Classic server: type "rfp" for read four pixels. The center pixel as normal and the one to the right of that one, plus the two directly beneath those pixels.

Tiptilt: (BEFORE you do this, read Observing below) Run TT at 20ms (40hz), and use NSUM at 3-5. After both of these are done, make sure you do a DBIAS for B5B6 when either the covers are closed, or not pointing at a star.

Telescope guis: Set both Finder and ACQ at 32/15 or 64/15 and with the finder cover and M3/M5 covers closed run a bias with the Dim Star button pressed (Brightness 160, Contrast 100) and when the bias is done, open the covers and change brightness to 200.

Notes: There is significant drift (significant when you cannot really see the fringes anyway) of the fringes on many baselines. Make sure you have a check star with easy fringes very close (less than a degree is optimal) that you can use to get offsets and at least monitor for 10-15 minutes to calculate the drift direction and magnitude

Observing: Do not change the TT until after you align on a check star as it will be difficult to track bright stars with the same settings as faint ones. Once NIRO is aligned with on the 4 pixels, NOW change TT to 20ms with 3-5 NSUM and slew to the fringe finder if you are not already on it. Find offsets on the fringe finder.

Once you have the offsets, move to the calibrator. On the cal, you will want to take two files one at each speed once you have the fringes, first probably should be the 500 nondestructive as it is the easiest to find, and then take it at 250Hz destructive. This is to make sure that if you picked the wrong speed for the conditions you wont have to go back and redo the calibrator to make a real bracket. Take note of the starting and ending offsets for the fringes to determine the drift. If the cal is quite close to the object, they should be similar Take both files, when the second is completed, you can relock the star one more time and check fringe offsets at 500Hz, abort, and then slew to the object.

After locking, try either of the two modes in the following manner. Choose speed, press 2 beams new, wait for windows to pop up.

For really faint objects there will likely be NO fringes in the upper left window nor on the waterfall, so you will have to watch the power spectrum. To find the likely location, press the HOLD button which will start Power Spectrum summing. Wait about 30-45 seconds as the PS builds up. If you see nothing in the dark blue frequency window, press clear and move one step >> left or right. Repeat around the area where the fringes are likely to be +/- 200microns. You can press hold again to stop the summing of the PS.

When you have found the peak in the power spectrum inside the window, hopefully it will not be the 50Hz noise spike we found today, but should be low and probably broad. The camera settings should be changed to get rid of this already now. Once you find and start saving, you can again run the PS sum, clearing it out every 40-50 frames or so to make sure they did not drift out of the window. The typical data sequence does not recognize this as good data so you will have to stop the file manually with the STOP button probably when the total scans including the first shutters is about 390. You can save both 250 and 500Hz and decide which is more likely to contain fringes, and then close out both brackets.

After the first bracket, best to pop back to alignment star and adjust alignment as it will have definitely shifted. Restart the bracket and move the reference cart here, and start with the cal again. Sequence is essentially: Cal - Obj - Cal - Realign - Cal - Obj - Cal - Repeat....