PHD Losing the Guide Star

[JSeaman]

Last night I had a bit of an issue, after an hour or so, PHD seems to have lost the guide star and fired off for about 20 minutes in a given direction. Does anyone have any experience of this? Extract from the log is attached

 

[kens]

Can you attach the full guide log? Could be something simple like a cloud passing over.

[JSeaman]

Sure thing, see attached. It happened again last night although slightly different (about half twelve), from the pictures last night it looks like it may have been cloud but I don't know how you tell for sure?

PHD2_GuideLog_2018-04-20_211857.txt

PHD2_GuideLog_2018-04-19_213844.txt

[kens]

The one from last night does look like cloud as the star SNR  drops off gradually till you get a star lost.

The previous one looks different in that it drops off very suddenly.

However last night you started off with an SNR of 13 to 14 which dropped off to 8 then suddenly dropped to 4. The night before you started with an SNR of just 7 before it suddenly dropped to 0. Also you were using exposures of 1.5s last night vs 1s the previous night. Maybe you were just underexposing. Try guiding with 2s exposures - even SNR 13 is on the low side. You could also try increasing the camera gain. 

After losing the star the night before last, 15 or 16 minutes later PHD2 locked on again with an SNR of 10. In that 15 minutes, assuming it picked up the same star, it drifted or jumped about 15 pixels in declination. If you have a sub (or subs) taken at that time (11:17pm to 11:35pm) you should be able to see just what happened.

[JSeaman]

Thanks very much for taking the time to look at that. I am glad we concur on last night, simply clouds looks to be the answer.

You've also pointed me in the right direction, I had noticed that the stars were very dim in PHD2 and didn't really think much more of it, cranked up the exposure a little to compensate (even up to 3 s at one point) but it had little effect. I haven't tweaked the gain before but just found the setting in PHD (not in the camera settings) and have gone from 12% to 20%, will try that out when we get another clear night - probably Winter some time! 

Thanks again for the help

[newbie alert]

I'd  suggest cranking up the gain or exposure to get a higher snr on the guidestar

[MarsG76]

Even if a cloud passed over the guide star, the mount should still be tracking of it own accord and surely once it went away PHD would resume guiding. 

[JSeaman]

True as long as it didn't drift much during that time where it missed the star.

The bit I'm struggling with is that the mount appears to have been driving one of the axes because I got drastic star trails in the window where it wasn't guiding. 

[kens]

Could be drift, flexure, cable drag.....

Can you post the subs that cover the period where it wasn't guiding? Jpegs would be ok

[JSeaman]

Just checked but don't seem to have them unfortunately, it was a star trail indicative of a huge polar alignment error or holding the RA/DEC motor on

[kens]

My guess would be a cable dragging or snagging

[MarsG76]

21 hours ago, JSeaman said:

True as long as it didn't drift much during that time where it missed the star.

The bit I'm struggling with is that the mount appears to have been driving one of the axes because I got drastic star trails in the window where it wasn't guiding. 

How accurate is your PA?

I increased my guidestar region to 50 pixels so that even if clouds did come and the star drifted it had a chance to be within the box and would be guided back into place. Another thing I do is lower the gamma so that when the clouds come keeps the background black, noise free black and PHD doesn't lock in on some noise, "guiding" the mount in random directions at full guide pulse lengths.

I've had situations where I was imaging, everything was flowing along swimmingly and I went to sleep, leaving the system to do its thing.... during my slumber clouds came for around half a hour and odcourse, the subs during that time were a lost cause, but when the clouds went away, I had one sub with streaks (PHD was moving the star back into place) and the rest until the sun came up were perfect....

What I'm trying to say is that if your PA is close to perfectly aligned, that even if you lose a guide star due to clouds, the system can track well enough (even with the PE) and resume autoguiding when the coat is clear.

 

[JSeaman]

All the cables are loomed up but there are a couple which could have the potential to snag I guess so I'll start there - thanks for the input again

 

Polar alignment seems pretty good, it's on a fixed pier and I check it regularly. I have seen it deal with clouds well a few times previously (as per your description) so this is unusual. When you say gamma, do you mean the slider next to the brain etc? I didn't think that affected the guiding at all, just the GUI?

[JSeaman]

Sorry to drag this back up but we had another clear night last night so I got to try out the gain fix. On the plus side, with the gain as high as 40% I was getting a better contrast and it guided all night. The downside is I just looked at the guide log and, despite it being quite flat most of the time, I have another anomaly at 23:20 ish, any thoughts?

 

PHD2_DebugLog_2018-05-02_215427.txt

[kens]

You need to attach the Guide log. Thats the debug log.

[JSeaman]

Sorry! Attached...

PHD2_GuideLog_2018-05-02_215427.txt

[kens]

Either a cable snag or the guide scope or camera shifted in its mounting 

[JSeaman]

Thanks, really appreciate the feedback. How did you reach that conclusion out of interest?

I checked over all the cables when I saw the log and don't think could get snagged so will have a play with the mounting of the guide scope later

[kens]

Its always a guess but some of the factors:

• It isn't periodic but mostly in RA so unlikely to be gearing related 
• The jumps are mostly short and sharp. And they do not persist. Most of the time the guiding looks pretty good. So its not a balance problem
• The biggest jumps are to the west, requiring east guiding pulses. This counts a little against cable snags which you would expect to pull the mount east. But there are several smaller eastwards jumps
• Most issues like this are cable drag or flexure
• A lot of the spikes after the first couple of hours could well be PE related. So the big spikes presumably occurred nearer the horizon. That might give you a staring point where to look. 

Unfortunately the log does not show the declination nor hour angle. But I'm guessing dec was around 65 deg. And a session of almost 6 hours suggests you started near the horizon and stopped before a meridian flip. 

 

[JSeaman]

Ok thanks again for the information, I'll have another look around cables. This was imaging Bodes so about 69 degrees dec