Barbell shaped stars - is this Periodic Error or Backlash ?

[jm_rim]

Going through my images from the last two clear nights, I see that many of the frames had small barbell shaped stars.
The direction of the elongation is following the RA axis. This is not on every frame, in some, the stars are as perfect as I can get them, see the images. 

My setup: SW EQ5 with RA motor, SW 150PDS, and a Nikon D5300.
I don't have auto-guiding (yet) so this is not the problem - only using a single RA motor.
My mount is on my balcony - quite protected against the wind - it was  -12c to -15c and there was absolutely no wind. So I don't think it is due to "external" factors causing vibrations.
I polar aligned using SharpCap and took successful 200sec test shoots, but used 120sec during - just to be safe.

Anyone have an idea of what can be the problem - backlash, Periodic Error in the RA, or something else...? And how to fix/minimize it?

[malc-c]

Could be any or a combination of the above.  Also be a balance or CofG issue, precision of the motors and gearing or a limitation of the mount.  Two minutes unguided, even with a good polar alignment may be on the limit of the mount, especially if the centre of gravity is off, or the scope out of balance...

Only thing to do is re-balance the scope and set the CoG correctly and repeat what you did with the same target.  If you get the same result then adjust the backlash on both axis to remove the slop but not bind the gearing, and try again.  If after testing and adjusting everything you still have issues then it just might be the case that the existing setup isn't ideal for unguided imaging.  It might be a case that you need the higher precision steppers used in the pro goto version of the EQ5 as the current drive motors are more suited for visual tracking rather than imaging?   

[pete_l]

Looking at the photo, a star seems to move (relative to the camera) from one position to another. The movement looks to be quite rapid as there is very little trailing between the two positions. I assume that the direction of movement is the direction the mount is tracking.

The sort of thing that would cause that are a "wobble" in the mount itself or the telescope - something loose. Possibly the gear meshing is not tight enough. Alternatively it could be that a cable is dragging getting caught temporarily or due to stiffness and then "giving" and flipping the position of the telescope to where the motors want it to be.

[vlaiv]

That is periodic error.

To be 100% sure - see if it is aligned with RA axis - if it is, then it is periodic error.

To minimize it - shoot objects closer to NCP/Polaris. PE is the worst around equator (at DEC close to 0).

Since you are using single RA motor - you don't have this option, but there is this thing called periodic error correction. It is designed to lessen effects of periodic error. It is available in some mount models, or via EQMod / computer

Shorten your exposure length. PE is usually the worst in some parts of worm revolution - if you shorten exposure, you'll throw away less subs.

Of course - discard the worst subs.

In the end - it might help to "tune" your mount - which is basically term for stripping it down, cleaning, applying suitable grease, possibly changing bearings for quality ones and then reassembling everything carefully so that everything fits tight but does not bind.

[jm_rim]

2 hours ago, malc-c said:

Could be any or a combination of the above.  Also be a balance or CofG issue, precision of the motors and gearing or a limitation of the mount.  Two minutes unguided, even with a good polar alignment may be on the limit of the mount, especially if the centre of gravity is off, or the scope out of balance...

Only thing to do is re-balance the scope and set the CoG correctly and repeat what you did with the same target.  If you get the same result then adjust the backlash on both axis to remove the slop but not bind the gearing, and try again.  If after testing and adjusting everything you still have issues then it just might be the case that the existing setup isn't ideal for unguided imaging.  It might be a case that you need the higher precision steppers used in the pro goto version of the EQ5 as the current drive motors are more suited for visual tracking rather than imaging?   

 

I will try to re-balance the scope again tonight, hopefully it will have an effect

 

2 hours ago, pete_l said:

Looking at the photo, a star seems to move (relative to the camera) from one position to another. The movement looks to be quite rapid as there is very little trailing between the two positions. I assume that the direction of movement is the direction the mount is tracking.

The sort of thing that would cause that are a "wobble" in the mount itself or the telescope - something loose. Possibly the gear meshing is not tight enough. Alternatively it could be that a cable is dragging getting caught temporarily or due to stiffness and then "giving" and flipping the position of the telescope to where the motors want it to be.

yes the direction of movement is the direction the mount is tracking

with something loose and knowing the direction of the movement - does that mean i can narrow it down to the RA axis?

