North American Mosaic -- NGC 7000
08/12/21 -- Crazy idea -- since it's relatively low in the sky early in the evening... attempt a giant mosaic of the North American Nebula (NGC 7000) with the venerable Orion CN10 at F4.7, 1200mm. It will take a crazy number of panels to complete... but the result should be remarkable.
Started on night #1 with no moon, good predictions for seeing... but hot & humid. Started drift alignment at 8:50pm at ~350'. After a major over correction, brought the polar alignment to 1.65' after several minutes of drifting.
After plate solving with the Canon T3I to establish the current field of view and camera rotation, used the Kstars/Ekos scheduler to generate a 2x3 mosaic with 5% overlap to cover the nebula. The scheduler launched the first collection at 9:51pm...starting with autofocus (updated to use a 9 second exposure...to differentiate from plate solving which is left at 10s).
Found that the FCUSB light was turning off after the first attempt at moving the motor by the autofocus routine (which was definitely not focusing up stars). Tried re-arranging USB ports to put the FCUSB on it own (USB3) port (overkill) along with the Zwo ASI 120MC-S on the other USB3 port. The Canon T3I (being USB2) and the serial adapter used for the CGEM DX are on a passive USB hub. Changed the focus routine to use a sub-frame...which speeds up the process of getting frames from the T3I enormously. This did not ultimately help. It may be that the FCUSB draws too much current for the Odroid XU4 to supply...perhaps a powered USB hub will be needed. For tonight, used the Accufocuser hand controller to achieve manual focus and removed the autofocus component from the job. More diagnostics on a cloudy night for the FCUSB is needed.
Ultimate (around 11pm) plate solving complete and guiding working...capture started. Guiding RMS hovered ~3.35" RMS. Clouds rolled in shortly...limiting collection for the night. The first panel collection was satisfied, but the system failed after attempting to slew to the second panel with plate solving errors. The last frame captured for plate solving had zero contrast...could have been heavy cloud. The mount auto-parked at 4:25 (dawn with 15 minute buffer) and was parked correctly.
 |
| NGC 7000 6-plate mosaic configuration in KStars / Ekos for the 1200mm F4.7 Orion CN10 with Canon T3I DSLR using the RCCI coma corrector. |
Round two of mosaic collection didn't commence until 09/02/21 due to clouds and vacation. Started setup around 8:30pm. Positioned in the driveway to provide an initial view of Polaris through the polar scope. After an hour of working with the focuser (still not reliably working), started drift alignment with phd2 still running on the Odroid (no luck running it on the server and then connecting to the indi server on the odroid). Initially found to be ~30' off (DEC drifting up). Over tightened the right azimuth adjustment knob and went back and forth a few times until settling on ~4' alignment error. Loaded the mosaic schedule which sorted the plates in order from 6,4,2,5,3,1. Started collection ~10:40pm. Guide calibration went quickly (internal guider in Ekos/Indi) with guide RMS values of ~1.41" RMS noted after 5 minutes+ of guiding. Ultimately found 65 frames for panel 6 (across two sessions) and 40 for part 2. In the AM, captured 20 flat frames by pointing the OTA towards blue skies at zenith and then 20 flat dark frames with the OTA pointed at the ground with the Canon T3I viewfinder covered.
The initial integration for frames for panel six of the mosaic with 9120 seconds (2.53 hours) from frames collected on 08/12 and early morning on 09/3/21 showed issues with strong dark current... suggesting a new set of darks may be needed (little nebulosity could be brought out).
