Tilt issues return!

[04Stefan07]

Took a few subs the other night and upon inspection of my images I saw some focus/tilt issue.

Been ironing them out tonight and now looks like theres tilt. I have the ZWO tilt adapter but it's a nightmare to use. To make adjustments you need to disconnect from the PC, unplug, adjust the tilt, put back on, connect all back up, refocus and take a sub!

Is there a quicker way to deal with tilt? Can confirm it is? Have attached links to the subs and a screenshot.

120s Ha, equipment (main rig) in my sig.

Currently using 55mm back focus.

Edited July 5, 2022 by 04Stefan07

[symmetal]

The push/pull tilt adapters are awkward to adjust and they usually have the adjustment screws on the wrong side for easy access. If you can mount it further forward in the imaging train with spacers between it and the filter wheel to allow allen keys to access the screws without having to disassemble everything it certainly helps.

The ZWO MkII adjuster looks like it's reversible which is a better design. The easiest to use tilt adjusters are the Gerd Neumann ones though they are espensive and M48 is the smallest, and with adapters to M42 may be too long to fit. Their adjustments are 'push' only mounted around the edge, as the 'pull' uses strong springs.

Can't see any links to your subs though. 🤔

Alan

Edited July 5, 2022 by symmetal

[04Stefan07]

2 minutes ago, symmetal said:

The push/pull tilt adapters are awkward to adjust and they usually havethe adjustment screws on the wrong side for easy access. If you can mount it further forward in the imaging train with spacers between it and the filter wheel to allow allen keys to access the screws without having to disassemble everything it certainly helps.

The ZWO MkII adjuster looks like it's reversible which is a better design. The easiest to use tilt adjusters are the Gerd Neumann ones though they are espensive and M48 is the smallest, and with adapters to M42 may be too long to fit.

Can't see any links to your subs though. 🤔

Alan

Hey mate.

I saw that fancy tilter but yeah a little expensive.

Ah, just uploaded them now. They took a while.

I did one sub as normal and the second one with the camera rotated 180 degrees. Tilt remained the same so that's telling me there's tilt in the camera itself.

I came across these shims which are a great idea.

https://astrodymium.com/products/rotating-tilt-adjusters

I got an awesome image of the Running Chicken Nebula that you can see on my AstroBin with my setup but I removed the stars so it wasn't a big deal (and it didn't seem as bad the tilt). 
Currently imaging NGC 6188 and want to keep the stars in so hoping to get this issue sorted.

ha_55mm.fit ha_55mm_180degrees.fit

[symmetal]

I ran your images through CCDI and it looks to be purely tilt left to right in the camera itself. The 55mm spacing looks good as the 3D view shows just a tilted flat plane. If the spacing is wrong the corners of the 3D image tilt upwards. As the 1600 is not a large sensor the spacing isn't that critical anyway.

The tilt rings you linked to look like a lo-cost solution, and are quicker to adjust than the push/pull screw tilt adapters. A possible issue is that as they force the threads to be tilted when done up the threads are only tight on two opposite spots with the areas inbetween 'floating' between threads. As long as they are done up tight enough and you don't feel any rocking they should be OK. Worth a try I would think. 🙂

As the tilt is in the camera, this means you can't rotate the camera 90 degrees for framing without altering the tilt adjustment as well, which would be a pain.

Good luck. 😊

0 deg

180 deg 

3D view 0 deg

Alan

[04Stefan07]

Thanks Alan.

Had the same results in ASTAP. What I can't work out is when I captured the Running Chicken my subs didn't seem as bad.

Have attached a few subs over a few nights when I captured that target. Same setup and same 120 secs (I think from memory my spacing was 56mm but my flat may have been off by a bit). The tilt appears to vary a bit across different filters. but not as bad as now.

I also checked the placement of the filters in the wheel and they looked good.

So if it is camera tilt which is what the tests confirm, wouldn't I be getting the same tilt across all filters regardless of back focus spacing? Stupid question, could seeling conditions contribute to the variations between filters?

ic 2944_Ha_120sec_1x1_Ha_0004.fit ic 2944_OIII_120sec_1x1_OIII_0008.fit ic 2944_OIII_120sec_1x1_OIII_0012.fit ic 2944_Ha_120sec_1x1_Ha_0006.fit

[symmetal]

Hi Stefan,

Running your latest images posted through CCDI they are much better. The min and max FWHM figures are much better than the previous two images implying the seeing was poor then. The top right of the 4 images below has the best FWHM, evenly distributed and has no tilt indicated.

