Waimea, Hawaii, 6-8 November 2007

Recommendation #1 Review of systems
Recommendation #2 WIRCam edge crosstalk
Recommendation #3 Data distribution
Recommendation #4 Data Assessment in future Large Programs
Recommendation #5 Material change in strategy of a Large Program
Recommendation #6 Extra observing time for recovery from failure for CFHTLS
Recommendation #7 Photometric calibration of the Wide component of the CFHTLS
Recommendation #8 Time allocation for future Large Programs
Recommendation #9 New instrumentation beyond 2012

The meeting was attended by SAC members Pierre-Alain Duc (vice-chair), L. Ferrarese, C. Ftaclas, J. F. Gonzalez, C. Gry, H. Hoekstra, R. Jedicke, Dae-Sik Moon, D. Mourard and N. St-Louis (chair). Jeremy Lim and Shiang-Yu Wang, from the Academia Sinica Institute of Astronomy and Astrophysics in Taiwan also attended the meeting. The CFHT director C. Veillet, D. Salmon, Kanoa Withington, Sarah Gajadhar, Kevin Ho, Loic Albert, Nadine Manset, Jean-Charles Cuillandre (for CFHTLS), Olivier Lai, Chris Pritchett (for SNLS) and P. Martin participated in presentations and discussions.

1. Technical Activities Report

SAC suggests the following development priorities:

A Elixir WIRCam/WIRCam*
A MegaPrime*
A Elixir – ESPaDOnS / Queue ESPaDOnS
C Observatory Automation
C ESPaDOnS from Prime

Derrick Salmon reported on the technical activities of this semester.

  • Following last fall’s earthquake, there was some concern about the centering of the dome bogies. After a study with a consulting firm, the observed decentering was judged acceptable.

  • The week-long planned shutdown to re-aluminize the primary mirror and perform various other tasks in August 2007 was very successful.

  • The site restoration including the lightning protection grounding grid is now completed. After the installation of the underground conduit for OHANA, the next step is to install the thermally-insulated above-ground conduit which has now arrived on the island and should be installed before the end of the year.

  • A new humidifier for the computer room has been ordered and the MegaCam-WIRCam chillers are being relocated in the basement to avoid sending heat in the building.

2. Instrument and Project Reports

· MegaPrime

Sarah Gajadhar described the two major failures suffered by MegaCam last spring and how they were addressed jointly by CFHT staff and a team from CEA.

First, she explained why the shutter and jukebox came to collide. Basically, to prevent interaction between these two entities, which share a common volume, a pair of switches had been installed. In the early days of MegaPrime, one of the switches failed (the ‘shutter-init’ switch) and was hard wired since extensive disassembly would have been required to reach it. The incident occurred when the other switch failed (the ‘shutter-closed’ switch). Essentially, the staff thought that the shutter was in a closed position and that is was thus safe to move the jukebox up, which was not the case. A full camera re- initialization was done which drove the jukebox up in the shutter.

The shutter was replaced by a spare and this constituted the bulk of the recovery work, essentially because of the flatness requirement and the delicate nature of the material out of which it is made. Naturally, both switches were put back into service but also more robust redundancy was implemented in the camera controller logic to prevent similar failures in the future.

Then, she described how the i’ filter was destroyed when the MegaCam filter clamping latch system failed. The jukebox and filter slide rails were also damaged in the incident. The culprit was identified as the filter ID tag system (or software) which mistakenly reported 8 filters in the jukebox when in fact only 7 filters were in the jukebox and one (the i’ filter) was still in the beam. When the initialisation was started after a failure, the jukebox moved down and the filter arm moved to its out-of beam position (ready for loading). This caused the i’ filter to be pulled out of the beam and released in a position partly in the camera filter slide, above the jukebox. When the jukebox moved up during the next initialisation step, it collided with the protruding filter.

The filter and jukebox slide rails were machined in house or refurbished and installed in the camera. A new i’ filter was ordered and put back into service in October. The reliability of the ID tag/reader system was improved but an improved radio-frequency filter ID tag system is being sought.

As the recovery from these two major failures involved a major disassembly and inspection of the camera, CFHT staff, with the agreement of CEA staff itself, decided not to carry out the annual maintenance disassembly planned for September. SAC agrees that this was a reasonable decision.

