Weird PHD guiding graph - any ideas ?

[glowingturnip]

I had the first few proper sessions with my new AZ EQ6 GT Skywatcher mount last weekend and was getting this guiding graph:

The Dec graph looks ok, but the RA graph is weird - it is below the axis and all the adjustments in RA seem to be in the same direction.  If I stopped guiding, selected a new star and started guiding again, then the RA graph would immediately drop that inch or so again and carry on below the axis.

The grouping in the little bullseye scatter graph looks nice, and the resultant stars look decent enough, but what worries me most is that if PHD is always adjusting RA in the same direction to stay on target then that implies to me that the new mount is not tracking at exactly sidereal rate and that PHD is having to do the heavy lifting to keep it on target.

If that is the case, I suppose I could train the PPEC when I've got a better PA and use the Sidereal+PEC to stay on target better and then guide from that, but that seems beside the point if I really do have an expensive mount that doesn't track properly.

So it's an AZEQ6GT, PA wasn't perfect, was out by about (1',10'), connected via ASCOM through the Synscan handset and RS232/USB converter, Celestron ASCOM Driver 5.0.23, QHY5liic as guide camera, 36 degrees latitude and -13 declination.  PHD calibration was uneventful and nicely orthogonal, the star was almost back to where it started.

Has anyone seen anything like this before, and have any ideas ?  Really hoping I don't have a duff mount... 

Cheers,

Edited July 17, 2015 by glowingturnip

[msh1]

i have the same mount i couldnt get phd2 to work same sort of graph ,went back to phd1 works perfectly, i can go 30 min subs 

mark

[rwg]

What sort of trace do you get if you turn the guiding output off? Does the RA drift off rapidly in that case? (that would tie in with your suggestion of tracking at the wrong rate and PHD working hard to try to fix the problem). If it doesn't drift when guiding is switched off then I'd almost suspect that the mount was ignoring the guide pulses, excep then how could it calibrate successfully.

Try changing things around - for instance use the ST4 cable for guiding (and if you are ST4 guiding already, try pulse guiding via ASCOM). Possibly try EQMOD (getting an EQDirect cable).

cheers,

Robin

[Zakalwe]

You have the DEC guiding output set the North. Change that to Auto and see what difference it makes.
Setting it to North (or South) means that PHD can only issue guide commands in one direction.

[glowingturnip]

@Robin - yes, you're right, more experimentation needed (I just wanted to carry on imaging at the time, since I was getting decent enough data anyway).   I'll certainly see what the guide trace looks like with the guiding output turned off, I hadn't tried that but v good idea.  I haven't visually noticed any drifting when unguided.  Was also going to give it a go with ST4. 

I need to reduce the calibration step size to 400 from 800 actually (I changed it from x0.5 guide speed to x1.0 guide speed but forgot to  change the calibration steps, so this calibration finished in about 5 steps each in W and E.  I definitely could see the guidestar moving in all four directions when it was calibrating though and in equal-ish amounts, returning back to where it started.

It's got me thinking though - my previous run had been with a guide speed set on the mount of 0.5x which gave me a strange calibration - the moves in W were very small and it took about 50 of those but the moves back E were much bigger so it completely overshot.  The N and S moves were normal.  The resulting guide graph looked the same.  Once I changed to 1x guide speed though, it calibrated normally as I said, so I assumed that had fixed it.  However, it's got me wondering if the same kind of thing is still happening in the background when it's guiding, if the pulseguides in one direction are making the mount move in a much smaller amount than in the other direction meaning it has to play permanent catch-up.  Maybe that Celestron ASCOM driver isn't up to the new mount, and I should look at ST4...

@Zakalwe - nah, it's not that.  The Dec guiding is ok.  The polar align wasn't perfect but not far off, so I was seeing a small drift being corrected for by the dec guiding in the North direction, but a spike every time it crossed the axes and tried to correct back South, so i turned the South off in this occasion.  Maybe I should look at the other dec algorithms though rather than just turn one direction off, especially when using dither could leave the guidestar on the wrong side.  After all, this mount should I guess be backlash-free with its belt drives (?).  The problem is with the RA though.

[michael8554]

Your RA Osc figure of 0.01 is very low.

I don't have the PHD2 instructions to hand, but I seem to remember it should be about 0.5.

