2011 December 18 (E. Kirby): The CCD for the red channel has been replaced. On 27 October 2011, the old red detector (1024 × 1024, 24 µm pixels) was replaced with an LBNL deep-depletion CCD (4096 × 2048, 15 µm pixels).

This upgrade increased the red channel's throughput by a factor of 1.5 shortward of 8000 Å. Longward of 8000 Å, the throughput increased drastically, and you can expect excellent sensitivity up to and beyond 1 µm.

The longer detector also increased the spectral range of a single exposure by a factor of 2.5. However, the camera was not designed for such a large field, and the reddest 1000 pixels may have unacceptable optical quality. Please decide whether to use this section of the chip based on the requirements of your program.

This thick chip is an excellent detector of cosmic rays. It is recommended that you limit your exposure times to 20 minutes or less.

The pixels on the new detector are smaller. The new plate scale is 0.293 arcsec/pixel. You may wish to bin by 2 pixels in the spatial axis.

The dichroic filters have second-order light leaks which affect wavelengths longward of 9100 Å. For observations in this range, the use of the order-blocking filter RG610 is recommended.

2015 May 17 (C. Heffner): An emission line around 6708Å is sometimes present in dome flats taken with the high lamp. Note that when these flats are used in data reduction, it can create a spurious absorption line at that wavelength. A normalized plot of the emission line in a dome flat is shown in the figure below.

6708Å emission line in 40sec dome flat, 1.0" slit.


Original document by Eric Bloemhof 1998, updated/revised: August 3, 2008 by Jeff Hickey. For all Palomar instrument related questions please contact P200 Instrument Support

This OVERVIEW gives a brief outline of the DBSP (double spectrograph) for the Palomar 200" telescope, including some of the theory of operation and information needed for advance planning of an observing run. For a step-by-step guide to operation at the telescope, go back to the DBSP homepage and click on the DBSP COOKBOOK button.

The photo below shows DBSP in its handling cart. Entrance slits are at the top, and gratings are at the bottom; the grating angle adjustment knob for the red side is visible at the center of the black circular region. The red camera and dewar are mounted on the left, while the blue camera is partially visible at the right.

...photo of DBSP in its handling cart...


The diagram below (adapted from Oke & Gunn, PASP, vol. 94, p. 586, 1982) shows the light paths through the initial parts of the red- and blue-side optics. The slit lies in the f/15.7 Cassegrain focal plane of the Palomar 200" telescope. A dichroic splits the light into red and blue components; the observer chooses (once for each run) one of four dichroics, with dividing wavelengths of about 4800, 5200, 5500, or 6800 Å. (Transmission curves for the 5200 and 5500 Å dichroics are shown in Section E of this Overview).

The light then passes through field lenses, into prisms, and through filters before entering collimators through holes in the collimator primaries. (Filters may be changed by remote control during observations; a table of standard choices is given in the DBSP Cookbook, Section A). The collimator secondaries are attached to (transparent) fused quartz plates that are optically flat.

...Fig. 1 from Oke and Gunn...

From the collimators, the collimated red and blue beams proceed to the two gratings and then to the two cameras. The red camera currently consists of transmissive optics (lenses), while the blue side has a reflecting f/1.5 Schmidt camera with a field-flattening lens immediately in front of the detector. The schematic below shows the grating configuration and defines the angles relevant to wavelength calculations. The quantity directly set by dials on the side of the spectrograph is called the grating angle, and is the angle between the incident optical beam (from the collimator) and the normal to the grating.

...diagram of one side, collimator and grating...(click on it for expanded view)...


The equation in the box immediately above (click upwards slightly to view it) is used to solve for the grating angle to which a grating must be set to place the desired wavelength at the center of the CCD. An important point is that the sum of the grating angle and the angle made by the outgoing diffracted beam is fixed by the geometry of the instrument. For the red side, this sum is 35.00 degrees; for the blue side, it is 38.50 degrees.

NOTE There is a new grating calculator available on the mountain. The calculator is an IRAF script (by R. Burruss) resident on the workstation used for DBSP. It is called dbangle and is located in the dbsp tasks. Use the web-based calculator to fill out your green sheet, and use the IRAF script to fine tune your settings if necessary.

