Ron,

Do you have checked and with value?
Only start if guide deviation <
and
Abort if guide deviation >
?

There seems to be a good amount of pixel change between the loss and subsequent regain


[2022-03-01T20:36:14.006 EST DEBG ][     org.kde.kstars.ekos.guide] - "findGuide   #   x      y      flux    HFR  SNR   score"
[2022-03-01T20:36:14.006 EST DEBG ][     org.kde.kstars.ekos.guide] - "Star        0  889.5  311.6  16126   4.09  89.8 0"
Loss
[2022-03-01T20:36:18.228 EST DEBG ][     org.kde.kstars.ekos.guide] - "findGuide   #   x      y      flux    HFR  SNR   score"
...
Regain
[2022-03-01T20:36:39.313 EST DEBG ][     org.kde.kstars.ekos.guide] - "Select      #   x      y      flux    HFR  SNR   score"
[2022-03-01T20:36:39.313 EST DEBG ][     org.kde.kstars.ekos.guide] - "Star        0  887.9  295.8  15530   4.30  88.1   215"

Another
[2022-03-01T20:37:05.196 EST DEBG ][     org.kde.kstars.ekos.guide] - "findGuide   #   x      y      flux    HFR  SNR   score"
[2022-03-01T20:37:05.196 EST DEBG ][     org.kde.kstars.ekos.guide] - "Star        0  885.6  290.1  15177   4.21  87.1 0"

Loss
[2022-03-01T20:37:09.665 EST DEBG ][     org.kde.kstars.ekos.guide] - "findGuide   #   x      y      flux    HFR  SNR   score"
[2022-03-01T20:37:14.169 EST DEBG ][     org.kde.kstars.ekos.guide] - "findGuide   #   x      y      flux    HFR  SNR   score"
[2022-03-01T20:37:18.659 EST DEBG ][     org.kde.kstars.ekos.guide] - "findGuide   #   x      y      flux    HFR  SNR   score"
...
Then Regain
[2022-03-01T20:37:31.988 EST DEBG ][     org.kde.kstars.ekos.guide] - "Select      #   x      y      flux    HFR  SNR   score"
[2022-03-01T20:37:31.988 EST DEBG ][     org.kde.kstars.ekos.guide] - "Star        0  885.5  299.9  16444   3.97  90.7   215"

Gene

Gene,

Yes, I use the default value of 2.0 to start and 3.0 to abort.

I'm not sure how to read the log, but when the SNR drops to zero it often jerks the OTA. I can sometimes see star streaks in the images because of the movement. It's like it just goes crazy for a short bit, then tries to get back on track. I'm really baffled as to why the SNR drops out when using the internal guider, but never does when using PHD2.

Thanks,

Ron

Ron,

The guider-jerking-the-OTA-around bug was fixed a couple weeks back in the nightlies.
invent.kde.org/education/kstars/-/merge_requests/531
Try that out if you like.

The SNR thing is an indication of the guider having trouble with its main guide star. Here's a long description of that: indilib.org/forum/stellarmate/11284-snr-...al-guider.html#80596
By itself, having the SNR jump around isn't necessarily an issue, but it could be an indication of bad sky conditions.

Hy

Ron,

If you have it set to abort at 3 and the regain is showing a changed of 311 to 295 and again 290 to 299 in DEC, try upping the default 3 to some larger value. If the search box is only a 6 pixels diameter and DEC is moving fast between guide images, it is not finding the guide star in the search box (?) Now why DEC is moving fast, what is you plate scale, arcsec/pixel of your guider, PE, puff of wind???

Hy,

Okay on the OTA jerking.  I'm currently using stable.  I've tried using nightlies before, but have always had problems.  Perhaps I'll back this one up and give it another try.

On the SNR... The night was very clear and seeing was average.  I monitored the guide images and there was no difference between the images with high SNR and those with zero SNR.  It's not that the SNR is low... it is zero... which doesn't make sense.  See the attached screen capture of a portion of the "Analyze" screen.  It's like EKOS is just not "seeing" the image.  I'll read your attached link to see if it provides me some insight.

Edit: After reading your reference, I'm wondering if the SNR drop out is necessarily the problem. Perhaps it is all related to the OTA jerking issue.  I'm running an imaging session right now using PHD2, but will try to do a nightly update tomorrow and test it.  Thanks!

Thanks!

Ron

 

Gene,

I use 128 for the box size when using the internal guider. As I'm watching it, the star is staying in the search box.

I think Hy may have hit on the problem with the jerking. I'll have to see about getting on nightlies.

I use an OAG on my Celestron 8" HD Edge. With my ASI533MC Pro, I get a pixel scale of 0.42"/pixel. When I use PHD2, I usually get around 0.5 to 1.0 RMS arcsecs. I get the same using the internal guider... until the SNR starts to zero out.

Now maybe if I get the jerking to stop, perhaps the SNR bottoming won't matter? I wonder when the next stable will be released?

Thanks for your help,

Ron

Hi Ron,

Understood you have a 128 box but you are also telling it to abort on greater than 3 pixels.
The logs show no star found, not a low SNR from what I can tell, a star not found displayed as 0 SNR(?)

The logs do show the loss then the re-acquire 10 or more pixels away after loss, 21-22 seconds between loss and re-aquire.

Way back when I had an lx200 classic that every once and a while would accumulate whatever on the main RA gear, tracking through the 'whatever' would cause a jump that could not be PEC'ed out and would cause a guider to struggle.

www.astrogene1000.com/DIY/lx200_internals/index.htm

Gene

Gene,

I have EKOS set to abort capture on deviation of 3.0 arcseconds... not pixels. Which, in my case, would be around 7 pixels.

Based on your analysis, there seems to be a couple of things going on. First, I don't understand why it's losing the guide star. As I monitor the guide images and the star is clearly visible in the box. However once it "loses" the guide star, there is movement in the mount... which I think exasperates the guiding issue... because now the star falls outside the box (I think?). In my view, if it can't "see" the star... it shouldn't send any pulses. I'm hoping that Hy's fix of the mount jerking may help eliminate some of this problem. However, I still don't know why it's losing the guide star. Particularly since PHD2 rarely loses the guide star.

Last night I used PHD2 for guiding and it worked fine. I did have a cable snag (my bad)... and I can clearly see the results as a rising RMS curve, until it releases and gets back on track. This is not what is happening with the internal guider. It's more like an on/off switch. You can see it in my previous post's image of the analyze screen.

I don't think it's the mount, as this behavior doesn't happen with PHD2. Using PHD2 everything works fine... except for meridian flip. Although last night it worked great... go figure! Only change I made was not to use the periodic autofocus. hmmm.

So... my current plan is to figure out a way to get Hy's fix on my system and continue to use PHD2. Through experimentation, I'm beginning to sense that autofocus may be causing an issue with PHD2... but am not certain. In future releases I will try the internal guider again, as I think it is fundamentally better and cleaner.

Thanks for all your help!

Ron