Subject: Re: [infrastructure 18947] Re: [sloan-25mlog 4948] 2.5m obslog
From: Conor Sayres
Date: Fri, 7 Feb 2020 15:18:06 -0800
That message looks like the MCP=E2=80=99s DRIFT command is no longer =
replying with the current axis trajectory, just the OK?
> On Feb 7, 2020, at 3:12 PM, Elena Malanushenko  =
wrote:
>=20
> Hi John,=20
>=20
> What is the recipe for recovery?=20
> How to get mirror numbers?=20
> Now, while  I have the same message about failing again, what to do? =20=
>=20
> 22:55:26 CmdStarted: OBSERVER.elenam 28 tcc track 121,30 mount
> 22:55:26 OBSERVER.elenam 28 tcc f text=3D"Slew failed: Sub-command(s) =
failed: DRIFT: For command 'DRIFT' expected 2 replies but got ['DRIFT']; =
DRIFT: For command 'DRIFT' expected 2 replies but got ['DRIFT']; DRIFT: =
For command 'DRIFT' expected 2 replies but got ['DRIFT']"
> 22:55:26 CmdDone: OBSERVER.elenam 28 tcc track 121,30 mount
>=20
> with tel position in stow:=20
> 23:08:03 OBSERVER.elenam 41 keys_tcc i =
AxePos=3D120.999749,30.003162,-0.002457
>=20
> Elena
>=20
>=20
>=20
>> On 02 /06/ 20, at 12:32, John Downey  wrote:
>>=20
>> Jeff and I managed to duplicate the problems from yesterday's remote =
checkout.
>>=20
>> As I suspected, the problem started when a tcc track 121,30 mount was =
commanded with a stop button in and the telescope not quite at 30 ALT.  =
We duplicated the problem with v1.10-6 today by putting the telescope at =
28 ALT and putting in a stop button, then trying a tcc track 121,30 =
mount - as expected it crashed the tmtcc_alt process which also hung the =
SDSSDC semaphore.  Jeff has fixed the tmtcc code to not crash when a =
requested slew cannot happen, but it will still loop the reposition =
errors until you do a tcc axis stop.
>>=20
>> This problem was not seen during Monday or Tuesday checkout, probably =
because the telescope was close to 30 degrees ALT those days when the =
tcc track 121,30 mount slew was commanded.  This was a good bug to =
catch, because I suspect usually the conditions would not be right to =
cause it - we might have had it from the outset and only just now =
triggered the problem.
>>=20
>> The TCC appears to be executing correctly, so presumably one could =
still do a tcc track 121,30 mount to get the mirror numbers even if the =
telescope is not at 30 and a stop is in.  One would just have to =
understand why they are getting a reposition error from the MCP every =
few seconds and do a tcc axis stop when finished.  In general, however, =
I believe that we should avoid this in the future.  =46rom what I =
understand of the process, commanding the tcc to track simply ensures =
that collimate is running - why not enable collimate without a track =
command?  Or if we want to do it with a track command, axes should be =
live and the telescope should move to exactly 30 and hold there while =
getting the mirror numbers.
>>=20
>> Jeff and I both suspect that this reposition error bug is probably =
the old out of closed loop bug - we may have to modify things like the =
procedure for recording mirror numbers if we want these bugs fixed =
correctly.
>> ____________________________________________________________
>> Common APO Hardware and Software
>> =
https://trac.sdss.org/tracmailman/browser/private/infrastructure/index.htm=
l
>=20
> ____________________________________________________________
> Common APO Hardware and Software
> =
https://trac.sdss.org/tracmailman/browser/private/infrastructure/index.htm=
l