PHD2 Drift Align - trend line doesn't change

[kens]

This certainly an interesting conundrum. Everything looks fine. Calibration looks good even with the smaller steps. It is slightly odd that the brain calculator does not give the right step size so double check the parameters. A recent guide log may be informative - especially if it contains the calibration and drift alignment sequences.

One possibility, although it seems unlikely, is that you are not aligning on Polaris. Or that the inverted image is not being taken into account. There is a tutorial on the EQMOD Polar Alignment at http://www.myastroscience.com/eqmodpa

You could also try other methods like SharpCap (which uses plate solving) or QPPA http://www.astronomyforum.net/telescope-mounts-forum/228041-free-polar-alignment-program-qppa2-new-easier-use-version.html 

 

[frugal]

Just now, kens said:

You could also try other methods like SharpCap (which uses plate solving) or QPPA http://www.astronomyforum.net/telescope-mounts-forum/228041-free-polar-alignment-program-qppa2-new-easier-use-version.html 

 

I have been trying to use SharpCap's Polar Alignment, but unfortunately with only a 0.9 x 0.6 degree field of view in my guide scope / camera setup it does not get an accurate plate solve. Without moving the scope, the software platesolves successive frames to different locations. I thought about using the finderscope that came with my ED80 and my lodestar just for alignment with Sharpcap to increase the field of view, and then move the lodestar to the guide scope for actual guiding, but I can not figure out what connector I need to connect the two together (certainly nothing I have in my box of spare bits fits the finderscope  )

[Thalestris24]

Hi frugal

Um, thinking about it, your symptoms have all the signs of being locked onto a hot pixel rather than a star... Do make sure you've created a bad pixel map in phd2 - it doesn't take long. Guide cams can be quite tricky to focus so make sure you're close to good focus and you can see a number of different stars. Adjust exposure time accordingly. If you can get a screen grab of your guide cam view in phd2 then post it here if you're still having trouble. 

Louise

[frugal]

2 minutes ago, Thalestris24 said:

Hi frugal

Um, thinking about it, your symptoms have all the signs of being locked onto a hot pixel rather than a star... Do make sure you've created a bad pixel map in phd2 - it doesn't take long. Guide cams can be quite tricky to focus so make sure you're close to good focus and you can see a number of different stars. Adjust exposure time accordingly. If you can get a screen grab of your guide cam view in phd2 then post it here if you're still having trouble. 

Louise

I am fairly certain it is not a hot pixel as I can see the FWHM chart of the star fluctuate due to the seeing. Plus as I am moving the Azimuth bolts so much, quite often the guide star is going off of the screen and I am having to pick a new one. I know I have a set of recent darks, but I can not remember if I made a bad pixel map, I shall check.

Next time I get some imaging done I will make up a set of images of the various parts of the operation and add them all here. Unfortunately the weather forecast does not have a clear night until after Easter...

[Thalestris24]

12 minutes ago, frugal said:

I am fairly certain it is not a hot pixel as I can see the FWHM chart of the star fluctuate due to the seeing. Plus as I am moving the Azimuth bolts so much, quite often the guide star is going off of the screen and I am having to pick a new one. I know I have a set of recent darks, but I can not remember if I made a bad pixel map, I shall check.

Next time I get some imaging done I will make up a set of images of the various parts of the operation and add them all here. Unfortunately the weather forecast does not have a clear night until after Easter...

Ok - but if you're adjusting the azimuth bolts then I can't see how the phd2 trendline doesn't change... Do you see the purple circle change size as in the tutorial ?

NB you only monitor the DEC dift!

[frugal]

1 minute ago, Thalestris24 said:

Ok - but if you're adjusting the azimuth bolts then I can't see how the phd2 trendline doesn't change... Do you see the purple circle change size as in the tutorial ?

That is the problem. I move the star to the edge of the purple circle by using just the Azimuth bolts, and when I go back to drifting, the circle is the same size after 2 minutes of drifting. I could understand if it doubled in size or shrank to nothing, but not changing is doing my noggin in. It must be something daft I am doing, I just can not figure out what.

The next time I have a clear night I will run through all of the suggestions the people have put forward,  and write down (with pictures) exactly what is going on.

[frugal]

1 hour ago, kens said:

 A recent guide log may be informative - especially if it contains the calibration and drift alignment sequences.

As I mentioned earlier, I had been working on the assumption that you did not need to recalibrate every time, so the other night's guide log does not have a calibration entry in it.

The first 23 sections in the file are all the drift aligning attempts. 24 onwards are actual guiding logs.

