Night Time operations Time Charging

This section written by Matthew Shetrone
with minor updates in October 2020 (SJ)

Helpful links

• The current RA night report is: ​http://nossy/noss/RA/NightReport/
• Previous TO night reports are at: ​http://nossy/noss/TO/dateselect.php
• The public night report reader is: ​https://het.as.utexas.edu/HET/hetweb/vlrweb/RR/reader.php
• Amy's time accounting summary page is: ​https://het.as.utexas.edu/HET/hetweb/apps/nightdata/index.php

Duties

Although time accounting is a responsibility shared between the RA and the TO it ultimately falls to the RA to make sure that the accounts are correct since it is the RA who must respond to any problems encountered in a PI's time accounting. The best way to keep problems minimized to communicate during the night about when certain actions occur (such as the end of stacking or when idle time begins), to double check your report for strange accounting values (such as large overheads or huge PR time lost values) AND to check the time accounting page at the end of the night: ​https://het.as.utexas.edu/HET/hetweb/apps/nightdata/index.php

Time Accounting Basics

The beginning of the night accounting is defined as beginning at the first science exposure or 18 degree evening twilight, which ever comes first. Thus if the TO is doing mirror alignment before 18 degree twilight then the alignment time is not counted against any science time. If the telescope is idle from the beginning of the night, the RA should use the 18 degree twilight time as the time for the first "Idle" activity entry in the NR. Time lost due to weather comes from the RA making an Idle activity entry and sometimes also changing a science/engineering target's acceptance status to "W" which is rejected due to weather.

Time lost due to PR are any entry in the PR section of the TO report (even if it does not have a PR number) as long as it follows the proper formatting. Any observations rejected for equipment failure ("E") in the RA report must have a corresponding entry in the TO report’s PR section and should probably have a PR filed to the Bugzilla system. PR time lost is only accumulated during official science, engineering or instrument commissioning time. Thus a PR reported in the TO report before science operations begins or during idle time is not counted in the Time Accounting sense, but is important for the Day Staff so they know to go fix/monitor it.

All actions should begin and end in the same minute to properly account for the overhead. Thus target 1 ends at 10:20 and target 2 begins at 10:20 and Stacking might end at 10:20 and the next target should begin at 10:20.

Idling and Splitting Science and Engineering Time

Idle time is really a misnomer; we don’t sit around idle watching the “Beverly Hillbillies” when conditions don’t allow normal science operations. If the priority for the night is science but the clouds are too thick to allow science operations then you are idle. There are some engineering tests that can be conducted under nearly any open dome conditions. Normally we save these tests for idle conditions. This does not mean we should account this time as engineering since we would be doing science if we could, we can’t so we account the time as idle. If you do these engineering tests, mark them as done in the queue, report the results to the appropriate person and watch the weather in case we can start back on the priority for the night.

If you find that the next few science targets have a low priority and there is a higher priority engineering test waiting in the queue then consider stopping science to begin engineering. In that case, make an entry in the RA NR to account the time as engineering since you could have been doing the priority for the night (assuming it is science).

Under conditions where a night is completely lost due to a problem and bad weather has rolled in the time should be accounted according to what occurred first. If bad weather rolls in and then you discover the tracker is broken the night was lost due to weather (assuming that the weather is bad all night). If the next night the weather is also bad and the tracker is still broken then the entire night (18 degree to 18 degree) should be accounted as lost to a PR and you never enter “Idle due to weather”.

Clear Weather- Normal Operations

Shutter Open time is charged from htopx (using "Done") while overhead is charged from the RA night report (calculated from start to expstart). It is important that the night report program number be correct in order to have the right program charged. It is important that the htopx program number be correct (i.e., 3 digits in the program number, UT03-2-XXX) in order to have the right program charged. The night report code should be an “A”. Targets should be observed within about 0.1 arcsecond seeing of the image quality observing constraint and within about 0.1 mag/arcsec2 of the sky surface brightness brightness observing constraint. Targets should be observed with the following priority

1. Targets on the priority list
2. New priority zero targets or newly activated synoptic targets.
3. If the conditions are good then observe targets with tight observing constraints (Dark time, good seeing, many visits).
4. Highest priority targets

Note this means that a dark time target with P3 might beat a P1 bright time object IF the bright time object is not on the priority list for some reason (synoptic, setting at beginning of the night, requires exceptionally good seeing).

