Declination Guiding Problem

[Dann]

Apologies for starting another 'guiding problems' thread, since starting to guide I have been unable to achieve good guiding in the declination axis and I’m hoping someone may have some ideas.

Below are the guiding graphs for three separate nights, the top two were in a similar part of the sky, M33 and M45. The last one was in the direction of the Dumbbell.

Each night was very still so I doubt it had anything to do with the wind and I’m in a reasonably quiet village so no heavy vibration. At first I was poorly balanced in declination, I’ve since solved that and I’m happy that everything is nicely balanced now, although it doesn’t seem to have solved its tendency to imitate a ride at Alton Towers. Should the RA and dec be deliberately and slightly unbalanced at all?

I’m guiding a 750mm focal length 150p with Orion’s 50mm guidescope and I’ve read I should up the calibration stop value from its default, so I’ll be giving that a try. At the moment it takes around 30-40 steps to calibrate each direction, could that be the cause? Any other settings I should look into? Currently I have everything as default apart how quickly the camera takes an image, set at 3sec.

Oddly last night’s guiding was great to start with and I managed to take this single 8 min test sub of M27 (longest so far) but after 20 mins or so it went mad again.

Any thoughts or ideas I can try?

Thanks

Danny

[Merlin66]

Double check your polar alignment - if the mount has good PA then there should be no need to make correction in Dec...

A slight out of balance -to the east for RA, either for Dec will help maintain worm/ wheel contact. Doesn't need much.

HTH

[A320Flyer]

Hi Guys,

I certainly dont want to hijack this thread but I too would also like to know the answer to this. I tried giuding for the first time last night and got pretty much the same graph as Danny. RA and Dec seem to guide very well then, suddenly, Dec takes a dive then recovers slowly. All is ok for a couple of minutes then the dive again.

Could this be anything to do with backlash? and if so, how to rectify?

Cheers

Bill

[Dann]

Thats interesting Bill, pretty much the same oscilations I have. Ill keep you in the loop with anything I try in the hope we can find a solution.

[James]

It's probably partly backlash but two things need to be done - one, as merlin66 says you'll need to tweak your polar alignment somewhat, the better the polar alignment the less of a problem this would be anyway and secondly you need to seriously increase the Max Dec duration. When I've guided using a 432mm focal length scope I've had the Max Dec value up at 1500ms

With a much higher Max Dec value PHD can issue a long enough slew command needed to take up the slack created by the backlash.

James

[Black Knight]

Funnily enough, I get the same behaviour in Dec. I shall be trying out James' suggestion tonight (hopefully)!

[bamus]

I have got same issue and it’s not due to bad polar alignment (spent 1 hrs yesterday doing drift method).

With guiding with DEC turned off I got 3 to 4 minutes perfect tracking and then out of nowhere dx/dy graph starts sharply moving from middle line.

Same with guiding ON. I can get 10 minutes perfect tracking and then it start again, DEC graph is all over the place. Increasing Max DEC to 2000 has improved graph slightly, but its still far from perfect.

I am in the stage now where RA tracking is perfect and only thing limiting my ability to take longer exposures is DEC.

I suspect few things can cause such a behaviour:

- Telescope is not balanced well enough. Its not possible to achieve good balance with my WO72 as mounting foot is too short – I will have to find longer dovetail for it. But on the other hand my setup is light enough (3kg) and my mount (EQ5) should handle this without any problem

- Cable from guide camera and main camera are causing some pulling force which affects guide graph (but it does not affect RA graph, and after attaching cables properly to mount issue still existed)

- Mechanical error either in DEC motor, or DEC shaft/bearings – but why so many other people are getting it?

- Backlash – this is my main suspect. I think PHD does not handle clearing backlash properly. It takes ages to clear backlash and during N-S calibration star is not coming back to original position

Have you got any other ideas?

Edited September 28, 2011 by bamus

[Merlin66]

bamus,

re-read your message....

What you are saying is that the mount, not the guide set up is allowing movement in Dec:

With guiding with DEC turned off I got 3 to 4 minutes perfect tracking and then out of nowhere dx/dy graph starts sharply moving from middle line

If your PA as as good as you say, then it's got to be that the dec axis is going "sloppy"..

the usual fix is to pre-load the dec axis ( some out of balance load) so there's no need for reversal/ dead band.

What do you think??

[bamus]

yes Merlin66, I am almost sure its something to do with "sloppy" dec axis.

