This is intended as an aide de memoire in the development of the PN Spectrograph and as a check that all operational aspects hang together coherently. It is not intended as an imperative document but to stimulate discussion.
The PN Spectrograph
The instrument will consist of:
- two cameras and a collimator connected together, weighing about 200 kg
- an interface unit for each of the TNG and WHT
- (possible Halpha arm)
- a third camera, including its own cryostat
- a large (190mm x 270mm) dichroic in a mount
- one (or possibly two) grating mounts
- several large (195 mm) diameter [OIII] filters
- some sort of electronics interface box - initially the old Roden electronics box but later perhaps something better
Storage
The PN Spectrograph will be stored as a complete unit with the optics in place (but with [OIII] detectors removed) either at the WHT or at the TNG (Checklist 6). The Halpha camera will remain attached (assuming that it exists). It will need to be moved from one dome to the other up a steep hill and attached to the appropriate telescope.
Interface
Data will be collected via the data acquisition of each telescope since the [OIII] cameras will belong to the respective observatory. It may be that the Halpha camera has its own controller and data storage facility.
The interface box will house power drivers for the actuation of the shutter (unless we use small iris shutters attached to the dewars, in which case they can be run from the observatory), and the calibration matrix. Calibration lamps, on the other hand, probably belong to the telescope.
Assuming that the two systems do NOT have an interface, a typical calibration observation might consist of:
| sequence | PNS CONSOLE | TELESCOPE CONSOLE |
|---|---|---|
| 1 | ||
| 2 | Switch on lamps | |
| 3 | Engage matrix mask | |
| 4 | Clear CCD (readout) | |
| 5 | Open shutter 10s | |
| 6 | Readout CCD |
With an analogous procedure for observing the target.
Focus
- Laboratory adjustment of collimator
- bright white-light source
- matrix mask
- through [OIII] filter
- autocollimation (i.e. mirror in place of grating)
- Insert grating
- adjust angle(s) if necessary
- Adjust camera
- at telescope
- tilt/piston shift on cryostats
Filters
Can only be changed off-line, and cannot be removed from the beam during the night. To cover a range of 67Å, we will have three filters: 5001/36, 5023/36, 5046/36. From experience the flat top on these filters will have a width of about 0.63 FWHM or 22.3Å.
Calibration
For our purposes we would like to have three lines within any given filter bandpass in order to get a reliable and unambiguous dispersion solution. At the WHT, CuNe and CuAr lamps can be used. Here is the relevant part of the combined spectrum (the major lines in the 5000-5050 part of the spectrum from L to R are:
5009.3
5017.2
5031.4
5037.8
He and Ar lamps together have the following emission lines (of which some are very faint):
4921.929 HeI 4965.0795 ArII 4972.1597 ArII 5009.3344 ArII 5015.6779 HeI 5047.738 HeI 5062.0371 ArIIThe following lines are medium-strength in the (UES) Th-Ar catalogue:
4985.372 5002.097 5009.334 5017.163 5028.636? 5039.23 5044.719 5047.043 5050.784 5062.037Possibles from the Al/Ca/Mg-Neon catalogue:
4957.033 4973.538 5005.16 5031.3484 5037.7505 5041.620
The matrix mask will include both pinholes and short slits perpendicular to the dispersion direction. When illuminated by an arc lamp, these will provide an excellent check on image distortion and on dispersion over the field. It should be done while the telescope is pointing near the target, so avoid effects caused by differential bending when the telescope is slewed.
This process will provide a well-defined zero point for the velocity/offset relation between the two arms, when two arms are in use. If it is performed carefully it will also give the zero point in one-arm operation.
To correct for small shifts in the image centre between integrations (recall that there may be no easily-visible point sources with which to align individual frames) we should be able to use the dispersed images of foreground stars, which look like trails (Ken has tested this). In any case, it may be useful to reobserve the matrix mask at regular intervals. We have no totally reliable method of correcting for telescope errors except for the star-trail method.
KEN FREEMAN'S NOTES (email 29/7/98):
ALIGNING IMAGES WITH THE FILTER PROFILE. This worked really well, even with my filter which does not have very sharp cutoffs. I guess it works because the filter is typically 40A wide and we need to align to about 0.2A. There is plenty of signal in the stellar spectra, so this is not so demanding. It needs corrections for the color of the individual stars (derived from the spectra themselves) - this correction is easy and small. I also had to correct for variations of the passband over the filter - this was measured from the matrix mask data and is a problem that will be *much* smaller with the PNS - my filter is not so good and was near the focal plane, while for the PNS the filter is near the pupil. Despite all this, I recovered the absolute velocities of galactic PN to 15 km/s.
ILLUMINATING THE FIELD WITH WHITE LIGHT. This is partly for flatfielding without the matrix mask (we have the filter in there to reduce the bandwidth), and partly with the matrix mask for registration as discussed above. The white light observations through the matrix mask are used to measure variations of the passband over the field. We could manage with dome flats though something a bit hotter than the usual dome flat lights would be nicer. But we do need a good source of He illumination for the matrix mask. John thinks this doesn't contribute much to the cost, and it would be only a small thing to add our own hot white light source as well. return