CKM Testbeam Tiggger/DAQ Logbook
                           Peter S. Cooper
                              Fermilab


Address Book

Mike Behnke     FCC3   MS:368 x2700      "Mike Behnke"      <behnke@fnal.gov>
Bill Finstrom   FCC3   MS:368 xyyyy      "Bill Finstrom"  <finstrom@fnal.gov>
                                                     <offsite_loans@fnal.gov>
Bob Forster     FCC3E  MS:368 x2535,4298 "Bob Forster"     <forster@fnal.gov>
Vince Pavlicek  FCC3E  MS:368 x3257      "Vince Pavlicek"    <vince@fnal.gov>
Dave Slimmer    FCC2   MS:369 x4334      "David Slimmer"   <slimmer@fnal.gov>
Margaret Votava FCC2   MS:369 x2625      "Margaret Votava"  <votava@fnal.gov>

Thu Apr 18 11:42:38 CDT 2002	Begin Work
----------------------------

Dave Slimmer's T917 webpage:  http://www-ese.fnal.gov/ods/ode/t917_da.html

created ckm01:/ckm/daq to work in
filing mail on the daq mail folder

Downloaded t917 code from /afs/fnal/files/home/room2/slimmer/t917.tar.gz
Downloaded sjyLX from http://fermitools.fnal.gov/abstracts/camac/abstract.html
Printed PN540 - font sucks but readable - MSword version NFG on wincenter

Visited Lab-6 this morning.  Clear plan is to clone that system.  Its
layout is in a 5 foot relay rack like:

         +--------------+
         +              +
         +Visual Scaler +
         +Visual Scaler +
         +              +
         +              +
         +--------------+
         +              +
         +  NIM Crate   +
         +              +         +-----+
         +--------------+         |Linux|      /-------\
         + CAMAC       7+  SCSI   | PC  |      |Monitor|
         + Crate       3+<------->|     |<---->|       |
         +             A+         |     |      |       |
         +              +         |     |      \-------/
         +--------------+         |     |    M  /-----\
         +              +         |     |       \-----/
         + Fluke HV PS  +         |     |
         + Fluke HV PS  +         +-----+
         +--------------+
         +              +
         + Berkley Cow  +
         +              +
         +--------------+
     
Need a SCSI adapter for CKMxx.  Dave Slimmer suggest I check with Mike Behnke

Will go shopping at FCC after lunch

Thu Apr 18 15:09:36 CDT 2002 - back from FCC
----------------------------

Got LSILOGIC SCSI card and cable from Mike Behnke ($70 on PAL)

Mike suggests I just file a PREP request form on the web
Even the Berkley cows are back in the barn!

Spoke with Vince.  He'll coordiante with Bob Forster about get me a rack
and space on FCC3.

Called DAve Slimmer to ask where fnca is (sjyLx/examples/cnaf.c)

Built sjyLX library and examples.  
  Had to hack sjyLX/src/sjy_interface.c to #include "/usr/include/scsi/sg.h
  Otherwise no "make" problems. 

Trying to build t917 - need to get rid of histoscope references.  
Make a copy in ckm/daq/ckm to play with
made setup containing  export SJYLX_DIR="../sjyLX"
After some hacking histoscope.h, hist.c and friends are gone.
rundaq now builds and runs 

<ckm01> ./rundaq 
Usage: ./rundaq <run_number>

ckm10 appears to be the obvious candidate for the DAQ machine.  
 - It has no assigned users yet.  
 - Installed SCSI controller - no problem

ckm10:~# date; lspci -vvv -s 00:0a.0 
Thu Apr 18 17:05:16 CDT 2002
00:0a.0 SCSI storage controller: LSI Logic / Symbios Logic (formerly NCR) 53c860 (rev 13)
        Subsystem: LSI Logic / Symbios Logic (formerly NCR): Unknown device 1000
        Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B-
        Status: Cap+ 66Mhz- UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR-
        Latency: 32 (2000ns min, 16000ns max), cache line size 08
        Interrupt: pin A routed to IRQ 11
        Region 0: I/O ports at d000 [size=256]
        Region 1: Memory at db022000 (32-bit, non-prefetchable) [size=256]
        Capabilities: [40] Power Management version 1
                Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
                Status: D0 PME-Enable- DSel=0 DScale=0 PME-


 -  make DAQ account - usual access
 -  copy everything over to daq/daq, put daq_log-book in ~daq

Installed cern software in /fnal/usr/products
copied .bashrc and .bash_profile form my account on ckm01
forwent setup /ups/upd.  Added cern/bin to path in .bashrc 
pawX11 runs paw properly

Fri Apr 19 17:43:51 CDT 2002  - got rack with CAMAC/NIM from Bob Jones, 73a
----------------------------

Bob Jones stopped by this morning and agreed to let us have the 6ft rack from 
the 9th floor with a CAMAC and NIM crate already installed.  Brought it up to
Sasha's old office for now.

Rebuilt kernel for ckm10.  Now sees SCSI 

<ckm10> cat  /proc/scsi/scsi /proc/scsi/sym53c8xx/* 
Attached devices: none
General information:
  Chip sym53c860, device id 0x6, revision id 0x13
  On PCI bus 0, device 10, function 0, IRQ 11
  Synchronous period factor 12, max commands per lun 32


Stopped by Mike Behnke and PREP to pickup a 73a crate controler and RFD02
lam latch module. 

