CanariCam latest news


icon Index



icon Latest News


23rd Apr 2013
Subject: Instrument cooling issues

Since taking CanariCam into operation the instrument has suffered a number of problems with its cooling system. The detector temperature has to remain low enough so as not to contribute too much to the background signal and noise. In the worst case, when the temperature rises to about 11.5 Kelvin its dark current starts to contribute strongly to the noise component, and at slightly higher temperatures the dark current becomes so high that it saturates the pixels even in the shortest of integration times, thus rendering the detector unusable for science.

The instrument temperature is maintained by a closed-cycle cooler. Last year and this year on a number of occasions this closed-cycle cooler lost efficiency which caused the instrument to gradually become too warm. Each time through a costly and complicated intervention we managed to get the system to operate again sufficiently well, but with significant down time for the instrument. Moreover, these interventions are not without risk to the instrument.

The last few months the situation again has worsened and consequently CanariCam has not been in operation for almost three months. We are currently taking measures to try to get the cooling back to acceptable levels, but given the history to date and the cost and risks involved, we are also considering a more serious re-design of the cooling system with the goal of providing a more permanent solution to the problem.

Update June 2013: the cold-head of CanariCam has undergone a repair and the instrument is back into operation.

Back to top


27th Mar 2013
Subject: Chop-Nod discrepancy

We have recently detected that there is a discrepancy between the chop and nod throws in CanariCam observations. Such a discrepancy is approximately 1 CanariCam pixel (0.08”) for a chop/nod throw < 15”, but it can be as large as 3 pixels (0.24”) for a chop/nod throw of 45”. This can be seen in the following two figures, which correspond to CanariCam observations of the standard star HD12929 taken with the Si5-11.6 filter with a chop/nod throw of 15” and 45”, respectively.


Offset of the star centroids in a CanariCam image of taken with a chop-nod throw of 15” along the detector X axis. The on source time was 5 minutes and the saveset time was 6 seconds. Each point in the plot represents the centroid measured in a saveset. The offsets are calculated with respect to the centroid of the first saveset in the Nod A. Filled circles represent the savesets in Nod A while plus symbols represent the savesets in the Nod B. The upper panel shows the offsets in the detector X axis, while the lower panel shows the offsets in the Y axis.



Offset of the star centroids in a CanariCam image taken with a chop-nod throw of 45” along the detector X axis. Each point in the plot represents the centroid in a saveset. The on source time was 5 minutes and the saveset time was 6 seconds. The symbols represent the same quantities as in the previous figure.


In both figures, the plus symbols and the circles in the upper panels have, on average, different abscissa values, due to the fact that Nod A and Nod B star positions are not coincident. Clearly, the separation between Nod A and Nod B centroids is larger for a 45” throw than for a 15” throw.

This is a systematic effect but its impact is not higher than the actual jitter of the PSF in the current CanariCam images. The advantage of being a systematic effect is that it can easily be corrected during the data reduction by summing all savesets in Nod A separately from the savesets in Nod B and then aligning both nod sets. However, this reduction method would be impossible to use in the case that the sum of all NodA savesets (or Nod B savesets) does not yield enough SNR as to proceed with the alignment between Nod A and Nod B savesets. This will not be the case for the large majority of programs, though. Alignment between Nod A and Nod B savesets might be difficult also in the case of observing extended sources which do not have a well defined feature as a reference for alignment.

Even if the Nod A and Nod B savesets can be aligned, a degradation there will be a degradation in the sensitivity that will add up to the degradation due to random PSF jitter even in each individual nod position. The following table shows a comparison between sensitivity estimates using two different reduction methods, namely shift and add and simple accumulation of savesets.


Reduction method Chop/nod throw (“) FWHM (“) Radius (“) Flux (ADU) Std. Dev. (ADU) SNR Sensitivity (mJy)
Shift & add 15 0.38 0.28 1.0780e+09 3016.0 57616.2 1.11
Accumulation 15 0.38 0.32 1.1480e+09 3546.0 45663.4 1.50
Shift & add 30 0.41 0.32 1.0100e+09 3355.0 42461.4 1.46
Accumulation 30 0.47 0.44 1.3730e+09 3990.0 35298.8 2.39
Shift & add 45 0.45 0.40 1.1820e+09 5104.0 26131.4 2.88
Accumulation 45 0.54 0.48 1.3960e+09 3842.0 34166.6 2.60

Sensitivity comparison using two different reduction method of a set of images with increasing chop/nod throws. Data correspond to images of the standard star HD12929 taken on January 30, 2013. Column 1: Reduction method. Column 2: Chop/nod throw in arcseconds. Column 3: FWHM of the PSF in the reduced image. Column 4: Radius in arcseconds of the aperture used to estimate the SNR. Column 5: Measured flux in ADU from aperture photometry. Column 6: Background standard deviation from aperture photometry. Column 7: Signal-to-noise ratio in aperture. Column 8: Sensitivity estimated as a 5-sigma detection in 30 minutes on source.