Above or below that required an adjustment to one of the settings, sorry can't remember which.

Michael

[Pompey Monkey]

i have the same mount i couldnt get phd2 to work same sort of graph ,went back to phd1 works perfectly, i can go 30 min subs 

mark

I set up my new AZ-EQ6 GT on Friday night with PHD 2.5 and it worked absolutely fine.

.....Until I started experimenting with the dither function in APT. But that's another story. LOL

[glowingturnip]

Your RA Osc figure of 0.01 is very low.

I don't have the PHD2 instructions to hand, but I seem to remember it should be about 0.5.

Above or below that required an adjustment to one of the settings, sorry can't remember which.

Michael

yes, I noticed that, but then I think that's because it's not crossing the axis at all, but is always trying to pull back towards it, so not oscillating around the axis.

I need to do some more experimenting - at the moment, I'm thinking that connecting ST4 should see me right.

Pompey, how are you connecting your pc to scope, if u don't mind me asking ?

[Pompey Monkey]

yes, I noticed that, but then I think that's because it's not crossing the axis at all, but is always trying to pull back towards it, so not oscillating around the axis.

I need to do some more experimenting - at the moment, I'm thinking that connecting ST4 should see me right.

Pompey, how are you connecting your pc to scope, if u don't mind me asking ?

I go via the SynScan handset. I'm too much of a skinflint to buy the HiTech cable

I don't use the ST4 port on my QHY5, I just let PHD do the work on the laptop.

Interestingly, I am using a bog-standard Cat 5e cable from the handset to the mount. I can confirm that this method works for at least a 20 m length.

[glowingturnip]

are you using EQMOD, or an Ascom driver ?  If a driver, do you know which one ?

[Pompey Monkey]

are you using EQMOD, or an Ascom driver ?  If a driver, do you know which one ?

EQMOD and ASCOM

I just followed the instructions on the EQMOD page. I use the same setup for my HEQ5. The only driver issue that I recall was associated with my QHY5.

[glowingturnip]

cheers mate

[tony210]

Hi-below axis RA might represent a mechanical hitch with the mount - I certainly had this with a cabling prob- it is always very frustrating if RA drifts down - best wishes Tony

[michael8554]

I watched this youtube today by a PHD2 developer:

Its 90 mins long but from it I gleaned:

Try Guiding Assistant in Tools menu. This turns off the drives, but PHD2 carries on showing star movements, so will show you what the RA drive is doing, and Trendline will show how good your PA is.

Guiding Assistant offers settings for Min Mo after each run, but I found it would alternate between the new and old settings, so not helpful in my case.

RA above or below the line - this is due to under-correcting, it's not keeping up with errors.  Shorter exposure time, more Agression, less Hys, less Min Mo,  are the settings to play with.

Balance slightly heavy so that the backlash is taken up.

ASCOM should be better than ST4.

Michael

[Zakalwe]

I go via the SynScan handset. I'm too much of a skinflint to buy the HiTech cable

Doing this can introduce signalling delays. EQMOD is designed to work with a direct connection to the mount, not through the handset.

When troubleshooting always try and remove as many variables as possible....perhaps the 20m cable and handset connection are part of the problem?

[glowingturnip]

I watched this youtube today by a PHD2 developer:

[ ]

Its 90 mins long but from it I gleaned:

90 minutes !  yikes, tl;dw  ;-)

thanks very much for the pointers though, I will certainly try them all and report back.  I should be getting some more scope time in a couple of weeks, so let's see

[Zakalwe]

Look at post 4 and 6 in this thread:

http://stargazerslounge.com/topic/219759-eqmod-through-eqdir-or-handset-pc-direct/#entry2361717

[glowingturnip]

I've not actually used eqmod really yet, may have to look into it in more detail if I can't resolve my problems another way

[Pompey Monkey]

Doing this can introduce signalling delays. EQMOD is designed to work with a direct connection to the mount, not through the handset.

When troubleshooting always try and remove as many variables as possible....perhaps the 20m cable and handset connection are part of the problem?

I'm not the one having the problem!

[glowingturnip]

sshh, I didn't like to mention that  ;-)

[Astronome]

I have this problem also... The guide rate of RA is 1/3 that of DEC even though they are set to 0.9 on both in EQMOD?

Something wrong.

John