SC4 and Maxim v6.4

One more thing. I think this is minor but since I'm on the SC4 topic here goes - Is there a possibility of an internal LED showing up on the top and side edges of a the SC4 image shown in DLImaging AO Control / Locate ? The attached screenshot show s bright area on the lower right side of the image and a lighter area on the top of the image. This is true for every SC4 image. Occasionally, the guide star selector gets confused by the right image which always shows with a Brightness of 119.

John

 

That is exactly what I see every time I try to calibrate bin 2x2 or bin 3x3....some random spot ALWAYS to the top left.

 

Hi Colin. More testing was done last night. Logging was done in Verbose+ mode last night to give a slightly different data set. Unfortunately the mount bumping issues do not show up in this logging mode. It would be useful to include everything in Normal mode plus the PA of the rotator in the log. I can confirm the issue observed two nights ago. The mount will track and bump correctly ONLY for the PA used for calibration. If a new guide star is selected with a different PA the mount switch settings will be incorrect and the guide start will eventually be bumped off the sensor. Once bumping starts in the wrong direction it will continue until it completely fails. This was tested at three different PAs (0, 30 and 272 degrees). The Calibrate function in the Camera Control panel was used at a PA of 0 degrees. This yielded Guider Settings / Manual Calibration / Angle of -179.68 degrees. FWIW I also tried the ACP guidercalibration.vbs script with yielded an Angle of 1.3 degrees. Tests were run with this setting manually set at 0 and -179.68 degrees with no difference in the result. Mount bumping failure was observed in each test.

Testing was also done at Bin1, Bin2 and Bin4. Selecting Bin1 or Bin2 didn't seem to make any difference in Calibrate mode. When Tracking Bin1 was unusable, with the target star jumping all over the box resulting in large X, Y excursions. Bin2 settled the excursions to normal. Bin4 dampened the guider star movement significantly (as one would expect with a much larger effective pixel size).

Hope this helps.

John

 

Hi Colin, I am using Maxim for the tests. Bob Denny's advice is to run ACP's calibration only after Maxim DL is properly calibrated. I tried ACP on the first night. ACP calibration showed a relative PA (between guide camera and primary camera of 1.3 degrees which seemed correct to me. It too had issues with the rotator so I went back to Maxim only to make sure the PA was being handled correctly. The relative PA was a problem we saw six months ago when calibrating an AOX / FW8 (with internal guide camera) in Chile. We had the PA incorrect and saw the same bumping issue. By trying different relative PAs (at 90 degree intervals) we eventually got to a good result. I've tried the same process in this installation but that didn't seem to do anything. The log file does not appear to contain rotator angle. For tests like these that information would be really handy. Is there any way for you to add this information? I'll take a look at the ACP logs to see if this information is available there. I will try ACP calibration tonight and report back. Thanks for your quick response.

John

 

Hi Colin, how would you like me to proceed? What tests can I run that would assist in determining a fix for this problem? Is there anything in the setup configuration that could cause this bumping error? I'm happy to help in any way and am fine with taking the production CDK24 out of service to get to a fix. But we should have some set of tests in place.

Best,

John

 

Hi Colin, many thanks for the detailed plan. The mount is a PlaneWave L-600 which operates like a fork mount with just one side supported. It is an Alt-Az mount with an equatorial wedge. In the setup it is designated as an equatorial mount. Perhaps it should be designated as an Alt-Az?

John

 

Hi Colin,

The test / installation sequence that you outlined has been completed. The installation is a CDK24 (4m focal length), L-600 mount, Moravian C3 imaging camera, Moravian filter wheel and AOX / SC4. Here is what I've found.

1. The guider calibration and tracking must be done using BIN 4. The required arc-sec/pixel is about 1 arc-sec/pixel. For the SC4 / CDK24 the pixel size is 4.8 and the telescope focal length is 3950. The 0.7 reducer in the SC4 changes the focal length to 2765mm. The BIN 1 resolution is 0.36 "/pixel. A BIN 3 selection might be usable but the BIN 4 (1.44 "/pixel) is what was successfully used. Previous tests were done at BIN 2 which resulted in oversampling and "ringing" where the guider would chase the star. Bottom line the 1 "/ pixel resolution minimum is critical.

2. Guider calibration was done using Maxim DL (no ACP) at PA of 0, 90, 180 and 270 degrees. Maxim DL will track successfully for any PA provided a calibration is done first and the PA does not change. That is, Maxim does not track PA after calibration and only uses the most recent calibration information. A careful read of the Maxim DL documentation (and a lengthly discussion with Bob Denny) confirms this observation. An attachment shows the tests which were run.

3. After a successful Maxim DL calibration at PA=0 DLImaging was reconfigured for no AOX the ACP guidercalibration.vbs script was successfully run. See attached screen shot. The relative PA setting was 179.6 degrees (correct) and the guiding switch settings were calibrated. See attachment. Using ACP the SC4 / AOX guiding works with any position angle without further calibration. Once again, the guider binning must be set to 4. With the AOX/SC4 having a fine tuned FOVI model for guide star location is essential. After setting RA, Dec and PA center the guide star in the Maxim DL cross hairs using the Maxim / Telescope jog feature. When centered note the new RA and Dec. Those are the numbers which will be used for ACP imaging scripts.