Spoke to Bob Forester.  We should probably move back into the space we had 
for the E949 RSTDC teststand. on FCC3

Hooked by 73a scsi to ckm10.  Still no devices seen.

Fri Apr 19 18:02:59 CDT 2002 - reboot system
----------------------------

Now we've got it!
<ckm10> cat  /proc/scsi/scsi                        
Attached devices: 
Host: scsi0 Channel: 00 Id: 01 Lun: 00
  Vendor: JORWAY   Model: 73A              Rev: 208 
  Type:   Unknown                          ANSI SCSI revision: 02


<ckm10> export PATH=$PATH":/home/daq/daq/sjyLX/example/"
<ckm10> cnaf 
Usage: cnaf <branch> <crate> <slot> <subad> <function> <data> <type: serial=0,parallel=1>
<ckm10> cnaf 1 1 8 0 0 0 2
CDCHN Error opening /dev/sga
CDCHN: No such file or directory
cdchn error: No such file or directory

Who is asking for /dev/sga?

<ckm10> grep /dev/sga daq/sjyLX/src/* 
daq/sjyLX/src/sjy_cdchn.c:static char *devtble[] =   {"/dev/sga",   /* scsi generic devices */

Try a link for now: - hot damn we have CAMAC!

<ckm10> cnaf 1 1 8 0 0 0 2
CMSG called from cnaf:, data=10188

Tue Apr 23 09:56:53 CDT 2002 - daq works!
----------------------------
Summary of past days:

Hacked daq code and moved RFD02 to N=2.  DAQ now runs
Hacked offline - now runs
Reorganized directory structure to keep Makefiles happy
Made logic diagram daq_trigger.xfig
Moved CKM10 into top of rack
Will download all documentation files to /home/daq/ckm/doc in case we fall 
off the net.
Made local webpage with all local files in /home/daq/ckm/doc.

Wed Apr 24 09:28:22 CDT 2002 - Webpage and documentation built.
----------------------------

Moved this log-book to webpage directory [/home/daq/ckm/doc]
I also have a copy at home on pcooper:/home/pcooper/ckm/doc/

Fri Jul 19 11:56:04 CDT 2002 - rebuild ckm10
----------------------------
ckm10's disk died about a month ago.  I rebuild the system from parts of the now defunct 
ckm01.  The system disk (/dev/hda) is the old ckm01 drive reloaded with ckm06 files.
Memory upgraded to 3/4Gb
Ordered a new 120Gb IBM disk for /dev/hdc as system and data backup
New monitor on system (Micron 17FGX) now works OK (with help from Jason)
Upgraded local webpage with Lecroy 4413, 4418 and 4508 modules
Time to design a CAMAC/ECL trigger system.
 
Thu Jul 25 16:23:05 CDT 2002 - Signal source definitions
----------------------------		

System			label		   signal	HV	signal cable
					type	cables	cables	type	length

Tagging Counters	T(n)  n=1,61	NIM	32	--	l-l	32ns
					PS754

Pair sepctrometer	P1{L,R}n n=1,8	NIM	16	--	l-l	16ns
			P2{L,R}n n=1,8	NIM	16	--	l-l	16ns
					Caen413A

Hole Vetos		V{L,R}n n=1,2	anode	4	4	N-l	60'
					dynode	4	

misc trigger cntr	R		anode	1	1	N-l	60'
					dynode	1

horiz cntrs		Hn  n=1,3	anode	3	3	N-l	60'
					dynode	3

finger cntrs		Fn  n=1,2	anode	2	2	N-l	60'
					dynode	2

PBG block		D0		anode	1	1	N-l	60'
					dynode	1

VVS prototype		Dn n=1,8	anode	8	8	N-l	60'
					dynode	8

Cable summary:

	60' 	SHV-SHV		19 + 11 spare = 30
	60' 	BNC-lemo	38 + 12 spare = 50
	16nsec	lemo-lemo	32 +  8 spare = 40
	32nsec	lemo-lemo	32 +  8 spare = 40

Fri Jul 26 13:49:43 CDT 2002 - PBG/cables from PC4
----------------------------
I went shopping in PC4 and found the following:

Cables
  54	95' 	SHV-SHV		Need 24
  10	60' 	BNC-BNC		Need BNC-lemo
  35	80nsec 	BNC-BNC		Need BNC-lemo  (80nsec = 52')
  40	16nsec	lemo-lemo
  40	32nsec	lemo-lemo

Bob Jones agrees to re-terminate the 60' and 80nsec to BNC-lemo if we bring 
them up to wh13sx.  I'll need some help to pull 24 95' HV cables and relabel 
and  box everything in PC4.

PBG
  A semi-infinite supply of 15x15x41 cm^3 PBG blocks with 5" tubes and bases.
I need to get 2 of them over to lab6 for testing.


Tue Jul 30 14:07:47 CDT 2002 - PREP Module Request
----------------------------

From pcooper@fnal.gov Tue Jul 30 14:06:10 2002
Date: Tue, 30 Jul 2002 14:01:16 -0500 (CDT)
From: Peter S Cooper <pcooper@fnal.gov>
To: Adam Walters <awalters@fnal.gov>
Cc: Peter Cooper <pcooper@fnal.gov>, Bob Tschirhart <tsch@fnal.gov>,
     Erik Ramberg <ramberg@fnal.gov>
Subject: PREP Electronics for CM teambeam run at Jefferson Lab.