PHD2_GuideLog_2017-04-08_210502.txt

[Thalestris24]

Had a quick look at your log:

 

Clearly your DEC is drifting very badly, amongst other issues. It seems likely that either you're very unbalanced or your PA is way out. What happens when you try to adjust ALT/DEC (east/west) in your PA routine?

Getting too late now zzzzz....

Louise

[kens]

Curiouser and curiouser. The drift in each of the unguided graphs tallies with what PHD2 reports at PA error. The drift is consistent at around -0.1" per second.

Even loading the guided graphs under PEMPro log viewer shows a similar level of drift when guiding pulses are removed - of course in opposite directions on each side of the pier.

What happens when you adjust the altitude bolt? I wonder if altitude PA error is swamping everything else?

[kens]

4 hours ago, frugal said:

I have been trying to use SharpCap's Polar Alignment, but unfortunately with only a 0.9 x 0.6 degree field of view in my guide scope / camera setup it does not get an accurate plate solve. Without moving the scope, the software platesolves successive frames to different locations. I thought about using the finderscope that came with my ED80 and my lodestar just for alignment with Sharpcap to increase the field of view, and then move the lodestar to the guide scope for actual guiding, but I can not figure out what connector I need to connect the two together (certainly nothing I have in my box of spare bits fits the finderscope  )

0.9 x 0.6 should be plenty as long as you have the appropriate index files. I plate solve my imaging scope which is only 0.63 x 0.48 and it solves fine. The fact that it resolves each frame to a different location could be a clue. That could be due to the amount of drift. You could try having guiding on whilst running SharpCap. Best to turn it on only in Dec whilst pointing at the pole. That should not affect the SharpCap routine. If that does not work, try the QPPA method I linked to. 

[frugal]

7 hours ago, Thalestris24 said:

Had a quick look at your log:

Clearly your DEC is drifting very badly, amongst other issues. It seems likely that either you're very unbalanced or your PA is way out. What happens when you try to adjust ALT/DEC (east/west) in your PA routine?

Getting too late now zzzzz....

Louise

I am slightly unbalanced to the East and the North as that was recommended for guiding to try and avoid backlash. Not so much that it fall off to one side, but just enough that it is easier to push one way than the other. I will try to get it perfectly balanced next time to see if that helps.

I am fairly certain that he polar alignment was awful at that point due to all of the faffing around earlier.

[frugal]

5 hours ago, kens said:

0.9 x 0.6 should be plenty as long as you have the appropriate index files. I plate solve my imaging scope which is only 0.63 x 0.48 and it solves fine. The fact that it resolves each frame to a different location could be a clue. That could be due to the amount of drift. You could try having guiding on whilst running SharpCap. Best to turn it on only in Dec whilst pointing at the pole. That should not affect the SharpCap routine. If that does not work, try the QPPA method I linked to. 

Unfortunately, SharpCap and PHD2 are using the same camera, so I can not guide whilst trying to polar align with SharpCap. I am not sure how SharpCap is doing the plate solve as it is using an internal process rather than calling out to PlateSolve2, or Astrotortilla. It picks 15 stars out of all of the ones it has found and tries to plate solve using just them (for speed I guess as it solves in under a second). However if it solves frame 1 using 15 stars and states that the North Celestial Pole is at the top left of the frame, it might pick a different set of 15 stars for frame 2 and then say that the NCP is off the screen to the bottom right, even though the scope has not moved.

Unfortunately SharpCap does not support Canon DSLRs as cameras, otherwise I would use my imaging camera  / scope for the SharpCap alignment as that has a field of view of 2.5 x 1.7 degrees.

[Thalestris24]

1 hour ago, frugal said:

I am slightly unbalanced to the East and the North as that was recommended for guiding to try and avoid backlash. Not so much that it fall off to one side, but just enough that it is easier to push one way than the other. I will try to get it perfectly balanced next time to see if that helps.

I am fairly certain that he polar alignment was awful at that point due to all of the faffing around earlier.

Hi 

You didn't say what happens when you do the dec adjustment to the east or west? If your alt/DEC PA is out, so might your az/RA be. You have to adjust both until both are ok.

Louise

[frugal]

I seem to recall that the Altitude alignment was a lot better, about 2-3 arc minutes so I left it. Although at that point I had pretty much reached the "I no longer care" point.

It looks like Thursday morning might be clear, so I can try the alignment even though there will be too much moon to do any imaging.

[michael8554]

It saddens me the amount of time and effort spent trying and failing to polar align using software.

Get a high power reticule eyepiece, or magnified LiveView with grid display,  and Drift Align.

Guiding not necessary, and Periodic Errors don't matter since it's only Dec errors that need to be corrected.

Michael

[Thalestris24]

1 hour ago, michael8554 said:

It saddens me the amount of time and effort spent trying and failing to polar align using software.

