Guiding Problems - Poor Polar Alignment?

[Ian McCallum]

I'm not sure exactly where to post this, so apologies if it's in the wrong place...

There was a couple of hours of clearish sky, promised last night.  I bought out the EQ5 Deluxe mount with Enhance dual axis motor drives (mouthful), used a spirit level to get the tripod level and used a compass app on my phone to get a bearing of due North.  I polar aligned the best I could, through the polar scope and added the mount.  Once I got everything attached and the ZWO ASI 120MM mini guide camera attached to the guide scope, I then set up the RPi (with Astroberry loaded).

I focused to tracking the Orion Nebula, but it was a bit hit and miss with clouds rolling by.  I don't think my guide was good enough, perhaps due to polar alignment. 

Could someone take a look at these screenshots and logs and let me know what you think, please?🤔  Please be gentle, I'm new to guiding and AP.😛😂

PHD2_DebugLog_2022-01-29_200609.txt PHD2_GuideLog_2022-01-29_200609.txt

[kens]

Your problem looks more like dec backlash.

Your calibration is also not great. Did you calibrate on your target? The rates indicate not but that could be due to any number of things. The dec rate vs RA rate indicate calibration at a declination of around 33 degrees

A good calibration is essential for guiding. Yours is orthogonal (good) but based on only a few data points. Ideally you want about 10-15 points along each leg. That is determined mainly by the calibration step size. If you use the profile wizard to set things up it normally calculates the step size correctly as long as you input the correct guide rate. If you didn't use the wizerd then set up a new profile using it and try again. Otherwise, recalculate the step size in the brain. You should be using a step size of around 400ms (vs 1200ms) to get a reaonable number of calibration steps.

A reasonably good polar alignment is also needed for calibration but I don't see the telltale signs of very poor polar alignment in your calibration (a vee shaped RA calibration leg). Since you have both Astroberry and PHD2 you can ditch the polar scope and use any of the excellent PA methods in those tools. As a starter I'd suggest using the Polar Drift Alignment (PDA) in PHD2 as it is quick and easy. The EKOS PA too in Astroberry is also very good and more accurate but a little more involved. Both are easier and more accurate than using a polar scope and do not rely on a good calibration to work.

After you calibrate then run the PHD2 Guiding Assistant including the backlash measurement. That will tell you more about what is going on.

 

[wimvb]

I haven't exami ed the guide log, but in the screen shot it loons like backlash in dec. Run the guiding assistant for 10 minutes and determine how much backlash you have. If it's not too bad, PHD can compensate for it.

[Ian McCallum]

17 minutes ago, kens said:

Your problem looks more like dec backlash.

Your calibration is also not great. Did you calibrate on your target? The rates indicate not but that could be due to any number of things. The dec rate vs RA rate indicate calibration at a declination of around 33 degrees

A good calibration is essential for guiding. Yours is orthogonal (good) but based on only a few data points. Ideally you want about 10-15 points along each leg. That is determined mainly by the calibration step size. If you use the profile wizard to set things up it normally calculates the step size correctly as long as you input the correct guide rate. If you didn't use the wizerd then set up a new profile using it and try again. Otherwise, recalculate the step size in the brain. You should be using a step size of around 400ms (vs 1200ms) to get a reaonable number of calibration steps.

A reasonably good polar alignment is also needed for calibration but I don't see the telltale signs of very poor polar alignment in your calibration (a vee shaped RA calibration leg). Since you have both Astroberry and PHD2 you can ditch the polar scope and use any of the excellent PA methods in those tools. As a starter I'd suggest using the Polar Drift Alignment (PDA) in PHD2 as it is quick and easy. The EKOS PA too in Astroberry is also very good and more accurate but a little more involved. Both are easier and more accurate than using a polar scope and do not rely on a good calibration to work.

After you calibrate then run the PHD2 Guiding Assistant including the backlash measurement. That will tell you more about what is going on.

 

I did align it on the target and calibrated it. 

Regarding your other comments, I can't answer them as I simply don't know. ☹️

[Ian McCallum]

I'm getting to the point where I'm thinking about giving up, before I spend any more money on it. 

I'm not getting the chance to do any experiments with it or experience, due to the very poor weather we've had recently! ☹️

[kens]

Then I'd suggest to use the profile wizard to create a new profile. When it is finished check the calibration step size on the Guiding tab of the brain. It should be around 400ms for  a target around dec 0. In the calculatoryou need to enter the declination for each target as that adjusts the step size accordingly. 

  

[michael8554]

I see from your GuideLog that RA, Dec and Guidespeed are not being reported from the mount, so you are ST-4 guiding ?

With ST-4 you must Calibrate on every new target.

Calibration took 7 steps in RA and Dec, but Dec included clearing 15 steps of Backlash before Dec Calibration started.

It is well worth clearing that Backlash by pulsing north until the guidestar moves consistently, THEN Calibrate.

