Subject: Re: [sloan-25mlog 3632] Re: 2.5m obslog 20170905 (SJD 58002)
From: John Downey
Date: Wed, 6 Sep 2017 09:15:01 -0600
Come to think of it, this couldn't be anything in the relax procedure - =
because based on the mig readings the mirror does move, just the wrong =
way.
Basically, if I send E & F to the forward limit (+50,000) from 0 (mid =
range), the lateral migs change by + 0.4mm or so.  Upon relaxing, the =
revert back to roughly mid range.
So the actuator seems to be moving, but that seems to be the end of it.
Going to do some more testing vs. the migs this morning before swapping =
the other galil in for Conor.  Since the current galil is the one that =
we were using during mechanical collimation, but not the one used last =
year, I don't think its anything to do with galil configuration.  This =
unit worked up until Thursday night.
> On Sep 6, 2017, at 7:48 AM, John Downey  wrote:
>=20
>>=20
>> We ran into problems with E and F during homing again tonight. JohnD =
and Conor
>> will investigate further tomorrow.
>>=20
>>=20
>> =
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=20
>>                                  NIGHT
>> =
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D
>>=20
>> Home/Relax troubles
>> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>> We spent a number of hours trying to get M1 E and F to home properly, =
but they
>> were not behaving as expected. Conor and JohnD also tried to figure =
out what's
>> going on, but at this point there's still something weird going on.
>>=20
>> At first the homing would fail because "one or more axes did not hit =
reverse
>> limit." After running the relax process two or three times, then E =
and F
>> completed the homing, but the position errors were tens of thousands =
of
>> microsteps (acceptable values are usually +/-50). But later =
subsequent attempts
>> at homing failed altogether with "#HOME Failing: ?CHKMV one or more =
axes didn't
>> make it (SC<> 009)". When JohnD pulled the air off the lateral links =
then it
>> homed, but it was also noted at the time that E and F were at around =
-100k and
>> -200k microsteps, respectively, which is strange since -50k steps is =
homed
>> (reverse limit switch).
>>=20
>> After relaxing, E and F were moved to -50k and +50k microsteps and =
their
>> respecitve MiG positions were recorded at those positions. The MiG =
values
>> indicated that E and F only moved a couple tenths of a millimeter =
over this
>> range, when based on the scale, 100k total microsteps should result =
in 3-4 mm
>> change in the MiGs.
>>=20
>> Conor thinks it could be a mis-configured Galil, or it could be =
missing steps.
>> Tomorrow the other Galil may be swapped in so Conor can have a look =
at it's
>> configuration once it's on the network.
>>=20
>> Initial errors during homing were of this form (see the "one or more =
axes..."
>> line for the reverse limit switch below):
>>=20
>> home E,F
>> 0 2 > Text =3D "Homing Actuators: E, F"
>> 0 2 i homing=3D0,0,0,0,1,1
>> 2 2 w Text=3D"Galil is busy executing: A=3DMAXINT; B=3DMAXINT; =
C=3DMAXINT; D=3DMAXINT;
>> E=3D1; F=3D1; XQ #HOME, showing cached status"
>> 2 2 i orient=3D0.00,0.76,-3.07,-255.49,-225.30,0.56;
>> desOrient=3D0.00,15.80,-1.50,-84.30,-865.50,0.00; =
desOrientAge=3D88346.04;
>> actMount=3D10400.00,-900.05,6700.01,16150.01,0.00,-0.00;
>> cmdMount=3D9400.00,-500.00,7300.00,5850.00,5650.00,5750.00;