[jm_rim]

2 hours ago, vlaiv said:

That is periodic error.

To be 100% sure - see if it is aligned with RA axis - if it is, then it is periodic error.

To minimize it - shoot objects closer to NCP/Polaris. PE is the worst around equator (at DEC close to 0).

Since you are using single RA motor - you don't have this option, but there is this thing called periodic error correction. It is designed to lessen effects of periodic error. It is available in some mount models, or via EQMod / computer

Shorten your exposure length. PE is usually the worst in some parts of worm revolution - if you shorten exposure, you'll throw away less subs.

Of course - discard the worst subs.

In the end - it might help to "tune" your mount - which is basically term for stripping it down, cleaning, applying suitable grease, possibly changing bearings for quality ones and then reassembling everything carefully so that everything fits tight but does not bind.

It is aligned with RA axis 

Shoot objects closer to NCP/Polaris - how close? I mostly shoot above 30 DEC (currently between 45-60)

I have thought about "tuning" my mount as you describe - just hoped that i could avoid it 😀 
I have also thought of upgrading the mount with GOTO, possibly with a belt drive kit, as I can't afford a new mount yet. Would auto-guiding help with minimizing these kinds of errors? 

[vlaiv]

31 minutes ago, jm_rim said:

Shoot objects closer to NCP/Polaris - how close? I mostly shoot above 30 DEC (currently between 45-60)

I think that it depends on cosine of dec. If you think about it - at NCP you are not actually tracking the star you are just rotating the field of view and it stays pretty much in place.

At Dec 30°, rate is about x0.86 that at equator, while at Dec 60° it is x0.5 and so on (cosine function - at 90° it is equal to 0). This is "star motion" on sensor - not RA axis - it always rotates at the same rate - sidereal.

35 minutes ago, jm_rim said:

I have thought about "tuning" my mount as you describe - just hoped that i could avoid it 😀 
I have also thought of upgrading the mount with GOTO, possibly with a belt drive kit, as I can't afford a new mount yet. Would auto-guiding help with minimizing these kinds of errors? 

Of course - guiding is the best way to deal with this - it is exactly why guiding was designed for - to compensate for any tracking errors (both PA errors and periodic errors).

If you can - you should start to guide - that will make your stars round in all subs.

[pete_l]

2 hours ago, jm_rim said:

with something loose and knowing the direction of the movement - does that mean i can narrow it down to the RA axis?

I'd suggest giving the RA a bit of a wiggle while the mount is not in use, see if anything moves. It is not unusual for the gear meshing to change with night temperatures and that can be adjusted without dismantling the mount. However, that might not be the problem and gears that are adjusted to optimal in winter can easily bind up when the ambient temperature rises in the summer and the metal expands.
I'd also check if any cables are flapping loose. If they move while the mount is tracking, that can present an unexpected load to the motors which can take a little time to correct and then suddenly release. Plastic can stiffen considerably at low temperatures so this may not be obvious if checked at room temperature.

Also the same applies to the telescope. check if anything in the optical train is loose or prone to sudden movement.

[malc-c]

You are also experiencing the issues of imaging on a budget.  Don't get me wrong, the 150pds and EQ5 is a capable system, but  with just the basic motor drive you are limited to getting that full potential.  It's also why on any given time when people ask about getting into imaging the HEQ5 is the most recommended and proven mount as an entry point.  However lets see if there is anything we can suggest to get you up and running  with the equipment you have.  But you will be needing to spend some cash - Imaging isn't cheap

I was going to suggest you convert the finder into a guide scope.  However, unlike the 200P, the 150PDS comes with a 6 x 30 rather than a 9 x 50.  This makes attaching a camera to convert the scope a tad difficult.  It is possible to purchase a SW 9 x 50 finder, complete with convector and an ZWO 120MM guide camera from Rother Valley Optics for £185  RVO Finder Guider Package

The camera has an ST4 port and a USB port, so there are two options for guiding, but if your single RA motor drive is this one

Then you won't be able to use the ST4 option as that hand controller has no ST4 port or guiding options (at least not without a lot of DIY modifying).  So you will need to upgrade the handset.  Personally I would suggest going the whole hog and getting the goto option as it has better motors, and provides you with the option of using a PC/Laptop and control the guiding using EQMOD and software.  But that is an additional  £299.  So the other option is to get the enhanced dual axis motor kit for £120.  This has an additional motor for the DEC (needed for auto guiding) and a new handset that has an ST4 port, which allows you to connect the handset to your computer via a suitable adaptor ( £54 from FLO ST4 - USB ).  