But the guiding shows that RA and Dec are both "stiff", and so not responding to PHD2 commands - Dec took 2 minutes to correct at one point !

Both worms need adjusting.

PHD2 Dec Backlash Compensation was disabled, which doesn't help.

After a good Calibration, run the PHD2 Guide Assistant with Measure Backlash ticked, that will advise a starting setting for Dec Backlash Compensation.

Michael

[malc-c]

31 minutes ago, michael8554 said:

I see from your GuideLog that RA, Dec and Guidespeed are not being reported from the mount, so you are ST-4 guiding ?

 

Michael

@michael8554 Michael,  the OP states  

Quote

 I bought out the EQ5 Deluxe mount with Enhance dual axis motor drives

So he has no other option than to use ST-4 ports.  @vlaiv is normally the expert on PHD2 graphs, hopefully he'll chime in with some suggestions

[vlaiv]

Something is not right with this mount / setup.

Did you forget to tighten DEC clutch?

It almost looks like mount is not responding to DEC corrections. DEC should be fairly stationary on its own if everything is setup right.

Although it looks like backlash - there is simply too much of it to be real backlash. I mean - look at this:

That is 15" P2P DEC excursion over 4 minutes that is not corrected properly with like bunch of pulses.

Cable snag? Strong wind? Very poor balance?

Not even sure what to recommend as a first step.

I'd say - check mechanics of the mount and scope. Check balance, check if anything is loose to the touch - including guide scope (maybe this is just massive differential flexure in DEC?) and camera to guide scope connection.

 

[Ian McCallum]

8 hours ago, michael8554 said:

I see from your GuideLog that RA, Dec and Guidespeed are not being reported from the mount, so you are ST-4 guiding ?

With ST-4 you must Calibrate on every new target.

Calibration took 7 steps in RA and Dec, but Dec included clearing 15 steps of Backlash before Dec Calibration started.

It is well worth clearing that Backlash by pulsing north until the guidestar moves consistently, THEN Calibrate.

But the guiding shows that RA and Dec are both "stiff", and so not responding to PHD2 commands - Dec took 2 minutes to correct at one point !

Both worms need adjusting.

PHD2 Dec Backlash Compensation was disabled, which doesn't help.

After a good Calibration, run the PHD2 Guide Assistant with Measure Backlash ticked, that will advise a starting setting for Dec Backlash Compensation.

Michael

Yes, I'm using ST4 guiding. Can gear slop cause or contribute to backlash? 🤔

Here's a video I made earlier. The first one is the Dec and the second one is the RA. Judge for yourself which has the greatest amount of play in it. 

[Ian McCallum]

7 hours ago, vlaiv said:

Something is not right with this mount / setup.

Did you forget to tighten DEC clutch?

It almost looks like mount is not responding to DEC corrections. DEC should be fairly stationary on its own if everything is setup right.

Although it looks like backlash - there is simply too much of it to be real backlash. I mean - look at this:

That is 15" P2P DEC excursion over 4 minutes that is not corrected properly with like bunch of pulses.

Cable snag? Strong wind? Very poor balance?

Not even sure what to recommend as a first step.

I'd say - check mechanics of the mount and scope. Check balance, check if anything is loose to the touch - including guide scope (maybe this is just massive differential flexure in DEC?) and camera to guide scope connection.

 

The Dec clutch was tightened. The wind had dropped off, that evening, but guiding was hampered by passing clouds. 

There was no cable snags or anything like that. 

The only thing lose was the play in the motors and gears, even though they are tight on the shafts. 

 

 

[vlaiv]

Ok, if everything is mechanically sound - you need to tune your backlash for both axis.

Here is short video on how it's done:

https://www.youtube.com/watch?v=_s8941MvYXg

You can also search for other resources that show this procedure on EQ5 - it is fairly easy but always sort of fiddly business as you need to make sure that slack is minimal and there is no binding all the way around the revolution. Often, because of precision of manufacturing - there will be one part of the turn that is "tighter" and make sure that it is not binding.

Do it on both axis.

[malc-c]

IMO that is a fair amount of backlash on both axis - It takes time to hit that sweet spot between too tight a mesh and free rotation but with no slop, just do small adjustments at a time and I'm sure you'll get there

[Ian McCallum]

Here's what I've found... 

Firstly, that RA knob was slipping on the shaft a bit, as it appeared to be a bit bent. Looking at the rotation of the shaft, it looks okay. I think the knob was partially rubbing against the bearing adjustment nut during it's rotation. I sorted that out and it's turns more freely now. 

I also found that the two grubs screws holding the brass cog to RA motor shaft had came a bit loose. Again, that was soon sorted. 

The Dec was also sorted to make it more free running. 

If you compare the video below with my previous video, you can see there's less play in the shafts. 

