CFHT, Instruments, Imaging, FOCAM, Observing Procedures.

Canada-France-Hawaii Telescope

A User's Manual for the CFHT Visible Imager: FOCAM

Observing procedures

CCD set up

The first task to be undertaken on the sky is to check the CCD orientation and the correspondence of the bonnette field and CCD centers.

First ask the TO to point to a faint SAO star near the meridian and equator with a magnitude fainter than 9 (so not to saturate the CCD; the minimum exposure time allowed by the CCD shutter being ~0.15 sec (for MOCAM and the UH 8K mosaic, the exposures must be longer than 1 sec. Shorter exposures will results in non-uniform illumination due to the slow shutter speed). Bear in mind that this is within the capability of the shutter but defenitely not recommended for photometry). Start running a "FOCUS/analytical" sequence (on GENIII system) or a "FOCUS/video" sequence (on GENII system; see the section Focussing) with a typical exposure time of 1 sec. You should adjust the focus raster to be centered on the star, with a typical raster size of 200x200 pixels. Then, using the telescope handpaddle controls, move the telescope in declination, and check which way the star moves on the CCD. For a CCD aligned North-South, usually corresponding to a bonnette rotation angle of 0 degree, the star should have moved along the columns. If you feel the star did not move as expected, rotate the prime focus bonnette on the telescope control panel (ask the telescope operator) until the alignment is satisfactory.

Your next step is to center this same star on the CCD. Then by asking the TO to move the bonnette mirror to the central field, identify the XY location of the star on the bonnette acquisition TV, corresponding to the center of the CCD. When you are done, ask the TO to zero the telescope coordinates to the current position, this will reset the telescope coordinate frame close to the selected TV/CCD location. This point is a time saver: by moving a star/object to the location known to correspond to the CCD center, you will save time both in upcoming focussing sequences and in centering your astronomical objects (see the sections Focussing).

These tasks are generally done by your support astronomer on the first night of your run.

Observing Sequence

The imaging sequence of observations with FOCAM can be divided into 4 sequential steps for a given field:


Overview: Focussing is probably the most critical task in direct imaging applications. A careful focussing will pay off for your observations giving you a better image quality and S/N. It needs to be done at the beginning of each night, and then for telescope motions greater than 30 degrees and/or every ~2 hours depending on the temperature variations in the dome.

We have designed a Computer Aided Focussing tool, CAF, which we have found to be more reliable and faster than even the most experienced observers.

For quick focus checks, a FOCUS/video and a FOCUS/analytical options are provided (note that FOCUS/video is not available with GEN III systems at the time of this writing December 1995).

Once a set of filters has been focussed at a given sky location, focus offsets between filters can be used reliably on other fields. Therefore, after moving to a new field, you only have to refocus one of your filters, and subsequently apply the computed focus offsets for the others.

Here are the typical filter focus positions for the CFHT, U,B,V,R,I filter set at prime focus:

Prime focus: There are two important reasons for careful focussing:

Since it is only possible to focus the TV guide field by using the telescope focus control, you need first to focus the TV guide field (you can also do it with the central field, but there is a slight defocussing between the two), and then focus the CCD field by using the prime focus bonnette focus control.

CAF Computer Aided Focussing: The computer aided focussing sequence at prime focus is the following:

It is illusory, and time consuming (and nonsensical from optical considerations) to try to achieve an accuracy better than Z=0.05. Larger steps will define a nice parabola, with a well defined minimum. Too small Z steps will results in a noisy plot, difficult to interpret. For a sequence of 8 exposures of 10 sec. each, 200x200 raster, the total elapsed time is 4 min from start to finish (focussed filter).

Prime focus focussing summary:

The following figure (shown previously) shows the Graphic output from CAF, FWHM (pixels) in X and Y vs focus position.

"Manual" focussing: The focus can be checked easily at any time by activating the "FOCUS/video" or "FOCUS/analytical" options.

The "FOCUS/video" can be started by selecting "video" in the "FOCUS" form, specifying the exposure time, and "accept". It will continuously take focus frames of the size defined in "FRASTER", until you tell it to stop (click in the working icon button). "FOCUS/video" is only available on GEN II systems. It has not been implemented in the GEN III system, and there is no plan to do so in the near future. Use "FOCUS/analytical" with GEN III systems.

The "FOCUS/analytical" can be started by selecting "analytical" in the "FOCUS" form. It will take one focus exposure, and compute the image quality of the star in the frame. Your focus raster should isolate a single star in order to produce meaningful results. A window will open with the image quality parameters, the most useful of which, for focussing purposes, are the X and Y FWHMs, and the peak intensity. The computed values can be selected before starting a focus sequence in the "IQE" form. If you need another focus frame, FOCUS/analytical allow you to continue (click "continue" at the bottom of the parameter display form), e.g. after having changed the focus to a new position.

Cassegrain focus: At Cassegrain, the process is somewhat reversed as compared to prime focus. Now you need first to focus the CCD with the telescope focus, and then you will focus the TV bonnette guide field.

The Cassegrain telescope focus is not currently computer controlled. Therefore a fully automatic focussing procedure is not possible. However "CAF" can be used in a semi-auto mode, as in the sequence described below.

The focussing sequence at the Cassegrain focus is the following:

Cassegrain focus focussing summary:

For manual focussing see this paragraph on Manual Focussing.