 |
| Plate 6 from NGC 7000 mosaic with old darks collected during hotter weather. Processing to bury residual noise bands resulted in almost no visible nebulosity in the resulting integration. |
After collecting a set of dark and bias frames at the same temperature (~61 degrees outside) the subsequent night and reprocessing, immediately obtained a workable result:
 |
| Plate 6 from NGC 7000 mosaic after processing with recent darks collected at current ambient temperatures |
Collection session #3 occurred on 09/06/21. Aligned without the benefit of the polar scope (couldn't position the scope to see Polaris over the trees). Started with some experiments to get the Shoestring Astronomy FCUSB autofocus controller talking to Ekos/Indi. After unplugging & plugging back in, was successful after connecting with a laptop and testing Kstars/Ekos/Indi from the drive way. Finally, used Kstars to a star near the meridian & equator and then stopped Kstars / Ekos and Indi on the Odroid to start a drift alignment with phd2 running on the Odroid (using native drivers). After several iterations achieved ~3 arc second error. Set up the mosaic job (leaving plate 6 off the schedule) and collected successfully / autonomously (with one issue with exposure timeout from the guide camera) through ~5am when the twilight safety buffer came on. Collected several frames manually until ~5:30am and then parked. Collected ~20 flats pointing the OTA at zenith as well as ~20 dark flats w/ the OTA pointed at the ground... and disassembled at 7 am for the day.
Collection session #4 started on 09/07/21 with the mount & OTA assembled before dark. Did a rough manual focus with the Accufocuser hand controller at 8:30pm. Started polar drift alignment at 8:45pm with PHD2 running directly on the odroid (no indi drivers). After many iterations back and forth (and some distractions), completed alignment at 9:45pm with sub-2 arcsecond alignment error. Started working on the autofocus routine... no success initially. Worked hard to get the internal (Ekos) autoguider to work... no luck with calibration. Rebalanced the OTA on the mount...which improved response in DEC and allowed the internal guider to calibrate successfully (although still rough... issues with elevation may remain). After enormous struggles... including focusing on Denab by hand... was basically functioning at midnight. Started the scheduler for Part 1, 3 and 5 (in that order in the scheduler). Plate solving frames look awful with significant elongation. The internal guider was again unable to calibrate. Suspect the mount shifted on vibration pads after leveling...or it's possible that although the mount is level, the elevation shifted... causing issues. While polar drift alignment was done in azimuth, it was not repeated for elevation because of limitations seeing close to the horizon. Adjusted the mount with the level and found a small degree of error. This appeared to help... resulting in a successful internal guider calibration. Restarted the scheduler, which ran successfully over night with fully automated collection on two different plates through 5:15am when the mount auto-parked.
Collection session #5 started on 9/09 with the mount & OTA assembled before dark. Took special care to level the mount after mounting the OTA (which did cause shifts). Utilizing a new set of vibration pads to avoid slippage seen with the old worn out set. Used the analog inclinometer to examine the OTA and noted a 51 degree slope (given latitude of 39.02 degrees). Hoping this removes the need to adjust elevation via a drift alignment leaving only azimuth as a concern. Remarkably, the Shoestring Astronomy FCUSB autofocus controller connected to the Orion Accufocuser motor was working immediately. Noted remarkable shift in focus relative to 09/07 with the image badly out of focus. 30x100ms of outward motion was required to get close to stars (11.7 HFR). This was close enough to attempt a sub-frame autofocus. Adjusted the initial step size setting to 500ms since 100ms seems far too small and also updated the tolerance setting to 5%. Noted that sub-framing worked successfully despite attempting the autofocus before performing the polar alignment (suggesting the initial placement was close ... by luck).
 |
| HFR plot from the autofous module on 09/09 |
Began drift alignment with PHD2 running on the Ordroid XU4 at ~9:15pm after performing a calibration. Indeed the initial positioning was lucky with polar alignment error of 25' (DEC drifting down). Noted with the OTA flipped to the other side of the meridian, adjustments are reversed (tighten on the right rather than the left in this scenario). Ended up iterating to ~2' of polar alignment error.
For grins, did one more round with autofocus to see if there was any significant change following the polar alignment activity. No significant differences were found with only minimal adjustments made during the routine (starting at 3.5 and ending at 2.5 HFR).
Finally set up the scheduler with the mosaic jobs for panels 5 and 1 (which have the least frames collected so far), but found the scheduler would not launch collection. It sat showing the job as idle.
Exited kstars and examined running processes. Found an additional instance of kstars running and killed it. Restarted and the scheduler worked as usual launching plate solving. Guiding calibrated successfully (internal guider in Ekos) and began capturing frames ~10:45pm. Total guiding RMS hovered around 2". Collection on plate 1 and 5 occurred successfully & automatically over night with frames collected until twilight at 5:05am. Ekos shutdown successfully and parked the mount.