Now for the open issues:

1. The locate function frequently results in Maxim DL halting. This happens when used manually or when called by ACP. It must be restarted by hand. This always occurs immediately after a guider image is taken and displayed. Log files from Maxim and DLImaging for June 6 are attached. Good examples of this can be seen at 06:29 and 06:41 . This happens at least 20% of the time when the Locate function is used. I have seen this also happen during Maxim guider calibration. I suspect the SC4 driver is the cause of this instability. This issue prevents the AOX / SC4 from being used in a production environment so it is a critical item to be addressed.

2. The Autoguider image is over stretched resulting in a large over exposed area (see screen shot attached) sometimes completely obscuring the guide star. Using the Maxim DL stretch tool will clean up the image and reveal the guide star. I haven't seen this with any other guider. This over stretched image happens when Maxim DL is stand alone or called by ACP. This over stretched image appears in both the Maxim DL window and the locate window. For diagnostic monitoring this is a real problem. If the over stretched image is used for guide star location this would explain loss of guide star on an otherwise perfectly clear image session. All testing was done at BIN 4. This is a significant problem which should be addressed.

3. The Tracking Error Graph is not populated when using AOX / SC4. The data simply doesn't appear although the graph axis can be set. Not as high a priority as items 1 and 2 but really would be a nice to have.

Final tests were run with 1800 second exposure times to verify tracking and bumping. Log files are attached - June 8, 2024. The threshold for bumping was set at 40% and the bump time was set to 0.1 second. The L-600 in equatorial mode has a good but not perfect polar alignment. Fine tuning the polar alignment will decrease the likelihood of the need for a mount bump. After all, the goal is to have no mount bumps at all. With these settings nearly all of the mount bumps occurred after about 900 seconds exposure time and nearly all were on the X axis (RA). This makes sense because RA is where all of the mass movement is (over 600 pounds of mount and OTA). Dec is easier with less mass. 0.1 second bumps resulted in short 1 arc-sec tracking variations (as observed using PWI4). All of the bumps were in the correct direction so in each case the guide stat recentered correctly. No guide star excursions. All of this was done at BIN 4 for the guider. (BIN 2 or BIN 1 is used for the main imager). Using this technique the tracking was nearly perfect - within a pixel at 1800 seconds (as close as I could measure). Log files from June 8 are also attached. They show several long (30 minute exposures). Plenty of loss of guide star errors in the AO log but it kept tracking and bumping correctly. Perhaps these errors are a result of the over stretched image?

I'm happy to run any other tests that would help address the two major issues that I've observed.

Cheers,

John

 

Thanks Colin and Adam. Some additional tests were done last night mostly to confirm that I could easily reconfigure the system back to the original external guide scope and then return to the AOX / SC4 by setting items in Maxim Configuration. That worked perfectly - all of the guider parameters were set correctly. A few items to configure by hand in ACP and the change takes only a few minutes. I did several 15 minute exposures - each with the PA moved 90 degrees with ACP controlling Maxim DL. The log files are attached. For most of the runs the guide star remained centered and never exceeded the bump threshold. On the first one done at PA=0 the mount did require some bumping near the end of the exposure - and those bumps were correct recentering the guide star. Over the course of the session no Maxim DL crashes were observed. The over stretched guider image was present for every guider capture.

A few other notes: It would be useful to add the PA to the Maxim DL log. The AO log sometimes logs the correct ADU and SNR but mostly those fields are zeroed.

John

 

Thanks Doug, I overlooked the fact the Maxim DL is not connected to the rotator when ACP is running.

John

 

Colin and Doug, I have something definitive for your team review. In an earlier post I mentioned that the guider image appeared to be over-stretched. Further testing demonstrated that the locate function guider exposure time is completely incorrect when used with ACP. That is I can set the exposure time manually with Maxim (say, 5 seconds) and easily locate a suitable guide star and then use that location for tracking. When ACP calls the Locate function and sets the exposure time that time setting is off by orders of magnitude - like a floating point number issue. An examination of the exposure time setting box showed numbers like 5.434e-47 and the like. So the guider was effectively taking a 0 second exposure - that explains the stretch issue and loss of the guide star. Sometimes that number is large enough for it to work (seems completely random) but the number itself is not what is passed by ACP. I've attached two screen shots showing the partial time box for a single locate call. Somehow the API for DLImaging is not taking the ACP exposure time and passing it correctly. Screenshots show the 5 second locate exposure done manually and the two parts of the ACP call to do a 5 second locate exposure. This, obviously, is a major issue.

John

 

Thanks for the quick reply. The system was converted back to the old guider (separate guide scope) last night which works with ACP / Maxim and completed the imaging for the night. But not with the precision that the AOX/SC4 offers. Please advise if you want any additional tests.

John

 

"One more thing. I think this is minor but since I'm on the SC4 topic here goes - Is there a possibility of an internal LED showing up on the top and side edges of a the SC4 image shown in DLImaging AO Control / Locate ? The attached screenshot show s bright area on the lower right side of the image and a lighter area on the top of the image. This is true for every SC4 image. Occasionally, the guide star selector gets confused by the right image which always shows with a Brightness of 119. "

Was there a reply to this observation? I may have missed it. I see this phenomena on my original SC3, and at first often thought there was a bight guide star in that direction (center right edge for me).

 

Hi Colin, I ran some further tests using ACP to call the Locate guide star function in DLImaging. All resulted in random numbers showing up for exposure time in the Locate guide star panel in Maxim DL. The ACP code functions perfectly for earlier generations of AO and, for example the FW8e guider. That code hasn't changed in years. It certainly appears to me that the DLIimaging API for the Locate exposure time is different than previous generations of guider code. It looks to me like a data format problem. This should be a very simple fix that will address the other symptoms I've observed.

Cheers,

John