2 days ago I have took apart upper part of mount head where DEC motor is, regreased everything and tightned all screws, but problem is still not solved. Could you explain what do you mean by :

"the usual fix is to pre-load the dec axis ( some out of balance load) so there's no need for reversal/ dead band"

(sorry need explanation - I am foreigner

)

Edited September 28, 2011 by bamus

[Merlin66]

That's OK - so am I - an exile from Oz -

Basically just make sure there's a degree of out of balance on the Dec axis at all times...this may mean sliding the OTA up/ down in the holding rings or by adding say a Tesco velcro wrist weight to the tube - the idea is to have a "load" onto the Dec axis.....

HTH

[malc-c]

That's strange... I don't have any suggestions, but my issue is with the RA slightly zig-zagging and a decent DEC trace....

[Dann]

Thank you. Ill tinker with the above tonight and let you know how I get on.

[Dann]

Ok, up and running. I've left it going for half an hour and hope to return to a better picture.

I've balanced the dec at horizontal ever so slightly mirror heavy and changed 'Max dec duration from 150 to 1000 and The Calibration Step from 750 to 2000.

Definitely made a difference to the calibration, it only took 15 steps in each direction as supposed to 40ish.

I await the red line of doom :-)

[malc-c]

The other thinng to do is balance with the tube upright. With the weight bar horizontal and level rotate the tube until its pointing straight up and see if it tends to fall. If so, rotate the tube within the rings so that the camera counter-balances the guide scope

[bamus]

Dann - how was it last night? any improvement in DEC graph?

[martin_h]

Its also a good idea to remove the Dec backlash before PHD starts its calibration routine. I usualy line up my target then slew the mount South a bit then bring it back North - the North movement takes out the backlash. PHD then only takes a second or two to "remove Dec backlash"

[A320Flyer]

Guys,

An update on my second night with PHD:

I spent ages getting good PA then I tried increasing the Max Dec Duration up to 1500 as suggested. This made the "saw-tooth" motion on the Dec graph worse. As I am using a 50x9 finder-guider on an eq3-2 mount, the slowest speed of the guiding motors is 2x sidereal. So, logically, I thought "if my motors are running faster than the ideal speed, I should reduce the Max Dec Duration so that the guiding is not so severe". I reduced this to 300 and saw an immediate improvement in the shape of the Dec graph. I then fine tweaked the settings to RA agr =5 0, RA hys = 10, Mn mo = 0.10, Max RA = 1000, mx Dec = 450 and Dec guiding left at Auto.

This gave me a fairly even graph and no noticable trailing in 5 min subs.

As I said, this is for a guider-finder on an eq3-2 mount. It works for me and I hope it might be useful to someone else.

Thanks for all your help.

Bill

[Dann]

Dann - how was it last night? any improvement in DEC graph?

Well I must thank you all for the help. Last nights guiding was much smoother.

All I changed was the Calibration step up to 2000 and the Max dec duration to 1000 and to 1500 towards the end.

Below is the graph for about 2 hours guiding, off out tonight to see if I can fine tune it a little more.

Thanks again.

[Starflyer]

I used to be plagued by this. It's caused by poor polar alignment, Dec backlash and having PHD set to guide Dec in auto.

Increasing Max Dec Duration sort of fixes it but setting the Dec guiding direction to North or South, depending on which way your guide star is drifting, completely removed the see-saw wave for me. Good polar alignment helps no end in combination with the above.

People on the Stark Labs Yahoo group have also commented a few times that this behaviour became worse after v1.11. I've stuck with the latest version as I like the added features and my fix works for me.

[bamus]

Tonight I will try longer focal length guide telescope (at the moment I am using 135 mm lens and will change it to 300mm refractor) and will see if anything will improve. But to be honest I don’t expect huge improvement as I think there is mechanical problem in DEC axis as looking on the PHD DEC graph (with poor PA) I have noticed at the beginning small North corrections for a while (2-3 minutes) than it went back to South, small corrections at the beginning , then bigger and bigger (up to 1000ms) and then out of nowhere it went to huge North corrections and in this moment see-saw graph was appearing. It was very regular so it was almost like periodic error (I know DEC PE doesn’t exist but it almost look like one). If I am thinking correct, with poor PA declinations corrections should be only in one direction all the time, right?

Another issue is fact that calibration in DEC is not looking good as well. First off all clearing backlash takes about 5 minutes. It takes about 30 steps in North direction (not very regular steps, first 10 moves star only about 3 to 4 pixels, then there is huge jump in one step of about 7-8 pixels, some steps does not move star at all and in some stars is going south during North calibration. South calibration hardly moves star back to its original position).