Session #6 began on 09/10/21. Started by planting the vibration pads under the tripod under the gravel in the drive ... rather than on top of the gravel...in an effort to get as much stability as possible to facilitate a better polar alignment and better stability / guiding performance. Positioned in view of Polaris in the polar scope...which was used for initial alignment (good agreement with inclinometer). Started KStars/Ekos to position the mount at the meridian / equator point and performed a successful autofocus (initial step size 2000ms). Moved to PHD2 running on the Odroid XU4 and measured an initial polar alignment error of ~25' (DEC drifting down). With the OTA on the far side of Meridian (already performed a meridian flip when autofocusing), tightened on the right to make the initial adjustment...resulting in an alignment error <1'. Just to compare guiding performance of PHD2 running on the Odroid vs the Ekos internal guider (nominally running on the main server), noted DEC and RA RMS values both around 2.5" over several minutes of guiding (~2.58 for RA and ~2.68 for DEC axes). Completed polar alignment and this test guiding session at ~9:45pm. Started Kstars/Ekos on the main server and restarted the indiserver on the Odroid. Started the scheduler with the full compliment of plates. Overnight, frames on plates 6 and 2 were collected. Frames from all three sessions on plate 6 were integrated (over 4.5 hours), resulting in this:
 |
| Plate 6 of the NGC 7000 mosaic after three imaging sessions totaling ~4.5 hours with the CN10. |
Used Astro Pixel Processor on frames thus far on plate 2 (two sessions totalling about 3.5 hours). Used the light pollution correction tool on both of these (plate 6 and 2) but applied the color calibration tool on plate 2, which produced a significantly different result:
 |
| Plate 2 of the NGC 7000 mosaic with frames integrated from two sessions totally ~3.5 hours. Applied both light pollution correction / background calibration as well as star color calibration in Astro Pixel Processor for plate 2. |
Session #7 began on 09/11/21. Had left the mount in the drive from yesterday... so started with an initial polar alignment with PHD2. Showed an initial error of ~25" (DEC drifting down). After several adjustments, achieved a polar alignment error of ~1". Manually focused on Deneb with the hand controller (noting that the focuser lock screw does work and does prevent the autofocus motor from functioning...doh). After loosening, was able to perform an automated autofocus routine. Started the scheduler on plates 3 and 5 at ~midnight and it collected through ~1:50am (having completed the plate 2 job of ~40 frames). It then switched to plate 5 and collected until ~5am (an additional 61 frames). Integrated these together in Astro Pixel Processor.
Session #8 began after a long break due to clouds and the moon on 09/29/21. Was able to level the mount and use the polar scope for a rough initial alignment by 7:30pm. Skies clear, no wind with the moon at ~37%...but not rising until after midnight. Seeing predicted to be poor, however. Ensured the OTA was balanced in RA and DEC.
First performed an alignment of the guide scope with the main OTA with plate solving to within ~40 arc seconds.
Then completed polar alignment with phd2 running over X from the Odroid XU4 (still segfaulting if running locally and connecting to remote indiserver). Sequence of corrections include:
34.21 (dec drifting down...) <-- tighten slightly on right
7.7' (dec drifting down...) <-- ditto
19.9' (dec drifing up) <-- tighten on left
17.6 (dec driving down)
4.45' (dec drifting down) <-- done @ 9:38pm
Ultimately, nearly 40 part 2 frames were collected along with a large number of part 6 frames.
Session #9 began after another long break due to moon and clouds on 10/31/21. Weather started with light clouds (clearly by 9pm) and only light, occasional breeze.
Began setup at ~5pm with leveling the mount. Placed vibration pads under the gravel (similar to session #8) to minimize the chance for shifts. At ~7pm, aligned with the polar scope. Moved the OTA into position with KStars running on the laptop, then forwarded the X-session (with compression) from the Odroid. Initial polar alignment error was -14.7' after ~10 minutes of drift. Tightened on the left resulting in -16.7 (odd). Attempted an additional adjustment with tightening on the left, resulting in -22. Switched to tightening on the right (note... no meridian flip had been done) and the (final) result was -3.6'.
