News Highlights

News Items


CSU Fatal Errors (2023-Sep-20)
MOSFIRE is experiencing more CSU errors recently. Our mitigation strategies are to move the CSU at EL~67 whenever possible and to minimize (if possible) CSU moves. One option is to take science data using the alignment mask which removes the need to move the bars from alignment to science configuration. This is obviously an impact on the science, so it is up to the observers to decide if the tradeoff is worth it. We don't have great statistics on the fatal errors as they are still a rare occurrence, but in the limited set of incidents we have seen, doing this at low elevation (e.g. EL~35) seems to be problematic. Thus if observers are acquiring a mask at low elevation, we would encourage to either take data in alignment mask, or to minimize the number of alignment stars in the mask to minimize the number of bars moving. If observers chose to carry out observations with alignment masks, it is worth to design science masks exactly the sam as alignment masks, that way there is no need to take separate calibration sets for alignment manually.

From small number statistics that we gathered in August, CSU errors tend to happen at cos(elevation)*sin(instrument drive angle)>0.4 and cos(elevation)*cos(instrument drive angle)<-0.4. These values correspond to large rX and Y accelerometer values (>6000), where gravity pulls are higher. Avoiding those combinations of elevation and instrument drive angle may help reduce the occurrence of CSU errors.


Guider Update (2020-Feb-04)
The MOSFIRE guider had been experiencing a failure of one of the amplifiers over the last few months which resulted in the guider field being reduced to half of the normal field of view. We have replaced the detector in the guider with a spare which has re-enabled the full field of view of the guider. This change however, has resulted in some fringing effects due to the cover glass on the replacement chip. This should have no effect in most fields, but extremely faint guide stars may be impacted. We are currently looking to replace the detector with one that does not suffer from fringing, but this may take some time. A screenshot of the fringing effect can be seen below.

Example of finging


MAGMA Localization (2019-Jan-24)
It has been brought to our attention that there is a possible point of confusion when using MAGMA with certain localization options set on your computer. MAGMA uses the decimal separator specified by the settings on the computer (. in the US, but which may be , for others), but the example input catalog files use the US standard of the decimal separator. MAGMA users which have European-style localization, should be aware of this possible point of confusion.

MOSFIRE DRP Version 2018 Released (2018-Jan-05)
The 2018 release of the MOSFIRE DRP is out. This new version includes an easier installation process, python3 compatibility, and an updated documentation web page, along with numerous other bugfixes and updates. Check the online documentation for installation instructions.

MOSFIRE back on-sky and CSU operating normally (2017-Mar-05)
MOSFIRE is on sky, and taking science data as normal. All previous restrictions on CSU motions are no longer an issue. Observers should still be aware, however, avoid configuring the masks which have physical drive angles of 0 or 180 degrees.

MOSFIRE back on-sky and taking data (2017-Feb-13)
MOSFIRE was returned to science operations on 12 Feb 2017. The performance of MOSFIRE so far appears to be the same as it ever was. Due to weather, we were unable to complete all of the engineering tests we wanted finish, and plan to complete those at a later date. We expect that the throughput and sensitivity of MOSFIRE has not changed.

There is an operational concern with MOSFIRE that observers should watch when observing. The CSU bar positioning is not as reliable and we are finding more bars out of position in images and spectra. Observers should be mindful to check their images and spectra to ensure bars are okay. Please work with your SA to help monitor the bar positions and correct the issue in real time.

For the time being, we have modified the default rotator angle for acquiring flats to be -10 degrees on the drive angle (formerly -90). Observer should complete major CSU configurations at physical drive angles of 10, -10, -170 or -190 to avoid major CSU faults.


MOSFIRE DRP Release (2016-Dec-12)
An updated version of the MOSFIRE data reduction package has been released. This version removes IRAF as a dependency and replaced it with astropy and ccdproc python packages. More information can be found on the MOSFIRE DRP web page.

MOSFIRE Repair Mission Status (2016-Nov-23)
The MOSFIRE repair is planned for completion by late January, 2017 and is expected to be ready for science in the new observing semester beginning February 2017. The MOSFIRE nights currently on the schedule for January 2017 will need to be converted to different instruments and possibly different programs as needed.

