3.61
#...
·
 |
|---|
|
Hi folks, I started a new project last night. Going after the Witch Head Nebula. Wanted to share my progress here, and hopefully get some feedback. I'm not sure exactly how long this is going to take. I'm guessing maybe 20 to 30 hrs total integration time considering the small aperture size. This was first light for my QHY268C. Had about an hour of astro dark before moonrise, and managed to make the most of it. I'm still getting accustomed to the camera. It's my first cooled camera, but the APS-C size sensor is similar in format to my Fuji X-T100. So there is some familiarity. I used the Smart Histogram feature of SharpCap to determine exposure settings, but the sensor analysis data is somewhat skewed due to a bug. Bias was set to 12, whereas it probably should have been 20 or 30. Gain was set to 60, but should have been 56 or 0. Exposure times probably should have been longer than 40", but I was targeting highest dynamic range. And given that I only had an hour to work with, I opted for shorter exposures to yield more frames. Stack consists of 80/90 subs, 50 darks, 50 flats, and 100 bias. I used the recommended exposure time of 40.1", but probably should have rounded up 45" and built a darks library for reuse with that exposure time. Something to consider moving forward. This is just under an hour of data captured with a William Optics SpaceCat 51/250mm f/4.9 Petzval APO and Optolong L-Pro 2" filter.  Witch Head Nebula One thing I'm not sure about is what to expect from the Optolong L-Pro filter. I've noticed reflection artifacts in my images when using the filter, and thought maybe it was due to the filter being installed so close to the rear Petzval element. I've since moved it further away from the rear element using the adapter provided with the QHY268C, but I'm still seeing reflection artifacts on bright stars. I also need to adjust the camera tilt. I was rushed to set up, and forgot to check the field tilt. Will need to fix it before collecting more data.    This is the setup I'm using for this project.  |
0.00
#...
·
·
1
like
|
|---|
|
Nikkolai 20 to 30 hours seems a lot of integration time to me, 10 hours would be my target and from then your improvement will show diminishing returns. I agree that you need longer subs, I would try something in the 3 minute region. At Bortle 4 is the L-Pro really necessary? Just a IR/UV filter would seem to be fine with your sky especially if you wait for the moon to fade away. Cheers David |
|
3.61
#...
·
|
|---|
|
Thanks, David! Originally I added the L-Pro to help reduce gradients in my subs. There are light domes to the south and east of my location. But with the filter, there's the green color cast and the reflection artifacts. I think I may swap it out for an Astronomik L-2 tonight. Will also try 3' exposures like you suggest. I think the Smart Histogram had suggested something around 120" at Unity gain. Might also increase the offset from 12 to 20, and set the gain to 0, or Unity gain, using the High Gain readout mode. Need to re-do the sensor analysis today. Gains of interest bug has been patched in SharpCap. Should get better results from the Smart Histogram. Might also work on the mount a bit. Was getting 2.37" total RMS with the AVX. Which is still under the sample rate of this image train, but was seeing some elongation of stars in center field. Tilt needs to be adjusted, too.   |
|
0.00
#...
·
·
1
like
|
|---|
|
Nikkolai If your guiding is below par go for more, shorter subs. I would definitely go to 2" guiding exposures. Unity rather than 0 might help with this. The L2 is excellent, though mine has gone a little loose in its ring, I am awaiting a little tool to fix it, from Gerd Neumann. Best of luck. I am out tonight on M42, did 60 x 200" and 60 x 40" last night, I had satellite trails in 60% of the subs...damn Elon Musk. Hopefully they will average ourt in stacking using sigma clipping. I am aiming fo 100 of both sub lengths. HEQ5 Rowan mod (guiding at .7 RMS and less) Redcat 51, ASI071 and L2 Cheers David |
|
3.61
#...
·
|
|---|
|
Is anyone else rocking a QHY268 on the Cat 51? I had a lot trouble adjusting the tilt plate. The M48 adapter for the camera overhangs the adjustment screws, making it difficult to get at them. Even with a ball head hex wrench.  I'm not sure I actually improved on it much. Will try again tonight, but no more messing around. Will be using CCDInspector this time. |
4.72
#...
·
·
1
like
|
|---|
| I would go at least 5 hours, 8 hours max, and in bortle 4 you can probably use the L2 unless you're under 30 degrees from the horizon and there's light pollution in that area. I hate to say it, but the halo artifacts are probably from the filter, unless the QHY causes the artifacts from the glass window, but it's certainly not from the Redcat. The sub-exposure length probably should be at least 60 seconds, but not not much more since Rigel is in the frame. If you'd crop out Rigel you can definitely go longer. |
|
3.61
#...
