Star drift in RA

[almcl]

Something came up in a thread last week (thread) that’s been puzzling me and as I experience the same problem with much the same equipment as Lukebl (RA drift across an imaging session), I wonder if someone could explain?

More than one comment was made that the drift (which was exclusively in RA) was down to poor PA.  But I can’t fathom why this happens in a guided session?  Surely the guiding software keeps moving the star back to where it should be (slowing down or speeding up the drive as required?  So how do all the stars end up 30 or 40 pixels further up the frame after an hour or so of guided imaging?

I should add I'm looking for enlightenment rather than an argument?

[Davey-T]

The stars can drift and the guide scope will follow them.

You can use this as sort of drift aligning aid.

Dave

[almcl]

Sorry Dave, but I think that's the bit I can't get my head round?  

Why doesn't PHD2 move the guide star back to the start position?

[Davey-T]

PHD moves the scope to follow the star, it can't move the star 

If you were imaging M42 at this moment and each downloaded sub was moving north you could adjust the dec bolt to try to reduce it.

Dave

[Ibbo!]

Is this at the calibration stage ?

 

[almcl]

No, this is during the imaging run and the star is moving west (or east - I'll have to dig out the subs and look).  The guide graph is its usual bumpy (+/- 1' self)

[Ibbo!]

if the star is moving that much i would look at polar alignment, any flexing and balance

[Davey-T]

1 hour ago, almcl said:

No, this is during the imaging run and the star is moving west (or east - I'll have to dig out the subs and look).  The guide graph is its usual bumpy (+/- 1' self)

Where abouts were you imaging ?

Generalising a bit but if the stars move north / south it's polar alignment, if they move east west it's a mount problem not moving fast enough or too slow.

Dave

[almcl]

3 hours ago, Davey-T said:

Where abouts were you imaging ?

Generalising a bit but if the stars move north / south it's polar alignment, if they move east west it's a mount problem not moving fast enough or too slow.

I've just noticed the recent images I originally looked at had been dithered, so scratch the 'I've got the same problem' comment above, at least until I can find an un-dithered example (sure I've got some, somewhere...), but the other poster's image includes an asteroid, so just on the border of Gemini and Orion, I would guess? The bright star is HD 42088.

Your second para chimes exactly with what I thought understood, hence my surprise at the suggestion that it was a PA error that was causing the star to move west across the frame.

However I still don't understand why PHD2 doesn't correct this?  My experience of it is that does try 'to move the star' back on to the cross hairs.  If it was flexure would the movement be steady and continuous?  

[Davey-T]

I think PHD moves the mount to follow the guide star so if the mount isn't tracking properly for whatever reason PHD will try to correct this, obviously it can't move the star in the sky.

You can switch off guiding to see what it's trying to correct.

Dave

[almcl]

I think we may be using different words to describe the same thing; PHD adjusts the tacking to keep a star on the cross hairs?

Here's an image at the start of a 40 minute run overlaid with one from the end of the run, both showing exactly the same bit of frame.  In case it isn't clear in the first, the second image shows a much magnified view where the bright star appears to have moved 10 pixels or so along what I *think* is the RA axis (open to correction here) over this guided time. (The bright star is HD 201731.)

 

Shouldn't this and greater errors be corrected by the guiding?

 

[Davey-T]

The error is being corrected by guiding if the stars are round, if the stars / mount is drifting guiding can't stop it drifting it just follows the star.

If you switch off guiding and get egg shaped stars in the same direction this is what guiding is correcting.

Dave

[Oddsocks]

There are a few possibilities I can think of.

A:

There is differential flex between the guide scope and imaging scope, I can imagine that with large reflectors there could be quite a bit of movement during a long sequence of exposures, the guide software is keeping the guide star located precisely at the same coordinates on the guide camera sensor but the main scope and camera is no longer pointing to the same place it did at the start of the sequence (mirror flop, tube flexure, temperature changes etc), this should not be the case if using an OAG without a separate guide scope of course.

