Telescope and AGw Commissioning at LFBG with IRTC - 02-03 April 2009 MST
Observers: J. Hill (mtn), A. Rakich (Tucson-Safford-Tucson), J. Storm (Potsdam)
Telescope Operator: A. Ceranski
Telescope Support: J. Hill
Lost most of night to GPS problem. Hacked the time to get one hour on-sky without rotator motion. Got spare GPS from Tucson working just in time for weather to close us down for the last two hours.
We did get to verify Dave's new servo code on-sky. Although he already knew it was working from tests in the morning.
Made a measurement to check the adjusted azimuth scale, but I need to verify with Dave that the new scale was
still loaded after the DSP patches installed at midnight.
Dave A. fixed the servo code. This was the first part of last nights problem.
Discovered that GPS was not locked (and may not have been locked for a while). Dave A. and John H. (plus Mike G., Dan C. in Tucson) tried to fix it for 2.5 hours, but
finally concluded that there was a problem with the receiver unit. It acts like there is a bad preamplifier unit as it sees only 6 satellites with signal -171 to -172 db at best, while the minimum threshold is -170 db. We checked the lightning protection circuit on level 10. We power cycled the GPS unit - this was the fatal mistake as it reset the time to 1970. We brought the antenna down from the roof to test it on a short coax cable, and we hooked up the antenna from John's 4WD GPS (which gave the same result as the regular antenna). The spare GPS unit is in Tucson for rotator testing. After giving up for the
night at 20:30, John H. gets the idea that you can enter the time from the GPS front panel menus
. GPS unit will be a stable clock even if it doesn't have satellite input. This may have been the second part of last night's problem.
(story continues below)
I heard indirectly that Michele found some extra processes on lbtmu102 that may have been slowing down PCS last night.
That may have been the third part of last night's problem.
21:15 Aaron filling LBC dewars
21:45 Restart MCSPU which was complaining about rotators that don't exist. Perhaps some residue of today's direct Gregorian rotator testing.
21:55 Called Mike G., and we figured out that the RIM (rotator interface) module for Bent Gregorian Rotator was turned off. Lost 25 minutes.
22:10 Open dome to clear sky and wind less than 12 m/sec. T=0.0 degC, D=-8 degC.
22:14 Resetting Azimuth encoder.
22:19 pointto M5_0441 ACQUIRE gs=0 PAM-
EL and AZ servos dropped to velocity mode. Telescope will not slew to track.
Apr 2 22:21:08.713: LFBG RTracker:Slew done. state-->TRACKING, pos'n= 1.1 v= 0.01
Apr 2 22:21:08.714: LFBG RTracker:State changed from SLEW_TO_TRACK --> TRACKING
Apr 2 22:21:09.384: EL Tracker:Servo unexpectedly changed to velocity mode!!!
Apr 2 22:21:09.385: EL Tracker:State changed from TRACKING --> ENTER_VEL_MODE
Apr 2 22:21:09.467: EL Tracker:State changed from ENTER_VEL_MODE --> TRACKER_VEL_MODE
Apr 2 22:21:09.699: AZ Tracker :Servo unexpectedly changed to velocity mode!!!
Apr 2 22:21:09.700: AZ Tracker :State changed from TRACKING --> ENTER_VEL_MODE
Apr 2 22:21:09.782: AZ Tracker :State changed from ENTER_VEL_MODE --> TRACKER_VEL_MODE
22:22 Andrew leaves Tucson with spare GPS. Fast new sportcars are good for something!
22:50 fpga load (hoping that the script still pointed at old fpga code, but the hope was in vain)
23:00 editting lbt.conf (with vi) seems to have crashed the network server but not reflective memory on to-station
Tony's clever idea of editting the TAI conversion in lbt.conf works. I was able to dial the PCS - MCSPU difference
down to 3 ms, but in fact this is only accounting for the difference between the lbtmu102 and jet clocks.
After talking to Dave A. (now back home) it seems that Tom checks polys with jet system time, but DSPs check polys from
23:46 Adjust GPS-jet from -2.32 sec to -0.33 sec by using ntpdate -1 gps
to check the difference. GPS unit loads the time you enter when you press enter to confirm. But DSPs still seem to be rejecting our polynomials.
