Guide star on edge of field

375 views
Skip to first unread message

Joel Short

unread,
Jan 21, 2021, 10:43:42 AM1/21/21
to Open PHD Guiding
I have noticed that when PHD2 automatically chooses a guide star it is almost always on the very edge of the OAG field of view.  This hasn't been much of an issue for my refractors, but I now have a 12.5" iDK (f6.7 FL=2128mm) and there are far fewer stars to to choose from.   
Almost always, there are stars in the middle of the field that have a good profile for guiding but are never chosen by PHD2.  Any tips for adjusting how PHD2 automatically chooses a a guide star?

I'm not sure it helps in this case, but here is my last guide log:  https://openphdguiding.org/logs/dl/PHD2_logs_HLtj.zip

It may not be helpful because I was doing repeated autofocus runs to determine filter offsets.
joel

Dale Ghent

unread,
Jan 21, 2021, 11:14:58 AM1/21/21
to open-phd...@googlegroups.com


> On Jan 21, 2021, at 10:43, Joel Short <buckeyes...@gmail.com> wrote:
>
> I have noticed that when PHD2 automatically chooses a guide star it is almost always on the very edge of the OAG field of view. This hasn't been much of an issue for my refractors, but I now have a 12.5" iDK (f6.7 FL=2128mm) and there are far fewer stars to to choose from.
> Almost always, there are stars in the middle of the field that have a good profile for guiding but are never chosen by PHD2. Any tips for adjusting how PHD2 automatically chooses a a guide star?


In PHD2's event API, there's an optional argument to the find_star method where you can specify an ROI in terms of X,Y of the desired center of the ROI and the width and hight if the frame centered no that spot, but I don't think there is a user-accessible analog of that in PHD2's own options. It would seem that the only way to effect this is through the event API.

In NINA, we allow the user to set the guide frame ROI in terms of percentage of the full guide camera frame size, so the ROI is always centered on the actual center of the camera's sensor, but it still allows the user to shrink the size of the image and effectively force PHD to consider stars more towards the center of the sensor and blocks off any that are at the physical edge of it. The app which controls PHD2 over the API would need to support calling find_star this way (as well as getting the sensor's dimensions by calling get_camera_frame_size first).

A better solution would be for PHD2 to derive a list of candidate guide stars and the ones that are closer to the center of the frame carry a higher weight when PHD2 decides which one to use.

/dale

bw_msgboard

unread,
Jan 21, 2021, 11:23:53 AM1/21/21
to open-phd...@googlegroups.com

It doesn’t matter where the star is in the field – it won’t get lost just because it’s near the edge.  To get the best results from auto-find, you should do the following:

 

  1. Set the saturation level =  the largest ADU your camera can return
  2. Set the min-HFD level = the smallest usable star size based on experience or measurements with your system
  3. Use a dark frame that is reasonably current and is a reasonable representation of your current sensor performance

 

If you want us to explain a single auto-find result that you’re not happy with, you need to send a debug log file and a specific time when you got the result and why you’re not happy with it.  Poring through logs looking at single star-find results is time-consuming and tedious.  FWIW, I run a configuration much like yours in a remote observatory and I’ve never had the auto-find function fail to find an acceptable guide star and produce good guiding results – nearly 2 years of operation with an emphasis on imaging galaxies.

 

Bruce

 


--
You received this message because you are subscribed to the Google Groups "Open PHD Guiding" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-phd-guidi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/4a65093b-a8a7-4805-b81f-ccab5a8ca9e4n%40googlegroups.com.

bw_msgboard

unread,
Jan 21, 2021, 11:28:33 AM1/21/21
to open-phd...@googlegroups.com
We're moving toward using multi-star guiding as the default mode. The
candidate stars can be all over the field of view and their position doesn't
matter. What does matter is SNR and that's the property used for weighting.
I don't really see the point of worrying about the location in the field.
If the edge stars are affected by poor optical quality, that should be
reflected in their SNR values.

Bruce
--
You received this message because you are subscribed to the Google Groups
"Open PHD Guiding" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to open-phd-guidi...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/open-phd-guiding/8BCFD7DC-3D33-450F-97FB-8
88AF596D444%40elemental.org.

Brian Valente

unread,
Jan 21, 2021, 11:29:16 AM1/21/21
to Open PHD Guiding
Bruce just a quick clarifying question:

>>>Set the saturation level =  the largest ADU your camera can return

Do you set it to the absolute max, or slightly under that?

For 16-bit camera the max adu would be 65536



--
Brian 



Brian Valente

Richard Beck

unread,
Jan 21, 2021, 11:30:55 AM1/21/21
to Open PHD Guiding
Bruce,

What about "Minimum star SNR for AutoFind"?  I had some very low SNR stars selected and had lost stars.  Increasing this value seemed to help.

The stars that seemed the best have SNRs in the 30s and 40s.  I set this at 20 and seemed to have better results.  Is there documentation on this parameter that I must have missed?

