PRIMUS observing checklist

From NYU CCPP Wiki

This is a page about the PRIMUS project. See also PRIMUS IMACS instructions.

All exposures should be with 'fast' readout, unbinned.

We are using 60 second dwells, typically in sets of 16 (8 on, 8 off). The IMACS GUI refers to this as 8 cycles.

For alignment images, we use 20 seconds in clear, dark conditions. If conditions are worse (and particularly in twilight), use 40 seconds.

We do use subraster mode for acquisition images. Here, find the file, then hit the load, apply, and done buttons.

Afternoons:

1) Check focus using pinhole mask

2) Take 10-20 bias exposures

3) Insert the flat field screen and take ambient light direct image through each mask. Tun icbox to make 200 pixel boxes. Take care in icbox to click fairly close (a couple pixels) to the center of the alignment box; this improves the first ialign.

4) Take arc lamp with the prism in through each mask. It is easy to intersperse this process with running icbox.

5) Copy .sub files to .sub60 files and edit to make 60x60 boxes.

6) Take Ql lamp direct flats, no mask. 4 seconds B band. 1 second V band.

7) Take direct flats with no lamp (but with flat field screen in): 10 seconds in I band, 20 seconds in R band.

8) Remove the flatfield screen. Close the hatch; enjoy your dinner.

These are calibrations that Richard has been taking during the Oct06 and Dec06 runs. Please get these when possible:

1) Be sure the flat screen is in and turn on both He lamps. Insert the LDP, V-band filter, and one of the PRIMUS slitmasks. Take 5x3s exposures through the mask. The gui will prompt you since you are not using the Spectroscopic filter, so don't forget to acknowledge that box. These are helpful in diagnosing any changes in the halo shape or amplitude. These are not needed in every filter, but try to get 1-2 masks per night (but more is helpful).

2) With the flat screen in, turn off all of the lamps. Insert the LDP and the Spectroscopic filter. Take one 40s N&s exposure (8 cycles of 5s each) using the shuffle parameters we are using for this run, but make sure the nods are both set to zero. These are being used to help locate pixels that have intermittent problems. Try to get one of these through each mask to maximize the area of the detector we are covering.

3) Put in the pinhole mask. For each Spec, B, and R filters (I proved to be difficult, but try if you would like), we want to take 10 exposures such that the pinholes are not saturated, but well-exposured. We want 10 exposures to get better signal in the halos. These should only be done once per run. In Dec, I found the following seemed to work - Spec : 30s ambient light (please check the counts for this, it may have been with the low lamps) - B : 6s and high quartz lamps - R : 1s with low quartz lamps

Evening twilight:

1) Take 3 or 5 twilight flats in each of B, I, and R (in that order). Counts in the range of 15k to 25k are best. Above 30k is getting high. Above 35k should not be counted. Track the telescope during these observations, but no need to guide on a star. Offset 1' N and E between exposures. During the course of run, consider prioritizing R in front of I to ensure that we do get flats here.

2) Let the telescope operator set up the telescope pointing and focus.

3) If doing an evening standard star, do here, before setting up on first science field.

4) Set up on the first science field. Wait until about 18 degree twilight. Do a final alignment check just before starting the first science exposure.

5) We find that data before 18 degree twilight does not satisfy strict nod & shuffle, so we prefer to start a science exposure at 18 degree twilight. It is ok to take a 60x8 exposure from 16 degree to 18 degree twilight; we'll see if we can rescue these data.

For each mask in night:

1) Slew the telescope, acquire guide and S-H star. Wait for the S-H to update the focus and mirror shape. If necessary, update the ADC (but this now seems to happen automatically).

2) Before inserting the mask, take a R-band direct image (20 sec) using sub-raster (200 pix). While this is reading out, insert the mask and setup ialign. You can also peek at the boxes to look for patterns in the stars; this is helpful in crowded fields for selecting the correct star in ialign. If there are no stars in the boxes, then there is a more serious problem (wrong mask, wrong coordinates, wrong subraster file?); the telescope should point to better than 20".

