Barbell shaped stars - is this Periodic Error or Backlash ?

[malc-c]

I knew what  I meant... just dyslexic fingers !!

But don't forget that the data you linked to on the EQMOD site is for the sysnscan pro version of the mount... Not the basic RA drive...  If you look at the gearing on the synscan upgrade you'll see the cog on the motor is a lot smaller and connects to a split gear arrangement to give the presion.  

 

 

Edited January 14, 2021 by malc-c

[Seelive]

9 minutes ago, vlaiv said:

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.

 

That top trace certainty has a sinusiodal form to it, albeit distorted by lower order errors (a square wave is just a sinewave distorted by an infinity of odd harmonics). I would expect the dumbell effect to only appear if the exposure time exceded half the primary periodic error period.

[vlaiv]

3 minutes ago, malc-c said:

I knew what  I meant... just dyslexic fingers !!

But don't forget that the data you linked to on the EQMOD site is for the sysnscan pro version of the mount... Not the basic RA drive...  If you look at the gearing on the synscan upgrade you'll see the cog on the motor is a lot smaller and connects to a split gear arrangement to give the presion.  

 

 

In any case - every version of EQ5 has 144 teeth - that means that basic period is 10 minutes (1440 minutes in a day) regardless of motors being used.

This in turn means that whatever drive system is connected to worm - it needs to operate at 0.1rpm on output shaft. Imagine for a second that stepper motor is attached directly to worm without any reduction.

Steppers have 200 steps per revolution. That means full step would be triggered every 3 seconds. Even with micro stepping - it would be visible - this is why there needs to be reduction stage - less microstepping - more reduction. Simple motors don't have enough torque to operate on many micro steps - so 16 micro steps is common with them.

Notice that RA motor has output shaft that is parallel but not central with respect to housing:

This is because stepper itself has reduction built in.

[jm_rim]

23 hours 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. 

21 hours ago, vlaiv said:

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

 

I was able to do some testing last night. First, I ensured that no cables could be the cause. Then I polar align using SharpCap, and finally I balanced my mount a little east heavy.

I have, again, made a small animation of the first 4 hours from last night. Running east heavy seemed to have help quiet a lot - or is it just me hoping to see an improvement 🙂
The frames are drifting in the RA tracking direction, I this do to polar alignment error? And in the end, I look like the telescope went out of focus (hoping to my autofocus motor today).
There is still barbell/elongated stars, which I assume is most likely do to periodic error, as I seem to be a repeating pattern. 

 

So I guess that my next step is, to check for backlash and shorten my exposure time, for now. And when there comes a period with bad weather tune my mount. 

 

[vlaiv]

4 minutes ago, jm_rim said:

The frames are drifting in the RA tracking direction, I this do to polar alignment error?

Drift due to polar alignment error is always in DEC. Drift that you are seeing is due to imbalance you created with east heavy scope. Did you tighten the clutches well?

Imbalance needs to be small - just to keep small tension on RA drive train.

This sort of drift can also be due to worm gear itself - it is largest drive component, and spins one revolution per day. Often in a session we don't get to spin it more than 1/4 of its revolution. There might be side of it that is some what distorted from perfect circle - and on that side it will run slower or faster (actually it needs to do both so that average rotation speed remains the same).

Assuming one is imaging for 4h each session - above scenario would repeat each 6 imaging sessions (that sort of total drift).

[jm_rim]

2 hours ago, vlaiv said:

Did you tighten the clutches well?

I hope I did. My RA clutch is not working optimally, which one of the things I need to fix.

 

2 hours ago, vlaiv said:

This sort of drift can also be due to worm gear itself - it is largest drive component, and spins one revolution per day. Often in a session we don't get to spin it more than 1/4 of its revolution. There might be side of it that is some what distorted from perfect circle - and on that side it will run slower or faster (actually it needs to do both so that average rotation speed remains the same).

If it is because of the worm gear, can that affect the frames - besides the drift? Or do I have to change it?

 

[vlaiv]

2 minutes ago, jm_rim said:

If it is because of the worm gear, can that affect the frames - besides the drift? Or do I have to change it?

No, don't worry about it - such drift is so slow that it can't affect individual subs and is easily corrected with guiding.

In fact - if you don't guide -  it is good to have such slow drift as it provides "natural dither" of your subs. In very rare occasions it might produce (I'm still not 100% sure of relation between the two) - so called walking noise.

[vfastro]

Hello all,

I have exactly the same equipment : EQ5 (no GoTo) , SW130 PDS, Nikon 3300 DSLR, and exactly the same problem except that in my case I shoot 60 sec subs.  I understand that the problem is most probably related to PE and that EQ5 is designed mainly for observation and not imaging. However I would like to know if anybody has succeeded to improve the above setup by using guiding and with what camera / guide scope.