NOTE ON CENTERFIELD CAMERA There are two choices for the slit viewing camera. This camera is sometimes called the DBSP centerfield camera. This camera views the reflective slit apertures and is used to accurately put objects onto the slit. The default camera is a Xybion image intensified system. This camera has the advantage of being fast for brighter objects. The Xybion has the disadvantage of not being able to save images. The alternative that is available is a Finger Lakes Instruments (FLI) CCD. The advantage of the CCD is that you can save images of the field.

The FLI camera was purchased with Targets Of Opportunity in mind. The FLI has a filter wheel (g',i',r',ND,open) which can be used for Photometry. The fastest the FLI can update is about two seconds per image. Like the Xybion it can integrate for longer until the sky background is reached. The unvignetted field of view for both cameras is about 120"X 70".

CLICK HERE TO GET A CALCULATOR that will solve for the grating angle needed to put your desired wavelength onto the center of the CCD. This angle is a function of spectral order, choice of red or blue camera geometry, and grating lines/millimeter. All of the options are laid out in the calculator, or you may look at the table in section F) Grating Specs of the table of contents at left.

NOTE ON GRATING SELECTION: there is only one copy of each grating (with a given number of lines per mm and blaze angle), so you cannot, eg, put the 158 lines/mm one into both RED and BLUE sides. Any of the complement of 8 gratings may be put into either RED or BLUE sides, though.

To enable correction of mechanical offsets in the readout dials, blaze angle is requested to uniquely identify the grating. Blaze angle is not otherwise relevant to the grating angle calculation. Thanks to these empirical offsets [NYI], calculator accuracy is about 1 arcmin; due to their variation, answers for different gratings with the same lines/mm will vary by typically 10 arcmin. If this latter accuracy is OK, you could use any blaze angle; but it doesn't hurt to know what it is, either.

Also, the calculator will ask for slit width, which is used to compute the achieved spectral resolution, but not for the primary calculation of grating angle.

Grating angle = 0 puts the grating perpendicular to the incident beam from the collimator (the grating will be horizontal if the telescope is at zenith). To move to a positive angle, turn the dial counterclockwise (for both red and blue sides). This makes the grating turn its "face" toward the camera. In general the daycrew, or support astronomers, will set the grating angle for you based on the value you entered on the "Green Sheet"

By differentiating the grating equation, the spectral dispersion is found to be...

where f is the focal length of the camera (12 inches for the red camera; 9 inches for the blue). With details of camera pixel size and format, as given on the DBSP Specifications page available from the DBSP Home Page, it is straightforward to find dispersion per pixel and related quantities. The grating angle calculator above will give dispersion, resolution, spectral coverage, etc. for the wavelength of observation.

The dispersion is also presented in tabular form, go to section F) Grating Specs of the table of contents at left. The values there are representative, evaluated at the blaze wavelength, but dispersion is a very slow function of wavelength.


The DBSP may be operated with a single long slit of length 128", and available slit widths are 1/2, 1, 1.5, 2, 4, 6, 8 and 10". These different slits are on a wheel allowing them to be changed quickly during observations.

Instead of this set of standard single slits, the observer may elect to use a "multislit" mechanism consisting of eight individual slit segments that may be positioned independently on the sky in the direction perpendicular to the slit axis (see the figure below). Choosing between multislits and the conventional slit wheel described above is a setup operation that must be done in the daytime by the Palomar crew. Multislits are chosen following the specifications on the pre-run "Green Sheet" web-form that is filled out by successful proposers well before their observations. Similarly, positions of individual slits in the multislit are not changed during a run.

...schematic of multislit mechanism...

Each multislit slit segment is 15 arc seconds long, and the maximum spatial displacement is +/- 26 arcseconds. Four multislit units are available: two have 1.0 arc second apertures on glass, and two have 1.5 arc second apertures on quartz. There is a single long slit (2 x 120 arc seconds) on glass available for use in conjunction with the multislit mechanism.

Multislit slit segments are formed by a reflecting coating on a transparent coating, and so there will be reflection losses to the overall spectrometer throughput that are not suffered with the standard slits.


The observer may choose (in advance, on the setup Green Sheet) one of 4 dichroics that divide red and blue sides at roughly 4800, 5200, 5500, and 6800 Å. Curves for the middle two (5200 and 5500 Å) are given below. We have no curves for D48 or D68, but their quality is presumably comparable.

It is also possible to observe with no dichroic (giving maximum light into the red camera) or a mirror in place of the dichroic (giving maximum light into the blue camera, IF THE MIRROR COATING IS FRESH). These modes are not recommended, for the following reasons:

1) The no-dichroic mode gives a non-standard optical pathlength into the red camera, and it may not be possible to achieve optimum focus...i.e., best spectral resolution.

