Guiding issues with DEC deviations

[almcl]

Have been guiding for a month or two now and while waiting for Astro dark to come back and the moon to go away, have been checking out how well PHD, my EQ5 mount, ST80 guidescope and SW 200P have been doing.

The other night after about 12 minutes guiding, I noticed a sudden large deviation in the declination trace. Subsequently I noticed that I hadn't fully tightened the tube rings and the OTA was slipping down and imagined that this was probably contributing to the problem.

So, night before last I set about carefully balancing everything and then making things slightly tail heavy and set off guiding again. The image was going to be poor with a full moon but at least I would see the guiding performance.

This time, after about 10 minutes of small deviations, a repeating disturbance in the dec trace, which gradually gets larger, appears. Here's the trace from PHD labs analysis program:

Now I am fully aware that this combined load (11kg) on the EQ5 is probably beyond sensible, but while I save up for an EQ6 (or possibly and NEQ6) is there anything that might mitigate this declination error and could some kind soul explain where the error is coming from? I would find it easier to understand if it was in the RA trace, but DEC is just going back and forwards about the same points on the gears, isn't it? So periodic error shouldn't be an issue?

Here's a single unprocessed 240 second sub taken a bit later on, resampled to 46% of original size, in case it suggests the source of the problem. The bright triple star left of centre with diffraction spikes is HIP 106886 according to Stellarium.

[Uranium235]

Looks like a buildup of backlash leading to an overcorrection in PHD (hence the sawtooth pattern). If its anything like the old PHD (which I still use), you can try setting DEC guiding to work in one direction only, which in this case would be North (N pushes the dec graph down, S pushes it up). You can see where its drifting South (going up) before the North corrections finally kick in causing it to jump.

[geordie85]

I've been recently having the same problems. I find, as said above, setting the Dec guiding to just 1 direction (south for me but it looks like it'd be north for you) and lowering the Dec aggression to around 40 pretty much counters the problem and I get decent guiding again

[almcl]

Thanks for those thoughts!

My Dec guide mode in PHD is presently set to auto, I'll try setting it to north and see what happens.

[Uranium235]

Oh, just be aware that once you do a meridian flip, you may need to change it to South.

[Thalestris24]

Thanks for those thoughts!

My Dec guide mode in PHD is presently set to auto, I'll try setting it to north and see what happens.

Hi

I'm not sure, but it might also be due to using 'resist switch' rather than 'hysteresis', but I could be wrong...

Louise

[Uranium235]

Just seen the OP states that you have 11kg on the EQ5, that is far too much. You need to cut that in half in order to make life easier, start off by using a 9x50 finderguider. That will shed a few kilos. 

[almcl]

Thanks for those, Rob and Louise, much appreciated.

Louise, is the setting you mention the one accessed via the 'brain' icon? If so, the only options (on my version) are resist or low pass filter. Resist seems to be the default and what it is currently set to, but perhaps I'm not looking in the right place?

Rob, I see you use a 9 x 50 finder guider and get fantastic results, but while I have one (with webcam attached) won't the mismatch in focal lengths between my OTA (1000mm vs 180mm on the finder) produce other guiding accuracy problems?

Certainly going to try it this evening if the clouds stay away, but may be back with more questions...

[Thalestris24]

Thanks for those, Rob and Louise, much appreciated.

Louise, is the setting you mention the one accessed via the 'brain' icon? If so, the only options (on my version) are resist or low pass filter. Resist seems to be the default and what it is currently set to, but perhaps I'm not looking in the right place?

Rob, I see you use a 9 x 50 finder guider and get fantastic results, but while I have one (with webcam attached) won't the mismatch in focal lengths between my OTA (1000mm vs 180mm on the finder) produce other guiding accuracy problems?

Certainly going to try it this evening if the clouds stay away, but may be back with more questions...

Hi

Ah, I think you must be using PHD1 rather than OpenPHDGuiding v2.5.0. I vaguely remembered reading something about PHD2's resist switch algorithm having issues, at least, at one time. I suspect though, that your problem is most likely caused by weight/balance issues. An oag might help?

Louise

[Astrosurf]

I have no problems with my finder-guider and 200P. It's a lovely set-up. I'm interested in this thread as I find that Dec can be a bit flaky. It always seems a bit weak, but only in certain positions. A lot of the time (touch wood!) I get a reasonably flat graph, with occasional Dec blips.

I hear that webcams are a little too insenstitve to use for guiding, but your graph looked fine before the Dec issue. Can anyone say if a more sensitive camera, like the QHY5, would help?

Let us know how you get on.

Alexxx

[kalasinman]

I believe the sensitivity isn't much of an issue unless one chooses an OAG as I have. The the scale changes from way undersampled, possibly as much as 4:1 to in my case slightly oversampled at 1.03:1.3" This increases guiding accuracy, but the much narrower FOV means fewer guide stars, so the sensitive cam and smaller pixels really helps. It's the QHY5L-II mono that is the big improvement.

[Thalestris24]

Hi

I think it's a ) the rms guiding error in arc seconds vs b ) the imaging pixel scale in arc seconds that counts. I believe a ideally needs to be less than b. I find my qhy5l-ii mono with it's small pixels (3.75 micron) is a good and sensitive (QE = 74%) guide camera Having said that, I don't get brilliant guiding myself (130pds+1100d+AVX) as my imaging pixel scale is small - 1.49" whereas my rms error is greater... I've put it down to being at least partly due to the atmospheric conditions here. The end result is slightly soft stars... Using my qhy8L would be better as I'd then be imaging at 2.25"/pixel. I've not used the qhy8l lately as it's attached to my 150pds which has been gathering dust since I got the 130pds... Hmm.. I really should do something about that!!!

