The boards in the tower FEE crates are configured using another GUI opened from the main GUI. This GUI communicates with the radstone board in crate 90 via HDLC protocol. This allows the setting of things like latency, fine delays, LED and laser firing and much more. To open the GUI simply click on the orange "Detail" box in the "Tower Fee" region of the main gui. A rectangular GUI with a box for each crate is displayed.
There are now standard files for running these FEE crates and the thing most often done is to reconfigure the crates after problems or startup from these files. The box in the lower right hand corner links to the files as "current_beam_config". These point to links on eemc-sc in ~/mini/tcl/tower-#-current_beam_config.dat (where # is 1,2...6). These link to files in the directory by date with typically the latest in use (currently 01.28.2005). Right now there are only these standard beam configuration files. In the future we might produce other optimized for various led and laser runs. NOTE clicking save all will try to overwrite these files unless you change the name in the box. They should be write protected on the disk but please try to avoid testing this. Also if you make a new set, write protect them! (Note the files are simple ASCII files and can be looked at or edited if changes are required.
To configure all boxes click on "Load", "Write" and then "Read". Wait for each to finish before going on to the next. Text will appear just above the buttons so you can watch the progress. If there are errors you will have to enter the crate specific GUI by clicking on the box with the appropriate crate number, e.g. "Crate #1". You can try reconfiguring a second time with "Load", "Write" and then "Read" at the bottom of the page. This does the same as above but only this box. If this doesn't work then you will have to reset the box. Click on the purple "Resets" box to get another GUI. Try clicking on "Reset" first. This will reboot the box. Then try reconfiguring again. If this still doesn't work try the "hard reset". This reloads the firmware from ROM and will overwrite any special LUTs that have been loaded. Try reconfiguring. If this works you will have to reload the LUTs as well. If it doesn't work the only alternative is to power cycle the crate, reconfigure and load LUTs. (Note early in the 2005 run we are not using any special LUTs so they don't have to loaded until they are set closer to pp running.) Remember the power supplies go to pairs of crates so you will have to configure the other as well.
This reconfiguration process sets up critical timing parameters wrt STAR clocks and event timing.
Table of tower FEE parameters:
RHIC clock delay (ns Adjusts ADC timing for crate wrt RHIC cloc
LED latency Depth in digital delay to find LED data
LED amplitude Adjust LED amplitude (typically 2000)
LED Mask 1 (Green) Picks which towers have their LED enabled
LED Mask 2 (Blue) Picks pattern for firing SMD LEDs
LED Pulser delay (ns) LED delay wrt RHIC clock (fine adjust)
Cal Pulse Latency to set up electronics pulser
FIFO latency (all Boards) Depth in digital delay to find beam data (=19)
For LEDs the settings depend on the crate you want to read out and whether the LEDs are triggered in a different crate from the one the data is read out in. The green mask controls the LEDs for the the towers. The blue mask controls the pulsing of the SMDs.
Table of tower LED masks:
LED mask
subsector A 2 (Note: that these are additive. You can pulse morethan
subsector B 4 one at a time. However we do not think
subsector C 8 it is a good idea to fire all of the channels at once
subsector D 16 by setting the mask to 0x3F for example.)
subsector E 32
Taking an LED run requires only that you set up the masks as desired, the defaults fire all channels, and then using the eemc_led configuration to start a run on the run control computer.
The other information read in in the configuration files is the masking that assigns the towers to the various trigger patches. These appear in the "spreadsheet" in the middle of the page. The mapping is documented in a master spread sheet at
http://www.iucf.indiana.edu/U/STAR/eemc_proj/poletip_wall/Poletip_cabling/ tower_feetrig_cabling.xls
and graphically in
http://www.iucf.indiana.edu/U/STAR/eemc_proj/electr_trig_daq/ eemc_trigger_patches_annotated.pdf
but any changes should be made in consultation with an expert (e.g. Steve or Renee). This would happen if a tube goes hot and needs to be cut off from trigger.
Note the spread sheet also documents what towers are in what crates,
board and channel. A summary is
Table of tower subsector vs. FEE crate:
Crate 1 11TD-1TC
Crate 2 1TD-3TC
Crate 3 3TD-5TC
Crate 4 5TD-7TC
Crate 5 7TD-9TC
Crate 6 9TD-11TC
Tower bases are assigned to HVsys branches in a similar pattern. The splitting of sectors was required to make trigger patches match TPC sectors.
The pedestals are now loaded automatically with the configuration. The are stored in /home/online/mini/tcl/ped/ and reference by the link current which points to the latest set of files labeled by date (currently 01.27.2005). There the are separate files for each board in each crate. These are determined and loaded by an expert On the crate specific GUI you can inspect and change the peds buy clicking on the "ped" box near the bottom of the GUI. There is one for each of the 32 channels on each of the 4 boards in the crate. You will not normally change these. If you do and you decide to change make sure you do not overwrite the "current" files. Change the name or path. They "current" files should be write protected but don't tempt fate.
Along the right hand side of the window are buttons to set the crates to uninstalled or installed. If a crate were to die we would set it to uninstalled until fixed at an access. If the crates are marked installed they are regularly scanned to check voltages and temperatures. This does not detect problems with other crate configuration parameters unless the crate gets so lost it no longer communicates temps and voltages. At this time this scanning can interfere with communication to the MAMPT boxes. When writing and reading the MAPMT boxes all crates should be TEMPORARILY uninstalled on this page.