2) The mirror-in-place-of-dichroic has an old aluminum coating that has probably lost more blue reflectance than the D68 dichroic. If you are tempted to use the mirror to get best blue-side reflectance, use, e.g., the 6800 Å dichroic instead.

...5200 Å dichroic...click on it for a more readable, expanded view...

...5500 Å dichroic...click on it for a more readable, expanded view...

(F) GRATING SPECS (lines/mm, blaze, dispersion); GRATING CURVES (wavelength vs. angle)

Overview of Gratings:

1st Order

Useful Range (Å)

Red Camera

Blue Camera


Blaze (Å)

(to blaze 1/2-intensity)

Dispersion (Å/mm)

Dispersion (Å/mm)

158 [a]



201 1st [c]

135 2nd [d]

300 [b]




140 1st

316 [a]



102 1st


600 [b]




71 1st

600 [a]



54 1st


1200 [b]




36 1st




27 1st

36 1st




26 1st

35 1st

NOTE: there is only one copy of each of these gratings, so you cannot, eg, put the 158 lines/mm one into both RED and BLUE sides. Any of the complement of 8 gratings can be put into either RED or BLUE sides, though.
"Useful Range" gives the wavelength limits at which diffracted intensity drops to 1/2 of its peak value (2/3 and 3/2 of the blaze wavelength, by a common rule of thumb).
[a] silver coating...can be used only longward of 3500 Å.
[b] can be used only in blue spectrograph.
[c] first order.
[d] second order.

(G) POLARIMETRY; Specifications

All Stokes parameters may be obtained with a rotating-waveplate polarimeter (which must be requested prior to instrument setup). The waveplate and calibration polarizers are inserted above the turret, ahead of the slit, and a calcite beamsplitter is placed behind the slit but in front of the dichroic filter.

In this mode, the field of view is reduced to a circle 35 arc seconds in diameter. Each exposure produces two orthogonally-polarized spectra, one at each red/blue detector.

Either of two waveplates may be installed: lambda/2 (3200 Å to 10,000 Å) or lambda/4 (4000 Å to 8000 Å), but it is not practical to change during the night. The Stokes parameters are measured individually with two exposures each:

Measurement of Stokes Parameters with DBSP

Stokes Parameter


Position Angles



0 & 45 degrees



22.5 & 67.5 degrees



45 & 135 degrees

The calibration polarizers are (a) an HNP'B polaroid that covers from the UV cutoff (3200 Å) to 7500 Å, and (b) a polarizing cube that covers from 6000 Å to the IR cutoff (xxxx Å).

The waveplate and the calibration polarizers are controlled by a new LabVIEW GUI shown below. A VISTA-based data reduction package is reportedly available by contacting Marshall Cohen; it is not, however, maintained by the Observatory.

Polarimeter Installation (For Palomar Staff)

(added September 5, 2004 Steve Kunsman)

There are basically two pieces that need to be installed when the Polarimeter is requested with the Double Spectrograph. These are; 1. the Waveplate Rotation Assembly, and 2. the Beamsplitter.

  1. Gather the materials needed for installation

2. HOME the Cornell Offset Guider in the Data Room.

The entire DBSP needs to be lowered from the cass ring to install the waveplate rotation assembly. Bolt the waveplate assembly above the guider box. Find the four empty screw inserts. The waveplate assembly is secured with four Allen screws found in the Polarimeter wooden box.

3. Install the black braided cable right angle end into the west side of the DBSP mounting assembly, following the offset guider cables in the west insert hole, over the top of the turret assembly through the South side cutout, into the space available for attaching to the waveplate rotation assembly. It is easier to attach before lowering the waveplate assembly onto the mount.

4. Remove the covers on the optics. a) square cover on top of the waveplate; b) cover that slides off the UV Polaroid lens; c) two covers on the IR polarizing cube. The top cover slides off the end of the arm, the bottom cover then pulls off the bottom.

5. Align the rotation arms: calibration arm towards the SE corner of the opening; IR polarizing cube arm towards the NE corner of the opening.

6. The East side and West side mounting brackets of the waveplate assembly are clearly marked.

7. Slide the East side of the assembly under the guider guide rail, taking care not to bump the aligned arms for installation. Once you have the East mounting bracket resting on the East mount, attach the black braided cable to the assembly before lowering to the West side.