Field Acquisition

The field acquisition sequence is obviously based on two parameters: good coordinates, and good telescope pointing.

Good coordinates mean accuracy better than a few arcseconds. This is easy to obtain from many catalog, and even from the Palomar Sky Survey prints. It is your responsibility to bring with you an accurate list of coordinates, as we have only a limited set of catalogs at the summit.

Good telescope pointing also means accuracy better than a few arcseconds, and is easy to obtain provided that you follow strict guidelines. The CFHT pointing is based on a computer grid of the "virtual"sky obtained by interpolating pointing data on bright stars obtained on engineering nights. As it stands currently, the grid provides a pointing accuracy of 6.9 arcsec r.m.s. with the prime focus top end and 6.0 arcsec. r.m.s. for the Cassegrain top end. However, these numbers are "theoretical" in the sense that they correspond to deviations from the pointing model and not the sky. The pointing grid has a limited resolution, since we cannot determine the "whole sky" telescope pointing. Therefore, significant deviations from the mean pointing accuracy can occur at some locations (accuracy can be only ~15-20 arcsec on occasions). At the beginning of the night the telescope operator normally verifies the zero point of the grid by pointing a bright SAO star near the meridian and equator. You should then be able to get the nominal pointing accuracies mentioned above by means of regular pointing, i.e. entering your object's coordinates and order the TCS to move the telescope to it. If your program can tolerate this accuracy, the fastest way to proceed is to start taking "blind" exposures on the location where the telescope locked its pointing. However, most people using this pointing method, end up spending a significant amount of time moving the telescope around with the hand-paddle to make sure the field is correctly centered, prior to taking any scientific exposure. This is certainly a waste of time for the reason described below.

Fast and accurate telescope pointing: Providing that you have followed the set up steps described in the first section of this mosaic page, it is much more accurate, and all in all much faster, to use the "two step blind pointing" method. Accuracies of 1-2 arcsec r.m.s. are obtained routinely by using the following steps:

Our experience is that, if your coordinates are accurate (beware of the epoch), and using this procedure, there is no need to try to identify your field on the acquisition TV camera. Try this method a few times to convince yourself of its efficiency. You will end up finding that in the course of a night, and depending on your program (short vs long exposures), you may save up to one hour of exposure time, which is valuable indeed.

If you plan to observe an important number of objects, you can prepare an ASCII file in advance with object coordinates. Contact your support astronomer to have this list loaded into TCSIII. The format should be the following (assuming it is sent through E-mail):

From Fri Jul 26 16:54 HST 1991
Date: Fri, 26 Jul 91 21:27:26 EDT
Message-Id: <>
Subject: 1st list

Normal message stuff up here

NAME mystars
1	12:13:14.81	+00:05:30.5	"star 1"	1989.3
2	12:13:15.82	+10:05:30.5	star2		1990.4
3 	12:13:16.83	+20:05:30.5	star_3		1991.5
93	12:13:17.84	+30:05:30.5	star_93		1999.9


Once the telescope is pointing on your field, the telescope operator should start looking for a guide star. For this he does the following:

During this procedure, the TV gain has to be cranked up to the maximum, in order to be able to pick even the faintest star. The leaky memory sends the best error signal corrections to the TCS when the flux in the box is between 200 and 250 counts, the TV gain needs therefore to be tuned carefully. Be careful not to saturate the leaky memory ("Z" counts in excess of 250 will show as a steady "250" display), as it will "blind" the centroid algorithm. The ideal flux count is around 230-240, with the counter display oscillating rapidly between close count values. The leaky memory can still guide successfully for fluxes down to 100 counts if you are unable to find a brighter guide star.

Recently, the HST guide star catalog has been put on-line on the TO workstation. The TO can at his convenience choose a guide star from the catalog, using a graphical interface, instead of starting a raster search pattern. This is usually faster than the previous procedure.

On the bottom right part of the TCS screen, a "trying" message appears, quickly followed by a "guiding" message once the leaky memory is successfully locked on the star for guiding. It should remain this way during your science exposure.

Science exposures

Before starting a science exposure, check the following:

You may then start your science exposure by clicking in the "EXPOSE" button on the menubar, selecting the exposure time on the EXPOSE form, and clicking "accept".

A working icon will appear as a new window. It will display the remaining exposure time with a count-down counter. It also allows to stop, hold, restart, and abort an exposure.


Currently reported best performances at prime focus are provided with RCA2 (for all of the U,B,V,R,I bands) and are as follows. They have been computed for a 0.6" arcsec seeing, a star at zenith covering 25 pixels, and an average S/N=1 per pixel. The computations are based on average photometric coefficients computed with the RCA2 CCD (see Appendix F). An exposure time of one hour is given as a reference for deep, sky noise limited images. Sky values have been taken to 14000 e-/hr/arcsec2 in U, 127000 in B, 219000 in V, 328000 in R, and 410000 in I, and are based on average measured (not absolute) sky background with FOCAM on the RCA2 CCD.

Observations Logsheets

It is imperative, both for the CFHT archiving of observations on optical disc, and for your own data safety in case of loss, that you leave a copy of your observations log sheets with the telescope operator. This is a sample of the FOCAM log sheet that you should use during your observations, please check with the TO for available forms or have him make photocopies as needed.

Please send comments to Dennis Crabtree at