So to use off the shelf commercial means you're looking at around £360 as a minimum.  You may be able to save a little by purchasing second hand, but to be honest I doubt that these items come up for sale that often. 

What will this give you.  A dual axis PC driven auto guided mount...

PHD2 guiding software, ASCOM platform, planetarium software such as Stellarium and CdC, and APT ( https://www.astrophotography.app/ ) is all free, although it's worth paying the 18 euro for the full version of APT.

Are you still going to have issues... yeah, even those with more capable equipment have off days, but what you will have will be a foundation that should give you acceptable results 

 

 

[jm_rim]

1 hour ago, pete_l said:

I'd suggest giving the RA a bit of a wiggle while the mount is not in use, see if anything moves. It is not unusual for the gear meshing to change with night temperatures and that can be adjusted without dismantling the mount. However, that might not be the problem and gears that are adjusted to optimal in winter can easily bind up when the ambient temperature rises in the summer and the metal expands.
I'd also check if any cables are flapping loose. If they move while the mount is tracking, that can present an unexpected load to the motors which can take a little time to correct and then suddenly release. Plastic can stiffen considerably at low temperatures so this may not be obvious if checked at room temperature.

Also the same applies to the telescope. check if anything in the optical train is loose or prone to sudden movement.

Is there still are some low clouds - despite all my weather app was promising a completely cloudless night - I will try to go thoroughly through my set-up tonight. To at least rule out the more obvious or visible tings that could be the cause.

[jm_rim]

3 hours ago, vlaiv said:

Of course - guiding is the best way to deal with this - it is exactly why guiding was designed for - to compensate for any tracking errors (both PA errors and periodic errors).

If you can - you should start to guide - that will make your stars round in all subs.

 

1 hour ago, malc-c said:

You are also experiencing the issues of imaging on a budget.  Don't get me wrong, the 150pds and EQ5 is a capable system, but  with just the basic motor drive you are limited to getting that full potential.  It's also why on any given time when people ask about getting into imaging the HEQ5 is the most recommended and proven mount as an entry point.  However lets see if there is anything we can suggest to get you up and running  with the equipment you have.  But you will be needing to spend some cash - Imaging isn't cheap

1 hour ago, malc-c said:

Personally I would suggest going the whole hog and getting the goto option as it has better motors, and provides you with the option of using a PC/Laptop and control the guiding using EQMOD and software.

 

Guiding is definitely something I will begin to prioritise higher. 
I have been looking at Sky-Watchers own GOTO upgrade but also belt drive kits such as the EQStarProEQ5. Hopefully, I will be able to get a guide camera, guide scope, a GOTO upgrade, and the needed accessories for £600-750. 
I know that with my current setup I am pushing my mount to its limit -  I'm hoping to change my scope to a 500mm refractor after the summer. This will at least lower the total weight and make guiding a bit easier/more forgiving.
I was thinking of upgrading my mount, but I think I will with that until a can afford one load capacity greater than 20kg.    

 

[ollypenrice]

7 hours ago, vlaiv said:

That is periodic error.

To be 100% sure - see if it is aligned with RA axis - if it is, then it is periodic error.

To minimize it - shoot objects closer to NCP/Polaris. PE is the worst around equator (at DEC close to 0).

Since you are using single RA motor - you don't have this option, but there is this thing called periodic error correction. It is designed to lessen effects of periodic error. It is available in some mount models, or via EQMod / computer

Shorten your exposure length. PE is usually the worst in some parts of worm revolution - if you shorten exposure, you'll throw away less subs.

Of course - discard the worst subs.

In the end - it might help to "tune" your mount - which is basically term for stripping it down, cleaning, applying suitable grease, possibly changing bearings for quality ones and then reassembling everything carefully so that everything fits tight but does not bind.

I'm surprised you say this. I'd have suspected backlash. Clearly the mount spends most of its time in one of two positions - at either end of the 'barbell' - and so must move quickly between these two positions. To me that says backlash, but I don't claim I'm right.

