Kudos for ZWO!

starnet59

Normally I am quite critical of ZWO in my posts, for good reason I think, but when credit is due, I am very glad to offer it.
Yes, it took more than 18 months of constant complaining and asking for these basic but essential features, but I have to concede that ZWO has finally delivered on three major items, EQ mode, mosaic (or framing as ZWO calls it) and planning mode.
All these three are really big new features and as of the latest version they finally do work well.

EQ mode just works and as an owner of both a S50 and a Dwarf 3 I am happy to report that ZWO's implementation of it is better. Especially since visibility of the zenith for polar alignment is no longer required. I have both a Sky Watcher clone like this one https://www.amazon.ca/Latitude-Equatorial-Adjustment-Telescope-Accessory/dp/B0DQTM3CT5/ref=sr_1_1?crid=3RHH2A3M2K7J0&dib=eyJ2IjoiMSJ9.TypIxFzTF83JOhVO7Xlt4ccUCJRLaNPRkPwpLoO6ujW8TjHeX_hzCk_gC1U6LOr8712tudmmfsQBPV0oNKaGvJ35WWzu1zUyGxmLPCh5C-z4gRv08PzJT19pLt1dBodHq2jeWFMmzqUynAEJ8K7ftGHgKg56RJkasZPACZlB44yyGMG-K7NVZeRweK8lTVoU2h3jLMRNKszYXz9w2UUX3itoH8VLUr2HmAXoqkNd7Ks.VxSP3x28s_JyvS--SphoGDLQ2dbLoN8Ml6tHlRkOQ3U&dib_tag=se&keywords=eq+wedge&qid=1755138354&s=digital-text&sprefix=eq+wedge%2Cdigital-text%2C123&sr=1-1 and a better one which is twice the price https://www.amazon.ca/dp/B0CT2NLX1F?ref=ppx_yo2ov_dt_b_fed_asin_title&th=1 - they are not cheap but worth every dollar. If you cannot afford one, 3D print a fixed wedge, I published 90 of them in one degree steps on printables.com - either way get one, it makes a lot of difference.
Mosaic mode did not work at all when first introduced, I was never able to produce a single good image with it, one night was not long enough, and I completely gave up. I am very happy to report that the second version works much better and is actually quite usable. It takes now only 15 minutes of stacking time (which for me translates into 21 minutes of real time) to completely cover an entire 2x frame and then run AI Denoise on it (which refuses to run on an incomplete mosaic image for some reason). I normally do 2 hour exposures which increases the signal to noise ratio but even a 15-minute 2x mosaic AI Denoised looks quite good (that's M52 in the lower left corner):

I did not expect this to work so well now and I am really impressed.
Planing mode is also better than the Dwarf 3 one, in many ways. You can see your plan in SkyAtlas, you can edit and reuse an older plan, etc.
It took one year and a half but these three features have made a huge difference in the way I use and enjoy my S50. I am just bitter it took so long and so many sleepless nights. I now use the S50 exclusively in EQ and planing mode and since I just discovered that mosaic actually works now, I will start using it too.

Are things perfect, not really, so here are some constructive suggestions:

EQ Mode - I am quite happy with it but some people here have reported flaky behavior when you redo the polar alignment repeatedly with the strange workaround that it works if you exit and then re-enter polar alignment mode. Also, some users see unstable accelerometer and magnetometer values. Worth investigating and fixing if possible.
Mosaic mode - Since mosaic mode works so well now, with a 2x frame doable in 2 hours, I think ZWO should increase the maximum size of the mosaic frame to 3x, a capture time of 4 hours or so seems quite reasonable now. Another feature that the Dwarf 3 has and is missing in Seestar is independent size factors for width and height, the Seestar forces you to use a 9:16 form factor.
Planning mode is actually fine, my only complaint is that in EQ mode the frame is rotated by an arbitrary angle and what you see is not what you get. It turns out that this only happens if you are not connected to the S50, for example during the day. The workaround is to turn on your S50 and connect to it, then everything works as expected. This is fine as a temporary solution but there is no reason things should be this way and this should be fixed. Another thing, ZWO definitely needs to add 4K mode to planning - the way you can choose to use the LP filter or not, there should be another choice for 4K or not. Right now there is no way to get 4K if you image from a plan, you need to capture manually to use that feature, which brings back bad memories of sleepless nights for me, I am definitely not going there. Another very useful feature would be the ability to split a plan over multiple nights, especially with 3x mosaics that would take 4 hours each.

All in all, I am quite happy now with the Seestar/S50, especially with these three major new features (four if you include 4K mode). They took a really long time to implement but the telescope is finally usable now and not just a toy.

Now if I could get stacking efficiencies in the 70% to 80% range instead of the current 50% I would be really happy!

Daemon42

Are you seeing better stacking percentage in Mosaic mode, or just less likely to fail an hour in (as many of mine did, I think due to frame rotation in Alt/Az mode as it was working out near the corners near the end of completing a full frame)?
And did you see this improvement before v2.6?

FWIW, I'm running a 2x Mosaic right now and at 9 minutes of exposures it's barely moved out of the center. Certainly am not going to cover the whole frame in 15 minutes of exposures. That's why I ask if you're only getting improved results with v2.6 as I have not upgraded yet.

FWIW, I also own both of those EQ mounts and the 2nd one is 1000% better.

marvindj

Daemon42 Others have noted that with the V2.6/Firmware 5.34 that the mosaic capture times have been noticibly reduced as per starnet59 comments above. I haven't tried myself, but , yes, app/firmware upgrade to this level is required....

starnet59

Daemon42 My experience is different than yours. The image I posted previously is a 2x Mosaic, after just 15 minutes of stacking the image was already fully covered. Still quite noisy but AI Denoise took care of that. I let it run for 2 hours and I was able to complete four 2x mosaics last night, each one 2 hours and with about 50% stacking efficiency. With two more done the previous night I now have a 3x2 mosaic, each image a 2x mosaic so this is like a 6x4 one, I will try to stitch them manually today.

starnet59

Here is my 3x2 mosaic, each component itself being a 2x image, captured over two nights. The original is a 5800x7200 image, I had to downsample it by 2x and compress the JPEG to be able to post it here. If the weather cooperates I plan to add more images to it:

Daemon42

Seems to be quite a lot of banding everywhere in that composite mosaic.
And you still never really answered my question about v2.6.
I'll assume marvindj is correct and that's when the improvements started.
Skies haven't been clear enough for me test this week at all.

edit - Skies just cleared up tonight. I see where the banding comes from. The new mosaic capture pattern moves in quite large steps. About 40-50% overlap on the short axis and 25-33% on the long axis where the old pattern only moved like 15-20% on both, which took much longer to get a full pass, but got cleaner results when it succeeded. I find that, especially in higher Bortle areas) the edge of the frame can be a lot noisier than the center so the bigger the jumps, the more banding you get.

At 25 minutes real time I had 10 minutes of stacked exposure time so the exact same 40% success rate (in Mosaic mode) as before v2.6, it's just covering more ground. At 14 minutes of exposure time, it still hasn't filled in the top 20% of the 2x frame. I figure it'll need another 4-5 minutes of exposure time, to fill in what I can see and get the overlap at the edge of the frame.

Took 28:20 of stacked exposures to cover the whole 2x frame (including overlap along all borders to get a full clean frame) in 68 minutes of real time. That's with zero frame rotation. Unfortunately my target went behind my roof a few minutes after the full frame so had to stop at one pass.