Adam,

We (CKM) are preparing for a testbeam run at Jefferson lab in the period
Sept-Nov 2002.  We will need about a crate worth of NIM and CAMAC
electronics to instrument this test.  I already have a rack, crates and a
DAQ system.  Now its time to get the modules.  My shopping list at this
time includes: ([] are modules I already have. Most belong to EPP not
PREP.)

                                        required   spare   total
NIM Modules

  ECL-NIM_ECL  LRS 4616 or PS 726           5        1       6

  Discriminator            LRS 621         [3]               3
  Discriminator            LRS 623         [1]               1
  quad logic               LRS 364/365     [2]               2
  Gate Generator           LRS 222         [1]               1

  slots used                               12

CAMAC Modules

  Programmable logic unit   LRS 4508         1       1       2
  ECL Discriminator         LRS 4413         2       1       3
  Programmable delay        LRS 4418         2       1       3
  ADC                       LRS 2249a        2       1       3
  ECL TDC                   LRS 3377         4       1       5
  ECL Scalers               LRS 4434         4       1       5

  Quad Scaler               Jorway 84-1     [1]              1
  12 ch scaler              LRS 2551        [1]              1
  12 ch TDC                 LRS 2228        [2]              2
  DAQ LAM Latch             RFD 02          [1]      1       2
  SCSI Bus Crate Controller 73A             [1]      1       2

  slots used                                23

I've included one spare of each type.  We will be running parasiticly in
a radiation area.  We will have to wait to make access to repair or
replace anything until the primary beam users make access.  We will need
spares in our hands.

I'm setting up this system now on wh13e.  I'll move it to Lab 6 and then
ship it to Jefferson Lab around Sept 1.  It should all return around Nov
15.

Peter

Wed Jul 31 17:00:58 CDT 2002 - cables to Bob Jones
----------------------------

I madeup and checked 24 40'shv-shv cables in PC4 labeled A-X.  I brought
all cables from PC4 to wh13.  I left the long signal and HV cables with Bob 
Jones.  Then I realized that the ADC cables need extra delay for the trigger
latency.  How Much?

Bob T showed me a PMT base for the trigger counters.  The PMTs will be 
9954KB's.  The base as bridged BNC anode outputs and an SHV HV connector.
The label said the base drew 1ua at 2KV.


Fri Aug  2 10:22:31 CDT 2002 - got long adc delay cables
----------------------------

More PC4 shopping.  Got 24 130' RG58 BNC-Lemo cables (3 bundles of 8 each).
Brought them to Lab3.  Rang several 202 nsec delay.

Mon Aug  5 17:41:49 CDT 2002 - install PREP electronics
----------------------------

I picked-up the electronic from PREP.  I added a 2465 scope and a LRS 128
linear fanout to drive the led's.  Hogan and I moved the CAMAC crate to the 
top of the rack and I installed the modules as follows:

    slot    module  serial   #function
      1     4413    050154   trigger disc
      2     4413    541415   veto disc
      3     4418    036587   programmable delay
      4     4508    071308   trigger PLU
      5
      6     RFD02   515796   trigger lam
      7     2551    044261   scaller
      8     
      9
     10     3377    540621   TDC
     11     4434    044261   scaler
     12
     13     3377    540791   TDC
     14     4434    061704   scaler
     15
     16     3377    544766   TDC
     17     4434    042092   scaler
     18
     19     3377    540641   TDC
     20     4434    042093   scaler
     21
     22     2249a   019995   ADC
     23     2249a   023952   ADC
     24     73a     549360   crate controller
     25     73a



Powered back up.  Connected signals to N7 2551 scaler.  Readout works.

Wed Aug 28 22:22:30 CDT 2002 - review of layout and spares
----------------------------
CAMAC Crate
    slot    module  serial   #function			spare
      1     4413    050154   trigger disc		066675
      2     4413    541415   veto disc			  "
      3     4418    036587   programmable delay		36578
      4     4508    071308   trigger PLU		073653
      5
      6     RFD02   515796   trigger lam		016213
      7     2551    044261   scaler			049561
      8     
      9
     10     3377    540621   TDC			540618
     11     4434    044261   scaler			042085
     12
     13     3377    540791   TDC 			  "
     14     4434    061704   scaler			  "
     15
     16     3377    544766   TDC			  "
     17     4434    042092   scaler			  "
     18
     19     3377    540641   TDC			  "
     20     4434    042093   scaler			  "
     21
     22     2249a   019995   ADC			020000
     23     2249a   023952   ADC			   "
     24     73a     549360   crate controller		567546
     25     73a
NIM Crate
    slot    module  serial   #function			spare
      1     128L    014987   Linear Fanout		  ???
      2     621AL   015948   Quad Discriminator		015634
      3     365      18487   Dual Logic Unit		  ???
      4     4616     42184   ECL-NIM-ECL		  ???
      5
      6     4616     52621   ECL-NIM-ECL		   "
      7     4616     55255   ECL-NIM-ECL		   "
      8     4616     53414   ECL-NIM-ECL		   "
      9     4616     42183   ECL-NIM-ECL		   "
     10     222      24096   Gate generator		  ???
     11
     12     4616     62799   ECL-NIM-ECL		   "