The MOSFIRE Dewar was opened in a newly constructed clean room on the dome floor of Keck I in early November. An optical element in the collimator assembly was found to be dislodged from its mount. The optic was shipped back to UCLA for repair and will be reinstalled by early January. The MOSFIRE Dewar will then be closed back up, moved up to the Nasmyth deck, pumped and cooled by mid January where it will undergo daytime testing. An engineering night will be used in early Feb to confirm on-sky performance prior to reinstating science operations later in February.

This plan keeps semester 2017A science programs in tact but we'll be working with the institutions and individual PI's to change the remaining MOSFIRE nights in 2016B (January) to other instruments.


MOSFIRE out of service and unavailable for science operations (2016-Sept-15)
MOSFIRE was removed from science operations due to a serious problem that is present in the optical system. MOSFIRE will not be available for science use in October, and we anticipate the repair mission lasting a few months. MOSFIRE's image quality is too poor to continue science operations and we are currently scheduling repair activities. The degradation in image quality was induced during engineering activities designed to tune MOSFIRE's rotator bearing. This activity resulted in MOSFIRE vibrating and a loss in the optical alignment. At the present time, there are many variables that remain to be constrained to determine where in the optical path the misalignment occurred, and a deliberate approach is being taken to complete a repair. Thank you for your patience during this unfortunate time while we develop our plan and coordinate a repair mission. Future updates will be posted to the news page as the repair mission progresses.

Filter wheel one exhibintingexhibiting a problem (2016-May-25)
Filter wheel 1 has a repeatable failure mode. When we move to or from the J2 filter back to the open position, the filter wheel status goes to unknown. It appears based on the spectra that the filter wheel is in the correct position, but the switches are not read correctly to identify the positon for J2.For most observers, this should not be an issue as filter wheel 2 is the primary wheel that is used. Homing will get the wheel out of trouble, but there are software work arounds we must apply if observers want to use J2 for science. Please consult with you support astronomer if you would like to use a medium band filter like the J2 filter.

Work around for MAGMA saving error (2016-May-18)
A couple of possible solutions to the problem described in the previous news item documenting a MAGMA error are
  1. Access from the menu bar Tools -> Options and deselect the first box "Write HTML file when saving slit configuration."
  2. To run magma offline with Java version 1.6. A handful of observers had had success with this workaround. We are working toward updating MAGMA and will resolve this bug in the next release.

MAGMA experiencing an error on saving (2016-Apr-26)
Observers who have recently updated their MAC OS or JAVA are noticing that MAGMA will not save files correctly. Upon save, the following error is reported: 
"org.xml.sax.SAXParseException;systemId:http://keck.hawaii.edu/mosfire/msc.xsl; 
lineNumber: 1; columnNumber: 50; White spaces are required between publicId and systemId."
In addition, the startup script uses xdpyinfo which is an X11 script used to determine the monitor screen size. X11 is no longer included with the MAC OS as of Jan 2016. These issues do not appear to prohibit saving *.xml, so these errors are not preventing mask design and creation.

DVD writer no longer available for data backups (2016-Jan-26)
The Keck Observatory is phasing out the auto DVD writer sometimes used by observers for data backups. Options to launch the DVD autobackup gui are no longer available from the background menu. Because the Keck Observatory Archive (KOA) now ingests all raw data from every instrument at the observatory, observers are encouraged to use KOA to retrieve a copy of the data following your observing sessions. Instructions and links to possible methods of data backup including KOA, scp, and rsync are found at: BackingUp.

DRP updated (2015-Aug-3) An updated version of the MOSFIRE DRP is released for observer use. For details please see https://keck-datareductionpipelines.github.io/MosfireDRP/ .

This release provides a number of new capabilities and improvements as well as bug fixes. It is also highly enhances the automation of the reduction process. Among the new pipeline capabilities are:

  1. Automatic generation of the driver file
  2. Reduction of longslit observations
  3. reduction of long2pos and long2pos_specphot
  4. detailed log file and log output
  5. K-sigma clipping during wavelength fitting
  6. A bypass mode to autofit for a wavelength solution
  7. Package installation similar to the ureka package
  8. support for Numpy 1.9

MOSFIRE Drift corrected (2015-Apr-29)
MOSFIRE observers detected that objects drifted along the slit over time. With the help of the MOSDEF team and Dr. Chuck Steidel, we determined that the observed drift was due to flexure in the guiding system and differential atmospheric refraction.