 

B:

It might be to do with how PHD manages guiding and also whether you use dithering between exposures, I don't use PHD so can't say how it's internal programming works but I use other guiding/sequencing programs that use different methodologies 

In one of my guiding and acquisition programs that manages both guide and main imaging cameras, at the end of each main camera exposure, while the main camera is downloading then guiding is suspended, at the end of downloading then the guide camera re-aquires the guide star and begins to guide again but on it's current location, if there was some drift during the 30 seconds while the main camera downloaded then the guide system does not return the guide star back to it's previous position but continues from where it is now, so if PA was poor and drift in RA or DEC was high then the result is a gradual walking across the sensor over several hours of imaging the same object. With this program I have set it to plate solve and recenter between exposures if I want it to stay perfectly on target but since I use dithering I only plate solve with filter changes.

The other guiding and acquisition program I use does not do this, guiding is still suspended during main camera download but when guiding recommences then guide commands re-centre the guide star on its previous location, waits for the mount to stabilise and then starts a new main camera exposure, so no drift occurs.

In both software packages you can set dithering between exposures but in the first program upon recommencing guiding it takes a single guide exposure of it's current position, then moves the mount in a random direction for a set period of time that it considers will move the guide star x pixels, then it reacquires the guide star and begins to guide on it's new position, images show a random"dance" across the frame but superimposed upon a constant drift in one direction due to poor PA.

In the second package, the guide software first returns the guide star to it's original location, takes a test frame to confirm, then dithers the mount, then takes a new test frame, then begins guiding on it's new location. Images from the main camera show the same random "dance" but with no continual drift in one direction.

C:

In either of the above cases, if PA is poor and even if the guide program keeps the star on the exact same coordinates throughout the whole image sequence without interruption you would still have rotation in the main camera image, the guide star would remain fixed but everything else in the image would rotate about it. This would be particularly noticeable when using an OAG as the guide star would be outside the main image field of view.

Imagine also if you use a separate guide scope that is not perfectly axially alligned to the main telescope so that you choose a guide star far from the centre of the main camera image, if PA is not perfect there would be considerable rotational drift in the main camera image, for example say 30 pixels in one hour, for individual subs of two minutes each that might cause the FWHM to take on a slight oblate form, maybe for a bright bloated star you would see 8 pixels by 9 pixels, stretched in the direction of drift. You might find that shape of star acceptable given the other variabilities due to seeing conditions etc but look at a sub from the beginning of a run and compare it to a sub from the end of a run two hours later and the stars in the main camera image will have drifted 60 pixels during that time but the guide camera is still perfectly locked on to the same coordinates that it started with.

 

[almcl]

Now temperature is a distinct possibility.  (Case A)

Although my scope lives in an unheated part of the house and is seldom more than a degree or so above ambient at set-up, it got covered in frost on that occasion. Which was probably why I wasn't dithering (I have to manual dither) and it was just too cold to sit out with the laptop (solely guiding - no imaging functions) and the tablet (solely image capture).

[ollypenrice]

I do find this pretty odd and haven't seen it myself at first hand.

I have one idea for a test. Before guiding, put a brightish star in the middle of the guide cam window. Aim to guide on that, but before you do so take and save a sub from the guide cam in that position, long enough to have a decent set of field stars. Do a guided run for a while, the longer the better, and then at the end take another sub from the guide cam to compare with the first.

If you then compare the first and last guide cam subs you will be able to see if you still have the mysterious drift. This eliminates diff flexure. It also eliminates the effects of guiding on an off axis star. If PA is the problem you will have field rotation. If you still have drift, well, I'm stuck! However, I can think of no way you could have drift during this test if the guide star is never lost. That does not mean you won't get it, by the way, it just means that I don't know how!

I have had ST 80 flexure in three places all of which can be fixed. 1) The whole back end where it is held into the tube by three radial screws. 2) the focuser draw tube. Set the lockscrew nice and tight. 3) The bodged barlow bodies I used as extenders. I Araldited them together and into the drawtube!  I wonder of the lens cell might also be worth checking?

Olly

[almcl]

Thanks, Olly!

Quite a bit of home work there :-)  - I really must get to grips with the QHY 5L's software.  It worked so well 'out-of-the-box' that, to my shame, I don't yet know how to record images from it...

Flexure of the ST80 is definitely a possibility.  The draw tube, focusser and extender tube could hardly be described as rigid and while the aluminium bar it's usually mounted on (via scope rings) is pretty solid, I don't always remember to check its mounting bolts with a spanner before the excitement of another imaging sessions kicks off.

