Queue Scheduling
How does the queue work? What are the priority algorithms?
The MMT queue scheduling software used for MMIRS and Binospec queues is derived from an early version of the AstroPlan package, an affiliate package of astropy; it has scheduling customizations and integrations to communicate with the webpages and catalogs used by the PIs and the queue observers.
The queue schedule software is re-run every morning, with new input on which fields have already been observed and which are still to be observed. During the night, queue observers will re-run it in “dispatcher” mode to provide a ranked list of observing blocks at the current time.
In the software, there are a few boolean checks on the observability of every field. These include airmass (set to a maximum value of 2.5), rotator angle (set by the rotator limits of the instrument; not relevant for imaging), altitude (a minimum of 20 degrees with a maximum of 88 degrees), and Moon separation (MMIRS has wavefront sensing in the optical requiring that the object be at least 15 degrees away from the Moon, and for optical instruments such as Binospec, we require 30 degrees away to minimize moonlight in the science data). We generally begin and end observations (or set up for observations) at Nautical Twilight (15 degree twilight).
There are also three possible variable constraints that the queue software uses to decide the priorities. The variable constraints are: the ratio of time charged to a program over the total allocated time for that program (TimeAllocationConstraint), how far a target is from the meridian (MeridianConstraint, prefers lower airmass and ideally better data quality), and PI priority (PIPriorityConstraint).
How big are the individual observing blocks?
Currently, observations are divided into two-hour-maximum observing blocks (automatically created by the queue software). Multiple two-hour observing blocks are used to fill the requested time for a target.This typically makes a target easier to schedule than would a single longer observing block. It increases the chance that at least part of the requested time for a target will be observed. It also allows the observing block to be scheduled later in the night. Finally, it allows the scheduler to compare the priority of this observing block with other observing blocks more often as observing continues.
How will I know what objects from my program have been observed?
You will be able to check on the progress of your program by logging into the MMT Scheduler at any time. Emails alerting you to observed objects typically are sent the morning following the observation.
Can I be there while my data are acquired?
During the peak of the Covid pandemic, protocols did not allow on-site visiting observers. We can now accommodate visiting guest observers at some times, subject to the MMT schedule, the preferences of the mountain staff, and dorm room availability. If you have a program on the MMT and are interested in sitting in on observing, please contact Susan Wahl, swahl@mmto.org , and your instrument scientist (for Binospec and MMIRS, Benjamin Weiner) to work out a time when you can visit. Due to the fact that the queue is dynamic and weather is changeable, the target schedule changes each day and during the night, so we cannot guarantee that your program will be in the queue on any particular night. However, staff can make some attempt to observe targets of interest if the weather is suitable.
How can I get my raw data? Can I look at it during the run?
Raw data are available at any time from the same catalog page where you uploaded targets. If you have any issues, you can contact the Smithsonian Telescope Data Center (TDC) by emailing tdchelp@cfa.harvard.edu.
Do I reduce the data myself? How do I do that?
For MMIRS and Binospec, there are publicly available pipelines and an accompanying published paper. See https://bitbucket.org/chil_sai/mmirs-pipeline/wiki/Home for the latest pipeline, and https://arxiv.org/abs/1503.07504 for the paper. Most mask and long-slit MMIRS data are reduced by an automated pipeline. Typically reduced data for MMIRS and Binospec will appear on your catalog page within several days. If for some reason your data are not reduced within a few days of observations, you can contact the TDC.
Why didn’t I get all the observations I asked for? Why didn’t I at least get all my priority 1 objects?
There could be a few reasons why you didn’t get all the data that you requested.
Weather. Approximately one-third of scheduled observations are lost due to clouds, high humidity, or other weather factors. This weather loss currently is shared more or less evenly over all the programs in the queue. Wind speed and direction may make it desirable to observe an object that allows pointing away from the wind. This could mean a priority 2 object gets observed before a priority 1 object.
Facilities Issues. A very small percentage of time each semester (typically <2%) is lost to telescope, building, instrument, or other facility problems. While small on the timescale of several months, an 8 hour loss of time could make a significant dent in a queue program over the course of a 7-10 day run.
Faint targets. If you are looking at very faint targets and a test exposure by the queue observer reveals no signal to noise, they may choose to pass on your object (temporarily) until they can receive more information about finding the target. You can plan ahead and submit nearby bright offset stars to the database to help locate your object, along with finder charts. Searching for your object, whether or not it is found, does get charged to your time allocation. MMIRS longslits require finding charts, and if there is any question about whether your MMIRS target is faint, we recommend using an offset star and having the PA run the slit from star to object. In contrast, Binospec sets onto the specified coordinates using the metrology of its guide cameras, so your targets have to be in the Gaia astrometric frame, but finding charts are not needed.
Slit Masks. For MMIRS, there are 9 slots available for multi-slit masks. If the total number of masks requested by all programs for an observing run is greater than 9, a mask change would have to take place during the observing run. This takes a full dawn-to-dusk time period. It may be that a lower priority target requiring a slit mask cannot be observed if there are, for example, weather losses followed by a mask change which removes the mask needed for the observation. Which masks are put in during the change is decided by staff after an evaluation of the progress of the programs up to that point.