Tomorrow will try RFD01 module as a 222 reset.  
I can return the 2323 I got today and several other modules.  
I need the following spares:

    slot    module  serial   #function			spare
      1     128L    014987   Linear Fanout		need 1
      3     365      18487   Dual Logic Unit		need 1
      4     4616     42184   ECL-NIM-ECL		need 1
      8     RFD01     ???    Output pulser		need 2
     10     222      24096   Gate generator		need 1

I have the following chages to the loan request:

    slot    module  serial   #function		
      6     RFD02   016213   trigger lam	add to list
      7     2551    049561   scaler		add to list
     24     73a     549360   crate controller	add to list
            415B     12891   Fluke HV Supply    add to list
            84-1    017712   Quad Scaler        add to list
            2228a   24590    8ch TDC            remove from list
            2228a   24590    8ch TDC            remove from list
            4418    062771   programmable delay remove from list

Thu Aug 29 15:55:14 CDT 2002 - after modlue exchange
----------------------------
CAMAC Crate
    slot    module  serial   #function			spare
      1     4413    050154   trigger disc		066675
      2     4413    541415   veto disc			  "
      3     4418    036587   programmable delay		36578
      4     4508    071308   trigger PLU		073653
      5
      6     RFD02   515796   trigger lam		016213
      7     2551    044261   scaler			049561
      8     
      9
     10     3377    540621   TDC			540618
     11     4434    044261   scaler			042085
     12
     13     3377    540791   TDC 			  "
     14     4434    061704   scaler			  "
     15
     16     3377    544766   TDC			  "
     17     4434    042092   scaler			  "
     18
     19     3377    540641   TDC			  "
     20     4434    042093   scaler			  "
     21     RFD-01  535823   NIM/ECL Pulser             561820
     22     2249a   019995   ADC			020000
     23     2249a   023952   ADC			   "
     24     73a     549360   crate controller		567546
     25     73a
NIM Crate
    slot    module  serial   #function			spare
      1     128L    014987   Linear Fanout		015618
      2     621AL   015948   Quad Discriminator		015634
      3     365      18487   Dual Logic Unit		005274
      4     4616     42184   ECL-NIM-ECL		042189
      5
      6     4616     52621   ECL-NIM-ECL		   "
      7     4616     55255   ECL-NIM-ECL		   "
      8     4616     53414   ECL-NIM-ECL		   "
      9     4616     42183   ECL-NIM-ECL		   "
     10     222      24096   Gate generator		052571
     11
     12     4616     62799   ECL-NIM-ECL		   "

Mon Sep  2 11:07:39 CDT 2002 - update loan agreement
____________________________
To      : Bill Finstrom <finstrom@fnal.gov>
Cc      : offsite_loans@fnal.gov, 
          Mike Behnke <behnke@fnal.gov>, 
          Peter Cooper <pcooper@fnal.gov>
Attchmnt: 
Subject : Update to Jlab loan.
----- Message Text -----

Bill,

We're shipping our gear to Jlab this week.  I've made some changes to the  
equipment list and have some additions for modules I already had.  Would  
you please update the loan agreement so we can keep up with what we are
actually taking?

Thanks,
Peter

Thu Aug 29 15:55:14 CDT 2002 - after module exchange
----------------------------
I have the following changes to the loan request: (N:nim C:camac S:spare)
NIM
    slot    module  serial   #function
     N1S    128L    015618   Linear Fanout      add to list
     N3S    365     005274   Dual Logic Unit    add to list
     N4S    4616    042189   ECL-NIM-ECL        add to list
     N10S   222     052571   Gate generator     add to list
CAMAC
    slot    module  serial   #function
     C6     RFD02   016213   trigger lam        add to list
     C8     RFD01   535823   Output pulser      add to list
     C8S    RFD01   561820   Output pulser      add to list
     C7     2551    049561   scaler             add to list
     C24    73a     549360   crate controller   add to list
            84-1    017712   Quad Scaler        add to list
            2228a   24590    8ch TDC            remove from list
            2228a   24590    8ch TDC            remove from list
            4418    062771   programmable delay remove from list
OTHER
            415B     12891   Fluke HV Supply    add to list

------------------------------------------------------------
        Peter S. Cooper       Email:      pcooper@fnal.gov  
        WH13SE                Phone:      +1-630-840-2629   
        Fermilab  MS 122      Fax:        +1-630-840-3867

Thu Sep 12 10:14:55 CDT 2002 - Detector alignment details
----------------------------

In the hall with Doug we agreed to the final alignment for the detector
stand:

  - Our beamline will be at 9.14 degrees to the photon beam.
  - The lead wall hole will be 16 inches along the beam with wall projectively
    parallel to the beam, 4 inches wide and 2 inches high.
  - The center of the hole in the lead wall is at 87.386 inches from the bend
    point of the pair spectrometer magnet - along th 9.14 degree line.
  - We want a magnification of 1.5 at the H1-3 scintillators ==> 131.1 inches
    The H1-3 is 4.5 inches from the front face of the stand ==> 131.1-4.5=126.6
    from magnet center to frame face.  We probably cant get this close by a
    couple inches.
  - The rear lead wall will be built around the PBG PMT base with vertical 
    limit switches set to protect the base.  The base bottom clear the stand
    by only ~1/2 inch.


Fri Sep 13 10:43:37 CDT 2002 - psc/BobT  First data
----------------------------

run-00015  2000 pulser events
run-00016  2000 pulser events
run-00017  2000 pedestal events H1 H2 and F1 HV off
run-00018  2000 cosmic events
run-00022 20000 cosmic events - beam trigger


Fri Sep 20 15:00:00 CDT 2002 - psc/Erik/BobT
----------------------------

Software training.  Erik and Bob succeded to logon, take data, modify
the analysis and analyse those data on ckm-1.  README was expanded by psc.



