Palomar LGS Engineering Summary 02/12/06 UT

Personnel: J. Roberts, V. Velur, R. Flicker, A. Bouchez, K. Dunscombe.

Conditions: Clear, V seeing 1.5-2.0"

1745 Beginning MGSU experiment.
1800 Trouble starting up ASCAM again.  'allsky' cannot connect to camera
driver.  Will have to trouble-shoot further in daytime.
1830 Acquiring new asterism with MGSU.
2300 MGSU observations unsuccessful tonight.  Trouble with temperature
controlled of one camera, and unexplained high noise in others.
2320 Handover to LGS engineering
2330 Beginning LOWFS performance test (dual-NGS mode)
2400 AO database failure, followed by 3 consecutive AOCP crashes (not
responding to commands).  Suspect this may have been due to a high rate of
acquisition camera telemetry (10 frames/s).  Lost 1 hr total.
0110 Acquired first binary (V=7.0,7.4; sep=7.6").
0140 TT performance on LOWFS very good.  Last night's apparently poor
performance probably due to incorrect gain settings.
0220 Unable to acquire fainter binary companions due to capture of primary.
0300 Measuring LOWFS performance on single calib. stars (V=9.7, 12.9, 16.0)
0350 Investigating apparent loop collisions, DM flatmap problems.  Resolved
most of last night's AO software concerns.  See details below.
0410 Successfully testing uplink TT control with new reconstructor.
0450 Measuring LOWFS-PHARO flexure using stimulus (dome closed).
0530 Abandoned test due to inability to close TT loop on fiber.

Thanks to everyone who helped out with this run, both on the mountain and on
the phone!  The MGSU effort was largely successfull, and we made a great
deal of progress understanding the performance of the LOWFS.

- Antonin


Detailed description of ao loop problems in LGS mode, from Jennifer:

So, we saw a couple of weird things last night:

>         3. Apparent conflicts between LOWFS and HOWFS loops, particularly
> when opening the HO loop while LO kept closed on LOWFS.
>         4. DM does not go to flat when opening HO loop (in dual NGS mode).

We saw this again tonight and discovered something interesting. The first
thing we discovered was that when the TT loop is locked on the HOWFS
and the LOWFS is turned on, that the TT offloading to the telescope was
not related to the TTM positions being reported in the HOWFS TTM positions
plot (wrong directions, unnecessary offloads).  When we turned off the
LOWFS, the offloading starting working properly.

Next we noticed that, while both the TT and DM loops were locked on the
HOWFS, the DM had gone to some strange shape and the TT mirror was railed.
We had been trying to acquire the LOWFS for dual NGS mode.  Turning the DM
loop off and applying flatmap did not seem to have any effect.  We tried
opening and closing loops both by using TAO and by using the command line
with the same effects.  But, with the TT still locked on the HOWFS, we
turned off the LOWFS and the loop offloaded to the telescope.  We still
had to close the DM loop to fully recover, though.  We reapeated this a
few times.

So, it seems that when the TT loop is locked on the HOWFS and the LOWFS is
turned on that the offloading numbers are coming from the open loop LOWFS
instead of the HOWFS (which I don't understand because I thought that
TTM_POS were always TTM_POS).  It also appears that the weird shape we saw
on the DM wasn't really on the DM, but was something to do with the tilt
aberrations introduced by the TT loop.  Also, the DM does try to take up
some of the tilt which could explain some of the weirdness that we're
seeing.

The DM does go to flatmap when we open the loop for the most part.  But in
this mode where the tilt has gone crazy, we can't really tell what it's
done.

-Jenny