 counts
(100ns), a delay of 120ns should be fine ... obviously there is some
Voodoo involved in the jumper settings of the 7186 ... currently it
works ...
20021017:
flat inter results
20021018:
flat inter results
20021021:
flat inter note results
2002-10-23 12:17:57 INFO jjv As I said in my previous e-mail, the laser
diode light output has changed significanly so that Mayank did not
detect any PMT pulses. Apparently, some time on Friday. Yesterday, I
have doubled the diode power, and today I have almost trippled it
compared to the initial setting. The question is if the light output has
changed abruptly, or gradually since switching from 10kHz to 20kHz a
week ago. Anyway, let's hope that the diode will work like this at least
until the work for IEEE is over. Now, that we have a new power setting
it will be interesting to check the timing resolution, as we did not
probe this corner of a phase space before. Doing all this, I noticed
that a LEMO cable on ch. 8 (PMT designation) came out of its crimp. I
have inserted the cable back into the connector, but I would be
surprized if it will work like this. It can be recrimped on Monday. Jerry
2002-10-23 15:21:25 INFO jjv Mystery solved. I went down there and
opened the box. It turns out that the moving stage got stack at the
bottom position, i.e., the spring was not strong enough to lift it up.
As a result we really did not shine into pad #7. Instead, because of no
signal, I have increased the laser diode power enough to get simply
stray light shining off the lens edge into the PMT from side somehow. I
realeased the stage, and immediatelly had to reduce the laser diode
power to original setting. We are getting a normal rate from pad #7.
Tomorrow, I will talk to Matt to see if we can increase the spring
force. In the mean time it should all work, except, as I said, pad #8,
which has bad cable comming from amplifier. Jerry
2002-10-25 14:49:07 INFO jjv I am afraid, not very good news. Mayank
came and said that he thinks the stage is stuck again. I asked him to
move the screw all the way up. I opened the box and indeed the stage did
not wind up. It was at the bottom. I touched it, and the spring did not
bring it all the way up, it came up only to ~80% height. So, something
happened to winding mechanism. We plan to look at it with Matt on
Monday. He has another stronger coil. In the mean time, we should make a
scan only in one direction, i.e., going down, without any calibration
between rows. The data taken so far is suspect to some degree. Jerry
2002-10-28 16:48:07 INFO jjv The scanning setup is operational again.
Today we managed to do the following: 1. Don has fixed the LEMO
connector on channel 8. Signals are OK (I checked). 2. Matt has added
another spring to increase the force to help the stage to come up
easily. It is not absolutely clear why it was getting stack. 3. We glued
a ~20 micron square hole about 7/16" below the center of the 2 inch dia.
reference tube. 4. When we opened, the fiber was caught by the
x-direction controller. This I have not noticed a few days ago when I
opened. So, this was definitelly another problem. It was fixed by a
cable tie. 5. I have been screwing around a bit with the controllers to
make them move in y-direction to a max. height. Not sure that they will
not need some in situ intervention. Do not seem to find a basic manual.
Jerry
20021028:
MISSINGFILES
flat inter note results
20021031:
flat inter results
2002-11-01 16:33:55 INFO jjv I did rotate the H-8500 PMT this afternoon.
It seems to be all running again. Just remember, the maping is different
since I did not rotate the stage... Let me know if there is a problem.
Jerry
2002-11-01 16:33:55 INFO jjv I did rotate the H-8500 PMT this afternoon.
It seems to be all running again. Just remember, the maping is different
since I did not rotate the stage... Let me know if there is a problem.
Jerry
20021102:
MISSINGFILES flat inter note results
2002-11-05 18:57:17 INFO jjv Given a certain urgency, I went down there
and moved the laser diode about a cm lower. This required to drill a
hole. It is not as pretty as Matt's holes but it is a hole, and the
setup is sound. Good luck. Cheers, Jerry
2002-11-06 17:33:30 NOTE hadig Dear all,
I just made a check to compare the efficiency/sensitivity of the 7186
chain and the 3377 chain.
As we have one pad of the PMT on the 7186 and all the others on the
3377, we need to know their relative efficiency.
* I moved the laser to pad 8 and measured the number of hits in pad
8 (on 3377) : 51345 hits in 500k events.
* I moved the laser to pad 7 and measured the number of hits in pad
7 (on 7186) : 40123 hits in 500k events.
* Then, I swaped the cable from the amplifier of pad 7 and 8. pad 7
(on 3377) : 46860 hits in 500k events.