Sun Sep 22 11:31:53 CDT 2002 - psc last upgrade before bugout to France
----------------------------

I fixed the scaler system so that display_scalers will run for any
directory.  Multiple display_scalers running at the same time are OK.
Had to modify scaler_read to get spill_scaler written in a fixed place 
[$CKM_SCALER] and modified display_scaler to plant a symbolic link
[spill_file] to spill_scaler in whatever directory it is run from.  The 
data flow now is:

                                                 <=Linked= display_scaler p1
[scaler_read] =writes=> $CKM_SCALER/spill_scaler <=Linked= display_scaler p2
                                                 <=Linked= display_scaler p3


I remade the scaler.ntup, writing $CKM_SCALER/example.kumac in the process,
to see if there has been any beam.  Not yet.


Mon Sep 23  EJR + BT
---------------

Beam is on our detector!  Hits in CLAS and the pair spectrometer are evident
on the CLAS screen.  Current in the magnet is 787 A, which corresponds
approximately to 5500 Gauss, which is approximately .8 GeV - not bad for our
purposes!

I loaded the default trigger and took some data to establish voltage levels.
I used the ADC output to set voltage levels.  Veto counters should be well
separated from pedestal, while we wish to see double tracks in the trigger
counters, so they should be slightly lower in HV.  Data up to run 32
establish the following settings for voltages:

V1 2200
V2 1950
F1 2025
F2 2125
H1 1950
H2 2050
H3 2075

We notice that the lead glass is way off scale.  We also notice that the
second set of pair spectrometer hodoscopes do not seem to be evident in the
TDC's.  We contact Arne to lower voltage in lead glass and check on pair 
spectrometer cables.  Here is his answer:

******************************
gang,
at 5:30pm est....
lowered the PbGl HV down to 1800V.

found a group of 16 unplugged inputs to a TDC and plugged them in....

Good luck,
Arne
********************

Heck of a guy!

Took another run (#33) of almost 10000 events (takes about a minute!).
We now see the downstream pair spectrometer in the TDC distribution.
Lead glass ADC's are still off scale.  I will ask Arne to lower it another
200 V.


24 September: 3:50 p.m.   EJR
-----------------------------
Beam has been off all day, as they indicated it would on Tuesdays.  Arne 
made an access today and lowered the lead glass another 200 V. to 1600 V.
I made changes to ckm_anal so that it writes a lot of information to an 
ntuple.  For some reason I run into memory problems I can't fix after
30000 events or so.  I'll keep trying.  Also, I wrote a kumac called 
'plot_all.kumac' in the ckm/offline directory.  It will plot a lot of 
relevant quantities, including the lead glass plot with all significant
cuts except for T counter tagging.


      8:10 p.m.   EJR
Lead glass ADC looks good now.  I just took almost 500,000 events in run
37 (ended with a segmentation fault).  I also made a kumac called
'cuts.kumac', which has cuts to use to select for good events.  I try to
analyze run 37 but get an end-of-file after 175000 events.  I don't
understand this, but will ignore it for the moment.  Using all the clean up
cuts, I see 6 events at pedestal for the lead glass out of a total of about
600 events.  Damn!  Can't get better than 10**-2 at this early stage. 


28 September  EJR
-----------------
00:30
We changed the trigger to (F1 .and. F2) .and. H1 .and. .not.H2.and..not.H3 
to reduce dead time.  Live time is still only about 6%, with about 2-300 Hz
data taking rate.  Run 38 was taken with this trigger - about 200,000
events before it crashed.  I analyzed this run, this time histogramming
the TDC value of the discriminated output from the lead glass.  It turns out
that for the 1% of events that pass all cuts which have no energy in the
lead glass, we still get an in-time TDC hit.  In fact, we only get no 
TDC hits about 10**-3 of the time, but there is energy in the lead glass
for those events.  I don't understand this.

15:00
Had problem taking data.  Got CDCHN error.  Peter rebooted computer and
indicated that documentation gives instructions for cold start.

Took 235000 events in run 39.  The rate was very low.  I can't read
scalers for some reason at the moment so I don't know what beam conditions
are.  Analysis of this is quite interesting.  If I analyze the whole
file, just histogramming the lead glass ADC, I get an enormous fraction
(33% !!!) that pass all of the cuts and have 0 energy in the lead glass
but still have a hit in the TDC.  I then analyzed the first 90,000
events separately and I find that there are NO such events!!  This 
seems very, very fishy.  I'm suspecting the ADC at the moment, but it
is hard to tell from this distance.  Clearly there is some non-constant
problem in the DAQ.

 
29-September    EJR
-------------------

00:00 

Took 6 runs of 30,000 events each:  runs 42-47.  When I look at the final
output, I see that two of the runs (45 & 47) have significant entries at
pedestal, but the other four runs have nothing there!  Not good.  Informed
Bob and Peter with an email.


30-September   EJR
------------------


11:30

For some reason the trigger rate seems low, even though the Master-OR at
Hall B is now at 50 MHZ.  The magnet current is the same, so I suspect they
have put in a thinner foil.  I'll send a message to Arne and ask.  

