- Gemini Home
- Telescopes and Sites
- Science Visitors at Gemini
- Observing With Gemini
- Partner Subscription
- Observing Modes
- Observing Overheads
- Large and Long Programs Announcement of Opportunity
- Proposal Submission
- 2014A Call For Proposals
- Phase II and S/W Tools
- Changing Approved Programs
- Advice for Band 3 Programs
- What to expect
- Telescope Time Charging
- Future Instrumentation
- Queue and Schedules
- Data and Results
- Image Library
Change page style:
The STAC has approved a remote eavesdropping observing mode, starting in Semester 2013A on completion of a trial late in 2012B. This document lists the main features of the mode as implemented for tests done in 2012B, with the intention to implement more fully in 2013A, and complete in all respects by 13B. PI Procedures for 2013A are detailed at the end of this page.
- Added value only. All observations must be executable without eavesdropping, no observation should rely on it
- Initiated by the observatory (advance warning of probability in the queue case, or definite date in the classical case, plus on-the-night notification by some method which would convey a password, probably one changed on an hourly basis)
- Ability to see, and comment on, target acquisitions
- Ability to assess data quality quickly enough to allow useful changes to be made in the observing
- Ability to understand what is happening in the observing at any given time
- Ability to see: sequence excecutor
- Ability to see: data reduction pipeline products
- Ability to see: acquisition
- Ability to request modifications to observation (change exposure time, re-arrange steps, change wavelength, add a filter, skip a previously included filter, etc). Must not significantly change the nightly queue plan in a way that the observer would not have done and must not redefine the observation in a way which might duplicate some other observation in the database. (guidelines will be needed for the observer in these cases)
- Ability to communicate: with observer (by phone, skype to a single observer account, or Polycom from a defined remote ops centre)
- Ability to see: TSD and/or ISD (this is non-essential)
- Ability to interact with raw data for their project (and associated calibrations) stored at the observator
Eavesdropping IS NOT:
- Ability to override: queue decisions taken by the summit staff and/or QC. All queue choices will be made in the same way as normal. This includes rapid Targets of Opportunity, which will continue to interrupt the current observation even if that observation is being eavesdropped upon. In a marginal case, e.g. very thin cirrus, and where there is no direct impact on higher-priority programmes, the PI may be asked if they wish to continue. If they say yes, despite the weather conditions being outside their formal requirements, a “PASS” flag will be placed on the data in the GSA.
- Ability to directly execute (start, stop etc.) the sequence
- Ability to directly control the instrument
- Ability to see the summit OT (because of confidentiality of other programmes)
- Ability of PI to communicate with QC (the observer will do this in case of need)
- Allowed to result in additional staff hours at the observatory. In practice this means that it will not be used in classical runs unless the classical programme has already sent the PI or a Co-I to the summit.
- Allowed to result in reduced efficiency – emphasis remains on pre-preparation of observations. No preparation is done on the fly at night, and all time spent inspecting acquisition images etc. is charged to the PI’s programme.
- Allowed to produce additional work on QCs or staff queue observers
- Applied to short observations (this is to avoid the overhead of contacting the PI, getting the eavesdropping session established etc. for an observation that may only take 15 minutes).
Eavesdropping is requested in their technical case by the PI. While it is not essential that the program would benefit directly from eavesdropping, if a program has complex acquisitions/fields, particular s/n requirements and uncertainty in target brightness, that should be stated because programs with such constraints would at least benefit from eavesdropping. Whether to allow eavesdropping is recommended case by case by NTACs, and subsequently assigned case by case by Gemini in discussion with the NGOs (acceptance criteria will be needed, both for the NTACs to recommend programmes and for Gemini and the NGOs to accept them; it is likely that there will be a maximum number of programmes allocated eavesdropping in a given month). Because the eavesdropping process will impact directly on summit observers, the final decision on allocations rests with the Observatory. In most cases, eavesdropping will be allocated to programmes using defined criteria before the semester begins, not in the course of the semester.
During the semester, QCs keep track of projects which have eavesdropping allocated and on a two-weekly basis (at quarter moon) send emails to PIs stating that their targets are available and requesting confirmation that they wish to eavesdrop.
On a given night, should the program come to the top of the queue and be identified in the plan as an eavesdropping program for the night, the PI or designated contact receives a phone call stating that observations are about to begin. They open skype a browser, connect to the summit and to a FITS storage sercer, and take part in the observation, including watching and commenting on acquisitions, inspecting data frames and commenting on whether to continue or terminate an observation. Note that the QCs will not always schedule a program for eavesdropping on a given night, even though it has that "status". For example, they may be guided by the weather forecast and we will also attempt to avoid scheduling sessions back to back to give the observer a breather. So data may be taken on an eavesdropping program without a call being made (this is also consistent with observers having discretion over calling, and with the first point in "IS", above).
To avoid confidentiality issues and debates between PIs In the course of any given night, only one programme will be invited to eavesdrop at a time. To ensure quick decision making and as easy as possible a technical interface, only one eavesdropping channel will be open on a given programme at a time.
In all cases, written records are kept (in the night log) of the key points of the dialogue between the observer and the PI, and the PI will have one day to confirm that these notes are correct, or request alterations. PIs will be asked to fill out a short report form upon completion of their session or sessions, to identify possible improvements.
Modifications resulting from the 2012B Test
Two main areas were found to need changes:
- the QC effort in contacting PIs and flagging programs for the night observers
- the wish of PIs to have more than one contact (either multiple phone numbers on different dates or multiple co-Is covering different dates)
Therefore in 2013A we are switching to a "PI self-service" mode in which all Band 1 PIs will be contacted and invited to take part; if they choose to do so then they will be directed to a google spreadsheet in which they can enter their range of dates for multiple contact numbers. Observers at the summit will check this spreadsheet to see if a PI has an available contact for the current night. There will be no twice-monthly email. If Band 1 demand is too low, PIs of programs in Band 2 will be conacted in the same way.
PI Procedures for 2013A
You will receive an email asking if you wish to take part in the 2013A eavesdropping (or wish to designate a Co-I to do so). You will be given 10 days to respond; the sooner you do so, the sooner any eavesdropping session can take place. If you respond that you wish to take part, you will be asked to send the contact details for up to three eavesdroppers:
- email address
- telephone number
Note that you can designate up to three sets of details; these may be used to give three different contact numbers for yourself, to be used at different times in the semester, three different co-Is to cover different ranges of nights,or some combination of these. We use these to set up our dial-up calls that will be used to make initial contact on any given night. The three contacts are labelled "a", "b" and "c" respectively; please refer to them as such in your emailed response.
Specifying contact availability
You will receive a link to a google spreadsheet in the initial email. This is where you tell us who is available and when. The data entry sheet looks like the following:
All you need to do is locate the "Availability" tab (bottom of the browser window) for the telescope you wish to eavesdrop at, and enter "a", "b" or "c" into each date on which your contacts are available (only one contact per night). Please be sure to consider the target observability and enter contacts, if possible, at the earliest possible date - your program may well be observed as soon as the targets become available.