During these events, the collaboration between CFHT and CEA staff was reported as effective and constructive. SAC is pleased with this situation and encourages a continued effort in this direction. We were also relieved to hear that the rotating valve problem seems to be solved with the manufacturing of a new version by CEA and delivery in October.

SAC congratulates CFHT and CEA for their excellent work during the recovery from these two major MegaCam failures.

SAC Recommendation #1 -- Review of systems

In view of the two recent major MegaCam failures, SAC recommends that systems and procedures be reviewed for all instruments to minimise the chances that this type of failure happen again.

Back to top

Work continues on the SNDICE (SuperNova Direct Illumination Calibration Experiment) project, which is aimed at improving the photometric accuracy of MegaPrime data. CFHT staff are working together with the SNDICE team to install the calibrated photodiode on the camembert in from of the dewar window.

· WIRCam

Loic Albert presented the many improvements that have been achieved this semester for WIRCam.

First, the 1.0 version of the ‘I’iwi data-reduction pipeline has been released and pre-processing of backlogged data from O6B and O7A has been executed and is continuing for the O5B and O6A semesters. The arrival of new manpower in the form of the recently-hired Canadian RA, Daniel Devost, was a great help in this regard. A lot of data has now been distributed to PIs and Terapix has now received images to be stacked, with a priority given to large programs.

Included in the new pipeline is the non-linearity correction, the reference pixel correction (bias), dark-exposure subtraction, dome flat-fielding, more robust bad-pixel masking, guide-window crosstalk masking and improved sky subtraction. In addition, the ‘I’iwi astrometry is much more robust as it uses the Terapix-produced ‘WIRCam.ahead’ file which describes the position of the WIRCam detectors based on actual observations. The solution is later tweaked detector by detector. Finally, the zero-point has been determined for all filters.

Another improvement that was made this semester is that the WIRCam cross-talk has been greatly reduced. The negative cross-talk was eliminated with a new wiring configuration of the detector controller and the positive cross-talk was eliminated with the replacement of specific video boards in the SDSU controllers. Only the edge cross-talk (which is much smaller than the two mentioned above) remains and is still a mystery. It is not present on the engineering chip. A software compensation has been implemented for now.

SAC Recommendation #2 -- WIRCam edge crosstalk

SAC recommends that efforts be continued to identify a hardware solution to reduce the edge crosstalk in WIRCam.

Back to top


Nadine Manset reported that this semester, most efforts were aimed at the integration of ESPaDOnS within the QSO mode. The entire chain of operation has been tested very successfully on the sky in late September. Although improvements are still required, all the QSO tools (phase 2 tool, queue preparation tool, observing tool and the electronic logbook) are functional. There should be no problem in using ESPaDOnS in QSO starting in 2008A, as advertised. However, it is expected that observing in QSO with ESPaDOnS will require much more real-time interaction with the PI compared with MegaCam and WIRCam. For spectroscopy, there are more complicated issues that need to be addressed compared to imagery. Also, the observer has to make sure he or she understands well the overall strategy that the PI wishes to adopt.

Some work has also been done on a new focus process and is reported as promising. Some minor work on improving the guiding techniques has also been carried out.

A lot of progress has also been achieved on the software side. First, concerning the issue of making ESPaDOnS work under the New Environment for Observing (NEO) framework, a new command scheme has been developed that is compatible with the software used for MegaCam and WIRCam. These will be used instead of the scripts already provided to control the instrument.

For the data reduction, in addition to the Libre-ESpRIT software, a quick-look software based on ds9 has been prepared (ds9espadons). It can be used during the night to view raw images and spectra. It has been tested by the CFHT staff and is working well. For the final data reduction, a software package called Upena had been developed. Before data are reduced by CFHT staff, consultation with the PI will be made to insure that it is tailored to his needs.

These routines are based on stable versions of the Libre-ESpRIT components. If there is a need to introduce changes to the latter, these will be discussed with the CFHT staff so they are introduced correctly in the entire chain of observing and data reduction.

SAC is satisfied that CFHT has gained sufficient control on the Libre-ESpRIT software to guarantee smooth operation of ESPaDOnS in QSO mode.

As for the cross-talk reduction down to a level of ~1%, not much work has been done on this issue this semester but the manufacturer of the triplet lens (which seems to be causing the cross-talk) are looking into the problem. Another possibility is to seek another manufacturer for the lens.