00:00 Aaron departs for basecamp to meet Andrew.
00:15 Dave loosens the time tolerances in AZ and EL servos, and they are working. But now rotator doesn't work, it drops back to hold mode when you issue a rslewtotrack.
00:25 reboot jet to unfreeze from Dave's command to check the system clock. Then rpcconfig start 10.144.0.20
00:45 Finished resetting Az and El encoders. This was the second time tonight that they worked correctly on the first try.
00:49 pointto M5_0441 ACQUIRE gs=0 PAM- timed out waiting for rotator ready.
01:08 Aaron takes possesion of spare GPS
01:22 Found star with offset IE=90 CA=-90
, but some funny collimation parameters. With everything cleared, IE=+50 CA=-115
Refined to IE=+52 CA=-110
after focussing, but still with +10 arcsec RXglobal. Azimuth was +310, EL=60. This seems to indicate that the time was off about 5 seconds+-.
Note that with telescope in equilibrium, the residual coma is small (2 arcsec without any correction). So we conclude that the IRTC collimation model is quite useable at EL=60.
Images in /Repository/AGW_Data/090403/. That directory also contains the first draft of a shell script that does ptmodify
commands to make a spiral search. This is at least a factor of two faster than making offsets from IRAF.
01:33 azslewtohold -48
We have verified the new DSP servo code for reading strip encoders on-sky. Tracking seems fine.
01:39 guider000428 After a full 360 slew the same star comes back within < 3 arcsec in CA (correction was moving CA from -110 to -107)
01:44 pointto M5_0912 ACQUIRE gs=0 PAM- timed out waiting for rotator ready. To get a lower elevation star.
Found faint star at IE=-73 CA=-122
01:51 azslewtohold 366
01:57 guider000622 - star is off by 7 arcsec in CA (correction was moving CA from -122 to -117) AZ 366 EL 39
02:00 azslewtohold 6
Humidity has risen to 90%, but sky is still clear.
02:06 guider000673 - star is off by only 1 arcsec in CA (correction from -115 to -116) (and maybe 2 arcsec in IE) AZ=5 EL=39
I'm supposing that this apparent hysteresis in the unwrapping correction is caused by drifts in open loop tracking
during the time it takes to slew 360 degrees. If Dave A. had the new Azimuth scale in place at this time, then the sense of the correction was OK, and not opposite. If not, then we need to repeat this test later.
Spare GPS arrives
02:12 Aaron arrives back on summit with spare GPS.
02:14 Open Left side vent door half way to stick the antenna out
02:15 Close shutter doors while we work on GPS.
IP address 10.144.0.6 (lbtmu201.lbt.as.arizona.edu alias gps)
subnet mask 255.255.255.0
Spare antenna on the old GPS produces no improvement.
Spare antenna on the newly arrived spare GPS has +20 dB more signal than the old one. This seems to confirm our hypothesis that there is an amplifier stage blown somewhere inside the old one.
Spare antenna sticking out the left vent door still requires a bit of driving the building around to get the 4 satellites needed for a lock.
03:32 spare GPS locked and working!!!!!!!
The good news is that the telescope is working on all axes. The bad news is that there is weather moving in from the west, so we can't reopen.
03:48 pointto M5_0912 ACQUIRE gs=0 PAM- succeeded with dome closed.
04:21 100% humidity - it is officially foggy outside.
04:26 HBS shuts down while slewing to park.
Forecast is for 60 mph winds and snow tomorrow night, so I have a list of closed dome projects to work on.
- Check LBC-Blue front baffle
- Check motion tolerance on M3 selector
- Test PMC (and PCS, ECS) with Tony's new version of TEL (Telemetry)
- Test patched OSS for the M3 selector update. (Paul already did this while we were working on GPS.)
- Use IR camera to check C-Ring and oil temperature.
- Finish the ptspiral shell script for faster spiral searches.
- Plenty of time to watch 80 telescopes around the world.
- 02 Apr 2009