Joel Short

unread,
Jan 21, 2021, 11:40:28 AM1/21/21
to Open PHD Guiding
Bruce, I was really just looking for some guidance, which you provided.  
I don't have empirical proof, but the edge stars seem to be more diffuse and the subsequent guiding seems to wobble more than stars that are more tight toward the middle of the FOV.  
1.  Somehow the saturation level was set at 255.  I have a 12bit camera so I changed that to 4096.
2.  I have the min-HFD level set to 1.3
3.  I use a dark library that is fresh. 
4.  I have been using multi-star guiding since it came out and I have seen a slight improvement.  

joel

bw_msgboard

unread,
Jan 21, 2021, 11:46:53 AM1/21/21
to open-phd...@googlegroups.com

The data being returned by your camera driver will be 16-bit – the pixel values are just scaled up from the internal 12-bit electronics.  So you should choose a value of 65000 or so.  That’s true for all these guide cameras – the interface is either 8-bit or 16-bit.  If you were running with 255, that’s probably your problem right there.

 

Bruce

 


From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of Joel Short


Sent: Thursday, January 21, 2021 8:40 AM
To: Open PHD Guiding

Subject: Re: [open-phd-guiding] Guide star on edge of field

 

Bruce, I was really just looking for some guidance, which you provided.  

I don't have empirical proof, but the edge stars seem to be more diffuse and the subsequent guiding seems to wobble more than stars that are more tight toward the middle of the FOV.  

1.  Somehow the saturation level was set at 255.  I have a 12bit camera so I changed that to 4096.

2.  I have the min-HFD level set to 1.3

3.  I use a dark library that is fresh. 

4.  I have been using multi-star guiding since it came out and I have seen a slight improvement.  

 

joel

On Thursday, January 21, 2021 at 10:23:53 AM UTC-6 bw_m...@  wrote:

It doesn’t matter where the star is in the field – it won’t get lost just because it’s near the edge.  To get the best results from auto-find, you should do the following:

 

1.      Set the saturation level =  the largest ADU your camera can return

2.      Set the min-HFD level = the smallest usable star size based on experience or measurements with your system

3.      Use a dark frame that is reasonably current and is a reasonable representation of your current sensor performance

 

If you want us to explain a single auto-find result that you’re not happy with, you need to send a debug log file and a specific time when you got the result and why you’re not happy with it.  Poring through logs looking at single star-find results is time-consuming and tedious.  FWIW, I run a configuration much like yours in a remote observatory and I’ve never had the auto-find function fail to find an acceptable guide star and produce good guiding results – nearly 2 years of operation with an emphasis on imaging galaxies.

 

Bruce

--

You received this message because you are subscribed to the Google Groups "Open PHD Guiding" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-phd-guidi...@googlegroups.com.

Joel Short

unread,
Jan 21, 2021, 11:48:48 AM1/21/21
to Open PHD Guiding
Thanks Bruce, I'll give it a try.  Learn something new every day!
joel

Dale Ghent

unread,
Jan 21, 2021, 11:49:20 AM1/21/21
to open-phd...@googlegroups.com

Understood about multi-star. Although ROI isn't a fix for this situation, I think we've seen some users run into situations where the star that was auto-selected was so close to the edge of the frame that it fell off later on in the session due to dithering. I think the prospect of this happening is what makes some users curl their toes a little when a star so close to the edge is selected.

Will multi-star pick a replacement and carry on with life if such a thing were to happen under its watch?
> To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/4308B09194B0437EAF9066128AA7034B%40HomeDesktop.

bw_msgboard

unread,
Jan 21, 2021, 11:54:09 AM1/21/21
to open-phd...@googlegroups.com

Hi Richard.  You’re using that UI field correctly but it also sounds like you know what you’re doing.  We put it in the dev build to see how things go and we haven’t added it to the documentation yet.  What we have to constantly worry about is people fat-fingering the user interface or entering values that make no sense and thus finding no stars to use. It’s a fine line to walk.  As you’ve seen, the values are very dependent on the properties of the guiding gear and we have to assume the user has otherwise set things up correctly with star-saturation level, min-HFD, focus, etc.  The default value of 6 is based on photometric theory but it can be too low for many setups.

 

Bruce

 


From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of Richard Beck
Sent: Thursday, January 21, 2021 8:31 AM
To: Open PHD Guiding
Subject: Re: [open-phd-guiding] Guide star on edge of field

 

Bruce,

 

What about "Minimum star SNR for AutoFind"?  I had some very low SNR stars selected and had lost stars.  Increasing this value seemed to help.

 

The stars that seemed the best have SNRs in the 30s and 40s.  I set this at 20 and seemed to have better results.  Is there documentation on this parameter that I must have missed?

 

On Thursday, January 21, 2021 at 10:23:53 AM UTC-6 bw_m...@earthlink.net wrote:

It doesn’t matter where the star is in the field – it won’t get lost just because it’s near the edge.  To get the best results from auto-find, you should do the following:

 

1.      Set the saturation level =  the largest ADU your camera can return

2.      Set the min-HFD level = the smallest usable star size based on experience or measurements with your system

3.      Use a dark frame that is reasonably current and is a reasonable representation of your current sensor performance

 

If you want us to explain a single auto-find result that you’re not happy with, you need to send a debug log file and a specific time when you got the result and why you’re not happy with it.  Poring through logs looking at single star-find results is time-consuming and tedious.  FWIW, I run a configuration much like yours in a remote observatory and I’ve never had the auto-find function fail to find an acceptable guide star and produce good guiding results – nearly 2 years of operation with an emphasis on imaging galaxies.