I changed the trigger configuration so that now it vetoes on the veto
counters.  Rate is an easily managed 150-200 Hz.  I took one run of 100,000
events - run 50.  The analysis shows the same problem - First 30,000 events
look great, then there is a bunch of pedestals.  I'm stumped for now.


15:25

Email from Arne:

Date: Mon, 30 Sep 2002 16:25:08 -0400
From: Arne Freyberger <freyberg@jlab.org>
To: ramberg <ramberg@fnal.gov>
Cc: Robert_Tschirhart <tsch@fnal.gov>, Peter Cooper <pcooper@fnal.gov>
Subject: Re: foils

ckm gang,

g7a agreed to the 10^-3 radiator.  pspec rate is now ~2.5kHz....so you
might get  a bit more rate...

Arne

> Arne - Thanks for the explanation!  That makes a lot of sense now.
> Erik
> 
> On Mon, 30 Sep 2002, Arne Freyberger wrote:
> 
> > Erik,
> > 
> > on friday, we figured out that  the photon beam was incident on the
> > frame of the converter.  We then put the 10^-4 converter in which
> > reduced the rate from 32kHz to 67Hz  (at 15nA of beam current.).  I've
> > requested the g7a gang to switch to the 10^-3 converter as the 10^-4 
> > doesn't  provide enough rate for luminosity monitoring....I will make
> > the request again....This would bring the total rate up to about
> > 2.5kHz(at 50nA), still a factor of
> > 10 below the frame....
> > 
> > Arne


16:00

Bob and I looked at runs 38 through 47.  There were a total of about 24,000
good events in those runs.  There were many of these events that had 0 energy 
in the ADC, but only 3 events in this set of runs had no TDC latch.  Two of
those had 800 counts in the ADC and the other one had 360 counts.  We
printed out the event numbers for those good events that had no energy in
the ADC.  Once this pathology starts in the DAQ, then these events occur
basically randomly throughout much of the rest of the run.  We are still
having DAQ crashes that prevent us from taking more than about 150000
events.  There is no beam now, but we are going to try and put in a wait
loop before reading out the ADC's, to see if that helps the situation. 


8 October   EJR
---------------

9:15

Bob and I looked at run 45 yesterday and checked to see whether the F2
counter had the same kind of problem as the lead glass.  We analyzed events
where there was energy restriction on the lead glass ADC and then looked at
the F2 ADC.  Despite the fact that this was a problem run for the lead glass
ADC, we did not see any of the problem pedestal events in the F2 channel. 
This prompted me to ask Arne to move the lead glass cable up one channel in
the ADC.  He did this last night around 5 p.m.  

I just ran 100,000 events with our more restrictive trigger in run 54.  I 
changed the analysis code so that it looks at data(23) for the lead glass
instead of data(24).  Same damn problem!  Lots of pedestals.

Wed Oct  9 13:13:09 CDT 2002  psc  changes to ckm_daq for ADC
----------------------------
Yesterday I modified ckm_daq to try help help the apparent ADC problems:
 1. removed EJR's delay loop since it didn't make any change
 2. zeroed the 12 word dataarray belfore each ADC DMA.  If any 0's are found
    in the data the DMA terminated early and the module isn't cleared by
    F2A11
 3. issued a separate F9 clear to each ADC after the F2 DMA readout to be
    sure the modules are cleared each event. 

I reloaded the default trigger and took 30K events

Run    65     30000 events - written to /data/jlab-data/run-000065

I analysed these data with the std ckm_anal.  The problem persists 101/1602
events have a tdc and no ADC energy in the PBG.  What now?


19:50  EJR

I now have evidence that the trigger rate (before busy veto) influences the
incidence of our ADC pedestal problem.  Run 71 has 50,000 events taken
at low trigger rates (~60 Hz, 75% live).  Run 72 has 100,000 events, with
the trigger rate changing in the middle to (~600 Hz, 35% live).  Sure
enough, Run 71 has no problem events and the problem events in Run 72 start
after 40,000 events.


20:30  EJR

More strong evidence for a trigger rate problem.  I changed the trigger
to the original loose trigger (H1.or.H2.or.H3) .and. (F1.or.F2).  This gives 
about 4 KHz rate and 6% livetime.  I took 2 runs of only 5000 events each
(runs 73 and 74).  Both of these runs have the problem from the very
beginning.  Contrast this with 90,000 consecutive events with no problem at
low trigger rate in runs 71 and 72. 


23:45  EJR

I tried run 75 with triggering on H2 instead of H1.  This lowered the
trigger rate down to about 120 Hz from 600 Hz.  When I analyzed the data,
however, it looked terrible, with the MAJORITY of events appearing in the
bad region of the TDC vs ADC plot.  I rebuilt the trigger and the analysis
code back to its original state.



Oct 11   9:40  EJR
------------------

We've been having problems with the daq after Peter rebuilt the code
yesterday, trying to solve the ADC pedestal problem.  I did a cold start 
according to the web page, including  restarting scaler_read.  This made 
it so that I can at least read scalers again.  No beam, so I can't test 
ckm_daq.


