Date: Thu, 12 Apr 2001 09:15:12 -0700
From: Russell E Owen
Subject: Re: 2.5m obslog 20010411 (MJD 52011)
At 6:56 AM -0500 2001/04/12, Eric H. Neilsen, Jr. wrote:
>Watcher errors shortly after pointing model abort. (The axis seem fine)
>AxisNoSlew:az (axis cannot slew: Apr 12 02:39:50 MDT)
>AxisNoSlew:alt (axis cannot slew: Apr 12 02:39:50 MDT)
I'm puzzled about the statement "shortly after pointing model abort".
The TCC log shows a pointing model just starting up at this time. the
initial slew goes fine, but the rotator does not slew, probably
because it's set to rotation type object, which isn't allowed when
the main axes track in mount coordinates.
2001-04-12.08:39:48.856 0 0 : Text="pointing error log file opened"
2001-04-12.08:39:48.948 0 0 > Started; Cmd='Track/Wait -32.27, 25.00 Mount'
2001-04-12.08:39:49.756 0 0 W Text="some axes will not be slewed";
AxisNoSlew="FFT"
This is a poor fit to the watcher error, as the watcher says that az
and alt cannot move.
I looked through the whole log for AxisNoSlew errors and there were
only a few. None showed az and alt failing to move while rot could
move. None (except the one shown above) appeared to be anywhere near
the right time. Here's the complete list. They fall into two
categories:
1) Instrument change when the telescope is not tracking. It tries to
adjust the telescope position for the new instrument. It should
probably be smarter about this and not issue a warning.
2001-04-12.07:04:00.410 0 0 W Modu="prt_ChangeDev"; Text="could not
initialize new device: TCC_ROT1"
2001-04-12.07:04:00.507 0 0 W Text="new position rejected"; AxisNoSlew="TTT"
2001-04-12.07:07:49.543 0 0 W Text="new position rejected"; AxisNoSlew="TTT"
2001-04-12.07:10:35.756 0 0 W Text="new position rejected"; AxisNoSlew="TTT"
2001-04-12.07:19:42.898 0 0 W Text="new position rejected"; AxisNoSlew="TTT"
2001-04-12.07:19:50.489 0 0 W Text="new position rejected"; AxisNoSlew="TTT"
2) Pointing error sequence starting up. There is an initial slew in
mount coordinates to get the telescope into the right quadrant. This
usually cannot turn the rotator. The pointing command sequence could
be made more clever to avoid the warning:
2001-04-12.07:49:13.941 0 0 : Text="pointing error log file opened"
2001-04-12.07:49:14.053 0 0 > Started; Cmd='Track/Wait -35.00, 25.00 Mount'
2001-04-12.07:49:15.358 0 0 W Text="some axes will not be slewed";
AxisNoSlew="FFT"
2001-04-12.07:50:30.862 0 0 : Text="pointing error log file opened"
2001-04-12.07:50:30.961 0 0 > Started; Cmd='Track/Wait -35.00, 25.00 Mount'
2001-04-12.07:50:31.615 0 0 W Text="some axes will not be slewed";
AxisNoSlew="FFT"
2001-04-12.08:39:48.856 0 0 : Text="pointing error log file opened"
2001-04-12.08:39:48.948 0 0 > Started; Cmd='Track/Wait -32.27, 25.00 Mount'
2001-04-12.08:39:49.756 0 0 W Text="some axes will not be slewed";
AxisNoSlew="FFT"
>Some other strange TCC watcher errors:
>Modu="mir o MoveOneGalil" (Apr 12 02:58:44 MDT)
>Modu="prt ReadReplySet" (Apr 12 02:58:44 MDT)
Looks as if this was real. The collimation process was being
alternately enabled and disabled around this time. This suggests that
folks may have been talking directly to one or more of the mirrors. I
suspect somebody left a bit of garbage when they disconnected (e.g.
an incompletely typed command or disconnecting before a reply could
be fully read).
2001-04-12.08:58:25.564 0 0 F Modu="prt_ReadReplySet"; Text="too many
replies on port LTA9:"
2001-04-12.08:58:25.658 0 0 I Modu="prt_ReadReplySet"; Text="the
following replies were read:"
2001-04-12.08:58:25.659 0 0 I Modu="prt_ReadReplySet"; Text="0"
2001-04-12.08:58:25.758 0 0 F Modu="mir_o_MoveOneGalil"; Text="could
not specify new positions for Prim mirror actuators"
2001-04-12.08:58:25.760 0 0 W Text="failed to adjust collimation and focus"
>TCCStatus:alt (clock_loss_signal: Apr 12 03:42:11 MDT)
>AxisStat:alt (clock_loss_signal: Apr 12 03:42:11 MDT)
>AxisStat:az
>TCCStatus:az
>AxisStat:rot (clock_loss_signal: Apr 12 03:42:11 MDT)
FYI, here are the full status words. Looks as if only the clock-loss
bit was set.
2001-04-12.09:41:57.405 0 0 W AzStat=-27.073569, -2.826483,
4493785348.79424, 0x00010000; BadAzStatus; AzDTime=-0.1
2001-04-12.09:41:57.408 0 0 W AltStat=55.960918, -0.652526,
4493785348.90609, 0x00010000; BadAltStatus; AltDTime=-0.1
2001-04-12.09:41:57.506 0 0 W RotStat=-22.098836, -2.290239,
4493785349.06623, 0x00010000; BadRotStatus; RotDTime=-0.1; RotID=1, 0
I was very pleased to see that there were no slowdown problems with
the MCP. There were some episodes of the MCP clock being off by about
a second (easily fixed by doing another AXIS INIT), but no
half-second-ish errors and no communication timeouts.
I'm also seeing a fair number of guider-related errors -- e.g.
"allocation error" and failure to select the camera. Is there a PR
outstanding for these? (I tried a query, but don't know what category
to use. I submitted a PR to add a guider category, though I wouldn't
be surprised if it exists and I just missed it.) I do have new guider
software available. Craig is doing some testing of it on the 3.5m. It
totally overhauls the way in which cameras are selected (internally;
there are no interface changes) and so may fix the problem.
-- Russell