Note that the comparative difference between the best and worst FWHM is fairly consistant on all six images posted.

2.17 / 1.79 = 1.21 Best image below

6.73 / 5.18 = 1.3 Worst image from previous post

The two previous 'tilted images' have been focused on the left hand side while the best images below are focused nearer the centre. The bottom two below are focused top right. If the two previous 'tilted' images were focused near the centre it's likely they wouldn't display any variations as tilt in ASTAP or CCDI.

Overall, I now don't think you have a tilt issue, and the 'tilted' images are the result of rather poor seeing and the best focus being off to one side. I assume you're using autofocus. I've found that autofocus will often choose to focus nearer one edge rather than nearer the centre, possibly due to the distribution of stars picked for the autofocus routine. It would be nice if there was an option to have the autofocus 'weighted' towards the image centre.

Also, your 56mm spacing is probably better than the default 55mm, as the filters, and the camera protect glass, will increase the required FF spacing distance by 1/3 their total thickness, so I should put it back to 56mm and see how things turn out when the seeing is better. 🙂

Alan

[04Stefan07]

4 hours ago, symmetal said:

Hi Stefan,

Running your latest images posted through CCDI they are much better. The min and max FWHM figures are much better than the previous two images implying the seeing was poor then. The top right of the 4 images below has the best FWHM, evenly distributed and has no tilt indicated.

Note that the comparative difference between the best and worst FWHM is fairly consistant on all six images posted.

2.17 / 1.79 = 1.21 Best image below

6.73 / 5.18 = 1.3 Worst image from previous post

The two previous 'tilted images' have been focused on the left hand side while the best images below are focused nearer the centre. The bottom two below are focused top right. If the two previous 'tilted' images were focused near the centre it's likely they wouldn't display any variations as tilt in ASTAP or CCDI.

Overall, I now don't think you have a tilt issue, and the 'tilted' images are the result of rather poor seeing and the best focus being off to one side. I assume you're using autofocus. I've found that autofocus will often choose to focus nearer one edge rather than nearer the centre, possibly due to the distribution of stars picked for the autofocus routine. It would be nice if there was an option to have the autofocus 'weighted' towards the image centre.

Also, your 56mm spacing is probably better than the default 55mm, as the filters, and the camera protect glass, will increase the required FF spacing distance by 1/3 their total thickness, so I should put it back to 56mm and see how things turn out when the seeing is better. 🙂

Alan

That is a fantastic analysis, Alan.

Part of me was thinking how could it be tilt if my images ic2944 images were a lot better.

Yep using auto focus with the ZWO EAF using SGP. I didn't even think that the way the software does the auto-focus as well. I will have to look into that a bit more to see what settings there are.

Tonight is looking good so I will change over the adapters back to 56mm now and get ready and report back with my findings in a few hours. Hopefully the seeing is decent.

Just to keep track, current spacing:

- Camera: 6.5
- EFW: 20
- Spacer: 5
- Spacer: 7
- Spacer: 1
- M42-M48: 16.5

Total 56mm.

Field Flat set to 7.1mm as per spec (or very close to it as it's difficult to get exact to 0.1 as the measurement increments is per 1mm).

Total back focus is 63.1mm
 

Edited July 6, 2022 by 04Stefan07

[04Stefan07]

Obtaining results now, will upload within half hour.

[04Stefan07]

Ha_120s.fitLooks to be the same results as last night. Tilt but from top to bottom (as the camera is in a different position).

[04Stefan07]

O3 

O3_120s.fit

[04Stefan07]

S2

S2_120s.fit

[04Stefan07]

So after a few hours I did a few configurations. 56mm, 55mm and 55mm all of those with the tilt. I tried SO many settings and could not get it.

Decided to do one more, back to the original 56mm. Still no good so for the sake of it I went 57mm and got this.

57mm_2.fitOfficially have given up on this. The ZWO tilt is such a crap design as it doesn't allow you to change it on the fly.