Fri Oct 11 22:31:08 CDT 2002 - high stats runs
----------------------------

Trigger - h1*f1*f2*not(h2+h3)
Run    95    204012 events - read from /data/jlab-data/run-000095
Run    99   1000000 events - read from /data/jlab-data/run-000099

Trigger - h1*f1*f2
Run   100   2000000 events - read from /data/jlab-data/run-000100
Run   103   3995000 events - read from /data/jlab-data/run-000103

ckm_anal -r 103 -h 103.hbk  -n 3995000 >103.log  | end of file is bad

Sat Oct 12 09:27:54 CDT 2002 psc - Initial analysis summary
----------------------------
I've analyzed the 6M events from runs 100 and 103.  103 ran for 3.995M
(~6.5 hours) of 5M requested before the network connection dropped in the
middle of the night.  The DAQ is clearly fixed. The data sample vs cuts
goes as:

    Events   Cuts
    5995000  none
     316778  pairs  - Good hits in the pair spectrometer
     299960  vetos  - + No hits in the veto counters
     296763  geant  - + Good hits in the trigger counters
     175232  all    - + no h2-3, high cuts in triggers cntrs

Unfolding the photo-statistics in the PBG by comparing the Gaussian width
of the peak from Vladimir's GEANT simulation at 1000 MeV and our observed
width:
                 mean   sigma  frac
                 [MeV]  [MeV]   [%]
     GEANT        937.   22.0  2.35
       ==>        800.   18.8  2.35

     deconv       800.   31.1  3.91

     1/0.0391^2 = 654 [pe] / 800 [MeV] = 0.82 pe/MeV

In the VVS prototype we expect 10[pe/vis-MeV] * 0.30[vis-MeV/MeV] =
3[pe/MeV]. This is 4 times the light yield of the PBG Cherenkov counter.

Using the ADC the inefficiency is zero for both the geant and all cuts
samples.  The lowest energy events are at:

     sample  events  Emin    Ineff
                     [MeV]
     geant   296763   20.    <3.5e-6 (E>20 MeV)
     all     175232   80.    <5.5e-6 (E>80 MeV)

Vladimir's Geant simulation predicts a peaking of the inefficiency below
20% of the incident energy (160 MeV in our case).  Our data do not
demonstrate this behavior.  Otherwise there appears to be nice agreement
between the simulation and the data.  Does this mean that we are not
sensitive to the hodoscope counts in our beam?

Looking at the TDC data from the PBG there are 211 / 175232 events without
a valid TDC hit.  62 of these are below the discriminator threshold of 410
ADC counts or 215 MeV.  There are an addition 149 events where the TDC
should have fired and did not.  This is an inefficiency of 149/175232 =
8.5e-4. Further study is require to understand if this is a discriminator
and/or a TDC problem.  The virtue of having two independent measurements
to beat against each other is demonstrated again!


Sun Oct 13 12:02:53 CDT 2002 psc - tdc-unpack bug is source of missing TDC evts
----------------------------

I've tracked the events where addition 149 events where the TDC should have 
fired and did not to a bug in the tdc_unpack routine.  Correcting this and
rerunning the same anaylsis yields:

    Events   Cuts
    5995000  none
     316778  pairs  - Good hits in the pair spectrometer
     299960  vetos  - + No hits in the veto counters
     296763  geant  - + Good hits in the trigger counters   
     175232  all    - + no h2-3, high cuts in triggers cntrs

Wed Oct 16 14:46:22 CDT 2002 Bobt - VVS pmt HV's
----------------------------
Date: Tue, 15 Oct 2002 11:46:11 -0500 (CDT)
From: R Tschirhart <tsch@fnal.gov>
To: ramberg <ramberg@fnal.gov>
Subject: Lab-6 VVS HV operating parameters.

Guys,
    VVS operating voltages to start with:

    Channel         HV
   --------      -------
      1          -1260V
      2          -1250V
      3          -1270V
      4          -1300V
      5          -1250V
      6          -1290V
      7          -1230V
      8          -1280V

The aggregate current draw for these 8 channels from the Fluke was 8 mA,
so I think we'll be OK with LRS HV.  Channel numbers are marked on the outer
vessel wall.  I would like to be present when the dynode cables are plugged 
into Sten and Wu's modules and the prototype.   There were a few non-obvious 
pin-1 issues I encountered when I plugged it all together at Lab-6.  Regarding the two BNC
connectors on the back of the PMT base:  These are NOT bridged outputs, one 
is the anode and the other is a QIE ground reference.  Good luck, see you 
Wednesday afternoon.

Thu Oct 17 11:41:33 CDT 2002  psc,er,bobt - prototype installed
----------------------------
Prototype installed and checked out.  Pulser LEDs drive only D7 and D8.
X motion stage checked out and working.  Commands like:

 echo "XOFF" >/dev/ttyS0  | turn motor off
 echo "XON" >/dev/ttyS0   | turn motor on (e.g. after hitting limit switch)
 echo "X0" >/dev/ttyS0    | set to 0 inches
 echo "X2" >/dev/ttyS0    | set to +2 inches
 echo "X-0.5" >/dev/ttyS0 | set to -0.5 inches
 echo "X-1" >/dev/ttyS0   | set to -1 inches   === Left it parked here

Thu Oct 17 13:33:49 CDT 2002 psc - register bit to ADC channels
----------------------------
patched N21b2 to 2249 N22A11
patched N21b3 to 2249 N23A11

Thu Oct 17 13:43:32 CDT 2002 BobT
----------------------------
Set and locked discriminator thresholds to -1.30V, corresponds to 4.0 mV 
threshold, 1 MeV charge.  Wheel setting is 2.25.