The first conclusion that can be extracted from the previous table is that, for a given chop/nod throw, the radially averaged FWHM of the PSF is worse in the case of reducing the data with simple accumulation of savesets than in the case of using the shift and add technique. The degradation in the PSF of the final reduced image increases as the chop/nod throw increases due to the discrepancy between chopping and nodding. The table also shows how the sensitivity can be improved by shift and add of savesets with respect to simple accumulation when a chop/nod throw of up to 30” is used. For larger chop/nod throws (45”) the effect of shifting and adding is the opposite, i.e. it will degrade the sensitivity rather than improving it. The reason for this trend is that the radiative offset correction is worse due to the discrepancy between chopping and nodding, and therefore the noise due to the background structure increases. The following figure shows the noise structure in the images used to create the previous table.


Comparison of the noise pattern appearing when images of a standard star are reduced using the shift and add and simple accumulation of savesets. The chop/nod throw increases from left to right. Each pair of images in the same column represent the same data but reduced with different methods, shift and add (top) and accumulation (bottom) of savesets.


The diagonal background structure due to a bad correction of the radiative offset becomes markedly worse in the shift-and-add reduction as larger chop/nod throws are used. The background structure does not seem to be affected by the chop/nod throw when data are reduced using the simple accumulation of savesets. However, if savesets are simply accumulated, the image quality is strongly degraded as it appears elongated in the nodding/nodding direction due to the discrepancy between the chop/nod throws (see next figure).


Comparison of the star peaks when images of a standard star are reduced using the shift and add and simple accumulation of savesets. The chop/nod throw increases from left to right and was defined along the detector X axis. Each pair of images in the same column represent the same data but reduced with different methods, shift and add (top) and accumulation (bottom) of savesets.


The discrepancy between chopping and nodding is currently under investigation and it is our top priority to solve it. Until the problem is solved, we recommend, whenever feasible, to use the shortest chop/nod throw possible, but avoiding the contamination of the positive image with the two negative images. We recommend not to use a chop/nod throw > 15”.

New tests performed on May 2013 after having made a small modification on the chopping scale show that the situation slightly better, but still the problem persists.

Back to top


15th Jul 2011
Subject: A first step for CanariCam science commissioning

Between June 19 and 25 an important first step was set on the path of commissioning the CanariCam mid-IR imager and spectrograph. The instrument worked very reliably and a wide range of tests could be conducted, ranging from imaging observations in various filters in the 10 and 20 micron atmospheric windows, spectroscopic observations at low-resolution, and even some first trials of polarimetric imaging observations.

As the proverbial proof-of-the-pudding we show a false-color image of the planetary nebula NGC7027, accompanied by a low-resolution spectrum. These are merely raw data, but already indicate the quality of observations that can be obtained.

Although the data analysis will take some time to be completed and further on-sky testing will be necessary to complete the commissioning process, we now feel confident that the instrument can be taken into operation soon and be offered in some limited modes to the user community from semester 2012A onwards.

This will be the second science instrument coming into operation at the GTC, and is therefore a huge leap forward for the scientific capability of the telescope.

Thanks are due to the CanariCam team, especially Charles Telesco, Chris Packham, Frank Varosi and Dan Li from the University of Florida for their relentless work and patience in delivering CanariCam.

The CanariCam commissioning team, together with observatory personnel investigating some of the results from the observations.


Image of the planetary nebula NGC7072, taken with CanariCam at a wavelength of 11.6 micron.



Low resolution spectrum of the planetary nebula NGC7027, taken with CanariCam in the 10 micron band.


More information here.

Back to top


8th Feb 2011
Subject: Preparing GTC for CanariCam

As we announced in an earlier news item about the CanariCam mid-IR instrument, commissioning of the instrument last year was severely hampered by weather and unexpected problems with the telescope. Since that time much work has gone into investigating and resolving the telescope problems. Very significant progress has been made in improving the robustness of the observing system under the demanding conditions required by CanariCam. Also problems related to the interface between the telescope system and the instrument have been tackled. Once this work is finished and the system is proven to be sufficiently reliable, we will proceed with the commissioning of the instrument at the shortest possible time scale and, if successful, start with the scientific exploitation of CanariCam.