Is it still tilt? Is there sag? Are my narrowband filters just crap? Is the camera sensor tilted? Is it seeing conditions and weather? This is what turned me off the hobby for a few months last year when I spent a month trying to get this right!

Just attached a third 57mm I did. ASTAP shows less tilt than before. I did a re-focus.

57mm3.fit

Edited July 6, 2022 by 04Stefan07

[symmetal]

Hi Stefan,

Your latest images are varied and all show rather large FWHM figures implying seeing isn't great. They have all got best focus in the top left corner.

The filters themselves won't cause issues like this. For checking tilt issues it's best to use short L exposures of like 15 secs to minimize guiding issues affecting the results. Also use an average of like 8 or so images as consecutive images are unlikely to be identical.

As far as focus is concerned what autofocus exposures are you using and are you binned 2x2 for autofocus to reduce exposure time. I use SGP like you and use 3s for L, 5s for RGB and 30s for NB.

Is your autofocus graph a well defined U or V curve and the first segment, top right part of the curve and not more horizontal. When it does its validation exposure is the HFR reported at 100%. If you have insufficient focuser backlash compensation then the final focus position will not be optimum and tends to favour one of the edges or corners like to seem to have. A screen shot of your autofocus final curve plot would help here. Check the autofocus backlash compensation is set to IN direction so it always finishes focusing pushing the focuser up against gravity.

Is the tilt adjuster set to no tilt with the small allen screws loose when the large screws are tightened up. It may be worth taking the tilt adjuster out and do some tests as I found with my initial cheap adjuster that it introduced tilt even when it was set for no tilt.

Under SGP Auto focus options there is the option to use HFD, with ASTAP installed, rather than HFR which may be worth a try.

As mentioned a screenshot of the final autofocus screen for each filter would be useful.

Alan

[04Stefan07]

Hey mate.

I just got the last batch of the auto-focus routine (doesn't say which filter they are for but this was the focus I was getting all night). The quality I noticed will always come back between 93% - 100%.

I ended up removing the tilt adapter as it was near impossile to adjust as you need to remove everything to make an adjustment and re-test. What I tried doing was removing the filter wheel, adding a 20mm spacer (acting as a dummy filter wheel) and the tilt adapter so I had access to the screws. After a few tweaks the tilt wasn't that bad so I put the filter wheel back on as it was, did another sub and there was massive tilt in one corner so that didn't work.

HFD with ASTAP, I will defs look into that.

If it's another clear night tonight I wll do another test but I am at breaking point now lol.

[symmetal]

I was wondering why your final focuser positions were changing so much but I assume these were taken with and without the filter wheel etc. for test purposes. Your focuser backlash setting and step size look to be OK, but with your quality not always being 100% you could try another 10 units or so of backlash compensation and see if you get more 100%s. Having too high a backlash compensation setting is not a problem.

Another thought is that your auto focus exposure settings are not optimum and they are being under or over exposed during autofocus. If you have noticeable vignetting then corner stars will be dimmer so possibly appear sharper than central stars which will be more exposed and throw off the autofocus results. Your O3 images tend to have the better results with the optimum focus more towards the centre compared to Ha. If you use the same exposure times for all your NB filters autofocus then the Ha autofocus images will be brighter than your O3 images with more central overexposed stars which are ignored, making the stars used for autofocus being more towards the edges. It may be worth trying lower autofocus exposure times for Ha and see if the results improve or are more consistant, hopefully focusing more towards the centre.

I think it's worth leaving the tilt adjuster off and making up the FF spacing required. Any tilt indicated will be the same for all filters so it's best to just use L for tilt assessments as it's quicker than waiting for NB exposures. Just try and get the NB best focus to be more towards the centre which I think will remove the indicated tilt on the NB images.

Alan

 

[symmetal]

Stefan,

I just noticed that SGP lets you centre crop the auto focus frames to stop edge stars affecting focus.

 

Here's the help file description

Quote

• • Auto Focus Crop:  This is designed to help those who image with severe field curvature at the edges of the image and will prevent the auto focus algorithm from using "bad" edge stars to determine focus.  This feature is specified as a percentage of the CCD's dimensions.  For instance, if you specify a crop value of 5% and your CCD dimensions are 800x600, auto focus frames will not consider any stars within 20 pixels of the left and right edges (40 pixels total) and within 15 pixels of the top and bottom edges (30 total pixels).

