Ran targets using yesterdays fullmtrlgy_071514_1.xml and script 2014_07_09_TargetsEMSetC_no7.lst
Results are in 07_17_14_04_35_18_TargetRun and look pretty much the same as before with pids 5, 6, 8 and 9 under-performing.
Going to run an adjusted map based on 07_15_14_09_06_05_TargetRun, 070914_set3_allmaps_adj6.xml
Loaded 070914_set3_allmaps_adj6 into MSIM and saved:
cmd_load_Xml_file_to_Database 070914_set3_allmaps_adj6
cmd_save_Database
Restarted MSIM and ran 2014_07_09_TargetsEMSetC_no7 again
Results did not look so great. Sorta see everything. Some got worse, some better, some unchanged.
Ran adjustmaps to produce 070914_set3_allmaps_adj7 off of this data since there were some very clear linear trends in the error vs request plots that hopefully can be taken out.
Loaded new maps into MSIM
cmd_load_Xml_file_to_Database 070914_set3_allmaps_adj7
cmd_save_Database
Started another run using 2014_07_09_TargetsEMSetC_no7
Aborted the test because we realized we could possibly save PID 9
THE RESCUE OF PID9
Realized that PID9 ontime is way too high. Confirmed in past ontime calibrations that the motor had changed performance on theta fwd quite a bit be comparing the ontime streaks.
left image: D:\PfsTests\07_15_14_15_36_35_onTimeTuning\Images
right image: D:\PfsTests\07_04_14_16_07_06_onTimeTuning\Images
Loaded 070914_set3_allmaps_adj7 into matlab
Changed PID9 ontimes
new = setARMval(new,9,'Link1_fwd_Duration',.04)
new = setARMval(new,9,'Link1_fwd_Duration_Slow',.04)
new = setARMval(new,9,'Link1_rev_Duration_Slow',.06)
new = setARMval(new,9,'Link1_rev_Duration',.06)
Starting with the script 2014_07_16_msimMotMap, created 2014_07_17_msimMotMap in which the following ontimes were changed:
- PID9 ontimes according to the changes just mentioned
- PID7 theta fwd .18 -> 0.2
- PID7 theta rev .2 -> 0.26
- PID5 phi fwd .06-> .1
- PID5 phi rev .07-> .1
Changed the rest of the ontimes in matlab also
new = setARMval(new,7,'Link1_rev_Duration_Slow',.26)
new = setARMval(new,7,'Link1_rev_Duration',.26)
new = setARMval(new,7,'Link1_fwd_Duration',.2)
new = setARMval(new,7,'Link1_fwd_Duration_Slow',.2)
new = setARMval(new,5,'Link2_fwd_Duration_Slow',.1)
new = setARMval(new,5,'Link2_fwd_Duration',.1)
new = setARMval(new,5,'Link2_rev_Duration',.1)
new = setARMval(new,5,'Link2_rev_Duration_Slow',.1)
Saved new XML with these ontimes as:
cobraCfg2xml(new,'070914_set3_allmaps_newOTs.xml')
Started run of 2014_07_17_msimMotMap
Results looked good except that PID9 Theta FWD stalled around 210deg. Further confirmation that streaks cannot be trusted as indications of what ontimes are necessary to complete full ROM with smaller moves.
Changed ontime for PID9 theta fwd to 0.06 in the new script 2014_07_17_msimMotMap_pid9fwd.lst and started it.
Updated config in matlab and saved a new xml, 070914_set3_allmaps_newOTs2.xml in D:\PfsTests\07_17_14_17_33_35_msimMaps\Log
cfg = setARMval(cfg,9,'Link1_fwd_Duration',0.06)
cfg = setARMval(cfg,9,'Link1_fwd_Duration_Slow',0.06)
cobraCfg2xml(cfg,'070914_set3_allmaps_newOTs2.xml')
had to tinker with processMsimMaps and artifically extend pid7 log file in order to get it to recognize and get past the sticky point on that positioner. But finally made 071714_setC_2.xml with good maps for all positioners.
Took the ARM_DATA structures for PIDs 5, 7 and 9 from 071714_setC_2 and merged them into 070914_set3_allmaps_newOTs2. Saved the result as 071714_setC_merged.xml
This is all in D:\PfsTests\07_17_14_20_55_09_msimMaps\Log
Ran targGenerator script to make new target files, this is just to get new PID5 targets since old ones were restricted by bad phi ROM.
Moved the pid5 target file to Targets directory and renamed the old one as OLD_TargetList_mId_1_pId_5.txt
Loaded 071714_setC_merged into MSIM and restarted MSIM
Ran 2014_07_09_TargetsEMSetC.lst which includes pid7 now