 |
Last Updated: 2007 Aug 30 Chris Benn crb@ing.iac.es, Sam Rix srix@ing.iac.es Rene Rutten rgmr@ing.iac.es |
Version 2007 Aug 30
Below is a summary of GLAS-related things that need fixing / looking at. A formal 'issue log' is being drawn up by Gordon, and this will be the main source of information on what needs doing, and what the status of each issue is.
List from crb:
GRACE
-----
- Risk to DM on inserting/removing LED. The DM cover should be in place
during this operation.
- Sooner or later, someone is going to knock the LGS WFS plate protruding
from the bench, which will presumably mess up the alignment.
Guard rail needed.
- The hardware (alignment etc.) is probably changing fast at the moment,
but once it starts to settle down, and particularly once we start
taking serious commissioning data, any changes in GRACE need to be
logged in the GRACE logbook.
- There are still a few Christmas trees in GRACE, lights need taping over.
- Labelling of the components on the GRACE bench and in the GLAS electronics
rack - someone needs to check everything is labelled with sufficient
clarity that a DE or observing support scientist can identify what's
powered on, or buzzing, or emitting smoke.
E.g. to adjust the laser range A and B, you have to press buttons 2 and 8
on one of the panels in the GLAS control rack, and I don't think this
is labelled.
sextans
-------
- Poor old sextans can't cope - the response on topgui can be very sluggish,
which makes it irritating to use, and will slow down commissioning.
TopGuiGlas
----------
- The slave CCD sometimes reads out with the orientation flipped across
a diagonal. I think (but am not sure) Jure said he's addressed this
problem and it won't recur.
- Button needed for taking offsets, i.e. to do:
WFS SetWFSoffsetsToCurrentCentroids slave
etc.
- When going to LGS mode, the WFS gain and the reconstructor should
be loaded automatically (cf auto mode 1 button for NGS mode).
GLAS gui
--------
- Some of the displayed data are incorrect. How about blanking out all
these fields until they can be filled reliably?
- Presumably the LSM buttons always steer in the same direction relative
to the 8-arcsec TV view - would be useful to indicate this direction
on the steering buttons e.g. 'up and left on tv' and 'down and right
on tv'
FITS headers
------------
- When I checked a few days ago, the GLAS (GL*), AO tip-tilt (AOTT*?)
and AOWFS headers were all missing.
This will make logging of commissioning activities onerous and
error-prone.
INGRID
------
- Now reading out, but fails in simplex mode after ~ 100s iterations.
Frank reported last night that AG7 is also failing in simplex mode.
Documentation
-------------
Documentation of technical aspects of the project, and of project status,
isn't easy to navigate at the moment. It would be nice if, with a few
clicks of the mouse (and without having to type in passwords), one could
reach:
- an index of technical documents, and the documents themselves (ideally
.html or .pdf)
- the commissioning plan, with bits filled in on a night-by-night basis
(either actual results, or status e.g. 'done')
- commissioning-night reports
- prioritised action list of things to be fixed / enhancements to be
addressed
Safety
------
- The powerful IR pump laser lives in the electronics room - should there
be a warning on the door that it's in there, and may be powered on?
Or is it zero-risk?
- When trained personnel enter the dome (or GRACE), they make a judgment
about when it's safe to remove the safety specs, e.g. after powering
off the pump laser. For less-trained personnel (e.g. observing
support scientists) it would be useful to document (and display)
clearly under what circumstances the specs can be removed.
More, from rgmr:
- LSM slow offload to negative lens needs to be implemented. - laser power meter logging needs to be installed. - Photodiode diagnostics needs to be wired up. - Laser on/off signal needs connecting up. - Access to the calibration source is scary, and even switching on/off a awkward. A quick fix would be to at least have an on/off switch on the power cord, but in the long term mechanizing the unit should be considered. - Current/voltage control to the calibration unit LED is needed; ideally remotely controllable (but not essential). - Look into vulnerability of calibration source. If it dies then GLAS cannot be used. Build a spare? - Possibly the de-rotator needs to be better aligned, and consequently the full LGS-WFS optical train. - L2 and L3 on the LGS-WFS needs redesigning (probably), and possibly L6/7/8 as well. - There was a question whether the 4x4 mode was working or not; needs testing.
More, from srix:
1.The filter wheel of the WFS keeps timing out. 2.The simplexing procedure is very slow e.g 1600 cycles took 353s. 3.Tim looked at the WFS X,Y centroids for a segment. With 10ms integrations, the average position stays about 0.0. However with 2.5ms integrations, the average position moves around by +/-0.03 of a box (e.g. 10% of on-sky variation) suggesting that there is lots of turbulence in GRACE. Is the air conditioning causing a problem? 4.To lower the load on the system, we used WFS SetPixelDecimate 20 to reduce the refresh rate of the displays (e.g image and alignment tool). (Note that both are sampled not averaged.) 5.Problems with vibrations when observing. Visible on LGS WFS, NGS WFS and in images. At one point, in a dense field, the TT star even jumped to a different one, meaning that the FOV shifted! We need documentation about the Pockels cells. Sync and anti-sync etc. I don't have good notes on this. Eg if you have it in anti-sync mode on sky you will get all of the plume except the bit you are interested in.
More, from don:
|
Last Updated: 2007 Aug 30 Chris Benn crb@ing.iac.es, Sam Rix srix@ing.iac.es Rene Rutten rgmr@ing.iac.es |