Olivier Lai presented an update on the status of the Ohana project. He informed SAC that the team is getting ready for an observing run on the Keck telescopes. During this run, studies related to the vibrations of the fibres and to the rotation of the polarisation vectors as they move down the fibres will be carried out in addition to science observations of YSOs and AGNs.

The next step in the Ohana project is a link between the CFHT and Gemini telescopes. Soon, all the conduits that will hold the fibres will be in place and the baseline will be measured next Spring. The Delay lines have already been laid in the CFHT Coudé room and soon flux and delay measurements will be made (in the IR) followed by auto- correlation and fringes. Once this is achieved, plans will be made to create a link with Gemini. On a much longer timescale, there are plans to link up with Subaru and UKIRT.

· FlyEyes

SAC was informed that FlyEyes saw its first light (loop closed on stars as faint as m=14.8) at the end of April 2007 during a 3-night engineering run. Work is continuing on reducing the readout noise and on the software communication interface between FlyEyes and the AOB.


Work on the seismometers and the two-color experiment for VASAO is continuing.

The seismometer was mounted on the caisson centrale and preliminary vibration measurements of the declination axis while the hour-angle axis was held fixed were made last spring. They were found to be small.

The hardware and software to carry out the two-colour experiment that will take place in November is being prepared and is nearly ready. In this experiment, two images of a bright star from the AO Bonnette are recorded on two CCDs (one operating at 400 nm and the other at 900 nm). This should allow determining at which precision and under which conditions the atmospheric tip/tilt component of the atmosphere can be obtained by measuring the separation of the two colours.

· DIMM, Skyprobe and IR AllSky camera

SAC was informed that the DIMM software (acquisition and guiding) has been tested at headquarters. This is working well. Full automation remains to be completed. On the hardware side, the CCD will soon be replaced with one capable of millisecond exposure times. The Skyprobe is now functional and has been gathering data for a month. CFHT is keeping informed on the progress of the IR AllSky camera being developed for Large Synoptic Survey Telescope. There are also plans to have a facility-wide seeing monitor, to be located on CFHT’s weather tower. This is still awaiting for approval from the Office of Mauna Kea Management.

· Telescope automation

A report was presented on potential gains and costs of the automation of the telescope. The potential benefits to be gained were described. These include, among others, improved reliability, reduced observing costs and improved remote diagnostics. The challenges to be met were also examined. These involve issues such as resources, implementation, changes in operation and logistic and finally safety. If, for example, at some point laser guide stars are in operation at the summit, the automation will become more difficult safety-wise. An update on the status of the project was also given. At the present time, in addition to the fact that the telescope is almost completely used in QSO mode, the staff has gathered all the required information on the technical requirements and have worked on generating high-level observing scenarios. They understand the scope of the project and have toured other facilities. The have now a good estimate of the overall effort that is needed in terms of resources, cost and time and have started the conceptual design process. It is estimated that 2 years will needed for implementation.

SAC is satisfied that there are sufficient reasons for examining this possibility in more detail. We recognise that even without remote observations, there already have been many benefits to the automation process. SAC would like to continue to receive a report on the progress towards this goal once a year at its fall meeting.

SAC Recommendation #3 – Data distribution

Given the work and costs involved, SAC recommends that CFHT eliminate the option of distributing user data on DLT tapes.

Back to top

3. CFHT – LS

· Progress report

The steering group chair, Jean-Charles Cuillandre, presented the status report of the Legacy Survey including science highlights for each component of the survey. Observations are continuing at a steady pace, thanks to favourable weather. Following the destruction of the i’ filter, the observing strategies of the SNLS and Wide components were adapted which resulted in no lost time on the sky. However, for proper calibration of the supernovae candidates to be observed in the new i’ filter, new calibration data will be required.

· Status

A detailed plan of the 2008A and 2008B time allocation for the Legacy Survey was presented by the Steering Group chair. A grand total of 228 hours in 2008A (below what is normally allocated) and 99 hours in 2008B are requested. In addition to normal, approved observations, this includes additional requests for the following elements.


The steering group is presently assessing the need for additional observing time to recalibrate the new i’ filter for the SNLS observations and are informing SAC that an additional 10 hours in 2008A (D1 field) and 30 hours in 2008B (D2, D3 and D4 fields) in good seeing may be needed to obtain references for calibration of the supernovae candidates observed with the new i' filter.

Patching of the Wide

In order to insure that there are no holes in the coverage of the Wide, The steering group is requesting additional 30 hours of observing time for patching (15 hours in 2008A and 15 hours in 2008B).