Noted that the guide scope is misaligned (mostly in DEC) by ~730" relative to the primary OTA. Did not take the time to realign it ... consider this likely 'close enough' for effective guiding.
Autofocus failed to make progress towards a focused image. Unplugged the Shoestring Astronomy FCUSB controller and plugged it back in (while resetting the Indi/Ekos server). This resulted in working autofocus. Completed setup steps at 9:46pm and launched the scheduler after deleting sequences for part 2 and part 6 from the list to be collected.
After obtaining 11 frames, re-examined the focus with an additional autofocus step. During autofocus, the HFR varies from 7 to 4 suggesting the focus motor and controller were making adjustments...however upon inspecting the focus motor and focus tube, noted that the focus lock was not loose and (with the hand controller) this prevented the focus motor from moving the focus tube inward. After adjusting and performing a quick manual focus on a bright star, tried the autofocus routine again. Initially a HFR of 3.0 was noted which was not improved upon by the autofocus routine. Restarted collection on part 4. Guiding statistics showed sustained RMS of ~3".
 |
| Ekos Drift Graphics for 10/31 with RMS of ~3". |
Session #10 began on 11/08/21 with great skies, very little moon and no wind. Started set up at 7pm with leveling the mount...with balancing and initial polar scope alignment completed at 7:40pm. Started drift alignment at 7:57 with initial plate solving. Rough focus required for a successful plate solve. After solving and using for Gamma Piscium for final fine focus adjustment (with the manual controller...no time to mess with autofocus problems tonight), the initial drift alignment error in phd2 shows as -14.7' (suggesting tightening on the right). Final phd2 polar alignment error of ~1' achieved. Started the Ekos scheduler for plates 5, 3 and 1 (which have no collection with the corrected camera orientation) at ~8:50pm. Autoguide calibration was successful and initial RMS errors were remarkably low at ~1.1" for RA and ~1" for Declination.
 |
| Meteoblue.com astronomical seeing for 11/08/2021. |
 |
| Ekos with the CGEM DX and CN10 oriented on the Horse Head nebula following plate solving. |
Dark Session: With colder weather setting in (lows in the 40's and even upper 30's), decided to collect an additional set of fresh dark frames during the evening of 11/09. Set up the camera and covered with foil and used the Odroid XU4 with gphoto2 driver with an intervelometer shell script to collect 30 darks.
Session #11 began on 11/10/21 with good skies (going to high level clouds late), moon at ~35% and low in the sky and no wind. Started set up at 7:30pm with leveling the mount...with balancing and initial polar scope alignment completed at 7:45pm. Ran into an Indi software issue with the gphoto2 driver...but an update seemed to fix the issue (random restarts, crashes). Started initial plate solving for drift alignment at 7:57pm. Performed fine focus adjustment (with the manual controller) on Gamma Piscium and rotated the camera back into position to be approximately the same orientation as previous NGC 7000 collections. With phd2 running on the Odroid XU4, initial polar alignment error was ~-5.3' (with a rough initial alignment with the polar scope). Final polar alignment ~3.5' with a single additional adjustment (tighten on the right....pre-meridian flip).
 |
| Meteoblue.com astronomical seeing for 11/10/2021. |
 |
| Montgomery County Astrophysical Observatory prediction from cleardarksky for 11/10/2021. |
 |
| Current weather from weather underground for 11/10/2021. |
Configured the scheduler to begin collection on plates 5, 3 and 1 and then 40x3 collections against NGC 2024 to compliment the Horse Head nebula collection from 11/08/21.
Guide calibration occurred swiftly and initial guide results were favorable with RMS of ~1.34" in RA and ~2" in DEC.
 |
| Example from Ekos guiding module from ~8:53pm. |
From the scheduled jobs, upwards of 33 frames were collected against part 5, 40 frames for part 3 and six frames for part 1 before hitting the horizon. Collection then switched to NGC 2024 with 60 frames total recorded.
Comments
Post a Comment