All things to have a look at.

[iapa]

5 hours ago, ollypenrice said:

I have had ST 80 flexure in three places all of which can be fixed. 1) The whole back end where it is held into the tube by three radial screws. 2) the focuser draw tube. Set the lockscrew nice and tight. 3) The bodged barlow bodies I used as extenders. I Araldited them together and into the drawtube!  I wonder of the lens cell might also be worth checking?

Olly

Re the ST80 - I'm assuming the device in question is the Skywatcher Star Travel 80 ( or the Orion variant).

With mine, there were 'wobbly' bits that left me with little confidence in stability (and as a noob, my standards are much lower that most  ), so I followed the steps in the Astronomyshed youtube tutorial - with the exception of black marker round the glass as my hands are not steady enough to risk marking the important bits.

Also replaced all the thumbscrews with some with a slightly larger diameter.

Result of this was improved confidence in guiding - and improved guiding, but that may be because I was getting more familiar with the process itself.

Personally, I would say these were all worth the effort - and minimal costs.

 

[ollypenrice]

In my book it makes a perfect guidescope but, like anything else, it needs to be put together properly.

Olly

[wimvb]

Maybe this thread can shed some more light (or confusion) on the matter.

I used the latest gap in the clouds to learn more about guiding. (The full moon stands in the way of imaging anyway.)

BTW, with my setup (ST80 with ASI 120MM piggy back), I can't rule out flexure. Will investigate this during the next gap.

[almcl]

Yes, I noticed your comments there and I have similar misgivings  about the accuracy of the PA using the Synscan handset.  It's reported zero Mel and Maz more than once but the tracking was still nothing special and the gotos weren't great, either.

However, while aligning the new style polar scope ring on Polaris has never been a problem, I've never paid much attention to the mount being level and I wonder how essential this is?

Also, have taken to looking through the finder scope while in the park position to see how close to the NCP the rig is, because I suspect eyeballing the park position may be the weakest part of my set-up procedure.

Have torn apart and reassembled the ST80/QHY5L setup and added shims to reduce the chance of flexing - now just need a clear night to refocus and calibrate...

[vlaiv]

There is another option for RA drift, apart from differential flex - everything can be stiff and keep relative position, but if guide scope is not pointing at the same place as main scope - there will be RA drift.

Just think of extreme case - guide scope pointing at the Polaris - it moves so slowly so corrections would not be even noticeable - it would be like there is no guiding at all.

So if guide scope is not aligned with main OTA - RA drift can happen for duration of the imaging session (I have about 10 - 20 px - 10" - 20" since my resolution is 1"/pixel or there about) for 4h sessions and I know for sure that there is a bit of misalignment between guide scope and imaging scope (small one).

[wimvb]

PA accuracy with the synscan can never be more accurate than the pointing accuracy, which according to the (V4) manual is 5 arcminutes RMS. As for accuracy in the home position, I set the RA scale to 0, and the DEC scale to 90. But I will always do a 2 star alignment, so the accuracy in home position isn't relevant. Even if I park my scope after alignment, for a break in setup/imaging, I will always do a realignment when taking it out of park. It's better to be safe than sorry, and star alignment doesn't take that long.

Goto accuracy has been quite good with my AZ EQ6 mount since I started using a barlowed short fl eyepiece for star alignment. I use 2 star alignment on the same side of the meridian as I will do my imaging.

[ollypenrice]

Dinosaur that I am, I always recommend (and use) drift alignment. It cannot lie...

Olly

[wimvb]

2 minutes ago, ollypenrice said:

Dinosaur that I am, I always recommend (and use) drift alignment. It cannot lie...

Olly

That is certainly the best and safest method. The problem is that THIS dinosaur isn't the fastest in the herd, and usually loses too much precious imaging time. (I have to set up my rig from scratch every time.) Maybe a polemaster is an alternative. But a new camera is higher on my wishlist.

[almcl]

+1 for the loss of imaging time. Most targets are only visible for a couple of hours in my location (trees, houses, street lamps &c), even when the weather is favourable, which isn't often.   

Not having an east or west horizon and only a limited view to the south are also constraints.  I've tried Alignmaster - trying to find a star thst matches its requirements was a nightmare - the digital image drift align method produced pretty flat lines but still didn't solve the problem...