The incomplete field coverage is due to the recurring problem of data being rejected by Terapix for not complying with the required seeing conditions, in spite of having been validated by Elixir. In its spring report, the SAC asked the steering group to provide a justification of why the Terapix and Elixir seeing measurements are consistently and systematically different. At the meeting, the steering group chair informed the SAC that the metrics by which the image quality is estimated by Terapix and Elixir are different, but the SAC was still unclear as to what exactly causes the difference, which measurement is more reliable, and how seriously this impacts the analysis of the data. Although SAC was still not satisfied with the justification, a decision was made to take the conservative approach and grant the requested 15h for patching of the wide in 2008A, rather than incurring the risk of compromising the legacy value of the survey.

Photometric calibration of the Wide

In order to bring the Wide component of the LS to a similar calibration level as the Deep, the steering group is requesting an additional 14 hours (10 hours in 2008A and 4 hours in 2008B) of observing time. They claim that because the observations have been obtained in a diversity of instrumental setups over the years, it is crucial to bring all data to a common reference. They claim that the data will allow a final optimisation of the overall photometric calibration of the wide and allow them to reach the 1% accuracy level.

Following these various additional requests, SAC has the following recommendations.

SAC Recommendation #4 – Data Assessment in future Large Programs

SAC recommends that for the upcoming Large Programs, a time limit is set, from the time the data are taken, within which the Program PI must assess the data quality, and report all problems to CFHT, if any. A standard form should be made available by CFHT and be used to report all such problems.

Back to top

SAC Recommendation #5 – Material change in strategy of a Large Program

SAC recommends that for the upcoming Large Programs, any change to the original allocation should be justified by a formal proposal to the SAC. Such changes might include, for instance, changes in the observing times or filter selection, even if prompted by instrument failures/malfunctioning or calibration issues. All such changes should be approved by SAC.

Back to top

SAC Recommendation #6 – Extra observing time for recovery from failure for CFHTLS

SAC recommends that the 10 hours in 2008A be awarded and that the 30 hours in 2008B be awarded if needed and based on a justification to be submitted by the CFHTLS Steering Group at the May 2008 meeting of the SAC.

Back to top

SAC Recommendation #7 – Photometric calibration of the Wide component of the CFHTLS

SAC recommends that this time be awarded.

Back to top

· Terapix activities

SAC was informed that since our last meeting, the T0004 dataset has been released to the LS community with a slight delay compared the planned date (July instead of May). This release includes all data for the Deep and Wide components from May 2004 to October 2006. The next release, T0005, is planned for May 1st 2008 and will contain CFHTLS data up to the end of 2007B. The Very Wide data (1st official release) is planned to be released 2 months later as recommended by SAC one year ago (within one year of the release of T0004). SAC believes that these slight delays are acceptable but encourages Terapix to attempt to comply with the planned dates as much as possible.


A report from CADC on the usage of CFHTLS data was presented. Two modes of data retrieval are available. The first, the Query mode, is when a user fills in a form to query the archive dataset. And the second is when a user knows the identifier of the dataset and can access the file directly without filling in a form (the API mode). The latter mode is increasingly popular. The CFHTLS dataset constitutes about 50% of the total CFHT data served by CADC and more files are accessed every year. 90% of the retrieved data are for ‘Elixir’ processed data although request for Terapix Stacks are increasing every year. When the stacks become available to the world community, the number of request approximately doubles. CADC is also providing access to CFHTLS stacks developed by Stephen Gwyn, which are made of all publicly available CFHT MEGAPRIME images.

4. Telescope Use and Scheduling

· DDT time

Christian Veillet prepared and presented a report on the usage of the 10 nights of the Directors Discretionary Time (DDT) at CFHT. Some 10 hours were awarded for various targets of opportunities and urgent requests. Once again, a variety of dynamic research programs are being carried out by the CFHT scientific staff. The projects span a wide variety of subjects from brown dwarfs to galaxy clusters and are carried out using all CFHT main instrumentation. SAC finds this report very interesting and would like to see them continued to be prepared once a year.