* I moved the laser to pad 8 and measured the number of hits in pad
8 (on 7186) : 44035 hits in 500k events.
* Then, I swaped the cables back. pad 8 (on 3377) : 51546 hits in
500k events.
Conclusion: The 3377 chain leads to 17% more hits than the 7186 chain.
2002-11-06 18:32:40 INFO jjv After getting a call from Joe this
afternoon it is clear that you guys did not understand my message about
drilling. It is simple. First, I understood Thomas that he would like to
have a size of the laser beam. To redo the pin hole properly would take
couple days. So I decided to move the stage down. Nothing could have
been done on the stage itself. However, one could drill another hole in
the laser holder and move it a bit down. That is what I did and you were
supposed to be "flying" yesterday afternoon. Did you ? Incidentally, one
can always go back to previous hole and have it like we had up to now.
2002-11-06 18:55:31 INFO jjv I did not have time to follow the
discussion about a relative efficiency of two TDCs. If i did, I would
have sent this meassage sooner. The two signal paths have different
discriminators. In case of 3377 we have a ~15mV threshold, in case of
7186 we go through a 30mV threshold. So, I think the problem is trivial
and not worth of spending too much time on it. Jerry
20021108:
flat inter note results
20021113:
flat inter results
20021116:
flat inter results
20021117:
flat inter results
2002-11-19 18:56:40 INFO jjv I have changed with Matt the setup today.
1. We have a new aluminum late with the new holes. Both top and bottom
holes should be reacable. I will post the detail drawings later. 2. The
plate has a 20 micron slit at the bottom in one dimension only. 3. There
are side holes with smaller diameter (0.013 inch. 4. I have measured the
focal length of the fiber with the lenses using a blue laser. As a
results I adjusted both the PMT photocathode and the slit to a correct
focus assuming that the laser diode is the similar as the blue laser. 5.
The new PMT #2 was installed with a nominal orientation. 6. All is
running right now. I have briefly explained the geometry to Mayank
already, however, I will post the details later. Jerry
20021121:
flat inter note results
* I went down there and indeed found one amplifier card offsewt by
one pin vertically. So, the scan should be restarted.
* Otherwise, the pulser is set to 20kHz Mayank. Jerry
20021124:
flat inter results
20021124b:
flat inter note results
20021126:
flat inter results
20021128:
flat inter results
20021130:
flat inter note results
20021202:
flat inter results
20021202b:
flat inter results
20021203:
flat inter results
20021204:
flat inter results
20021204b:
flat inter results
20021206:
flat inter results
20021209:
flat inter results
20021209b:
flat inter results
2002-12-06 14:51:01 INFO jjv I have removed the PMT-2 and replaced it
with the PMT-3. This one has a modified resistor chain (2:1:1:1...:1).
All is running again. Jerry
On Tue, 18 Feb 2003, J. Va'vra wrote:
>> Hello Mayank,
>> I have done a modification of the setup to see if the efficiency
>> around the detector boundaries can be improved by not grounding the
>> brackets. You can start the new scan any time.
>> Jerry
2002-12-18 15:49:35 PROBLEM mayankj By analysis of calibration2 files
for date20021204_time91344 and date20021206_time145650, DIRC PMT got
dead between 4th and 6th of Dec. So, we will have to look at the
hardware when we get time.
20021218:
flat inter results
20021220:
flat inter results
20030212:
flat inter results
20030213:
flat inter results
20030218:
flat inter note results
2003-04-15 16:41:26 INFO jjv The scanning setup is again ready for
testing. We have new amplifiers, which provide a gain of 100x.
Unfortunatelly, at this point they also give higher noise. I had to
increase the threshold to 40mV (from 15)!! We may have to itterate on
the banwidth once more to reduce it, but lets make a scan. It took me
most of the afternoon because of a confusion what is up and what is
down. I asked Mayank twice where is the pad#1. First answer was that it
is on the right side bottom looking from the outside observer. Indeed,
if I look at the picture on his web page, this is what I would have said
also. When I said that I do not see a signal, he said it is left bottom
looking from the diode side. After not seeing a signal again, I opened a
box and found that the stage is sitting on top right side looking from
the diode side. Although in software one can define whatever one wants,
in a hardware bottom tends closer to legs and top closer to head.
