Date: Mon, 13 Feb 2023 09:13:25 -0700
Subject: Re: [sloan-25mlog 6596] Re: 2.5m obslog 20230212 (SJD 59988)
From: "Daniel Oravetz"
>> ** I tried to initiate the auto mode during design 101684-118721-8810 in
>> preparation for the following field, but it automatically loaded the
>> preloaded design for the next field. The auto mode previously waited for
>> any exposures to conclude before loading any preloaded designs, so
>> something must have changed and I need to learn again how to properly
>> enter and exit the program without disrupting anything and costing
>> observing time. The 101684-118721-8812 design was reloaded manually
>> while the exposures were still running, so hopefully the data can be
>> associated with the correct field-design, but the configuration number
>> will be different. The auto mode never fully completed, so I am unsure
>> if Jose was able to fix the UNPARSEDTEXT error or not.
>>
>>
>
> Per Dmitry’s experience: the good sequence to enter the auto mode is:
> preload a design, load it to HAL, then toggle the auto button. One can
> exit the auto mode any time.
Dmitry's note is correct in order to avoid having the currently-loaded
design REPEATED when Auto mode gets toggled on, but that's different than
the timing of a preload/load happening when Auto gets toggled on.
When I used Auto mode last time I remember I was able to toggle Auto mode
on *without* it automatically initiating a preload (especially a load) -
there would be a message in the Boson HAL window stating "expose is
running. Waiting for it to complete." So it does sound like different
behavior now if a new design gets loaded right away as soon as Auto mode
is turned on.
> The loading of the current preloaded design occurred when the auto-mode
> initiated with about 5 minutes still left in the exposure sequence. This
> was not supposed to happen. I assumed the macro was supposed to be aware
> that exposures were running and wait until they are finished before
> loading and executing the preloaded design for the subsequent field.
The timing for a preload is supposed to be three minutes before the end of
the remaining 'expose' time (including BOSS readout), but a load shouldn't
happen at all until the 'expose' is complete. It kind of sounds as though
the actual remaining time left in the in-progress 'expose' wasn't known,
or somehow wasn't determined quite right, or somehow knowledge that an
'expose' was in progress to begin with was lost within hal. I would expect
that if Auto gets toggled on with less than three minutes remaining in
'expose' that a preload would happen immediately, but the timing doesn't
line up for this issue last night.