 

Bruce

 


From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of Joel Short
Sent: Thursday, January 21, 2021 7:44 AM
To: Open PHD Guiding
Subject: [open-phd-guiding] Guide star on edge of field

 

I have noticed that when PHD2 automatically chooses a guide star it is almost always on the very edge of the OAG field of view.  This hasn't been much of an issue for my refractors, but I now have a 12.5" iDK (f6.7 FL=2128mm) and there are far fewer stars to to choose from.   

Almost always, there are stars in the middle of the field that have a good profile for guiding but are never chosen by PHD2.  Any tips for adjusting how PHD2 automatically chooses a a guide star?

 

I'm not sure it helps in this case, but here is my last guide log:  https://openphdguiding.org/logs/dl/PHD2_logs_HLtj.zip

 

It may not be helpful because I was doing repeated autofocus runs to determine filter offsets.

joel

--
You received this message because you are subscribed to the Google Groups "Open PHD Guiding" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-phd-guidi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/4a65093b-a8a7-4805-b81f-ccab5a8ca9e4n%40googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Open PHD Guiding" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-phd-guidi...@googlegroups.com.

Joel Short

unread,
Jan 21, 2021, 11:58:43 AM1/21/21
to Open PHD Guiding
I guess I spoke too soon.  The ADU value doesn't stick but always reverts to 255.  The camera is an ASI174MM-mini.  I'm using "ZWO ASI Camera" in the connect equipment field.  I use two other ZWO cameras via the ASCOM driver so I need to use the native ZWO driver for the guide camera.  Any ideas?  
joel

bw_msgboard

unread,
Jan 21, 2021, 12:01:41 PM1/21/21
to open-phd...@googlegroups.com
PHD2 won't allow the primary guide star to be pushed out of the fov because
of dithering. But if you're dealing with a sloppy setup, the guide star
could be lost for that reason. The desired solution in all these cases -
and I'm specifically talking about apps like NINA - is for the imaging app
to take the required recovery action. That means pausing the sequence,
re-issuing the auto-find/start-guiding, plate-solving, and re-centering if
necessary. Anything less than that is pretty much just praying for rain
IMO.
https://groups.google.com/d/msgid/open-phd-guiding/17361B8D-B11C-4099-842C-1
1E95B244606%40elemental.org.

Brian Valente

unread,
Jan 21, 2021, 12:05:31 PM1/21/21
to Open PHD Guiding
sounds like you have the camera in 8-bit mode

bw_msgboard

unread,
Jan 21, 2021, 12:05:41 PM1/21/21
to open-phd...@googlegroups.com

I generally set it to less than the theoretical max just for some headroom.  The star brightness fluctuates between frames so you could see the max-ADU hit the theoretical maximum sporadically.  But it’s not real important, the big issue is 255 vs 65K.

 

Bruce

 


From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of Brian Valente
Sent: Thursday, January 21, 2021 8:29 AM
To: Open PHD Guiding
Subject: Re: [open-phd-guiding] Guide star on edge of field

 

Bruce just a quick clarifying question:

 

>>>Set the saturation level =  the largest ADU your camera can return

 

Do you set it to the absolute max, or slightly under that?

 

For 16-bit camera the max adu would be 65536

 

On Thu, Jan 21, 2021 at 8:23 AM bw_msgboard <bw_m...@earthlink.net> wrote:

It doesn’t matter where the star is in the field – it won’t get lost just because it’s near the edge.  To get the best results from auto-find, you should do the following:

 

1.      Set the saturation level =  the largest ADU your camera can return

2.      Set the min-HFD level = the smallest usable star size based on experience or measurements with your system

3.      Use a dark frame that is reasonably current and is a reasonable representation of your current sensor performance

Joel Short

unread,
Jan 21, 2021, 12:10:41 PM1/21/21
to Open PHD Guiding
It must default to 8bit, as I don't remember setting that, nor that I could set the bit depth.  I've changed it to 16bit and 65000 now and it sticks.  I'll run new darks and the guiding assistant next time I'm out.
Thanks Bruce!
joel

Bryan

unread,
Jan 21, 2021, 8:36:22 PM1/21/21
to Open PHD Guiding
Joel

I have a ZWO 174 mini as well.  Sensitive camera (picks up stars even on an OAG)...amp glow is moderate, but mitigated with darks.  I set the max ADU to 65535 in the Camera tab on the Brain (Advanced Settings).  It has never reset to anything else.

If yours is resetting, then that is happening outside PHD2

Bryan

Joel Short

unread,
Jan 21, 2021, 8:54:30 PM1/21/21
to open-phd...@googlegroups.com
Bryan, the ADU value is not resetting.  The issue was that the camera was set to 8bit in the driver.  Once I set that to 16bit the ADU now stays at 65000 where I set it.
joel
Reply all
Reply to author
Forward
0 new messages