So today, my final test night will be with longer focal length refractor, good PA (drift method) and I will check first how it will track with guiding in DEC off to reconfirm that problem has got mechanical nature, not software. If it is mechanical, then I will have to strip my mount completely down and put it back together. If it does not improve things I will throw the towel and put all my equipment in FOR SALE section (hope that I will not have to do it, but it will not make any sense to play astrophotography with this mount any more)

Edited September 30, 2011 by bamus

[mark7331]

bamus sorry if you know this approach already but I thought it might help to share my method for calibration. I understand your frustration, I've experienced it myself, but there should be a fix for reasonable guiding.

So:-

I started by working out a calibration step value that produces between 10 and 20 steps when calibrating. Leave everything else as default until you can produce this. If you're using EQMOD you can change the RA & DEC Pulse Guide rate to help reach the calibration "zone" you're looking for. If you're using the ST4 (Autoguider) port then adjust the guide rate to help - you may need to do this with your handset.

Once you've found a good calibration step/guide rate pair then take some exposures of increasing length and modify the MaxRA and MaxDEC durations. Once you have a reasonable image then tweak the other settings (RA agressiveness, hysteresis etc) to smooth further if you need to.

In my experience the guiding graph isn't really a useful indicator of good guiding. I've taken some long exposures where the stars are perfectly round but the graph has appeared quite bouncy. So my message is don't worry too much about the graph. Take some subs, note the settings you used to get them and compare to previous subs. You may find you're getting good images with a crazy looking graph.

Sorry if you've heard this all before but I hope it helps.

Mark

Edited September 30, 2011 by mark7331
added guide rate pair

[bamus]

Thanks Mark, any help/advise is more than welcome!

RA guiding is perfect. Calibration is done in about 15 steps and graph is within +/-1,5 arc sec, so there I have got no issue at all. I was doing yesterday 5 minutes subs and the one which have been in-between DEC graph peaks were perfect, other ones showed trails in Declination)

I am using EQMOD, so I will try to increase MAX DEC duration, is it done before calibration or after? (already have tried to increase DEC guide width to 150%, with little improvement). What about EQMOD backlash compensation. How can I measure how many ms I can use (at the moment it is on “0”)

[mark7331]

Bamus, glad to help mate

The EQMOD pulse guide rate should be set before you calibrate and if you need to adjust it again you should re-calibrate PHD because it's a major change to the responsiveness of the mount.

I'm not sure you're going to need to set the backlash value in EQMOD. Once your calibration is reasonable in DEC it may not be an issue any more (mine is also "0") but maybe that's because I have different mount? So I can't find a direct way of measuring the backlash but I wondered could you use the EQASCOM trace feature to work it out? Right click on the ASCOM logo in EQMOD to see the trace. If you watch/record the log while PHD is clearing backlash it might give you an idea of the number of ms you'll need. I checked the PHD debug log and I couldn't see any backlash related data.

Just to be thorough I have the EQMOD minumum pulse width set to 20ms. My RA guide rate is x0.30 and DEC is also x0.30.

I'm really interested to see if making these changes takes out your DEC bumps (as it should do) as it would be great to have a clear procedure for getting rid of these kind of guiding issues.

Keep us posted.

Mark

[bamus]

Many thanks Mark! I went to bed very angry and frustrated last night asking myself why do I bother wasting my time, but today I am bit calmer and will try to solve this problem again.

I will keep you updated with my progress

[chrisshillito]

I started by working out a calibration step value that produces between 10 and 20 steps when calibrating. Leave everything else as default until you can produce this. If you're using EQMOD you can change the RA & DEC Pulse Guide rate to help reach the calibration "zone" you're looking for. If you're using the ST4 (Autoguider) port then adjust the guide rate to help - you may need to do this with your handset.

You really shouldn't be changing the guide rate simply to achieve a certain "calibration" result. The purpose of calibration is for PHD to learn about the mounts response, not to change it. If the mount isn't moving far enough for calibration to succeed then change the calibration step. I don't see why worrying about the number of calibration steps matters. Because PHD only works by measuring pixels movement it is inevitable that different declinations will produce varying amounts of RA pixel movement during calibration and consequently more/less steps may be required to move a given angular distance. This is quite normal and doesn't require corrective action.

My advice in setting EQMOD pulse guide rates would be to keep them below 0.5 x sidereal. Using a fast rate will simply result in very short guide pulses which are less likely to be implemented accurately.

Chris.