· QSO Report

Pierre Martin presented the QSO report. The weather during 2007A was better than usual for a winter semester, which alleviated somewhat the time lost due to the two major failures with MegaCam. However, the flexibility of queue observing allowed to quickly change to WIRCam programs which resulted in more than excellent statistics for that instruments: 100% of all programs (A+B+C) were completed. Some 2007B programs were even started. For MegaCam, 10 nights less than usual were available and the loss of thei’filter made the situation difficult due to the fact that after the repair the camera had to be used even though the moon was bright. In spite of all this about 85% of A+B programs with MegaCam were completed. The global validation rate is still excellent at 92%.

The total amount of time lost to weather and engineering this semester was 35% for MegaCam and 23% for WIRCam.

Operational overheads for MegaCam are stable at a level of about 10-15%. This is not expected to become lower as they are limited by filter changes. In any case, this is the level that was expected before observations started with MegaCam. For WIRCam, the overheads are still dominated by the two-step focus sequence, guiding acquisition, pointing corrections and dithering patterns. The implementation of an automated focus model saves about 30 minutes per night, which reduces the overheads to about 20% of a night. Efforts are still being made towards reducing this number.

Following a question from SAC, we were informed that, at the present time, the length of a night is considered as follows: 6h-long for MegaCam and 6.5h-long for WIRCam as the nights are longer in the IR and the observing overheads are shorter. For ESPaDOnS, a value of 7h is adopted for the 1st semester of operation in QSO mode (2008A) but there it is hoped that this can be increased to 8h in the following semester.

5. Future of CFHT

· CFHT’s Golden age.

The CFHT director informed SAC that he will be presenting to the Board a new plan with an extension up to 2012 instead of 2010. The present agreement with Taiwan, which is valid until 2010, could be renewed. In addition, SAC was informed that a total of 5 nights per semester for 4 semesters (20 nights in total) are to be opened to the US community starting in 2008B. The proposal will be submitted to the Telescope System Instrumentation Program (TSIP). The observing would then be attributed to observers via the regular NOAO time allocation process. Finally, we were informed that options to finance new instrumentation beyond 2012 are being looked in.

Some observatory metrics were briefly mentioned. As of early November, a total of 75 papers based on CFHT data have already been published. Due to the technical problems with MegaCam this semester, it is clear that the goal of 4% for clear sky time lost to failures per run is not going to be met. However, regarding staff turnover, the numbers are very good, since none of the 40 staff left this year.

· Large programs in 2008B-2012B

A total of 15 Letters of Intent were submitted for possible Large Programs to be carried out in the period 2008B-2012B for a total of 1571 hours. 9% of these are to use WIRCam, 19% for ESPaDOnS observations and 72 % for MegaCam projects. We note that the Letters of Intent are mandatory and that no other Large Programs can be submitted.

SAC is pleased with this enthusiastic response. Following our recommendation number 12 from last spring’s 71st meeting and based on the proposed science, a suggested LPTAC membership has been submitted to the Board.

SAC Recommendation #8 – Time allocation for future Large Programs

SAC recommends that time allocation for future Large Programs be made in hours.

Back to top

Details of the proposal

We suggest that the proposals to be submitted by 1st February 2008 should be about 12 pages long and should contain the following elements:

  • A Scientific Justification – 5 pages long + 2 pages of figures and references
  • A Technical Justification – 3 pages long
  • The Observing Strategy – in the form of a standard table prepared at CFHT
  • A Data Management plan – including real-time analysis requirements, a data-analysis plan and a description of available resources – 1 page long.

External referees

In view of the proposed names for the LPTAC listed above, we feel that external referees are no longer necessary. However, if the Board decides otherwise, we suggest that referees should be requested for all proposals and PIs should be asked for a list of people they wish to exclude as referees.

· New instrument feasibility studies for future instrumentation

In response to the call for feasibility studies for future CFHT instrumentation, a total of 4 proposals were received. All appeared reasonable to SAC and could constitute viable options, some in the nearer future others on a longer timeframe.

The CFHT director has informed SAC that sufficient funds are available to carry out these four feasibility studies for future instruments. We note that once delivered, these feasibility studies will be the propriety of the CFHT Corporation.

SAC Recommendation #9 – New instrumentation beyond 2012

SAC recommends that the four feasibility studies be funded.

Back to top

6. Other items

· Next meeting

In order to facilitate travel to the LPTAC meeting which should be held just before the next SAC meeting, in a central location in order to improves chances to attract potential LPTAC members, we suggest that the next SAC meeting be held in Toronto, Canada on the 15 and 16 May 2008. We suggest that the LPTAC be held in the same location on the 13 and 14 May 2008.

Back to top