Software updates were applied to both MOSFIRE and MAGIQ (the guider). MAGIQ now corrects for the flexure in the system by adjusting the guide star slightly as the instrument rotates and the telescope changes elevation. MOSFIRE software now provides MAGIQ with values for the filter central wavelength and ensures that MAGIQ DAR corrections are enabled.

On sky testing during two nights appeared to show that objects appear to remain in the same location with their assigned slits. FITS headers will also indicate whether DAR is enabled and the central wavelength used for correction.


Telescope Offset Problem fixed (2014-Dec-15)
We have addressed a long standing observatory wide problem where commands to offset or nod the telescope were lost in the command queries to the telescope. In the past, the telescope would occasionally not nod to the desired location, resulting in two consecutive spectra with the objects at exactly the same location spatially in the slit. So a nod sequence may look like this: A,B,A,B,A,B,B,C,B,C,B,C.

To solve this problem we implemented a gateway keyword server. This is relatively transparent to observers. Most importantly, since installing the gateway server in July 2014, we no longer observe a skipped telescope offset. All nod command work as expected. This solution was deployed to all instruments at the observatory.

This issue was initially reported in April 2014 and was a problem for all observations until the 2014B semester.


MOSFIRE DRP now supported by Keck (2014-Dec-15)
The MOSFIRE DRP is now being served and maintained by the Keck Observatory. With this change there are some minor updates in the latest release which should be transparent to DRP users. A new issue tracking system is also implemented to help us better track issues, questions, and concerns. Please see https://mosfire-datareductionpipeline.github.io/MosfireDRP/ .

Drift in spectroscopic exposures (2014-May-15)
Observers are noticing a 1 pixel/hour drift along the slit while acquiring spectroscopic observations with MOSFIRE. The graph below shows the steady motion of a slit star in a mask observed by the MOSDEF team. Over a two hour observing sequence using a 4 point slit dither pattern, the star drifted by 2 pixels.

Working with Dr. Richard Ellis, Matt Schenker was the first to report the drift which was later confirmed by members of the MOSDEF team, Dr. Mariska Kriek in particular.

The drift does not appear to be present in all spectroscopic observations as some observations show a maximum drift of 1 pix per hour while other observations indicate that the star was stationary in the slit. We are currently investigating if the drift is correlated with telescope parameters such as elevations, azimuth, rotator position angle (physical drive angle), etc. We have exonerated the guiding software during engineering completed in mid-May.

The MOSDEF team and Dr. Ellis's observing group monitor the drift by designing their mask with a suitably bright enough star that may be detected in a single nod pair of observations. The star is bright enough that the teams can monitor the position and intensity of the star. Keck observatory suggests that observers who are concerned about slit drift include a relatively bright star comparable to those used for alignment (15-19 mag) in the science mask configuration. The magnitude of the star will depend on the length of individual exposures.

We are in the process of also porting over software used by the teams to help monitor the position of the star in real time so that observers may determine if a re-alignment is necessary.


Telescope Offset Monitoring software released (2014-Apr-14)
On occasion during a spectroscopic nod sequence, the telescope does not nod to the desired location. This will often result in two consecutive spectra with the objects at exactly the same location spatially in the slit. We have recently released software that monitors the requested positions and confirms telescope motion. If a move is suspected of not executing, TKlogger, will pop up a message on your display screen suggesting a course of action.

The typical course of action is to abort the dither sequence and then restart the sequence. However, the software is in its infancy. It is in the observer's best interested to confirm that the telescope did not move by examining the last image and the current image to asses if the targets offset as expected or are at the same location spatially in the slit. Please talk to your SA, if you have questions.


DVD writer for data backups will be phased out (2014-Mar-7)
The Keck Observatory is phasing out the auto DVD writer sometimes used by observers for data backups. Although there is no time line for the removal of the DVD writer, we will no longer troubleshoot or maintain the DVD writer. Because the Keck Observatory Archive (KOA) now ingests all raw data from every instrument at the observatory, observers are encouraged to use KOA to retrieve a copy of the data following your observing sessions. Instructions and links to possible methods of data backup including KOA, scp, and rsync are found at: BackingUp.