20030416:
flat inter note results
20030417:
flat inter results
20030418:
flat inter note results
20030425:
MISSINGFILES flat inter note results
20030430:
flat inter results
20030513:
flat inter results
20030521:
flat inter results
20030530: detailed scan of one pad.
flat inter results
2003-06-11 15:34:29 INFO jjv I have looked at the dead channels seen in
the last Mayank's scan. The channels are #18,#21,#26, and #53. Indeed, I
have veryfied that they are all dead amplifier channels. Not clear why.
Perhaps, it is related to a failure of the NIM power supply experienced
before this last scan. In any case, this is a serious problem requiring
a detailed understanding. Therefore, I will wait until Gholam comes
back, which is next Tuesday. Until then the setup is switched off. On
the positive side, I have received a new MCP-PMT, so we can scan as sson
as the system is ready again, which may be the next week.
2003-06-20 15:23:32 INFO jjv I have placed the amplifiers on the MCP-PMT
and started the setup. However, I do not see a laser diode signal,
presumably because it is sitting in another position. Mayank, can you
move it to pad #1 ?
2003-06-23 11:58:12 INFO jjv I looked at Pad 1 and it seems to be
reasonable, so Mayank, you can start the scan.
I also looked into conditions under which we can trigger the system into
the oscillation. They are: (a) improperly pluged LEMO cable, (b)
unconnected LEMO cable and wiggling with ground on our white LEMO cables
(at least with Pad 1). These crimps were done by Don and he complained
that he cannot do a good job because the cable is too thin for the
connector. Anyway, nothing we can do this second, but we should be
careful wiggling with cables.
I looked at the noise and it looks like that our thresold (~80mV) is
rather high to be affected directly. It will affect the timing
resolution though in future when we care about it more. So, we wil have
to set up a gate to kill it.
Finally, there is going to be a power outage tomorrow in our area.
Luckily, it will not affect the scanning setup (I checked experimentally
today with an electrician). However, it will affect the cosmic ray
telescope, so I switched it off.
20030623:
flat inter note results
2003-07-07 18:39:27 INFO jjv I have restored the system and it is now
operating. It turns out that one bad channel was indeed again a dead
amplifier, and the second one suffered lower gain because of a cold
solder joint, which made it intermittant. All dead channels were caused
by the same mechanism, we believe with Gholam. Because of a historical
reason, the protecting diodes were connected to the second board (10x).
When a large signal occurred, the diode was simply not fast enough to
protect the chip. All protecting diodes were therefore moved to the
front card (40x), and hopefully we will be sailing smoothly.
So Mayank, if you can you can start to scan the couple pads we were
missing. Let me know if you cannot do it. Let me also know when this is
finished, so I can put in the next tube.
20030708:
MISSINGFILES flat inter note results
2003-07-11 19:10:10 INFO jjv As Mayank told me that he finished a scan
of MCP-PMT#2, I have placed into the scan position MCP-PMT#3. I am told
that this tube is better, and indeed I see higher gain on pad #1, and
higher QE, so I had to reduce the laser diode power a bit. It is running
nicely. The new tube is running at the same volatge, i.e., -2.4kV.
So, Mayank you could start the new scan right away. Please, check that
all pads are working again. If not, let me know. Thanks.
20030712:
flat inter results
2003-07-20 13:22:14 INFO jjv I have finally managed to come to 403 bldg
and changed the MCP voltage from -2.4kV to -2.3kV. So, if either Mayank
or thomas would be kind enough and start the scan again. The plan is to
continue like this and determine the plateau curve for each pad.
20030720:
flat inter results
2003-07-24 18:17:28 INFO jjv I have changed voltage to -2.2kV (from
-2.3kV).
20030724:
flat inter results
2003-07-28 17:55:27 INFO jjv I have changed a voltage on MCP-PMT to
-2.1kV (from -2.2kV).
2003-07-28 18:36:55 INFO jjv I should add that we are presently running
with a 85mV threshold in all discriminators. In the lab, I could
actually run lower threshold, but in a system of 64 channels, and 403
bldg., I had to increase it at some point. This was because of an
enviromental noise.
20030728:
flat inter results
2003-08-01 18:14:50 INFO jjv I have brought three cards to Gholam's lab
to check if there is something wrong with channels 2, 50 and 38. They
are OK. We checked also neighbors to make sure I am not off by one or
so. So, whatever happened to channel 38, it must be something else.
After checking in the lab upstairs, I brought them down and connected
them. Channel 38 is fine when selftriggering. When I trigger on Pilas
laser diode I get counts but at small rate. However, I am not 100% sure
that Thomas has left it on the channel 38. Anyway, the system is running
and we could start another scan. The conditions are: -2.4kV and 65mV
threshold (I moved it down from 85mV). It would not hurt to do a quick
scan to check that I did not introduce a new bug.