Back to top


1st Oct 2010
Subject: Progress with CanariCam commissioning

The CanariCam mid-IR imager and spectrograph has made its first steps towards scientific operation. In September the instrument was mounted in the Nasmyth-A focal station of the GTC. The first indications are that the instrument is performing well and is reliable.

Unfortunately bad weather severely interfered with the progress of the night-time commissioning, and also problems were encountered with the telescope that will require attention in the coming weeks. The combination of these two aspects has resulted in that hardly any on-sky time could be exploited for the formal commissioning tests. But in spite of this negative outcome of the first commissioning run, the few on-sky images under excellent seeing conditions did show a glimpse of its future potential. The accompanying picture shows an image of a standard star in the Si2 filter at 8.8 microns, displaying the superb image quality of the GTC + CanariCam system, reaching the diffraction limit.


Back to top


26th Aug 2010
Subject: CanariCam Commissioning

Lately very good progress has been made in developing the funcionality of the telescope that allows us to start commissioning the CanariCam mid-IR instrument. The CanariCam instrument places strong requirements on the performance of the telescope in terms of image quality and in terms of control of the fast chopping motion of the secondary mirror. Relative phasing of the mirror segments has now been achieved to a fraction of the wavelength at which the instrument operates, reaching a level that meets the requirements of CanariCam. The chopping motion of the secondary has been tested and poven to work reliably for frequencies and chop throws that are relevant for CanariCam operation. An important functionality that has not yet been completed is that of fast guiding, but this is not seen as a show-stopper for the first commissioning tests of CanariCam. Therefore green-light has been given to commence the night-time commissioning of CanariCam in the Nasmyth-A focal station on September 22nd during 12 nights.

Back to top


26th Apr 2010
Subject: CanariCam status update

At the moment commissioning of the instrument is awaiting final preparation of some of the key functionalities of the telescope, in particular phasing of the primary mirror and characterization and commissioning of chopping and nodding motions in combination with guiding. These tests are under way and have been given the highest priority, but there is no firm time table yet as to when commissioning of the science instrument will commence. Given the uncertainties and the delays suffered, the consequence is that scientific use of the instrument, in the best of cases can only start towards the very end of the current semester.

Back to top


25th Nov 2009
Subject: First Light for CanariCam

We are pleased to announce that CanariCam, the GTC multi-mode thermal-infrared camera, has achieved first light ("first heat") at the GTC. This occurred at about 22.10 h (local time, La Palma) on 18 November. The first image was of the star 20 Ceti. Although the image was far from perfect, typical for such first light observations, it was as beautiful to the University of Florida and GTC personnel as a new-born baby is to its parents!

These first-light observations, which followed laboratory performance verification on 16/17 November and CanariCam installation onto the GTC on 18 November, indicate that CanariCam is working well, meeting or exceeding all expectations based on results of formal acceptance tests performed previously at the University of Florida (UF).

Although formal on-telescope commissioning of CanariCam will not begin until spring 2010, the joint UF-GTC team was able to test several critical features of the CanariCam-GTC interface and functionality. In particular, the chopping and nodding modes, fundamental to achieving CanariCam functionality, are operational, although considerable exploration of these modes must occur during commissioning. CanariCam was moved from the telescope to the GTC instrumentation lab on 19 November.

The UF team visited the GTC for nearly two weeks, and all testing and integration activities went smoothly, with the UF and GTC personnel working well together as a integrated team. The UF CanariCam team wishes to express its gratitude to our GTC colleagues for their hard work and gracious hospitality.

Charles Telesco
CanariCam PI

On behalf of the CanariCam team members at the GTC:
Greg Bennett, Kevin Hanna, Dan Li, Chris Packham, Frank Varosi; and the CanariCam Team members not present at the GTC:
Dave Ciardi, Jim French, Chris Ftaclas, David Hon, Jim Hough, Jeff Julian, Roger Julian, Margaret Moerchen, Robert Piña, Francisco Reyes



Some of the people involved with CanariCam commissioning, at the Nasmyth-A focal station of GTC. The instrument can be seen just above the "Gators" banner, flanked by two electronics cabinets.


A bare image of the pupil, taken with the dome closed. The only "star" visible here is the faint mid-IR glow towards the bottom of the image of one of the Team members standing on the dome catwalk.

Back to top


Last modified: 14 August 2014