8. Lower the West side over the mounting screw holes. You may have to slightly raise the East side again and square the assembly up before being able to lower both sides.

9. Align the screw holes and secure four Allen screws with 3/16 Allen wrench.

10. Re-install DBSP into the cass ring.

To install the beamsplitter, the procedure is the same as installing a dichroic. If you need to change or install a dichroic, this should be done at the same time.

*NOTE* For safety and to protect the dichroic and beamsplitter from damage, two people should be involved in this step.

11. While one person holds up the slit box assembly, insert the proper dichroic.

12. Insert the beamsplitter in the rectangular hole cut-out above the dichroic. It only fits one way. Secure the slit box assembly with three Allen screws and re-assemble the DBSP.

Polarimeter checkout;

(Jeff Hickey Aug. 17, 2006)

The old SAM polarimeter commands have been replaced by a new GUI. The new polarimeter GUI may be found after you start the VNC from the control room workstation. On the VNC window look for the DBSP folder. Open the DBSP folder by double clicking on the icon. The folder may be hidden by the Blue camera Main GUI, so look behind it. In the DBSP folder you will see an icon for the polarimeter GUI, double click that icon. The polarimeter GUI is shown below. The GUI operation is simple. First press the "Pol. Init." button, this initializes the polarimeter. Next press "CA Home" and "PA home". The GUI should now show "0" for PA angle and "HOME" for the cal position. See the table above for the proper position angles to get the Stokes parameters you desire. Changing position is easy, simply type in the angle you want in the "set PA" window and hit return. Changing the Cal location is also easy, type "HOME", "UV", or "IR", for the three Cal positions. To check position click on the "PA read" and "Cal read" buttons. The "current PA offset" window should always read "0" at startup. As of 8/17/06, I have only tepid understanding of how PA offset is used, I belive it is a calibration offset used after observing polarization standards. I will update this document when I find out more.

Taking arc images with DBSP will only show that the beamsplitter is in place. To test the polarizer calset positions, and the Stokes angles, you will need to take dome flats. The arc lamps are not in the polarimeter optical train. If you are unfamiliar with turning on the highlamp and opening the mirror cover, ask the daycrew or the support astronomer. All calset and PA commands echo position numbers to the GUI "current" windows, which should tell you that moves have been made. As a final check, the observer should observe a polarization standard star to confirm that the correct angles and polarizer elements are indeed in place.

Polarimeter tips; (for the observer)

I have taken the liberty of including some valuable observing tips from an email by Patrick Ogle to Eric Agol ~2003. (Jeff Hickey Sept.6 2004)

You can run either with or without the dichroic when running DBSP red side only. The choice mostly depends on what wavelength range you are observing.

Advantages of the dichroic are:

1) Camera was designed to focus with dichroic in.

2) Blocks out contamination from 2nd order light.

3) High (>90%) transmittance.


1) More limited wavelength coverage.

2) Some polarization calibration problems within +/-100 Angstroms of the dichroic split. This problem is not a big deal with D55, but is more severe with D52.



Setup of DBSP + polarimeter can be a lengthy procedure, so make sure you get there by noon and do the following:

1) Set grating tilt(s) to get right wavelength range (test with arc lamp spectra.)

2) Focus the camera(s) using arc lamps. Both of these steps are best done with one person in Cassegrain focus cage and one in the control room.

Necessary calibrations (The more the better!):

1)Arc lamps (wavelength)-- Finish before dinner.

2)Dome flats (flux) -- Run a set during dinner.

3)Spectrophotometric standard stars (flux.)

4)Bright star observed with cal polaroid filter (PA spectrum calibration.)This can be done during twilight to save dark time.

5)Polarized standard star (PA offset calibration)

6)2 Null polarization standard stars (Instrumental polarization.) Doing more than one is helpful to avoid stars that are variable or not-quite null.)

See Ogle et al. 1999, ApJS, 125, 1 for detailed discussion of polarimetry calibrations. Take note of Appendix A, a discussion of the significant instrumental polarization (1-2%). I also give useful lists of Null and PA standards.

You might want to talk to Marshall Cohen about spectropolarimetry reductions. He has a suite of SPOL procedures written in VISTA scripting language, or you can make up your own IDL procedures. I don't think anything is available in IRAF. It is best to have these available at the telescope so you can do quick-look reductions on the fly.