Poor Weather Operations, Priority < 4

When the observing conditions are worse than the requested observing conditions for a target it should not be observed except under unusual conditions (the target may never again be available in the period or no other targets are available). Should this happen then the RA should choose a target that has the greatest chance of being successfully completed:

1. Targets with the closest observing constraints to the current conditions.
2. The target with the shortest exposure times

Multiple additional exposures may be taken to bring the signal to noise up to roughly the desired level. Since the first exposure will have the largest overhead this visit should be marked as rejected due to weather (“W”) in the night report. This will have the result of reducing the overhead charged down to 0-2 minutes. When the RA believes that enough spectra have been acquired the target should be incremented (in htopx) for a single visit. Thus a five minute visit request could be done with 5 five minute visits but the PI would be charged for only a single five minute visit. Only a single entry in the nigh report should be marked with acceptance "A" and this entry should have a small overhead.

When targets are observed during variable conditions or degrading conditions it will be up to the RA whether the visit should be charged or not. If the observation appears to be a success but the conditions degraded to worse than the requested observing constraints then the RA should note this in the notes section of the night report, increment the target in htopx, and use acceptance code “B” for borderline in the night report. The borderline marker will appear as acceptable to the night report reader but gives the RA a reminder that if the PI rejects the spectrum there is no reason to investigate further. Should the PI reject a spectrum then the RA will need to:

1. Edit the target in htopx with the "Restore" button to remove the appropriate number of done visits. NOTE: Look at the night report and try to determine then correct number of visits to remove, i.e. it may be less than 1. Htopx keeps a log of all operations done through the code so you can scan back through it and look for the appropriate data entry.
2. Edit the night report and change the “B” to the appropriate rejection code (most likely “P” which is Rejected by PI for weather).
3. Contact the PI (cc the other RAs using the astronomer email account) and let them know what changes have been made

Options to exhaust before going idle for poor seeing

1. check htopx for any (including P4) targets with loose constraints on IQ
2. observe HPF telluric standard star (even if not requested that night)
3. observe LRS2B/R standard stars (even if not requested that night)
4. observe VIRUS standard stars (even if not requested that night)

Science Operations in an Empty Queue and Priority = 4

The purpose of priority 4 targets is to offer the RA objects that can be observed in poor weather conditions OR when the queue would otherwise have no targets. For targets of priority 4 the PI is not charged any overhead and the CCD shutter open time is charged at only 50% of the requested visit. To do this the RA simply marks the acceptance type to “4” and increment the number of visits done by 1. If multiple visits are taken to get the S/N up to near the requested level the RA can reject the first few spectra as “W” and accept the best as “4”. All of the accounting is done automatically within the software if the acceptance for the night report is set to “4” and the priority in the Phase II is 4.

If a hole develops in the weather filtered queue and no engineering or science targets are available then rather than keep the telescope idle some rather extreme actions are required. Targets may be re-observed from inactive programs, i.e., targets that have already been observed and marked as done or targets that are on hold for any reason. The observation should be marked as “I” which stands for Idle. The PI is not charged for the overhead. If the program had been placed on hold because it was out of time then the target should be marked as “Done” without incrementing the number done. This will keep you from observing it again but keeps the PI from being charged for the time. If the target was on hold because it was waiting for a specific date then just leave the Phase II entry unchanged but explain the situation to the PI in an e-mail.

Time accounting logic in Amy's "nightdata" code

The graphs/summaries produced by Amy's "nightdata" code (​https://het.as.utexas.edu/HET/hetweb/apps/nightdata/index.php) are used in monthly and trimesterly reports to the HET Board and users. Her code operates on each minute of the night between 18-deg evening and 18-deg morning twilight, and determines what category it belongs to. The following logical decision tree describes the categorization process:

• First, if the current program is Idle, or if a weather-rejected target was being observed or acquired, then that minute is counted as Idle.

• Next, if a problem has been reported, or if a target rejected for equipment failure was being observed or acquired, then that minute is counted as PR.

• Next, if a stack is reported, then that minute is counted as Stack, unless the stack has circumstance 'e' (engineering), then that minute is counted as Engineering instead.

• Next, if the current program is Engineering, or if an engineering target was being observed or acquired, then that minute is counted as Engineering.

• Next, if the current program is Commissioning, which it isn't, then the minute is counted as Commissioning.

• Next, if the current minute falls between a target start time and exposure start time, then the minute is counted as Overhead. If the current minute falls between exposure start time and the target end time, then the minute is counted as Exposure.

• Finally, if the current minute meets none of the above criteria, then the minute is counted as Unused.

Monthly statistics produced by this code can be found here: ​https://hydra.as.utexas.edu/?a=help&h=7