20030801:
flat inter results
2003-08-06 18:00:21 INFO jjv I have reduced the
disc. threshold to 50mV. It was 65mV in the previous scan. The MCP-PMT
voltage is the same, i.e., -2.4kV. The noise is well below the
threshold, so except an ocasional enviromental noise, we should be
OK. Jerry
2003-08-06 18:00:21 INFO jjv I have reduced the
disc. threshold to 50mV. It was 65mV in the previous scan. The MCP-PMT
voltage is the same, i.e., -2.4kV. The noise is well below the
threshold, so except an ocasional enviromental noise, we should be
OK. Jerry
20030807:
flat inter results
2003-08-13 17:45:37 INFO jjv I have the thresholds to 85mV
(we had 50mV in the previous run). The MCP-PMT voltage remains at
-2.4kV.
Jerry
20030813:
flat inter results
2003-08-19 13:44:37 INFO jjv I looked at the amplifier
ch. 38. It was acting up once a few days ago, showing a large
disturbance of the base line. Since then it works perfectly. I plugged
it back in. Let's try to make a scan of this pad only.
Jerry
20030819:
flat inter results
2003-09-11 15:41:49 INFO jjv I looked at the blue laser
diode and concluded that we may have to live with the time jitter. If
I go to a multiphoton mode, its leading edge becomes stable. However,
as I attenuate the light into a single photon mode, I start seeing a
large jitter. It may be feature of this "more expensive" diode. I have
to ask Sumiyoshi if he knows that. Anyway, I have it on loan for a
week only. I would suggest to start running with much larger time
window of acceptance, say 10-20ns. Right now the laser diode power has
been adjusted so that we are in a single photon mode.
Looking at the signals, we do get a noise from time to time. If we see
anything early, we should throw away an event.
Cheers, Jerry
2003-09-12 15:49:27 INFO jjv The jitter problem is
solved. It turns out that the blue laser diode cannot run well with a
substantially throttled down power. After putting an absorber, and
also playing with a connector, I get finally a small timing jitter at
a power setting of ~53%. Thomas, try to run it and let me know how it
goes. If an absorber is to fall off you would get for every trigger a
pulse. Hopefully, it will not happen.
Cheers, Jerry.
2003-09-13 20:26:03 INFO hadig So, it seems that Mayank's
program still runs and it found a shift of -12.54 mm on the top
controller. In the base, it finds an x shift is of -44.9797
together, that is -57.5mm ... about the whole PMT.
The base Y shift is 7.72904 mm.
Actually, it makes total sense, as the laser was standing at pad 1 at
the time of the power outage, the center of pad 1 has the coordinates
57.5, -8
I think the program has reestablished the old system, so I will start
the scans again ... as you said with a corser step size.
Thomas
20030913:
flat inter results
2003-09-15 10:35:46 INFO jjv Hello Thomas and Joe, I have
changed thresholds from 85mV to 50mV to see if the efficiency
changes. This happened at around 9:45 this morning. I propose to run
until tomorrow until ~9am. Lower threshold will bring more "paging"
noise, but hopefully we can throw it out by some clever cut on early
hits.
Jerry
2003-09-19 14:34:51 INFO jjv I have increased thresholds to
100mV (from 50). I looked at a cross-talk from pad #1 to pad #2. It
almost does not exist, that is using that "cheapo" little scope (it
may exist with a fast scope). However, what I did find is that a LEMO
cable on pad #2 is extremely sensitive to touching, and can cause an
oscillation of the entire system (Don did really a bad job on these
white cables...).
Anyway, if you run out of things to do, why don't you shut it down so
we do not burn the laser diode at least.
Thomas, I will update the bckg. file later today and send you some
comments based on today's background meeting.
See you on October 7. Cheers, Jerry
2003-09-19 14:36:55 INFO jochen I started a new scan with
100mu step size in x, 1mm step size in y from -11mm to 41mm, the usual
layout. -- Joe
20030919:
flat inter results
2003-10-08 14:22:25 INFO jjv I placed an amplifier board
back in. A good news is that the fault we have in the new PC board is
not present in our prototype board, i.e., it is not a design feature
but simply a mistake. A bad news is that I found the whole system
oscillating due to one bad crimp contact in a white cable, and, that a
channel #1, which was used the most heavily, finally failed for
good. Right now the system is on and "behaving" without the channel
1. Go ahead and do a test, but as soon as it is done, I would like to
replace all white cables. Cheers, Jerry.