>> desEncMount=3D9410.11,-494.54,7284.46,5829.00,5655.17,5764.22;
>> modelMount=3D9400.00,-500.00,7300.00,5850.00,5650.00,5750.00; =
maxIter=3D1;
>> homing=3D0,0,0,0,1,1; axisHomed=3D1,1,1,1,1,1
>> 0 2 w state=3DHoming, 0, 1, 0.00, 0.00
>> 0 2 w state=3DHoming, 0, 1, 27.10, 27.10
>> 0 0 w Text=3D"Device Command A=3DMAXINT; B=3DMAXINT; C=3DMAXINT; =
D=3DMAXINT; E=3D1; F=3D1; XQ
>> #HOME Failing: ?GORLIM one or more axes did not hit reverse limit  0, =
 0,  0,=20
>> 0,  1,  1"
>> 0 2 f text=3D"?GORLIM one or more axes did not hit reverse limit  0,  =
0,  0,  0,=20
>> 1,  1"
>> 0 0 w state=3DHoming, 0, 1, -0.18, 27.10
>> status
>>=20
>> Later on, we had errors of a different flavor when E and F weren't =
appearing to
>> move:
>>=20
>> home E,F
>> 0 2 > Text =3D "Homing Actuators: E, F"
>> 0 2 i homing=3D0,0,0,0,1,1
>> 2 2 w Text=3D"Galil is busy executing: A=3DMAXINT; B=3DMAXINT; =
C=3DMAXINT; D=3DMAXINT;
>> E=3D1; F=3D1; XQ #HOME, showing cached status"
>> 2 2 i orient=3DNaN,NaN,NaN,NaN,NaN,NaN;
>> desOrient=3D0.00,15.80,-1.50,-84.30,-865.50,0.00; =
desOrientAge=3D1586.33;
>> actMount=3DNaN,NaN,NaN,NaN,NaN,NaN;
>> cmdMount=3D9400.00,-500.00,7300.00,5850.00,5650.00,5750.00;
>> desEncMount=3D9410.11,-494.54,7284.46,5829.00,5655.17,5764.22;
>> modelMount=3D9400.00,-500.00,7300.00,5850.00,5650.00,5750.00; =
maxIter=3D1;
>> homing=3D0,0,0,0,1,1; axisHomed=3D1,1,1,1,0,0
>> 0 2 w state=3DHoming, 0, 1, 0.00, 0.00
>> 0 2 w state=3DHoming, 0, 1, 0.10, 0.10
>> 0 2 w state=3DHoming, 0, 1, 62.10, 62.10
>> 0 2 i encMount=3D9400.00,-500.00,7300.00,5850.00,NaN,NaN
>> 0 2 i orient=3DNaN,NaN,NaN,NaN,NaN,NaN; =
actMount=3DNaN,NaN,NaN,NaN,NaN,NaN;
>> cmdMount=3D9400.00,-500.00,7300.00,5850.00,5650.00,5750.00
>> 0 0 w Text=3D"Device Command A=3DMAXINT; B=3DMAXINT; C=3DMAXINT; =
D=3DMAXINT; E=3D1; F=3D1; XQ
>> #HOME Failing: ?CHKMV one or more axes didn't make it (SC<> 009). SC, =
or 0 if
>> not moved  000,  000,  000,  000,  003,  003"
>> 0 2 f text=3D"?CHKMV one or more axes didn't make it (SC<> 009). SC, =
or 0 if not
>> moved  000,  000,  000,  000,  003,  003"
>> 0 0 w state=3DHoming, 0, 1, 59.54, 62.10
>=20
>=20
> This problem was basically the same as the start of Thursday night's =
trouble, and that part of it could easily be missing steps getting the =
galil in a very confused state.
> The E & F actuators had impossibly large negative counts - -100,000 =
for one and -200,000 for the other (though not round numbers, I just =
don't recall the loose change).  The reverse limit for these actuators =
is -50,000 counts so the galil was clearly not handling a flaw =
gracefully.
>=20
> What I did upon arrival was to pull the lateral link air and home the =
E & F actuators.  This did get the actuators back in touch with reality =
and would then allow normal moves (the home is sometimes needed before a =
move to keep the actuator within its limits).  What I found then is that =
the actuators functioned normally, but the mirror basically did not move =
- only about a millimeter or less before the relax step and less than =
0.2mm after relaxing.  I suspect something is amiss with the relax =
routine and will be looking at the valves that control the belloframs =
this morning.
>=20
>=20
>=20
>=20
>=20
>=20
>=20
>=20
> ____________________________________________________________
> Logs of Nightly APO Activities
> =
https://trac.sdss.org/tracmailman/browser/private/sloan-25mlog/index.html