Olly

[malc-c]

Olly,

Looking at how corse the teeth are on the basic RA drive, could it be that the backlash between the cogs will cause this effect when the scope is in a set position

[vlaiv]

50 minutes ago, ollypenrice said:

I'm surprised you say this. I'd have suspected backlash. Clearly the mount spends most of its time in one of two positions - at either end of the 'barbell' - and so must move quickly between these two positions. To me that says backlash, but I don't claim I'm right.

Olly

Due to elongation and other things mentioned in post - it seemed that PE is likely candidate. There was no wind, and if backlash was issue - such perfect symmetry requires perfect balance and very uniform random swaying back and forth. Why only some subs and not all if there is such symmetry?

On the other hand - symmetry can easily happen with PE without any special requirements - shape of gear is enough to cause it - just think egg - it will create such symmetric shape - most of the circumference - radius is the same and stars are point like - when approaching hump - radius is increasing and mount is trailing - at hump itself almost no time is spent and then mount starts to catch up again - thus you have symmetry in resulting star shape.

Also - do we know many of SkyWatcher mounts that don't have very distinct PE? I don't think I've heard someone bragging not to loose a single sub without guiding - that just does not happen with these mounts.

In any case - backlash is easy to sort out by making scope east heavy.

@jm_rim

You know what you can do to see if it is indeed periodic error - take the subs you shot - in the order of time they were shot and do a crop on them - just a few stars - like in first post - but do that with each sub.

Make animated gif out of it.

If you get something like this:

(up down motion is due to periodic error and left to right is due to polar alignment error - notice that some frames have elongated / split stars)

then it is periodic error. If you get random jumps between round and elongated stars that does not follow nice path - it is most likely backlash - but I seriously doubt that.

 

[jm_rim]

14 hours ago, ollypenrice said:

I'd have suspected backlash. Clearly the mount spends most of its time in one of two positions - at either end of the 'barbell' - and so must move quickly between these two positions.

14 hours ago, malc-c said:

Looking at how corse the teeth are on the basic RA drive, could it be that the backlash between the cogs will cause this effect when the scope is in a set position

13 hours ago, vlaiv said:

In any case - backlash is easy to sort out by making scope east heavy.

@jm_rim

You know what you can do to see if it is indeed periodic error - take the subs you shot - in the order of time they were shot and do a crop on them - just a few stars - like in first post - but do that with each sub.

Make animated gif out of it.

Okay, I have made a small animation of the first 4 hours from the last two nights.
The elongation is in the RA direction, but looking closely, it seems to "wobble" in both directions. Could this mean that it is backlash do to the RA being in perfect balance and random swaying back and forth? 

 

[ollypenrice]

I'd have thought that backlash on that scale would be easy to feel by wobbling the C/W bar. Since it's in RA the first test would be to run the system east heavy. 

Olly

[vlaiv]

32 minutes ago, jm_rim said:

Okay, I have made a small animation of the first 4 hours from the last two nights.
The elongation is in the RA direction, but looking closely, it seems to "wobble" in both directions. Could this mean that it is backlash do to the RA being in perfect balance and random swaying back and forth? 

This is perfect example of two most common causes of star elongation

- periodic error

- polar alignment error

Wobbling up and down is periodic error, it repeats every few minutes - mount trails a bit then it leads a bit and back again ...

Motion from left to right is in DEC direction and that is due to polar alignment error.

Most people blame elongated stars on polar alignment error - but it is usually much smaller in magnitude than periodic error - you can see that in your animation - drift to the right is fairly slow compared to constant up and down of stars.

 

 

[jm_rim]

53 minutes ago, ollypenrice said:

I'd have thought that backlash on that scale would be easy to feel by wobbling the C/W bar. Since it's in RA the first test would be to run the system east heavy. 

 

Going through my setup, I could see that it was almost perfectly balanced. So tonight I will try to test if running it east heavy will help. 

 

51 minutes ago, vlaiv said:

This is perfect example of two most common causes of star elongation

- periodic error

- polar alignment error

 

Going through my setup, I could see that it was almost perfectly balanced. So tonight I will try to test if running it east heavy will help.
Else it seems that guiding is properly going to be needed to compensate for these errors - based on your earlier comment @vlaiv

 

[vlaiv]

20 minutes ago, jm_rim said:

Going through my setup, I could see that it was almost perfectly balanced. So tonight I will try to test if running it east heavy will help.
Else it seems that guiding is properly going to be needed to compensate for these errors - based on your earlier comment @vlaiv

Yes, please try east heavy - just to be sure, but there is another thing to check when periodic error is examined.