This should help your NB autofocusing to stop focusing near an edge which may help.

If you do find that there is consistant tilt then it's easiest to make a test jig as shown here out of some odd pieces of wood and a cheap laser pen. You can put the camera, filter wheel and Zwo tilt adjuster on the jig and can remove the tilt adjuster to make any adjustments before refitting it and retesting. Much quicker and easier than having to do it outside with test images. Using the tilt rings you mentioned would likely be even quicker. Can be done indoors during the day too. 🙂

Alan

[ollypenrice]

This thread tells you how to make a camera tilt adjusting jig so that testing on the stars isn't required.

Olly

[04Stefan07]

Alan,

Thanks for the feedback. I will give that setting a crack. Just managed to get out tonight to do a few more tests. I actually setup NINA as I hear it has a much better auto-focus routine than SGP just to try out. 

Alan/Olly, thanks for the link to the tilt jig. THat looks like it will come in handy esepcially for testing indoors during the day! Luckily im a decent handy man haha.

 

[04Stefan07]

Here are a few subs I took using NINA. There isn't as much tilt like there was before so it must have been how SGP was doing the auto-focus (that setting you mentioned, Alan).

The stars arn't perfect but it's an improvement. I think there is very minor tilt.

I don't think seeing conditions were the best as well.

I am going to take some subs with the changes in SGP. Got some light cloud cover which just rolled in.

_Ha_8.50_60.00_1x1_0000.fits _OIII_8.50_60.00_1x1_0000.fits

Edited July 10, 2022 by 04Stefan07

[symmetal]

Stefan, the latest images are better and more consistant. Note the min and max FWHM are much closer. Yes, there does does look like a little top to bottom tilt so I should make a tilt test jig and get it fixed. I've found on all my setups that tilt occurs mainly in the camera and getting it right on the jig improved the results greatly and it puts your mind at rest too. 😊

Alan

[04Stefan07]

6 hours ago, symmetal said:

Stefan, the latest images are better and more consistant. Note the min and max FWHM are much closer. Yes, there does does look like a little top to bottom tilt so I should make a tilt test jig and get it fixed. I've found on all my setups that tilt occurs mainly in the camera and getting it right on the jig improved the results greatly and it puts your mind at rest too. 😊

Alan

Yep it is looking much better! So it looks to have been crappy focus this entire time mixed with some minor seeing conditions.

I am going to have a go and building up that rig.

Is there a recommended laser to use in terms of power? I hear sometimes it can damage the camera sensor. I found a 650nm 1nW Red on ebay for cheap.

[symmetal]

1 hour ago, 04Stefan07 said:

Is there a recommended laser to use in terms of power? I hear sometimes it can damage the camera sensor. I found a 650nm 1nW Red on ebay for cheap

I think you meant 1mW. 😉 A 1mW laser is fine, 5mW should be OK and 10mW is risky. When you get it check that it is 1mW. It will likely say 'Class II less than 5mW'. Class II is actually up to 1mW. I ordered a 1mW laser pen and it had Class III on it. Class III is up to 500mW though the pens were  likely 10mW by its brightness compared to another actual 1mW I also bought. ebay blocked some sellers who were selling Class III lasers as 1mW.

Alan

Edited July 11, 2022 by symmetal

[04Stefan07]

Ah yes! Luckily I didnt say 1.21 gigawatts!

[04Stefan07]

Alright, so I returned to this issue on the clear nights I have been getting (only has been a couple in the last month!).

I purchased a few tilt shims which took 5 weeks to arrive but are meant to be easier to use than a tilt adapter. So far, they are so much easier to use.

Every shim adds around 0.4mm back focus so I have compensated for that. Ran a few subs and I am still getting tilt but its all on the left hand side. Since the corners are reversed, I put the shim on the right hand side (looking from the back) of the camera and the tilt does not change.

Is it possible there is tilt elsewhere in the imaging train?

At this rate I think I will need to leave it and start imaging.

[tomato]

This might be heresy to some imagers, but if I had that sensor analysis result, I would start imaging, especially if you are not blessed with a lot of clear nights.

Modern post processing software will take care of those aberrations. 
 

Now for the backlash from the pixel peepers…😉