·
|
|---|
|
Thanks for the feedback! Yeah, I think the reflections are from the Optolong L-Pro filter. No reflections prior to using the filter. After swapping out the L-Pro for the L-2 filter, the Smart Histogram was suggesting roughly 60" exposures at Unity gain (0 gain). But when targeting high dynamic range, it was suggesting 10" exposures at gain 56. These recommended exposure times are shorter compared to with the L-Pro filter, but I'm probably getting more light pollution with the L-2. I really want to keep Rigel in the frame, but it's so bright the challenge is not over exposing it. I just started a sequence using the recommended 10" @ gain 56 w/ offset 12. Aiming for 90' total tonight. It's nice and clear out right now.  I used CCDInspector for the first time tonight to evaluate and adjust tilt. It's still not perfect, but I think it's better than it was. Will try to do better tomorrow using flats. I also ran three PEC training sessions, and uploaded the averaged data to the mount. With PEC enabled, guiding seems to be doing a little better. Had some gremlins tonight. Trouble connecting to cameras, first the Player One Mars-M, and then lost connection to the QHY268 later on. Rebooting seemed to fix things. Think I also had an issue with plate solving. Mount lost its orientation, and then had RPC issues preventing further plate solving. Had to shut everything down and restart again to recover it. Everything seems to be running smooth now, though. |
|
3.61
#...
·
|
|---|
|
I wasn't confident in my tilt adjustments, so I reset the tilt plate before imaging last night. These are the first frames of two sequences taken on either side of the Meridian. Not a whole lot of difference after the flip. Which tells me the weight of the camera isn't impacting tilt as much as I thought it might be. Sensor is rotated so the left side of the frame is closest to and square with the dovetail.  Also had a chance to redo the Sensor Analysis on the QHY268C for read modes 0, 1, and 3. Results look good now that that gains of interest are applied.  Having switched from using the Optolong L-Pro to Astronomik L-2 filter, the issue with reflection artifacts appears to be resolved. Exposure times suggested by the Smart Histogram in SharpCap are a bit shorter due to the increase in sky brightness with the change in filters, but the difference in light pollution under Bortle 4 skies is almost unnoticeable for this target. The stars look great. I'm encouraged by the results so far.  |
|
0.00
#...
·
·
1
like
|
|---|
|
Nikkolai How much integration time so far. Looking good. I am imaging this target as I write, looking for four hours total, RASA 8 and ASI 071, 120" subs; I like the Redcat but love F2. Cheers David |
|
3.61
#...
·
|
|---|
|
Hi David, The stack from last night was only 56 minutes total (28x 120" subs). Thinking about increasing the exposure to 300" tonight. I haven't tried combining all the sets yet. I'm sure I'll be throwing out some of the data due to reflection artifacts and tilt. I had captured another 90 minutes worth of light frames after the Meridian flip last night, but the frame dimensions of the second set didn't match my calibration frames. This is due to a peculiar aspect of how SharpCap saves frames for QHYCCD cameras between Live and Still Frame capture modes. Which I had completely forgotten about it until I realized my two sets of light frames had different dimensions. Something to keep in mind. How is the field tilt/curvature on the RASA 8? From what I've heard, the critical focal distance is even tighter at f/2. Last night, I knew my field wasn't flat, so I used the Bahtinov mask to try and average the focus across the field by checking focus of stars on both right and left sides of the frame. I may use this technique tonight to help make adjustments to tilt. Then confirm the adjustments using CCDInspector. Since brighter stars can throw off the FWHM average, I think it might be a little easier to see where stars are in focus and by how much when using the Bahtinov mask. When they're all dialed in, fast scopes are an absolute delight. Good luck with the RASA, tonight! Clear skies, Nikko |
|
0.00
#...
·
·
1
like
|
|---|
|
Nikkolai My tilt is not bad, but I don't worry about it. I am an amateur and do this as a hobby that I never thought would be possible when I started my interest in astronomy in the 60s. I use APT for capture, Sharpcap for PA. Cheers David |
|
3.61
#...
·
|
|---|
Had another go at tilt adjustment last night. This time using a Bahtinov mask and the 4-corner method. This is prob about as good as it gets. Remaining tilt appears to be due to flexure under the weight of the camera. Like before, samples below are captured from either side of the Meridian. Might need to take some time here soon to work on the mount again. Still need to swap out bearings and replace grease. There are a few areas I need to pay closer attention to when reassembling to minimize eccentricity in guiding. I'm still seeing star elongation, even at 3.10"/px. Here are the Guiding Assistant results with the mount's PEC playback enabled. Training data consists of 3x averaged training runs. Might need to add more training data.  I increased the exposure time last night to 300". Here's the initial stack of last night's data (3h 20m total).  Captured 40x 300" subs at 0 gain in High Gain read mode. Moved from the backyard to the driveway to be able to image further West without utility lines getting in the way. Appear to be picking up more light pollution, but should be able to clean it up. |
|
3.61
#...