Thank you in advance. 

[malc-c]

Using a guiding system, either guidescope / camera or off axis and camera will normally improve tracking to the ability of the mount.  The minimum option is to use an ST4 interface cable between the camera and the control box (needs to be the enhanced controller for a plug and play option).  It is possible to modify the basic handset, but that involves opening up the handset and soldering wires to  it.

I've not used this personally, but have a look at the Hi-Tek astro ST4 autoguider here This allows you to connect the mount to a computer and then use their software to handle the guiding using the ASCOM platform.  You basically use an USB cable from the PC to the autoguider  and then plug the autoguider into the ST4 port on the controller.  You can then use any guide camera  that is ASCOM compliant with a USB cable and the software provided.  Being ASCOM compliant it should also work with PHD2.

 

[900SL]

Also check the grease on your gear train and shafts. At - 15c this might cause stiction and give the dumbell pattern (bind... release) 

[neil phillips]

I am no expert but looks like the gears need tightening to me. A dead giveaway would be at high focal length work. If at high power when you focus. the image moves sideways. This is a big indication. I had this recently with my mount.  Upon checking i could actually feel the slack. Moving ever so slightly side to side. Once tightened the problem went. I was imaging tonight with a 10" Newtonian unguided. at 1600mm focal length on the EQ5 Single RA drive. and although the stars are far from the level of a good setup for this kind of thing. Certainly no where near what your showing. At a much higher focal length and weight.  Than you are using. Here is a lucky imaging example. But anyway its only a suggestion for you to rule out. I could be totally wrong. 44 subs 40 seconds each Exposure=1.000s

 

Edited December 11, 2021 by neil phillips

[alacant]

On 13/01/2021 at 22:56, vlaiv said:

PE is likely candidate

I don't think so. That gives a uniform line.

Our vote goes to stiction or backlash due to an almost perfectly balanced telescope. The gear wheel is rarely round so at certain angles will need more nudges to overcome stiction whereupon it overshoots.

 

. 

[malc-c]

39 minutes ago, alacant said:

I don't think so. That gives a uniform line.

Our vote goes to stiction or backlash due to an almost perfectly balanced telescope. The gear wheel is rarely round so at certain angles will need more nudges to overcome stiction whereupon it overshoots.

I always thought PE was more sinusoidal - Looking at graphs produced by EQMOD tools

I would have thought that stiction would be more random as the CoG and mass shifts over the course of time. ?? - (just a hunch - no real expert in these things)

[vlaiv]

1 hour ago, alacant said:

I don't think so. That gives a uniform line.

Our vote goes to stiction or backlash due to an almost perfectly balanced telescope. The gear wheel is rarely round so at certain angles will need more nudges to overcome stiction whereupon it overshoots.

 

. 

It would only produce uniform line if it was "sawtooth" pattern - constant speed in one direction, then constant speed in other direction.

Periodic error is often very non uniform and consists out of few sine waves superimposed.

Here is recording of my PE on HEQ5 prior to belt mod:

Left to right is due to polar alignment, and obviously up and down is due to PE. Here you can see that stars don't spend equal amount of time along the curve - there are places where star is almost still and other places where it accelerates rapidly. In fact - if you look at some frames - you'll see exact same pattern - two bright spots interconnected with fainter line in between.

[michael8554]

Periodic Error is rarely a smooth sinusoidal shape.

You can see the repeating pattern in vlaiv's PE, hard to follow the video, but looks roughly like this:

 

Michael

[alacant]

Mmm. We find it produces elongated stars. Maybe the theory predicts otherwise.

Cheers

 

[malc-c]

31 minutes ago, alacant said:

Mmm. We find it produces elongated stars. Maybe the theory predicts otherwise.

Cheers

 

Looking at the PHD2 graph, with a decent PA I would expect that.  DEC, until the end is fairly flat, but RA is tracing back and forth,  Any deviation in DEC is so slight that all it produces is slightly wide trails.  If PA was off and DEC was also Moving back and forth then the mount has the potential of tracking a figure 8 pattern resulting in dumbbell stars... at least thats how I interpret things... I could be wrong ??

[vlaiv]

44 minutes ago, alacant said:

Mmm. We find it produces elongated stars. Maybe the theory predicts otherwise.

Is that image with autoguiding running?

In that case - it is possible to have little lines. OP did not auto guide and barbell shapes were due to PE alone.

[alacant]

2 hours ago, vlaiv said:

OP did not auto guide

Neither did we. That's why we got lines.

If you want circular stars from a mount with this error, just activate autoguiding.

Cheers

 

 

Edited December 12, 2021 by alacant