MagE Commissioning

MagE Commissioning

MagE Commissioning, 13-28 November 2007 OCIW Staff present: Christoph Birk, Scott Burles, Jennifer Marshall, and Ian Thompson

Summary Over the past two weeks we have successfully deployed MagE on the Magellan Clay telescope. Thanks to the mountain staff for their help and support!

Details of the MagE commissioning are given below. A complete description of the instrument for observers is available in the User Manual.

Technical details about the instrument including instructions for maintenance, datasheets for components, and mechanical drawings of the instrument are found under “Technical Documentation” on the main MagE

Please send any comments or requests to Ian Thompson or Jen Marshall.

October trip

On a previous trip to the mountain (12-19 October), Scott Burles and Jennifer Marshall unpacked the MagE instrument and returned the instrument to the state it was in before being shipped from Carnegie.
We successfully assembled the instrument and the camera, pumped and cooled the camera, focused and aligned the spectrograph, and weighed the instrument. The instrument was balanced for motions in elevation.
The dewar was left on the pump in the auxiliary building for most of the next three weeks. Patricio Jones worked on the electronics box and PLC programming during this time.

The new cable wrap was installed and tested during this period.

Installing MagE on the Clay AUX2 folded port

Upon our return to the mountain, we finished the cabling on the instrument. Note that some of the cables on the instrument are tied to the instrument cover. These will have to be cut and moved in order to open the instrument, however, the instrument should not need to be opened very often. We also noted that the vacuum level of the camera while on the pump was acceptable (~1e-7 mbar). We held an extensive walk through of the instrument and its capabilities for the mountain technical crew, and a second the following week for the other shift.

On 20 November we installed MagE on the telescope, turned on the ion pump, and cooled the instrument. Everything functioned as expected.
The glycol coolant connections at the MagE junction plate leaked, the leak was promptly fixed. MagE was removed from the telescope to  check on the FP2 collimation. MagE was reinstalled on the telescope on 23 November.

Telescope Balance

On the night of 20 November we had some problems that seemed to be associated with the new balance of the telescope under moderate wind loading. The telescope was rebalanced on 25 November and no further
balance problems were noticed for the rest of the run.


The MagE GUI has been installed and may be used from either observer computer. The GUI is similar to the MIKE and B&C GUIs, with the exception of a new feature that indicates the wavelength and order number at the position of the cursor in the Quicklook tool. The GUI is described in detail in the MagE manual. The MagE program has been  distributed to all four CCD computers in the Clay equipment room.
One of the two spare CCD computers in Baade will be moved to Clay in January. (This was not done now because the way the CCD servers are launched on these computers is being changed by Christoph Birk and
Jim Hughes. The rest of the computers will have the new code installed in January.

AUX2 port

We noticed that when M3 was first moved to the AUX2 port the stars had large tails. Dave Osip worked on this problem over several nights and determined that there is some sort of problem with the offsets in the
Shack-Hartmann, i.e., it is not a problem with the guide cameras themselves. Once S-H is turned on, the correction seems to take care of the problem across the field. We did not notice any image quality problems with MagE, and measured image FWHMs through the slit that were consistent with the seeing.


The MagE slitviewer has good image quality over a field of 120 arcsec.
Operation of the “sky” and “sub” commands on the guiders does not work because of a bug in the code for full frame, 2×2 binned images.
Jorge Estrada and Steve Shectman have tracked down the problem,  and the new code (available from Jorge) should be installed. This will help locating faint objects on the slit.

On the night of 26 November we had some problems with the appearance of the images on the slitviewer camera. A large halo appeared on the image and light from bright stars scattered across the image. In the
morning we determined that the cause of the halo was condensation on the camera and slitviewing optics. Dave Osip is working on correcting this problem.

We also determined the center of the slitviewer camera in pixels.
This number will have to be remeasured before the next observing run since the camera has been removed and reinstalled.

Flexure tests

We did a series of flexure tests on the instrument. These are written up in more detail in the online manual. In brief, there is some flexure in the instrument, particularly when moving in elevation. We measured 0.7 pixels of linear flexure (mainly along the spatial direction) when moving from the zenith to an elevation of 20 degrees. The amount of flexure seemed to depend on the rotation of the instrument on the port: at its nominal home position (+15 degrees) flexure was small, while the most flexure was measured with the instrument at +/-180 degrees. There is also some flexure between the instrument and the calibration system. However, the amount of flexure should not be a major issue for most observations, especially if the instrument is used at the parallactic angle.


We measured the MagE throughput using several standard stars and the 5″ slit. The peak throughput of the instrument including the telescope is 22% at about 5600 Angstroms. The complete set of throughput measurements is available in the MagE manual.

Science Verification Observations

We took science data on a wide collection of objects, including supernovae, a DLA quasar, low-metallicity stars, massive binaries, extreme subdwarfs,  and a star forming galaxy. These data will be used to help write the data reduction pipeline.

Outstanding issues

We are in the process of finishing up the online manual. This will be completed before the first MagE observing run in January (shared risk observing by Paul Schechter on January 5 – 7).

We are in the process of determining a procedure for what to do with MagE in the event of a power loss. The ion pump in MagE is required to maintain the vacuum in the dewar, and if turned off for more than
about a day the dewar may need to be pumped. The procedure in case of a power outage will be added to the Technical Documentation section of the manual.

A script to measure the readnoise and gain of the CCD is being added to a MAGE subfolder of the IRAF distribution on the observers workstations.

We still need to write a script to focus the spectrograph. For now, the MIKE focusing script may be used to focus MagE.

Work will start soon on a data reduction pipeline, and we hope that there will be a quick look reduction package available at the telescope. Recommendations on how to take flat field and fringe removal data will follow as we gain more familiarity with MagE science data.

The MagE team will return to the mountain to observe with MagE and to finish any remaining tasks in late January, a science observing run is scheduled for January 31 – February 4.

Table of Contents