PHD2 Drift Alignment & Calibration - end of my tether now

[Notty]

You gurus out there please help, I'm getting ready to throw in the towel. I recently had some issues getting PHD2 to calibrate properly, there would always be an error message reporting questionabke results, or RA differing from DEC by an unexpected amount etc.

I've had my HEQ5 belt modded and professionally taken down relubed and tuned with a clean bill of health. I suspected my PA may have been the cause of my distressing guiding graphs previously so I've been trying the PHD2s inbuilt drift alignment. I initially PA using the automated routine in EQMOD so I know it HAS to be in the right ballpark. I find a star close to the meridian and CE as specified. I attach a shot of an example calibration error

[Notty]

Sorry it posted before I'd finished, here's one of the errors, and yes I agree RA and DEC should be the same at the equator!

During drifting itself I can't understand what is going on, the tutorials have nice steady straight lines sloping up or down but look at the volatility of my corrections what on earth can I do with this!

When I run it for Altitude it's just as crazy

I'm convinced my sawtooth graphs are something to do with the weird calibration errors, but I'm completely stuck now. I even swapped scoped (200pds off 80ed on) more than halving the weight to see if that was an issue but it had no effect. There was zero wind too. Additionally if using the same star (when it let me) the drift curve would sometimes change from positive to negative even without me adjusting anything! I can't see any snagging wires or other whoopsies.

Please please can someone tell me why this is happening? I'm getting seriously demoralised now.

Many thanks

[Notty]

...and just to add, before the calibration error about the RA and DEC amounts during calibration it did its west and east steps but just stopped after that ie none of the normal north South steps?!

[Owmuchonomy]

Your SNR is much too low. Could be a number of reasons for this. Your exposure duration is also low. I would put this at 3 or 4 seconds. Use Alt S to auto pick a star and make sure it is of SNR 30 or more. Then you should calibrate and then finally run the drift alignment routine. Come back to us if you are still stuck.

[Zakalwe]

Agreed. A 1 second loop is far too short and you are "chasing the seeing". I use 2.5 or 3 seconds which seems to work for me.

[Notty]

Thanks Chris but I don't think I've ever seen a star through phd with a SNR of 30! The tutorial I was using that everyone recommended suggested 1 second exposures on a SNR target of 10 and I was sometimes struggling to get that. The best I got was 12 or so. Do I therefore assume it's better to wander as far away as necessary from the CE/meridian to get the brighter star in favour of the mathematically more correct position?

I've also just remembered that I still had PHD2 set to the focal length of my finferguider (focal length 180mm) but last night I was using the main scope (200PDS focal length 1000mm doh!) How much would that have thrown everything out? That's the trouble with a hobby I can attempt for a couple of hours at a time once every 3 weeks aaaarghh....

[glowingturnip]

your Dec guiding is turned off (at bottom-right) which is probably the cause of several of the errors - see how in the first screen grab it took 13 RA steps but 5 million Dec steps ?  You'll need the Dec guiding turned on to do the polar align routine and it's probably confused itself with the Dec off.

Also in the last picture, where you've got really low/no SNR and it's auto-stretched the screen to try to see anything, you've got a grid showing, looks like a bayer grid.  What guide camera are you using, and are you sure you are using the correct choice for it in the camera chooser dropdown thingy ?  Wrong choices in PHD can often show a bayer grid for a colour guidecam

[glowingturnip]

Thanks Chris but I don't think I've ever seen a star through phd with a SNR of 30! The tutorial I was using that everyone recommended suggested 1 second exposures on a SNR target of 10 and I was sometimes struggling to get that. The best I got was 12 or so. Do I therefore assume it's better to wander as far away as necessary from the CE/meridian to get the brighter star in favour of the mathematically more correct position?

I've also just remembered that I still had PHD2 set to the focal length of my finferguider (focal length 180mm) but last night I was using the main scope (200PDS focal length 1000mm doh!) How much would that have thrown everything out? That's the trouble with a hobby I can attempt for a couple of hours at a time once every 3 weeks aaaarghh....

the focal length setting in PHD won't affect anything.  It changes the suggested calibration steps that it gives you, and also the arcsecond scale on the guide-graph, but it doesn't do anything to affect guiding itself. 

Theoretically you can drift-align on any star, but since you've got two things to adjust (alt and az) it's difficult to know which one to adjust.  The reason for starting with a star near to meridian and CE is that that isolates any drift to being errors in az only.  The second star over towards the E/W horizon could drift due to errors in both alt and az, but it's assumed you got the az sorted already in the first step, so you only need to adjust it for alt.  For finding a star to drift-align on, close enough should be good enough, I don't think you need to be exact.

[Notty]

your Dec guiding is turned off (at bottom-right) which is probably the cause of several of the errors - see how in the first screen grab it took 13 RA steps but 5 million Dec steps ?  You'll need the Dec guiding turned on to do the polar align routine and it's probably confused itself with the Dec off.

 

Thanks Chris, but are you sure about that? The tutorial says that the DEC guiding is automatically switched off. Though I'd love for the answer to be something as simple as that!

[glowingturnip]

Pretty sure, yeah, no mention about PHD turning the Dec guiding off for drift align in the manual - http://openphdguiding.org/man/Tools.htm#Drift_Align

When it's doing the drift align, it's doing its usual thing of staying on the guide star by sending adjustments in RA and Dec, but it also produces the drift lines.

It's also why, of course, it didn't step through the usual N and S calibration steps for you as you mentioned in another post.

(Don't call me Chris, that's my dad's name !  Made me feel quite strange reading that )

[Notty]

Sorry about that Stuart I don't know where I got Chris from, my apology.

