Two Crate Data Sharing Setup Two crates fully

Two Crate Data Sharing Setup • Two crates fully populated • Call Crate 0 and Crate 1 • Seven 68 -pin data sharing cables • To connect Crate 0 RC’s to Crate 1 RC’s • Currently not a “barrel” but still works - needs to be fixed • See next slide for connections • Four 50 -pin data sharing cables • Only RC 0 and RC 1 need them normally • Configuration not proper either • See next slide for connections • One MCC Crate with CIC timing set up • Cables from CIC to two calibrated CCCs • Two JCCs with all 6 data cables attached to JSC, both clock and control cables attached to CCC P. Klabbers, Two. Crate. Tests, 23 January 2007 1

Cable Arrangement Pam’s Arrangement (wrong but easy to assemble for me) Crate 0 Crate 1 RC 0 50 -pin RC 0 68 -pin RC 0 50 -pin RC 1 68 -pin RC 1 50 -pin RC 2 68 -pin RC 3 68 -pin RC 4 68 -pin RC 5 68 -pin RC 6 68 -pin Two 50 -pins are shown as one black line. Unconnected arrows of the same color and type connect to each other (one 68 pin cable per connection). Card numbers in orange, regions as hollow numbers. + Configuration for crates = Upper 50 -pin connector, J 3 & = Lower 50 -pin connector, J 5 P. Klabbers, Two. Crate. Tests, 23 January 2007 2

Operation with the MCC Clock and Control Cards must be setup to receive a clock from the TTC. The toggle switch on the front must be in the down position (the LED should be off) Plug in the double cable to the TTC IN Clock (lower) and TTC IN Control (upper) near the middle of the CCCs. Screw in the connectors to make good contact. First setup the CIC using the package vmedia to execute a simple script before operation of any RCT crates: • This script will have to be updated to use the correct base address for the CIC. Currently it is e 00000 (hex) - I think the card there uses f 80000. • Make sure the SBS driver is installed • Subdirectory ~cmslab/MCC: • vmedia ‘read cicb_init. txt’ • This needs to only be done when the crate is turned on. • If the crate is left on, it doesn’t need to be repeated for every run. P. Klabbers, Two. Crate. Tests, 23 January 2007 3

Package Description and Use Written as part of the package JCCTest • For data sharing test is done in two steps (then repeated for other direction) • RCTTest. With. JCC -t 20 • In RCTJCC. cc do. Cable. Share. Send. Test() • Function called Cable. Share. Send. cc • File produced with all shared data (Shared. Data. txt) • RCTTest. With. JCC -t 40 -n 10 • In RCTJCC. cc do. Cable. Share. Recv. Test() • Function Cable. Share. Recv. cc • Writes random “electron-like” data to the receiving crate • Reads in the file (Shared. Data. txt) with the Shared Data written to other crate • Predicts output - written to pointer and returned to RCTJCC: : do. Cable. Share. Recv. Test() • A reset is sent to both crates via the MCC using system() • Function EGJCC() called to read out JCC • Prediction compared to result at JCC • Repeated readout possible with -n flag (above = 10 reads of JCC) P. Klabbers, Two. Crate. Tests, 23 January 2007 4

MCC, RCT, and Reset is sent by executing a shell script resys. txt A vmedia script is run from this shell script (I needed to change to a third CAEN controller in the MCC) • In subdirectory ~cmslab/MCC • resys. txt - vmedia script • Need to update the base address for the CIC in resys. txt • resync - performs vmedia and logs • Writes to a logfile resetlog - take care that this doesn’t get too big. • Remove CAEN environment variable • Create links to these in JCCTest directory • system() call in programs does the work • Will have change this to use ssh (slide after next) when using SBS, since will be on a separate PC • Single crate tests can be run (with J 4 silencers) by using with the flag -mcc in Test. Crate. With. JCC. com • Look for the system() calls and update them for the SBS • Maybe get environment variables to change how this is done P. Klabbers, Two. Crate. Tests, 23 January 2007 5

Operation with the CAEN Controller The CAENs’ PCI cards are all in one PC so I was able to script the process using environment variables to change crate control: First top crate receives, bottom crate sends export CAEN_LINK=0; Select bottom crate’s CAEN controller RCTTest. With. JCC -t 20; export CAEN_LINK=1; Select top crate’s CAEN controller RCTTest. With. JCC -t 40 -n 1; MCC Reset to RCT done here Then bottom crate receives, top crate sends RCTTest. With. JCC -t 20; export CAEN_LINK=0; RCTTest. With. JCC -t 40 -n 1; MCC Reset to RCT done here The reset to the MCC crate controller is integrated into a script called with system() in the RCTJCC: : do. Cable. Share. Recv. Test() P. Klabbers, Two. Crate. Tests, 23 January 2007 6

Operation with SBS With the SBS, several PC’s need to be used along with ssh remote command execution. • ssh user@login command • Either in system() or in scripts • Use keys to enable “password free” login • You will have to replace some of the program calls and the CAEN_LINK environment variables P. Klabbers, Two. Crate. Tests, 23 January 2007 7

Upgrades The whole test can probably be simplified • write all data to both crates first • then check sharing for both Cabling should be redone as well to be more “barrel like” • Next Slide • Good idea to do this before expansion to 4 crates Expand to 4 crates • Two possible crate arrangements • Two Eta halves, two Phis (see slide after next) • One Eta half, four Phis (see two slides after next) P. Klabbers, Two. Crate. Tests, 23 January 2007 8

Cable Configuration “Barrel” Barrel style Crate 0 Crate 1 RC 0 50 -pin RC 1 50 -pin RC 0 68 -pin RC 1 68 -pin RC 0 50 -pin RC 1 50 -pin RC 0 50 -pin RC 1 68 -pin RC 0 68 -pin RC 1 50 -pin RC 0 50 -pin RC 2 68 -pin RC 3 68 -pin RC 2 68 -pin RC 4 68 -pin RC 5 68 -pin RC 4 68 -pin RC 6 68 -pin Two 50 -pins are shown as black lines Unconnected arrows of the same color connect to each other (one cable) + Configuration for crates = Upper 50 -pin connector, J 3 & = Lower 50 -pin connector, J 5 P. Klabbers, Two. Crate. Tests, 23 January 2007 9

Expanding to 4 Crates Sharing between crates loops around, becomes a small “barrel” To lowest crates 68 -pin cables shown by small double arrows and …same color arced arrows connect to each other - one 68 -pin cable per pair. 4 “Y” Cables “Y” cables on each RC 6 for a total of 4, Short end connects to a long end 50 -pin connections do not make sense in this configuration P. Klabbers, Two. Crate. Tests, 23 January 2007 To highest crates 10

Expanding to 4 Crates 2 “Y” Cables – To lowest crates + 2 “Y” Cables To highest crates Similar to slide before, but now we can use 50 -pin cables to share corners and left-right. The magenta cables are corner sharing and the indexes go to the same index on the other end. The black cables are the left-right sharing. A total of 8 50 -pin cables will be used in this configuration. The same number of 68 -pin as before. Configuration of the – crates will be slightly different, as well. TP_LF, a bit for the EISO ASIC will change. P. Klabbers, Two. Crate. Tests, 23 January 2007 11

Coming Soon I’ll need to write something up on how I group the towers to do the electron algorithm I’ll give you the format of the Shared. Data. txt, since there is some compression of data to avoid writing towers twice. Also of the Tower. Patterns. txt which is a sort of ASCII viewer to look at the patterns in the crate. TP_LF configurations for different etas. P. Klabbers, Two. Crate. Tests, 23 January 2007 12