20031009:
flat inter results
2003-10-13 17:14:40 INFO jjv The cosmic ray telescope HV
was switched on by me and Francisco. Thomas, could you start data
taking ? Francisco will try to look at the data and see what is
different. Later he will try to find an edge of a bar. Everything
seems to work except one PMT (S4-3) is very noisy. We left it as it is
to see if it will quiet down over night. It is otherwise a candidate
for replacement in future. Thomas, could you add Francisco to the list
of the log book users ? Another bug I seem to see is that when I ask
for a CRT log book, I actually get a Scanning setup log book. Cheers,
Jerry.
2003-10-15 16:38:47 INFO hadig Hi,
I added Aakash and Francisco to both log books.
A reminder on how to get to the logbooks:
For PMT scanning setup information:
http://kosmic.slac.stanford.edu/cgi-bin/hadig/logbookpmtscanelogbook.pl
For CRT information:
http://kosmic.slac.stanford.edu/cgi-bin/hadig/logbookcrtelogbook.pl
Both pages are available from SLAC computers only.
Ciao
2004-01-07 17:42:08 INFO jjv Today, all 64 amplifiers were
finally ready, so I have started the setup. The laser diode was
sitting on the DIRC tube, so I have adjusted the rate using
it. However, neither me nor Joe seem to know how to position the laser
diode on Pad #1. Joe tells me that the old recipies do not
work. Hopefully, Thomas will rescue us from Germany. Sorry,
Thomas. Otherwise, we should consider the initial run as an
engineering run - there could be mistakes. So we should start just
with a coarse scan. I do see a HV noise on pad #64 & 63. Burle may
have a problem in that region. But, I need to spend more time on
it. Jerry
2004-01-13 15:48:36 INFO jjv After Thomas positioned the
laser head on pad #1, I went down and tried to start it. It did not
work, because I managed to get confused when installing the amplifier
boards (they were swapped). It is all set now, ready to do a scan.
Two additional observations: 1. We seem to need more laser power
now. This will need to be investigated. 2. Channel 64 is indeed close
to the HV connector, where I seem to observe noise pulses. In my log
book I have an error in the picture, which I will correct.
Jerry
mcp3_20040113:
flat inter results
Good uniformity of 100%...90% (75% on pad 50), peak efficiency at 65% of Photonis.
2004-01-19 13:44:58 INFO jochen Lowered HV on MCP to 2.3kV,
all other parameters unchanged. Started new scan.
mcp3_20040119:
flat inter results
Peak efficiency at ~55% of Photonis.
2004-01-23 10:26:55 INFO jochen Lowered HV from -2.3kV to
-2.2kV. Left all other parameters unchanged. Started new scan.
mcp3_20040123:
flat inter results
Peak efficiency at ~50% of Photonis.
2004-01-27 14:58:44 INFO jjv I placed a new MCP-PMT tube. I
call it #4. It runs at -2.4kV at the moment. So far, these are the
tubes we tested according to my accounting:
1.MCP-PMT#1, 2x2, S/N 08290202 (no scan, only timing tests),
2.MCP-PMT#2, 8x8, S/N 01150302, laser at 28%, 3.MCP-PMT#3, 8x8, S/N
05220304, laser at 28%, later 41% after a blue laser run, 4.MCP-PMT#4,
8x8, S/N 09130305, laser at 47%, running now.
I started to record the necessary laser power to get ~10% efficiency
on pad#1. It increased after the tests with the blue laser from some
reason (either some sort of pollution from a ND filter, or the laser
itself has changed). In a setup in my trailer I need less than 10%
typically.
Jerry
mcp4_20040128:
flat inter results
Good uniformity of 100%...75%, peak efficiency at 35% of Photonis.
2004-02-01 16:08:35 INFO jochen Jerry checked on pad #1 that
the rates on MCP #4 looked reasonable after the end of the last
scan. He verified that the rate looks fine at the again slightly
higher power of 47%. The new MCP in the setup has serial number
07030301 and is MCP #5 in our consecutive numbering. Started new
scan.
mcp5_20040201:
flat inter results
Poor uniformity of 100%...50%, peak efficiency at 35% of Photonis.