F-counter setting of -1.0 corresponds to the 1-inch wide F1 counter spanning 
1.0 to 2.0 inches INBOARD of the VVS inner corner.  

Negative direction =  INBOARD motion with respect to the VVS inner corner.  
Postive direction  = OUTBOARD motion with respect to the VVS inner corner.  

The VVS inner corner is vertically aligned in the second of five H-counter
bins, where bin #1 is the highest, and bin #5 is the lowest.  

Motor is left activated.  

Sun Oct 27 09:45:38 CST 2002 psc - 1st try at X dependence
----------------------------
The beam reappeared last night so I tried 4 1M events runs at different
X positions.  Something failed after 1/4M of the firsst run.  The script
which tried this is ckm/offline/bin/xs.  I'll hack it and try again.
As Erik reported magnetic field has changed.   The energy has gone up from
840 MeV to 1165 MeV.

 echo "X0.5" >/dev/ttyS0
 ckm_daq  -r 220 -f         -d 0 -n 1000000 
 ckm_anal -r 220 -h 220.hbk -d 0  >220.log
 =>Run   220    229425 events - read from /data/jlab-data/run-000220

 echo "X-0.5" >/dev/ttyS0
 ckm_daq  -r 221 -f         -d 0 -n 1000000 
 ckm_anal -r 221 -h 221.hbk -d 0  >221.log
 =>Run   221     14834 events - read from /data/jlab-data/run-000221

 echo "X1" >/dev/ttyS0
 ckm_daq  -r 222 -f         -d 0 -n 1000000 
 ckm_anal -r 222 -h 222.hbk -d 0  >222.log
 =>Run   222     10321 events - read from /data/jlab-data/run-000222

 echo "X-1" >/dev/ttyS0
 ckm_daq  -r 223 -f         -d 0 -n 1000000 
 ckm_anal -r 223 -h 223.hbk -d 0  >223.log
 =>Run   223     59306 events - read from /data/jlab-data/run-000223

Sun Oct 27 22:37:08 CST 2002 psc - 1st try at X dependence
----------------------------
 echo "X0.5" >/dev/ttyS0
 =>Run   224   1000000 events - written to /data/jlab-data/run-000224

 echo "X-0.5" >/dev/ttyS0
 =>Run   225   1000000 events - written to /data/jlab-data/run-000225

 echo "X1" >/dev/ttyS0
 =>Run   226    349928 events - read from /data/jlab-data/run-000226

 echo "X-1" >/dev/ttyS0
 =>Run   227   1000000 events - written to /data/jlab-data/run-000227

 echo "X0" >/dev/ttyS0
 =>Run   228   1000000 events - written to /data/jlab-data/run-000228

Sat Nov  2 21:32:13 CST 2002 psc - more data
----------------------------
 echo "X-1" >/dev/ttyS0
 trigger_load h1f1
 ckm_daq  -r 229 -f  -d 0 -n 4000000

 stopped after 761622 events

Sun Nov  3 22:21:16 CST 2002 psc
----------------------------
 ckm_daq  -r 230 -f  -d 0 -n 4000000 # completed successfully
 modified ckm_anal.F to add -s option - summarize -no analysis
 wrote ckm/offline/bin/summary script to make run_summary.log

Fri Nov  8 23:03:05 CST 2002 psc - more data
----------------------------
 echo "X-1" >/dev/ttyS0
 ckm_daq  -r 231 -f  -d 0 -n 4000000

Wed Nov 13 00:00:30 CST 2002 psc - Got pair magnet set to 500amps
----------------------------
 ckm_daq  -r 303 -f  -d 0 -n 10000   | quick confirmation <E>=538 RMS 42.8 
 ckm_daq  -r 304 -f  -d 0 -n 4000000 | seg fault
 cz 1 1 1
 trigger_load h1f1                   | reload trigger
 echo "X-1" >/dev/ttyS0              | std position
 ckm_daq  -r 305 -f  -d 0 -n 4000000 | try again

Wed Nov 13 06:35:12 CST 2002 psc - wrapup
----------------------------
Modifed ckm-daq to skip tagger readout on -T switch
Took about 2M events at -500 a (runs 310 and 311 1M and nearly so)

<ckm-1> ckm_daq  -r 310 -T -f  -d 0 -n 1000000
Run   310   1000000 events - written to /data/jlab-data/run-000310
STOP ckm_daq - done statement executed

<ckm-1> ckm_daq  -r 311 -T -f  -d 0 -n 4000000
Segmentation fault00 polls 90000

Copied data , scalers and log back to ckm00

DON"T FORGET TO SHUTDOWN THE MACHINE!

<ckm-1> [logon as root]
<ckm-1> shutdown -h now

Mon Nov 18 11:29:53 CST 2002 psc - as found HV settings
----------------------------

We just logged back into clasrun@clon01 and transcribed the HV setting for 
our counters.  Relative to the settings from above in this log-book they are:

	Sep 23	Oct 16	Nov 18
 V1	-2200		-2200
 V2	-1950		-1950
 F1	-2025		-2025
 F2	-2125		-2125
 H1	-1950		-1950
 H2	-2050		-2050		
 H3	-2075		-2075
 D1		-1260	-1260
 D2		-1250	-1250
 D3		-1270	-1270
 D4		-1300	-1300
 D5		-1250	-1250
 D6		-1290	-1290
 D7		-1230	-1230
 D8		-1280	-1280

There are no HV exceptions - they ran as we initially set them.