I now hope that helps solve the problems, somewhat. 🤞

 

[Ian McCallum]

I wonder if this evening's effort was any better?🤔

PHD2_DebugLog_2022-02-06_103137.txt

[malc-c]

7 minutes ago, Ian McCallum said:

I wonder if this evening's effort was any better?🤔

PHD2_DebugLog_2022-02-06_103137.txt 5.05 MB · 2 downloads

Ian, that's a debug log, you need to upload the guide log file

[Ian McCallum]

34 minutes ago, malc-c said:

Ian, that's a debug log, you need to upload the guide log file

Whoops. I'll get that done. 

[Ian McCallum]

42 minutes ago, malc-c said:

Ian, that's a debug log, you need to upload the guide log file

I hope this is the right one.🙄

PHD2_GuideLog_2022-02-06_103137.txt

[malc-c]

Bit of a wobble at the start which has skewed the peak values somewhat, the rest of the trace of the longest guide session looked quite reasonable  to me for the equipment used.  The drift suggests a 16.1 arc minute polar alignment error so better polar alignment might improve the guide accuracy a little more.  Hopefully @vlaiv will be able to provide a more detailed in depth diagnosis of the log.

[vlaiv]

I will look into that at first opportunity.

My computer died on me, so I have now new machine with band new install of Windows 11 (that was adventure to install and get going), so missing all the software needed to analyze the data.

Hopefully, I'll install it all in next few days.

[Ian McCallum]

1 hour ago, malc-c said:

Bit of a wobble at the start which has skewed the peak values somewhat, the rest of the trace of the longest guide session looked quite reasonable  to me for the equipment used.  The drift suggests a 16.1 arc minute polar alignment error so better polar alignment might improve the guide accuracy a little more.  Hopefully @vlaiv will be able to provide a more detailed in depth diagnosis of the log.

Some of the wobble may have been me trying to realign the image on to the laptop screen, whilst outside, moving the controller buttons - waiting for the image to update on the screen.

[Ian McCallum]

1 hour ago, vlaiv said:

I will look into that at first opportunity.

My computer died on me, so I have now new machine with band new install of Windows 11 (that was adventure to install and get going), so missing all the software needed to analyze the data.

Hopefully, I'll install it all in next few days.

I've been pulling my hair out with Raspberry Pi, as I've never had one before.  I'm more used to Windows, so it's a steep learning curve.  I'm finding KStars and Ekos rather daunting, as they look very complicated.😟  I've been very tempted a few times to give up on it.

[vlaiv]

2 minutes ago, Ian McCallum said:

I've been pulling my hair out with Raspberry Pi, as I've never had one before.  I'm more used to Windows, so it's a steep learning curve.  I'm finding KStars and Ekos rather daunting, as they look very complicated.😟  I've been very tempted a few times to give up on it.

This was rather interesting experience.

I hoped to just "transplant" my SSD into new setup and Windows 7 is usually resilient enough to withstand hardware changes like that with few driver updates - but new hardware is not compatible!

There are no drivers for Windows 7, UEFI won't even see SSD as bootable disk - it needed some compatibility setting turned on - but since I used integrated graphics instead of discrete - it was not an option since graphics would not work in compatibility mode, so I had to do fresh install of Windows 11. That required fiddling with bios - turning on secure boot, TPM and what not just to be able to install it.

Windows 11 is rather bloated with preinstalled applications so I took quite bit of time to setup (turn off) everything I did not need/want.

In any case, papa's got a brand new machine now - it is literally at least x4 as powerful as last one (6 core / 12 thread vs 2 core / 4 thread, 32GB ram vs 8GB ram, 6.4GB/s NVME SSD vs 550MB/s SATA 3 SSD and so on ).

[scotty38]

5 minutes ago, vlaiv said:

In any case, papa's got a brand new machine now - it is literally at least x4 as powerful as last one (6 core / 12 thread vs 2 core / 4 thread, 32GB ram vs 8GB ram, 6.4GB/s NVME SSD vs 550MB/s SATA 3 SSD and so on ).

 

No excuse to not run PixInsight now Vlaiv 🙂

[malc-c]

31 minutes ago, vlaiv said:

 

In any case, papa's got a brand new machine now - it is literally at least x4 as powerful as last one (6 core / 12 thread vs 2 core / 4 thread, 32GB ram vs 8GB ram, 6.4GB/s NVME SSD vs 550MB/s SATA 3 SSD and so on ).

Sounds like a nice spec'd machine.  I use Maxim Reflect to take regular images of the drive as backups.  Makes restoring a PC a doddle - I had issues after installing NINA on the observatory PC and rather than try and figure out how to role back various C++ run time files etc to try and revert the machine back into the pre NINA install it was quicker to just re-image the drive back to the pre-install state. - Took 20 minutes.  If you place an old image on a new PC the only thing that is required is to re-activate windows as the change in hardware gets detected.  I can't recall the last time I built a windows install from scratch and installed all the programs needed.