·
|
|---|
|
Combined the data from the last 6 imaging sessions for a total integration time of 9h 38' 44": 1) 2022-01-21: 83x 40.1" @ 60 gain, 50x dark, 50x flat, 100x bias, Optolong L-Pro filter 2) 2022-01-23: 242x 10" @ 56 gain, 50x dark, 50x flat, 100x bias, Astronomik L-2 filter 3) 2022-01-24: 377x 10" @ 56 gain, 50x dark, 50x flat, 100x bias, Astronomik L-2 filter 4) 2022-01-25: 60x 120" @ 0 gain, 36x dark, 36x flat, 100x bias, Astronomik L-2 filter 5) 2022-01-26: 25x 300" @ 0 gain, 36x dark, 36x flat, 100x bias, Astronomik L-2 filter 6) 2022-01-27: 35x 300" @ 0 gain, 36x dark, 36x flat, 100x bias, Astronomik L-2 filter Camera sensor cooled to -10C for each session. Using read mode 1 (High Gain).  I'm still not satisfied with the result. I might try reprocessing the sessions individually before combining the data. Think still need more data, though. |
|
3.61
#...
·
|
|---|
|
I don't know if I'm going about this right, exposure wise. Looking over the histograms of the FITS files for each exposure setting. They don't appear to be clipping black levels. So I think I'm managing offset correctly, but I'm not so sure about the gain value. The histogram peaks so far to the left when viewed linearly. Looks a bit better when viewed logarithmically. 1) Linear stretch function  2) Log(x) stretch function  There does still appear to be clipping with bright stars, but I think that might be unavoidable. Just a matter of how far I'm willing to push it. Been trying to avoid it as much as possible, but I'm also struggling to surface any detail of the faint nebulosity in the region. Even with the logarithmic stretch, all you can see are stars. Thinking maybe I'm should be pushing it further. I don't know, though. I'm already seeing bloat from stars with 5min exposures at gain 0. I'm going to start setting up again here soon. Will see what the Smart Histogram suggests. |
|
3.61
#...
·
|
|---|
I think I now understand why there is so much more curvature in one axis than the other. If I understand correctly, there appears to be more curvature in one axis than the other because the working distance with Petzval optics is dependent on the focal distance, since all four lens elements move together with focus adjustment. As stars get further out of focus due to tilt, they would also get further out of focus due to the difference in working distance from the corrector element to the sensor. The difference in apparent focus/correction or FWHM would then appear to be non-linear across the field with any amount of tilt. Again, that is if my understanding for how this works is correct.After inspecting the camera rotator assembly of the SpaceCat 51, I found I was able to tighten the index/tension ring another 35° of rotation. The rotator is pretty stiff now, which I'm hoping eliminates most if not all the flexure that I believe may have been the source of difference in tilt between each side of the Meridian.  |
|
3.61
#...
·
|
|---|
The SpaceCat is on its way back to Taiwan for servicing. During the inspection of the camera rotator, I found strips of plastic that had come loose and had bound up in the groove of the rotator's flange insert. My suspicion is that these are strips of PTFE tape that were adhered to the outside of the flange insert to act as spacers and plain bearings. Only one of the three strips was still adhered to the flange insert.   Curious about what these strips actually were, I reached out to William Optics for more info. I described the issue I've been experiencing with tilt and flexure, and asked if they could provide some more insight. They said I wouldn't be able to fix the issue with the scope myself, but didn't elaborate much. Partially because they would need to inspect the scope themselves. They did offer to service and repair it for a reasonable fee. In attempting to resolve the issue of tilt, I did manage to use the laser method outlined by Starlight Xpress to guide the correction adjustments made with the tilt plate. However, this was done through the OTA instead of with the sensor by itself.  The reflection pattern is a bit more complicated due to the additional optical surfaces, but it still works. I took some time to identify each of the reflections before proceeding with adjustments to tilt. The 4 large diffuse spots are reflections from the 4 lens elements. Then there's the filter, sensor window, and cover slip for the sensor. The surrounding grid are diffraction spots from pixels. When rotating the sensor to evaluate tilt, the rotation should center on the cover slip reflection.    I was still getting similar results, though. There still appeared to be flexure affecting tilt, and making the adjustments to tilt while the scope was positioned horizontally on the bench incorporated the tilt due to camera sag into the correction. For this reason, it's probably better to continue using the 4-corner Bahtinov method for guiding tilt adjustments, described in detail by Alberto Ibañez in his article on Sensor Tilt Adjustment. First capture below was made near the zenith, East of the Meridian. Witch Head and then Rosette were captured West of the Meridian, and were captured roughly 2hrs apart. Focus was not evaluated or adjusted between targets. Temperature dropped over the session from 45.7 °F at 8PM to 38.7 °F at Midnight.  Below are focus and field flatness evaluations done at the beginning of the session using a rich star field in the Milky Way captured near the Zenith w/ and w/o a Bahtinov mask. Each tile is 500x500px selected from raw FITS files: 4 corners, 4 edges, and center.   Results have been fairly consistent despite my efforts. I'm not sure that the camera rotator of the William Optics Cat 51 is actually the source of flexure. I'm not even sure that the scope actually needs servicing. But I thought, since the option was available, it might be worth a try. |
