CFLT Calibration Instructions

Pre-Calibration

There cannot be any beam in the machine during a calibration
You need control of the CAL and CFLT during a calibration
Total time of calibration is about 40 minutes. You will need the calorimeter readout for about 20 minutes and the CFLT for the same 20 minutes.
Call the control room and tell them you want them to take the CFLT offline and use the CAL
It's a good idea to turn off noisy channels before taking a calibration.
Important calibration output files:

delays_out.dat

lowgain delay values which will be downloaded into the hardware

for050.dat

a list of channels which have bad timing (something wrong with pulse shape)

cflt_constants.dat

Contains the 'calibration constants'
This file is in a format which is nice for downloding into the hardware.
lowgain overflow value for each of the eight channels of every TEC = (8 channels) x (14 TECs) x (16 crates) = 1792 numbers
highgain overflow value for each of the eight channels of every TEC = (8 channels) x (14 TECs) x (16 crates) = 1792 numbers
pedestal value for each TEC = (14 TECs) x (16 crates) = 224 numbers
Total numbers in cflt_constants.dat: 1792+1792+224 = 3808

calib.dat
- Contains highgain and lowgain overflow value for each channel plus pedestal plus other information.
- The file has one row of numbers for each channel and is easily turned into an ntuple.
- Use this file to analyze the results of a calibration.

Calibration

Login into necessary computers (open 3 shells)

shell_1$> telnet cflt02
username: cflt
password: *****

shell_2$> ssh cal_rc@teucer

shell_3$> ssh -X cal_user@teucer

Take CFLT Offline

CFLT_CFLT02> offline

Kill the CAL readout

cal_rc@teucer> kill_readout

Lowgain values

CFLT_CFLT02> make_lowgain_delays

The following files are created:

delays.dat (this is the lowgain input to cflt_analog.btl)
time.chan
delays_out.dat (this is the file actually downloaded in the trigger)
timing_check.dat (a list of bad (lowgain) timing channels)
timing.paw

Check that lowgain delay values are good. (average times are reasonable and not too many bad timing channels.) If the timing numbers are OK, then type
CFLT_CFLT02> install_time

If you are curious to learn more about the bad timing channels (not essential)
CFLT_CFLT02> badtiming
This will generate a PS-file with a profile of each bad timing channel. You need to be in the lowgain delays work directory to run this command, because it looks for the file check.lis which contains all the bad channels.

Start Readout in DAQ Mode

cal_user@teucer> start_readout DAQ

Start Cal Readout Control Mangager (CALRC)

cal_user@teucer> setenv DISPLAY your_computer:0.0 (ignore this if you used ssh -X)
cal_user@teucer> start_RC

Click on IDLE in CALRC to get it started
Start a run in DAQ mode using the configuration file trigger.dat:
If you are asked to restart the readout in DAQ mode,
cal_user@teucer> restart_readout DAQ
The CNF file for this configuration should be tony_save.acf.
Click on SETUP in CALRC
Wait until CAL status is setup and the READY button turns blue
Follow the arrow in CALRC to click on ACTIVATE.

Now DAQ run starts and triggers are firing.

Note: During this process, look carefully in the bottom LOGVIEW window, sometimes it shows error messages like transputer ping timeout, just be patient, usually it will be fixed by itself. If you can't get the DAQ run started, restart CALRC and the readout, if still no good, contact CAL people.

Highgain Values

While the triggers are firing
CFLT_CFLT02>make_highgain_delays

The following files are created:

all_6gev.time
delays_out.dat
for050.dat --> mean = about 27,27,24 ns

Somebody's questions: [number of bad channels = about 280-330? doesn't mean much. are high gain delay values good? How do you tell?]

Calibration constants

While the triggers are firing
CFLT_CFLT02> make_constants

Should take about 10 minutes for make_constants to finish

End Run

Click End on the CALRC.

End Readout in DAQ Mode

cal_user@teucer> kill_readout

Restart Readout in ZEUS Mode

If you are giving the runcontrol back to the control room
cal_rc@teucer> restart_readout ZEUS

If you give it back to the DQM people, it doesn't matter how you return it.

Install new constants:

CFLT_CFLT02> install_constants

Put the CFLT online

CFLT_CFLT02> online

Check that CFLT is actually online
CFLT_CFLT02> wholink

Tell control room CFLT is back online.

Post-Calibration

Check the Calibration Constants

Note: Lines do not show up on plots, must be fixed

In the calibration work directory on cflt02 ( set def CALIBRATION_ROOT:[MAKECONST.WORK] )
The most recently created calibration is loaded automatically
CFLT_CFLT02> paw

On paw panel:

Note: RC (foudas) numbering is used 0-15,0-13
Click on high gain overflows.
Most of the points should be between the lines.
Click on low gain overflows.
Most of the points should be between the lines.
Click on very bad channels.
You will see a list of channels for which either a high gain or a low gain calibration constant could not be found for some reason. The value of the constant is reported as -9.0000. A -9.0000 gets set to zero in the file which is actually downloaded into the hardware.
Click on quit
An example of the plots with notation can be found here

ToDo List

Show differences between two successive calibrations.
Check if badchan low and high have been made already.
If you are looking at old cal constants what happens when you try to make a bad channel list?
Maybe put time stamp on serial_shuts.dat file so it can be used when a user is looking at old calibration constants????
Create paw panel to look at delay values and differences in delay values. shouldn't be hard to do and might be useful. I think Brian did this. We just need to look in his area...

Don't know when this list was created

Link to old instructions

Updates:

P. Ryan 2004-12-03
CFLT 2000-01-16
A.S. 07-10-2003
L.Li 05-06-2004