What was your sub duration? Period of EQ5 mount is 10 minutes - see if pattern repeats every 10 / sub duration frames. If you shot 2 minute subs then pattern repeats every 5 subs.

[malc-c]

Jesper,

I personally don't think guiding is the answer.  I'm sure Olly or Vlaiv will correct me, but I think the PE is due to the the choice of drive you have.  I feel they are designed to provide an aid to visual observation so that the target simply stays in the field of view for a given eyepiece, and thus precision isn't that important.  Any backlash that creates a PE wouldn't be noticeable to the naked eye.  

IMO the pro goto upgrade would give you a lot more options, and with the microstepping employed with the NEMA motors the PE should be less noticeable, and what is detected can be compensated for as you can train the handset to record the PE and then use that data to compensate for the error.  If you then opt for using the mount with an EQDIR cable you could then recoup some of the cost by selling the handset.  EQMOD and and EQDIR cable would make guiding a lot more simple as there would be no need for any ST4 connections.  

Anyway, that's my take.  

[vlaiv]

15 minutes ago, malc-c said:

I personally don't think guiding is the answer.  I'm sure Olly or Vlaiv will correct me, but I think the PE is due to the the choice of drive you have.  I feel they are designed to provide an aid to visual observation so that the target simply stays in the field of view for a given eyepiece, and thus precision isn't that important.  Any backlash that creates a PE wouldn't be noticeable to the naked eye.  

Periodic error is produced by mechanics of the mount rather than motor. If motor is driving unevenly - that will show in similar fashion - but it will be much quicker.

This is because motor makes one revolution every 2 minutes - there is 5:1 reduction between motor and worm. Each sub would be then effected by whole motor periodic error - provided that error is of mechanical nature  - then it would be tied to motor rotation period. If it was electronic in nature - then motor would simply rotate either slower or faster than sidereal - that would show as constant trailing / leading, or it would be random if it was electronic noise - and not periodic.

I'm not against changing the drive train - it will be needed for guiding. I just don't think it will help with mechanical periodic error.

[Seelive]

The period of the final worm wheel periodic error on an EQ5 would be 10 minutes and since periodic error normally approximates a sinusoid I can't see, given the 2 minute exposures, how you would get the 'dumbell' stars. Surely, depending upon the when the exposure occurs during the periodic cycle, at the zero crossing of the 'sinewave' you will get a long trailed star whereas at the crest you will get a short trailed star but never a 'dumbell'.

[malc-c]

47 minutes ago, vlaiv said:

there is 5:1 reduction between motor and worm.

Looking at the gearing on the RA drive I would beg to differ... As far as I'm aware (and Jesper can confirm) the gear shown in the image above fits on the end of the worm shaft and engages directly with the gear on the motor.  There is no way the worm turns 5 times to one of the motor. 

I use a 4:1 ratio between the motor and worm drive on my belt driven HEQ5, and the motor pulley is tiny compared to the one on the worm shaft (15t  motor and 60t worm)

Edited January 14, 2021 by malc-c

[vlaiv]

33 minutes ago, malc-c said:

.  There is no way the worm turns 5 times to one of the motor. 

Other way around - 5 motor turns for one worm turn

http://eq-mod.sourceforge.net/prerequisites.html

Eq5 has total ratio of 704:1 and 144:1 is worm ratio

 

[vlaiv]

50 minutes ago, Seelive said:

The period of the final worm wheel periodic error on an EQ5 would be 10 minutes and since periodic error normally approximates a sinusoid I can't see, given the 2 minute exposures, how you would get the 'dumbell' stars. Surely, depending upon the when the exposure occurs during the periodic cycle, at the zero crossing of the 'sinewave' you will get a long trailed star whereas at the crest you will get a short trailed star but never a 'dumbell'.

Usually PE is far from sinusoid. If it were sine form - it would be easy to fix.

To understand how dumbbell shape forms - take a look at second part of this video - try to concentrate only on horizontal motion and "integrate" it in long exposure:

https://www.youtube.com/watch?v=S8EWKm2FwBo

In any case - I still maintain that it is PE - but happy to be proven otherwise.