Guiding goes well until...

[Lloyd Merrill]

This is my first post to the group.  I've run out of ideas on how to trouble shoot my problem.  This is my first attempt in trying to describe the issue that I've having.  Any suggestions would be greatly appreciated.

Using N.I.N.A, I can get and a number of well guided exposures in before I lose guiding all together.  The example that that I have provided is a sequence of ten, 300 second exposers of M106.  The sequence starts with the telescope pointing east and and should end before reaching the meridian.  I get 8 good images and then lose my guiding.

Equipment:

• The scope is a 152mm, f5.9 refractor.
• The mount is a Losmandy Titan fairly well balanced but biased to the east with a weighted pully system using Gemini-2.
• The camera is an SBIG STF-8300C
• The guide camera is a Lodestar X2 imaging through OAG-8300.

Log files below.

https://openphdguiding.org/logs/dl/PHD2_logs_vitz.zip

Once again, any suggestions would be greatly appreciated.

Lloyd Merrill

[Bruce Waddington]

Hi Lloyd, welcome to the forum.  You only sent us one guiding sequence but I assume you're talking about the unhappy event that occurred at its end (RA in red):

You can see that the guide star moved by a large amount in both axes, and we know the Dec motor wasn't running at the time.  I think it's unlikely to be something coming from the mount, it looks like an external event that is creating unwanted movement - a shift in the guide camera/guide scope or a cable snag of some kind.  If it happens at fairly predictable intervals, it's not going to be wind or something like that but either of the first two problems can exhibit this behavior because the mount is moving at a constant rate from east to west.   There's a long-winded discussion of these "gremlins" in this document:

https://openphdguiding.org/tutorial-analyzing-phd2-guiding-results/

These can be fiendishly hard to track down because the amounts of movement are so small - about 80 microns in this case, roughly the thickness of a human hair.  You can't simply go out and look at it for a few minutes and conclude everything is fine because you can't see what's happening at this scale of movement.  If it helps at all, the scope was pointing at Dec=47, 2.5 hours east of the meridian, with the scope on the west side of this pier.

Hope you can track it down,

Bruce

[Lloyd Merrill]

Bruce,

Thank you for your feed back.  And yes, I have reviewed this document.  The funny thing is that when you're reading it, it sort of makes sense and then you try and apply it to your own equipment, it doesn't make sense at all.  

This is not the only sequence that I have but they always end up the same.  I suspect that as the mount reaches ~75 degrees in altitude guiding goes wonky.  In the past I used to used MaximDL for guiding and had similar problems as the scope approaches the meridian.  The image train seems fairly secure.  The mount is bias weighted to the east.  Could a shift in DEC axis be causing the problem?

Lloyd

[Bruce Waddington]

It could be anything.  I don't know how this "weighted pulley system" works but if it's pulling on the mount in any way at all, it is a potential source of trouble as well.  80 microns, 80 microns, that's all it takes...

Bruce

[Bryan]

Bruce

This is DIY technique, especially for Losmandy mounts, that biases the weight to the east, REGARDLESS of pier side.

https://www.cloudynights.com/topic/607823-ra-worm-string-bias-trick/#entry8391537

Bryan

[Lloyd Merrill]

My system is very similar and also biases east regardless of side.

Lloyd

[a_pet...@yahoo.com]

Hello Lloyd,

I had a functional imaging setup using a Celestron CG4 and 6" reflector.  That was the ultimate test for finding system errors  :)

I tapped into the hand-box box and added a cable for dec/ra commands.  

I had exactly the same issues to what was shown above : Random, Huge Dec steps. It was caused by the cables slowly loading the Dec axis

until they overcame the friction in grease/bearings and created a quick, huge error as the dec axis 'slipped through the deadband'...

PHD compensating, and the whole process repeating 10s of minutes later ...

Possible Solution ( or worthy of a test )

1) Force a Dec Drift in only One Direction by Polar aligning and then shifting in Azimuth a few degrees.

2) Wrap a stiff Bungee to load the Dec Axis similar to the string you have for RA .. but since the total motion will be small, a weight is not required.

The Bungee needs to be set to load the Dec against the slight drift. ..

A Stiff Cable can unload either axis through its backlash dead-band  and cause the PHD control system to go nuts in compensation ...

( Wind Gusts can do the same if strong enough )

Cheers,

Alex P

Winter Springs Fla

[Lloyd Merrill]

Alex,

Thank you for your response.  Are you suggesting I do another polar align of my mount then force a 2 degree error on the Azimuth?  I had worked pretty hard to get it to within Polar Alignment Error of 3.3'.  If I'm misreading your statement please correct me.

Thanks,

Lloyd

[a_pet...@yahoo.com]

Hello Lloyd,

Two things :

1) Dec correction may be improved by forcing a small,slow predictable single directional error.

2) Some physical external thing is causing the quick Dec shift

1) The problem with a well aligned Dec is that the error will probably shift slightly N and then S,  This initiates a huge motor correction to

force the mount through the gearing deadband ( backlash ) so it can move the physical scope position just a few arc-secs... back & forth, etc.

If there is a small misalignment in Az, the Dec will only drift in one direction and the gears will remain meshed, and a tick or two of the steppers will correct

a small ( constant direction ) error .. Either No Correction, or a Correction always in the same direction so Deadband passage is never required.

No precision in Az is required, just guess a few degrees and adjust.

2)  There is no 'optical' reason for a quick Dec shift .. something 'pulled' the mount through its gearing backlash.

I am surprised anything can move the Titan, but maybe a cable shift is doing it...

As the time passes and the RA rotates through many degrees, the angle of the cables shift and may tug on the

Dec axis and pull it through the backlash.. You have the RA weighted, so the cables will probably Not influence that axis

Alex

--------------------------------------------------------------------------------------------------------------------------------------------