Louise

[geordie85]

Hi

I'm not sure, but it might also be due to using 'resist switch' rather than 'hysteresis', but I could be wrong...

Louise

This actually completely solved my Dec problems.

No more dodgy guiding for me.

Thanks

[Astrosurf]

Great! That's useful to know.

[almcl]

Well, switching to a 9 x 50 guidescope hasn't completely resolved the situation, I am sorry to report.

Here's the trace from last night. The seeing wasn't brilliant, I was usng longer, 3 min subs an d half of them showed double (or in one case quadruple) stars.

I'll try switching from resist and see what happens. I'll also beef up my home made scope rings...

The half of the subs that were usable suffered other problems (speedily processed result below), but the guiding one is my main concern for the moment.

[ollypenrice]

Al,

Backlash in Dec is not as easy to kill by imbalance as backlash in RA certainly is. In RA you just run east heavy, be it weights east or scope east. Alas, being tail (camera) heavy in Dec may kill oscillation as you move towards the zenith but, on the zenith, the top of the scope is directly above the bottom and the effective imbalance has reduced to zero. Now you need a weight offset to one side. That's the problem. Finding the solution is more difficult!

Firstly it might be worth trying to tighten the dec worm-and-wheel mesh. Secondly, the advice to correct in one direction only is very sound. You have to experiment to find which way to guide and, as stated above, this will need reversing after the flip. I've done that successfully many times. Thirdly, and this might sound grotesque (as it is, really!) Dec oscillation can be attacked by using a long, low resistance elastic attached from a fixed point to the camera end just to pull the Dec axis into one side of mesh during a run of exposures. Call this a last resort, but it has been done.

I do think it's Dec backlash. The trace is good, then there's a big lurch which is well corrected, it's good for a spell then there's another lurch. The lurches could very well be provoked by a guiding correction firing the mount to the opposide side of backlash before gravity, or a new correction (or both) send it back.

Olly

[Uranium235]

Rob, I see you use a 9 x 50 finder guider and get fantastic results, but while I have one (with webcam attached) won't the mismatch in focal lengths between my OTA (1000mm vs 180mm on the finder) produce other guiding accuracy problems?

The 9x50 as already stated, is good for up to 1000mm. One thing you need to take into account is the minimum motion setting (the amount a star needs to move before a guide command is issued). That number should be quite low, 0.5 or less as the 9x50 has a fairly short focal length (the longer the fl of the guide scope, the larger min-mo should be).

Did you try guiding in one direction only? The multiple star images youre getting is from backlash still, you've dropped about as much weight as you can by using the 9x50. Try stiffening up the finder backet a bit, I used to pack mine out with masking tape - since the standard two screws and spring you normally get in a finder bracket can be a little wobbly. Velcro up all your cables too so there is no cable tension on you guide camera.

In think that HEQ5/6 is now long overdue! I went through the same pain with my CG5, finally losing my patience with it during one Summer session (where you can not afford to lose time fiddling).

Edit: Following on from what Olly said about tweaking the mesh... take your time with it and dont do it up too tight. Take it close to the limit (almost to the point where the motor struggles), mount your telescope (or something heavy) - then back it off a little more because the dec axis will behave differently when under load. It might take a few goes to get it right, but it improved the situation for me.

[Jannis]

 Try stiffening up the finder backet a bit, I used to pack mine out with masking tape - since the standard two screws and spring you normally get in a finder bracket can be a little wobbly.

I can confirm this. In my case i wasn't able to completly align the finderscope to the main scope before i had tightend the 2 screws so little that the spring was just barly touching the finderscope. After i had drilled and tapped more holes to get a total of 6 adjustment screws was i able to get a more much more stable guiding with the finderscope.

[kalasinman]

Low weight, low cost, kill flexure=OAG

[MarsG76]

I agree with the OAG statements... An OAG gave me the below graph which looks messy but both RA and DEC always return to zero mark and is rhythmical... I think the peaks are so high because of the high magnification, 2000mm in this case.

I'll play with settings and lower the aggressiveness somewhat and see if that smooths out the guiding graph but for now I'm doing 30-45minute subs with no drift... Perhaps smoothing out the correction might Titan the stars somewhat, time will tell.

[almcl]

Thanks all for the thoughts and info.

Definitely going to tighten up the gear train - there's a bit of slop on the motor gears, as well as on the worm, so reducing that will be first task. Not using the finderscope mount but a home fabricated set of tube rings and bar, piggybacked on the OTA's mounting rings. I'll try a slight modification to these although I don't think flex has been an issue so far.

I do like Olly's idea of an elastic tie down, must have a look at that one when I've worked out which way I should go!

[ollypenrice]

Have you thought of a belt drive conversion? I'm not sure if the kits are available for HEQ5 but they are for the 6. This eliminates primary drive backlash.

Olly 

[Thalestris24]

Have you thought of a belt drive conversion? I'm not sure if the kits are available for HEQ5 but they are for the 6. This eliminates primary drive backlash.

Olly 

Hi Olly

Yes, I have the Rowan belt drive for my Heq5

Louise

[almcl]

Olly

I am afraid my lowly EQ5 doesn't take the belt mod - adding the motors was a bit of a step up!

[ollypenrice]

This is where we all start spending your money, Al!  Don't listen to us...

lly