Dithering in SGP requires you to set a Settling Option. What exactly is it monitoring to confirm that it has settled? When I watch the PHD graphs on my A-P Mach1GTO mount it seems to start the next image once the RA and Dec have converged on the X axis with PHD reporting the settling distance as less than the threshold.
My question is, what does it do with a mount that only guides in RA? My portable setup is an AstroTrac which doesn’t have an Dec guiding. How does SGP/PHD tell that it is settled with a mount that only guides in RA? When I look at the PHD graph, there is some drift in Dec over multiple exposures because polar alignment isn’t perfect so if PHD is looking for convergence of RA and Dec on the X-axis then at some point drift will exceed the settling distance. Will it reach a point where the convergence between Dec and RA will always exceed the settling distance, and thus never settle? Or will SGP ignore Dec altogether (Dec guiding is turned off in PHD) and just look for RA to settle on the X axis?
…Keith
Keith,
PHD2 makes the determination of settling done and notifies SGP. PHD2 uses the criteria passed to it by SGP to determine when settling is complete.
Settling is done when the guide error stays below the tolerance setting for the specified time interval. The guide error is the distance of the guide star from the lock position, regardless of direction. So yes, you are right, it would be possible for the Dec drift to prevent settling when the dec error exceeds the settling distance.
We can look at changing the settling criteria in PHD2 for the case of RA-only guiding. I’ll discuss it with the other developers to make sure we’re not introducing any problems by making that change.
Andy
Ah, very interesting. Thanks Andy. I first was alerted to this possibly being a problem when I left my AstroTrac running unattended and SGP was shut down when I returned. The only thing I could see from the PHD graph that might have caused it was that the Dec line had drifted quite a bit over the hour or so it was running (not a problem over the short duration of the light exposure, but over multiple exposures if you don’t have a dead accurate polar alignment there is bound to be Dec drift).
Thanks for bringing this to the attention of the developers group, as it doesn’t really make sense to allow Dec position to affect settling on a mount that doesn’t guide in Dec. It would seem reasonable that when Dec guiding is turned off in PHD to just use RA deviation for settling.
As is I’m not sure what to do, as getting my polar alignment so perfect that there is no drift in Dec over the 2h run of the AstroTrac would be pretty daunting. I don’t seem to have a choice in SGP to turn off settling.
I’m leaving on a trip on Monday to California for a week and I’m planning to take my portable imaging setup. I’m worried that this issue is going to cut short my imaging. Can you think of any work around that would bypass this problem, Andy?
You could set the tolerance value very high, like 100 px, and set the settle time to a conservative value like 20 seconds. That will cause PHD2 to settle for 20 seconds and then notify SGP that settling is done. The distance will effectively be ignored since it will always be less than 100 px.
Tim seems to be right Andy. 2.5 pixels doesn’t give me a lot of leeway to counter polar alignment Dec drift over a 2 hour run. Might be enough though.
…Keith
That limitation almost certainly is an artifact of PHD1 which could only report distances up to 2.5 pixels. These days with PHD2 there is no such limitation, and it would be great if Jared or Ken could remove the 2.5 pixel limit in SGP.
In the meantime I think your options are:
disable dithering, or
manually stop & start guiding after each dither (or periodically) to keep the lock position close to the guide star. (Make sure the Sticky Lock Position option is not enabled)
We will be making the change in PHD2 to ignore the Dec distance for settling when Dec guiding is disabled. However, we are about to release 2.6.2 and do not want to make this change until after 2.6.2 is released.
Thanks Andy. I’m trying it tonight and have gotten such a good alignment with my QHY PoleMaster that drift isn’t much of a problem. If I can replicate this then 2.5 pixels will be sufficient until you release the PHD no-dec-guiding fix.
…Keith
I would definitely support this. With the Edge 14 and the QHY9, I’m at 0.4"/px with the reducer. 2.5px is barely enough - only 1" of total error, and it can take quite a while to meet that if seeing is bad. I can’t use the settling routine at all without the reducer.
It’s worse than I thought. Last night I tried again and this time didn’t readjust my polar alignment so my Dec was off by about 2.8 pixels per 2min exposure - more than the 2.5 pixel max settling setting. I thought that if I turned dither OFF by unchecking the box that I wouldn’t have settling problems. However, SGP still tried to settle before starting each frame.
Am I correct that SGP will try to settle the autoguider when using PHD2 regardless of whether dithering is on or off? That’s what it was doing with my setup. Which seems odd since what is it settling from if there is no dither?
Anyway, this restricted me to taking 1 light frame at a time as I had to pause SGP, stop and restart PHD2, then unpause and take the next frame. This was an extreme example of poor polar alignment but I was just trying things out for my California trip - normally I would have a better PA. Still it means that settling seems to occur regardless of whether dithering is turned on or off. Is there any way to turn off settling?
It doesn’t make sense that settling was happening with dither off. AFAIK
there’s no way to turn off settling, but you could try going with the max
error 2.5 with a settle time of 0 sec. I think that means that any time
the error gets at 2.5px or below, even for an instant, the sequence should
resume. But again, I too would have thought that if dithering is disabled
there wouldn’t be a settle time at all.
I just re-read the whole thread…unless I’m mistaken (could be!) I believe the settle TIME means the guide star has to stay below the error for that whole time before settling will be considered complete. So settling at 2.5px for 20 seconds would mean that the error has to stay BELOW 2.5px for all of 20sec. That’s well out of the range of what Keith needs to do.
Of course getting rid of that 2.5px limitation would solve the issue completely.
It shouldn’t try to settle if you have no dither set and for me at least, I have never seen that problem. Are you sure that you are turning dither off? If you change it in the equipment profile while you have the profile loaded, the change won’t go into effect until the profile is loaded again. A log might be helpful. However, in the meantime, there is an easy fix. If you don’t intend to dither, then just don’t connect to PHD2 in SGP. Select “None” for autoguider and start PHD2 separately. It will happily guide away while SP snaps frames.
You could be right, but I did restart SGP and I was still getting settling even though the dither checkbox was unchecked after opening, but maybe SGP was still using the setting in the profile - I didn’t Save the profile with dither off.
But of course, using PHD independently of SGP makes sense. No need to have SGP connected if I’m not dithering! If my polar alignment isn’t perfect, I have a form of dithering anyway due to my polar alignment drift.
At that point the profile is irrelevant. If you had dithering disabled in the control panel that’s the setting the sequence was using, not some profile setting.
Hmm, well I couldn’t get it to stop settling before it started a frame even when the dithering checkbox was unchecked in the control panel.
Anyway, if my PA proves to be too inaccurate for a 2.5 pixel deviation in Dec, then I’ll just run it with no autoguider in SGP and run PHD2 separately.
Yes, I know. That is very strange to me. And yes, an inelegant but simple
solution is to just guide with PHD outside of SGP altogether.
I was reponding to the earlier comment about the profile setting maybe
having an influence on the current sequence you were running. It does not
(or at least it shouldn’t).
“We will be making the change in PHD2 to ignore the Dec distance for settling when Dec guiding is disabled. However, we are about to release 2.6.2 and do not want to make this change until after 2.6.2 is released.”
I’m bumping this post from 6 months ago to ask if this change was made now that 2.6.3 is released? Will PHD2 now ignore Dec distance for settling when Dec guiding is disabled (e.g. on mounts that can guide only in RA such as my AstroTrac)?