3) Run ialign and use it to move the telescope. While the coordinated offset is occuring, you can setup the GUI for the next exposure.

4) Take a R-band direct image (20 sec) using sub-raster (60 pix). While this is reading out, set up ifalign. ( With mask in!!!)

5) Run ifalign and move the telescope if the offset exceeds 0.1 arcsec. This offset is nearly always needed.

6) Take a R-band direct image (20 sec) using sub-raster (60 pix). While this is collecting, set up ifalign. While it is reading out, insert the prism and change to spectroscopic filter.

7) Run ifalign and move the telescope if the offset exceeds 0.1 arcsec. This offset is usually not needed unless the previous ifalign was large. Ask the telescope operator to start rotator guiding.

8) Take science exposure(s). If necessary, update the ADC before each one (but this now seems to happen automatically).

We are experimenting with checking the guiding by removing the S-H lenslets and checking the nominal guiding position. This could be done in the 'nod' position if one computes the guiding position for that case. Or during readout. We haven't quite resolved exactly what criteria in this would cause one to trigger a re-alignment.

9) Mid-way through the science sequence, i.e., after 2 of 4/5 or 3 of 6, update the alignment. Always do this the exposure after meridian crossing; IMACS seems to displace a noticeable amount during this. This alignment check takes only about 60 seconds and is clearly worth the time. In detail:

When the science exposure is reading, remove the prism and change to the R filter. Take a direct image (20 sec) using the 60 pixel sub-raster. While collecting the data, setup ifalign. While reading out, re-insert the LDP and the spectroscopic filter. Run ifalign, and send the offset if it is large. While waiting for the guiding to be updated, restore the N&S and Full readout mode (don't forget this)! When ready, start the next science exposure.

10) After all science, insert the R band filter and remove the prism. Take a direct image (20 sec) with the mask, full frame. While reading out, remove the mask. Take a direct image (90 seconds), full frame. While this is reading, ask the operator to move to the next field. ( Do Arc instead after February 2007? )

Standard stars:

Observe a standard star in evening twilight, the middle of the night or at morning twilight. Because this will be a slitless image, you need to pick a standard that is in a relatively uncrowded field. The instructions below will set the dispersion axis east-west, so one must avoid contamination in that direction.

DA white dwarfs are our best standards because their lines are broad enough that we can set the wavelength solution. Standards without lines will be hard to calibrate!

We want to position the star in a portion of the chip with few defects.

In May, we used LTT 7987 and EG 274. The latter is somewhat crowded. We used 1 second exposures for EG 274 and 3 seconds for LTT 7987.

    LTT3218 is another good DA standard near LST 8 hours, use rotator setting -270 degrees in OFF mode.
    First offset is 360" north, no east/west offset. Use 2 second exposure times.  Offset by -10" +10"
    between subsequent exposures.  Only need to readout CCD 2.

At present, we are not trying to solve out the airmass terms using multiple stars, though if there is time then getting multiple stars at different airmasses is a good idea.

1) Ask the telescope operator to slew to the position in the catalog and then offset by 4' east and 6' north. Set the rotator so that the IMACS gui image has north up. This combination will put the standard on the middle of CCD 2.

2) Let the telescope operator select a guide star and S-H star. Let the S-H correct the telescope and focus; we need a reasonable PSF for the slitless extraction.

3) Set up for the LDP with the spectroscopic filter. If one is confident in which chip the star is in, then one can opt to read out only that chip. This doesn't save time, but it may slightly reduce confusion later.

4) Take a short slitless image. You should see the WD on the image; it is the longest trace (because it is so blue).

5) While reading out, offset the telescope by 10" north or south.

6) Repeat the exposure and offset again, if desired. We need at least two exposures on a given chip, so that we can subtract off the sky.

If there's extra time, one can offset the star onto other chips and give two exposures each. This can be done without S-H and probably without guiding (since the exposures are so short).

If one is doing this, one should probably read out all 8 chips in each exposure; we might use other stars too when setting the relative transmission and it cuts down on the chance of not reading out the correct chip in an exposure.

At the end of night, fill out night report with any slitmasks you want changed for the next night.