Funnily enough when I calibrated the first time last night it DID move in all the cardinal directions! but still gave me a warning about "questionable" results. I messed about with the max dec and RA values but definitely didn't switch anything off.

I was trying to use the "bookmark" method contained in the following guide:

http://www.cloudynights.com/topic/472874-drift-alignment-with-phd2-the-bookmark-technique/

It's on Part 3 to the guide he says that PHD automatically disables DEC guiding when you click drift. It would be good to know if this is correct or not, as I said I didn't turn it off!

This is a dark, dark art... If anyone's feeling generous and would care to share their PHD settings it would be really interesting to me.

Thanks again.

[glowingturnip]

hmm, it does say that doesn't it, and his graphs don't show any dec adjustments being sent.  I'm pretty sure last time I did it it was still guiding in Dec though - one way of finding out I suppose, try it again !  Maybe PHD has been tweaked since he wrote his tutorial.

The 'questionable results' dialogue boxes are an annoying feature of recent PHD releases, and I think they've got the nag level turned up a bit too high.  It's highlighting the fact that its measured RA and Dec are not sufficiently orthogonal to each other.  If you turn on the RA/Dec overlay in the view menu you can see what it thinks they look like - it will never be an exactly square grid even though theoretically it should be, but should be square enough and not too 'diagonal-ey'.  Sometimes PHD will turn off Dec guiding itself if it couldn't calibrate it, maybe that happened ?

[SteveA]

Yep....if you are running the drift alignment routine in PHD the dec drive is definatly turned off, it has to do this to determine the drift. I would agree with the others though....go for a longer exposure, I usually use between 2 & 4 seconds.

[Leveye]

There was a new version of ascom and PHD2 and a driver for my mount so personally I was using the latest for everything. I had to use auto exposure last time. Luckily the stars available were all way over 10. most 30-40 snr.I drift aligned a few days ago. I used over 1000 in RA and DEC max duration. Dec guide mode is in auto. History is X400 and Y: =/- 16. You have to let things settle quite a bit so for Az at least 30 seconds and in alt I let it go for at least 1 minute.

Be patient It eventually settles down and gives a reading. Sometimes I did get a star lost message but it picked it right back up and the magenta circle would do it's thing again.The DEC line will go up and down. Keep watching the error numbers at the lower left bottom of the graph screen. Calibration is key though you must do that first and be in the green guiding first before drift aligning. Your very close to success.

After I was done I went from being able to track something centered for only about a minute or so to being able to track Jupiter dead centered for over 2 hours until it eventually hits the trees and I have to stop if you can believe it. I'm a true believer in DA now.

[Notty]

Leveye, thank you that's all really useful. When you say "in the green guiding" can you just explain if that means some specific value in PHD?

[Astrosurf]

A simple statement here, and probably useless. Have you tried PHD1? I only ever use that and it's pretty much glitch-free. I'd never try PHD2 as it seems to cause problems for quite a lot of people!

Alexxx

[Notty]

No, but I'm desperate enough to try anything! Thanks for the tip I hope it's still available.

[Astrosurf]

Try here, if not, give me a shout and I'll see if you can upload from a disc if I send you one.

http://www.stark-labs.com/downloads-2/downloads.html

[salas]

Sorry if its been covered / ive missed it, but have you tried increasing the input pulse timing / duration in eq mod?

Steve

[salas]

Sorry if its been covered / ive missed it, but have you tried increasing the input pulse timing / duration in eq mod?

Steve

[Leveye]

Leveye, thank you that's all really useful. When you say "in the green guiding" can you just explain if that means some specific value in PHD?

No value. Just that the calibration is good and is working and you have a green crosshair. That PHD is guiding properly. Then you can move onto DA. Being connected to the mount using ascom really helps knowing where your pointed. Highly suggested.

Also I don't think reverting to PHD1 will help. I'm not sure the drift align tool was even part of the program back then.

[Thalestris24]

Just to reiterate - DEC guiding is definitely off when you drift align! It wouldn't work otherwise. You don't have to be that close to the meridian - they say within 1 hour of RA for best results. A decent exposure time is essential as is a decent snr. Make sure your balance is good before doing anything! Also make sure you enter your guide cam focal length and guide cam details correctly. Once you've got a good calibration you should be able to reuse that calibration in the future, if nothing changes.

Good luck!

Louise

[strutsinaction]

I found this article useful when starting out with PHD2 and drift alignment:

http://www.cloudynights.com/topic/472874-drift-alignment-with-phd2-the-bookmark-technique/

Regards

John

[glowingturnip]

this was bugging me, so I wanted to check before coming back. 

I stand corrected, even if you calibrate with auto dec guiding on (in simulation) then when you turn drift-align on, it does indeed turn the dec guiding off temporarily.  I think theoretically it shouldn't have to, since it could still integrate for the error term even while it corrects for it, but it does indeed turn it off.

Mind you...  if you calibrate with dec turned off, then theoretically PHD doesn't know which way dec is, unless it makes the assumption that it is orthogonal to RA (and there is a button for that, but I don't think I'd trust it).  So if it doesn't know which way Dec is, then the error it is showing is an arbitrary tracking error, not necessarily a dec tracking error.  Then again, even minimising an arbitrary error is going to help with correcting PA, since Dec error is directional but RA error is periodic.  Hmmm...

Anyway, for the OP, I would say definitely make sure that you calibrate with Dec guiding on, with calibration steps as recommended or close to those that PHD recommends, and make sure that you have a good calibration with sufficiently orthogonal RA/Dec before you try drift-aligning.  Let us know how you get on !

Cheers

[Thalestris24]

Yeah, calibrate with Dec guiding on but drift align is done with guiding off.

Louise