MOSFIRE DRP updated (2014-Mar-7)
Dr. Nick Konidaris has updated the DRP code to improve maintenance, eradicate some persistent bugs, and address some problems that observers like you have documented either via e-mail or at the the google docs repository forum.

The updated pipeline may be found at: http://www.astro.caltech.edu/~npk/mosdrp_releases/

Major changes include:


Flat Lamps Updated (2014-Feb-13)
New flatfielding lamps installed atop the HIRES and AO enclosures in Keck I replace the previous MOSFIRE dome lamp attached to the Nasmyth Deck. The old flat lamp restricted telescope access by Keck personnel. With the new lamps in use, the telescope no longer needs to be set to an azimuth of 7°; instead, the telescope and dome only need to be 90° away from each other. The MOSFIRE calibration GUI is now updated to use the new lamp, and includes revised default exposure times. Spectrophotometric observers using extra-wide slits will find that the standard exposure times on the calibration GUI (which are intended for typical 0.7 arcsec slits) will saturate the detector. Instead, observers with wide slits should acquire their calibrations manually (i.e., not using the Calibration GUI) and should use the new Domelamp Control GUI to reduce the power level of the lamp. With low-power (i.e., high attenuation) settings, spectrophotometric observers can obtain flats with the lamps on and better illuminate the spectra at all wavelengths. Click the image below to see the full-sized flat lamp gui.

To launch the new GUI via the MOSFIRE workspace menu, right-click on the background and select 
	K1 Telescope Status Menu > NEW Dome Lamp Control GUI

New Observing Mode Gui (2014-Feb-5)
The Observing Mode GUI was updated with a new layout. The new version removes unnecessary options and adds new methods for configuring to a dark filter. The new "quick dark" option efficiently moves the filter wheel to the nearest filter wheel combination to create a dark. Click on the image below to see the full-size version of the new ObsModeGui.


MOSFIRE DRP (2013-Dec-16)
Keck has revised instructions for installing the DRP in conjunction with the newly releases Ureka Python package supported by Gemini and STSCI. Please see: Keck DRP install instructions for the procedure. The official DRP code repository resides at http://code.google.com/p/mosfire

PIG Unleashed (2013-Dec-04)
We have added the new Program Interface GUI (PIG) widget to the MOSFIRE user interface. On split nights, observers use the PIG to indicate which observing program to associate with the current image. This permits the Keck Observatory Archive (KOA) to tag the images with the appropriate observing program and observer list, and ensures that the images are assigned the correct proprietary period.

MAGMA coordinate output feature (2013 Dec 1)
A post mortem of a failed slitmask alignment revealed that MAGMA has a feature in the mask center coordinates. Recently a mask was designed with the Declination seconds of arc listed as "60.00" in the mask design output files (mask.xml, starlist, etc.). Although the telescope DCS handled 60.000 arcsec the Slitmask Alignment Tool DSS query failed, resulting in a difficult coarse alignment. Please check your coordinate files to ensure proper formatting. Keck staff will update software to check the coordinates (SAT and MAGMA) in the near future.

New MOSFIRE FAQ Page May Answer Your Questions (2013 Sep 29)
We have compiled answers to a number of common MOSFIRE questions on the new MOSFIRE FAQ page.

'scent' Command Permits Reverse Moves (2013 Jun 21)
The scent command is used to offset the telescope such that a target at image location (x,y) moves to the slit center. However, sometimes observers may want to do the opposite operation; i.e., move a target from the slit center to a given pixel position. This is now possible by supplying the -r (reverse) flag when invoking scent. For example, an observer can place a star at pixel (1000,500) as follows: This technique is useful for putting a star on a wide (10 arcsec) slit for spectroscopy.

dp Command Displays Images to ds9 (2013 May 31)
The new dp command allows you to easily display MOSFIRE images on the ds9 image display, and will overplot the predicted slit positions if the image was a direct image. The syntax is quite flexible: This command is available on all MOSFIRE accounts when logged into the MOSFIRE host computer, nuu.