2004-02-02 16:38:17 INFO jjv I went down and checked a
noise in channel 64 of MCP #5, which is now running. I should say that
I did not find anything catastrophic with the tube. When I switched
lights on and connected channel 64, while triggering on the laser, I
saw the following noise (scope is at 10ns/div and 50mV/div, the
discriminator threshold is nominally 100mV):
http://www.slac.stanford.edu/~jjv/activity/dirc/noise_channel_64_1.JPG
It is high but it should not trigger the nominal threshold, unless
there is some internal threshold DC offset, or the discriminator has a
different BW response relative to the scope.
Channel 63 had aslightly lower noise:
http://www.slac.stanford.edu/~jjv/activity/dirc/noise_channel_63_1.JPG
After 10 minutes, I saw this noise on channel 64:
http://www.slac.stanford.edu/~jjv/activity/dirc/noise_channel_64_2.JPG
It is smaller, which may mean that balast in lights settler down or I
simply wiggled the cables.
Anyway, the bottom line is that I see a larger cross-talk in channel
64, which is not nice and which was reported to Burle, but I do not
see a show stopper from this particular point of view. In future, we
may have to set a different threshold for channel 64 (by addding an
attenuator) to take it into account its larger cross-talk noise
level.
Jerry
2004-02-02 17:26:19 INFO jjv A comment to clarify things
after I talked to Joe. It is not suprising to have variation between
channels in such a complex tube. If a DIRC tube would be segmented to
64 channels, we would also see all sorts of things. Anyway, what I was
looking for was a sign of MPT troubles (channels 64 is just next to HV
feedthrough, for example). I do not see that. Therefore, either the
noise is related to a noise from the laser firing, which is possible
(I am triggering off it), or it is a result of the cross-talk activity
from other pads (a fast pulse in other pads can induce a sligh
oscilaltion in cannel 64, as I have shown in the R&D meeting). If we
run into these problems in the final setup, we have to adjust a
threshold on the channel 64 with an attenuator. However, I agree with
Joe. At this stage it does not hurt to check MCP #4 again. Jerry.
2004-02-03 18:12:21 INFO jjv I swaped channels 64 and 63 at
about 6pm.
2004-02-04 11:57:27 INFO jjv I swaped channels 64 and 63 to
nominal positions today at 11:36am. At that point, I did not see a
noise in channel 64, which should trigger 100mV, unless something is
wrong with the discriminator. Jerry
2004-02-06 14:18:10 INFO jjv I looked at pulses from Pad
28, MCP#5. The gain is nominal, but we just have fewer pulses than pad
28 of MCP#3, which is now in in the setup. It runs at -2.4kV. Jerry
2004-02-06 14:29:35 INFO jochen Started a scan of 4 lines on
MCP #3 at 3,4,5,6mm with only 20k triggers in each position to check
if the loss of efficiency in red relative to the Photonis, observed
for MCP #4 and 5, is real and if MCP #3 has a noisy channel 64. Scans
should be done by 7pm.
2004-02-06 18:41:19 INFO jjv I have removed MCP#3 and
placed in MCP#6 (S/N 09050302). It is running at -2.4kV. The laser
power at 48%. It appears a tiny bit lower eff. cpmpared to #3. The
gain is nominal. Jerry
2004-02-06 20:57:28 INFO jochen Started new scan for MCP
#5. 20,000 triggers per point, adjusted measurement area by few mm.
mcp6_20040206:
flat inter results
Poor uniformity of 100%...50%, peak efficiency at 30% of Photonis.
2004-02-07 17:44:11 INFO jjv I have switched the ref. tube
power supplies at 17:40 today. Jerry
2004-02-10 11:56:20 INFO jjv MCP#7 is in and running. It
needs about the same laser power as MCP#6, i.e., 48%. Its serial
number is 08220302. The voltage is -2.4kV. Looking at ch. 64 wia a
T-bridge, one does not see noise which should trigger 100mV. This tube
was on Burle's list to be returned because it had slightly higher
leakage current. to me it looks good. Jerry
mcp7_20040210:
flat inter results
Poor uniformity of 100%...50%, peak efficiency at 50% of Photonis.
2004-02-13 11:38:06 INFO jjv MCP-PMT#8 is running at
-2.4kV. I had to tweak the laser power from 48 to 49% to get ~10%
rate. The serial number is 09130303. Jerry
mcp8_20040213:
flat inter results
PMT seems resonably uniform, overall efficiency ~25% or Photonis.
mcp8_pad23fine_20040215:
flat inter results
Noticed fairly large intensity fluctuations of laser diode that does not
see to 100% cancel with the calibration. Needs further study.
 |
| Last modified: Feb 15, 2004, |
by Joe |
|