Get a high power reticule eyepiece, or magnified LiveView with grid display,  and Drift Align.

Guiding not necessary, and Periodic Errors don't matter since it's only Dec errors that need to be corrected.

Michael

Well, phd2 does just that except it does it via a graphic display. Normally it's just a 10min job!

Louise

[michael8554]

Yes Louise, you've mastered the PHD drift align, but many struggle with it and other software solutions for far too long, instead of trying old-school drift align, which presents clear information that cannot be misinterpreted.

Michael

[Thalestris24]

3 hours ago, michael8554 said:

Yes Louise, you've mastered the PHD drift align, but many struggle with it and other software solutions for far too long, instead of trying old-school drift align, which presents clear information that cannot be misinterpreted.

Michael

Some people seem to struggle with doing PA by whatever means! You may be right about using software though generally PHD2 is very convenient and straightforward... Having said that, I think PHD2 works best when the mount is already reasonably close to alignment. I'm not sure what the op was doing previously. Of course, it's perfectly possible to get reasonably good PA just using the handset routine (I've no personal experience with it) though I've gathered you have to be careful about entering the data. I don't think I'd advocate using an eyepiece - it can be a lot of messing about undoing/replacing a camera.

Louise 

[Physopto]

On 09/04/2017 at 09:17, frugal said:

This is pretty much what I saw through the polar scope. Can the polar scope really be that much out of alignment?

 

Yes easilly. You need to adjust the initial setup of the polarscope to get it properly centred. There are videos on Youtube on how to do it. pretty simple and can best be done during the daytime in the light. Just use a really distant object to aim at. If it is too close it will introduce errors.

Derek

 

[frugal]

2 hours ago, Physopto said:

Yes easilly. You need to adjust the initial setup of the polarscope to get it properly centred. There are videos on Youtube on how to do it. pretty simple and can best be done during the daytime in the light. Just use a really distant object to aim at. If it is too close it will introduce errors.

Derek

 

I have double checked the centering of the polar scope against a distant object during the day. It is nicely centered. I am hoping that this evening might actually be clear enough to get the scope out and re-test everything

[Physopto]

2 hours ago, frugal said:

I have double checked the centering of the polar scope against a distant object during the day. It is nicely centered. I am hoping that this evening might actually be clear enough to get the scope out and re-test everything

I would suggest joining this open group on Google. Andy there has just been helping me. Very quick responses and as he is one of the coders of PHD 2 I would say you can not do better. You will probably need to download PHD2 Log Viewer from the PHD website.  As he will possibly want to view your logs.

https://groups.google.com/forum/#!forum/open-phd-guiding

Once you have joined just post a new topic question and wait for an answer.  It was Andy who answered my query and very quickly, within a few hours.

best of luck.

Derek

[daz]

On ‎11‎/‎04‎/‎2017 at 09:18, frugal said:

Unfortunately SharpCap does not support Canon DSLRs as cameras, otherwise I would use my imaging camera  / scope for the SharpCap alignment as that has a field of view of 2.5 x 1.7 degrees.

Themos on here has developed a utility (PPA I think) that does support the Canon camera...

Thread with more info -

 

[AndyKeogh]

On ‎19‎/‎04‎/‎2017 at 13:46, daz said:

Themos on here has developed a utility (PPA I think) that does support the Canon camera...

Thread with more info -

 

I would second this choice as I have been using it very successfully for over 2 years with my canon camera.  I use both local solving via Astrotortilla and occasionally nova.astrometry.net.

Have a read the whole thread you should find it enlightening. I can now get polar alignment to better than 30" in a couple of iterations.

Andy

[spillage]

On 4/18/2017 at 15:25, frugal said:

I have double checked the centering of the polar scope against a distant object during the day. It is nicely centered. I am hoping that this evening might actually be clear enough to get the scope out and re-test everything

Hi

I would suggest you use eqmod again to get pa before attaching you scope to the mount and once you have this correct then rotate the mount and see if polaris keeps on the line. If it drifts off badly thats one issue to tick off and fix.

I always find that letting phd drift run until the reading gets to the end of the screen and starts to cycle is best as I then notice a change in pa at this point. After a short while it settles and then I make my adjustment. I have seen a few tutorials that tell you to only move halfway to the circle but feel that this is due to adjustments being make to quickly and not allowing enough time for the drift to settle.

This one is not meant to sound patronizing and not sure how the gt mount differs from my neq6 but when making adjustments slacken off the mount adjust and then tighten the mount back up. This will eliminate any movement.

[frugal]

What do you have your horizontal axis set to in PHD when you do this? I generally use 100 readiness and 2 second exposures. I generally let it do at least 50 of them before I adjust (so 100 seconds)