Guider Communication Compromized (2013 April 24 )
The fiber optic cable on MOSFIRE's interconnect panel was damaged in the morning of 22 April. Fiber cables are used to communicate to the instrument and guider with the guider requiring two sets of fibers. The guider fibers were damaged preventing communication with the guider. All other instrument control behaved normally. After much troubleshooting the guider is in a usable state after troubleshooting the issue for two days. We now have full control over the guider detector but we have to manually focus the guider during the night. This will cause some minor delays at the beginning of the night. We plan to resolve the issue after MOSFIRE come off sky in early May.

Calibration Tool Updates (2013 April 5 )
The GUI and scripts used to automate calibration acquisitions for user designed slitmasks was updated and released for observer use. The new GUI (seen below) presents a couple of new options to the observer by giving them options for changing arc-lamp exposure times and acquiring a flat lamp off image of the dome. Default values for exposure times and quantities were updated. The acquisition script is much more intelligent and now ensures that masks properly execute, skips masks if no exposures are requested, and is in general, more robust to errors in mechanism moves. Like DEIMOS and LRIS observers, MOSFIRE observers should be able to setup their calibration sequence in the afternoon, initiate the script, and safely grab dinner.


CSU Fatal Error Avoidance Strategies II: The New Voodoo (2013 Feb 10 )
The configurable slitmask unit (CSU) has been experiencing “Fatal Errors” when reconfiguring to new masks. This can occur when you are changing fields or when you are reconfiguring from an alignment to a science mask. There are new strategies to avoid a fatal error. First, all afternoon calibrations are should be acquired with the instrument at a PA of -90°. Second, before changing masks, ensure that the drive angle on FACSUM is at least 30° from zero. If the drive angle is within 10° of zero or -180, please ask the OA to rotate MOSFIRE 30° from the present location and then initiate your mask move. Note that the drive angle is likely different than the current PA of MOSFIRE. Both angles are located in the bottom right on FACSUM.

TRUITIME vs ITIME (2013 Jan 30 )
Observers who require knowing the exposure time to a high precision should note that there are two FITS header keywords that specify the exposure time.

CSU Fatal Error Avoidance Strategies (2012 Nov 15 )
The configurable slitmask unit (CSU) has been experiencing "Fatal Errors" when reconfiguring to new masks. This can occur when you are changing fields or when you are reconfiguring from an alignment to a science mask. There are two strategies for observing that may mitigate the CSU errors that are recommended by the MOSFIRE team.These two strategies apply when there are significant rotator or telescope moves.

In addition to the night time observing strategies above, we also recommend power cycling the CSU just before starting your calibrations in the afternoon.

New recovery software has reduced the lost time to 20-25 min.most of the time. Less frequently, fatal errors may result in up to 1.5 hours of lost time.


MAGMA - slitmask design software released (2012 Aug 7 )
MAGMA is the software used to design slitmask for MOSFIRE. Version 1.0 was released today and may be down loaded for use following the links at MAGMA . A version of the software is available for MAC, Linux, Solaris, and Windows. We are still shaking out a few bugs, so if you observe an issue, please contact Marc Kassis.

MOSFIRE scheduled for 2012B (2012 June )
MOSFIRE is scheduled for regular use starting in 2012B, and all commissioning efforts are expected to be completed before the first night of observing in September. For observers, who are scheduled to use MOSFIRE in 2012B, you will need to prepare slitmask design files in advance of your run. The new slitmask design software is called MAGMA and will be released in early August. Please see the MOSFIRE observatory newsletter article for first light details.

MOSFIRE First Light (2012 April 4)
MOSFIRE first light images were acquired on the night of April 4, despite heavy circus clouds over Mauna kea. Images and spectra were acquired on the night of the 5 as part of the continuing commissioning of the instrument. Please check the web pages for information on the performance of the instrument. The observatory press release may be found at observatory news

MOSFIRE moved to RT1 (2012 Feb 22)
After complete unpacking from the shipping crate, MOSFIRE was lifted to the RT1 position on the Keck I telescope Nasmyth Deck next to LRIS.


MOSFIRE arrives at Keck! (2012 Feb 16)
MOSFIRE arrived safely at the Keck Observatory early afternoon on 16 February 2012 after traveling by boat from the mainland. MOSFIRE was uncrated and placed inside the dome for unpacking.