Differences

This shows you the differences between two versions of the page.

--- on_sky_testing [2026/06/03 12:09] – holtz
+++ on_sky_testing [2026/11/02 01:19] (current) – holtz
@@ Line 1: / Line 1: @@
+====October 31 2026====
+- QHY camera apparently recovered! although lots of ups and downs along the way, with it seeming to fail, but eventually got it working (not that software would keep turning it off until chiller fault was commented out). Replaced chiller with Coolance pump, at pump speed 8, seem to be able to achieve -20 C at cooler power ~68%.
+- investigated watchdog system and recognized that perhaps relay trigger needs to be on for more than one second to trigger watchdog reset.Changed to 5 seconds and testing...seems to be holding.
+- cal fiber shutter not working due to apparent relay failure. Replaced LCUS relay with Noyito relay
+- swapped channel 1 purple wire thermocouple for failed black wire thermocouple. Note that the half of the cell that wasn't being heated was not achieving temperature; both channels apparently need to be enabled (I had previously disabled channel 1 because of the erratic sensor readings).
+- checked lower dome, wouldn't open at first, but did after pushing on latch. On first two closes, didn't trigger stop switch, but after that, it did
+====October 20 2026====
+Catastrophic (?) failure of chiller. CCD held temperature about 10 minutes as thermocouples rose, then CCD power hit 100% and CCD temperature rose over 10 minutes, then perhaps power set to zero, then no response from CCD. Couldn't get CCD  to respond despite power cycles. Had Tracey unplug CCD after about 2 days.
+====August 2026====
+- compared flat field throughput with default fiber (2?) and extra fiber(3?), they seem similar
+- attempted to improve cal channel image quality, unsuccessfully
+- adjusted iodine cell to preserve spot position. Adjusted iodine_in stage position
+- refocussed guide camera
+- reset locations of cal spot, hole, stage positions
+- discovered failure of a mirror cover. Removed linear actuator and secured mirror cover open. Implication is that only one mirror cover will close. Following morning hooked up actuator off of telescope, got it to move short distance, repeated multiple times until I got it to open position, then it seemed to start work, so reattached to telescope
+- had telescope oscillation. Attempted to clean az encoder surface, no difference. Quitting PWI4 and restarting cleared it up
+- attempted some observations (mu Her), marginal throughput. Clouds arrived. Adjusted some guider stuff: expavg to 0.5s instead of 1, proportional term to 0.8 instead of 0.5 for subsequent observations.
+- subsequent night got one OK throughput on mu Her with good seeing. Had an azimuth disconnect.
+- Looked at Vega and could see the doughnut at the front of the iodine cell. Put a circular card over the cell and attempted to mark center of doughnut, then adjust position of cell, both downwards and laterally towards the iodine stage home position. Took series of images at different iodine stage positions and chose best. Iodine image quality was improved after this, but still not great. Doughnut is pretty large at front of cell!  Flipped the cell and repeated the process, and seemed to get significantly better image quality
+- did have to adjust iodine cell tilt a bit after flipping
+- proceed with some throughput measurements, but clouds came in, so difficult to interpret the poor results
 ====June 2026====
@@ Line 18: / Line 60: @@
 - took pixel flats with liquid cooled QHY600. This was achieved using a LED penlight wrapped in some bubble wrap and inserted into the upper light feed without moving anything in the spectrograph. Took 101 .0005s exposures
-- at end of run, inspected calibration frames and see apparent drop in flux from previous week, especially in flats, but also in ThAr, with drop in both cal channel and direct channel. Lots of investigation followed, determined that fluxes can change with reseating of fiber in cal source, although couldn't recover previous fluxes. Did fiber inspection and cleaning using APOGEE fiber microscope and cleaning device (with assistance from Jimmy Davidson), found some of the fibers pretty dirty, but didn't make dramatic changes in flux. Flipped the bifurcated fibers as one test. Direct fiber still much brighter in flats, but maybe not so much in ThAr? Final tests suggest significant change in flats but less in ThAr? Not clear what is going on, but maybe the illumination in the cal source is sufficiently non-uniform to make small changes in fiber placement significant, especially for quartz lamp?
+- checked lower dome, which wouldn't open, even though it hasn't been used since Ben last looked at it and claimed it was fixed and working. Perhaps there is some thermal effect. When it gets in this tightly wrapped state, a screwdriver can be inserted in the right side to push the spring loaded catch down, after which you can pull on the cable/push shutter to loosen the system, after which it worked. Experimented with software for safe use of lower dome shutter. When it does fail, it always seems to be in that it won't open, but closes fine, so perhaps the episodic failures are not a show stopper (Bill suggests that running the motor and having the couple slip won't damage the coupler). Otherwise, software seemed to work ok, although there was one instance where, after opening, software thought is was closed, so automatic closing didn't proceed as it should have. Probably should triple-check this!  Implemented some new supported actions in dome driver to allow for easier testing by setting open/close time for both shutters to a shorter value so that they don't have to go all of the way.
+- at end of run, inspected calibration frames and see apparent drop in flux from previous week, especially in flats, but also in ThAr, with drop in both cal channel and direct channel. Lots of investigation followed, determined that fluxes can change with reseating of fiber in cal source, although still couldn't recover previous fluxes. Did fiber inspection and cleaning using APOGEE fiber microscope and cleaning device (with assistance from Jimmy Davidson), found some of the fibers pretty dirty, but didn't make dramatic changes in flux. Flipped the bifurcated fibers as one test. Direct fiber still much brighter in flats, but not nearly so much in ThAr? Final tests suggest significant change in flats but less in ThAr? Not clear what is going on, but maybe the illumination in the cal source is sufficiently non-uniform to make small changes in fiber placement significant, especially for quartz lamp? But swapping bifurcated fiber didn't seem to change the relative brightness dramatically. Maybe some issues with quartz lamp?
 ====April 2026====