I did run it through the brain and got a 400ms step size, but for 12 calibration steps and a calibration distance of 25px. Online tutorials suggest more steps, so I will increase this to 20 steps and increase to 40px (assuming this is the “Calibration Distance” you refer to) next time out, for a Calibration step of the same 400ms - as the two changes offset each other.
- Is the “selection Box” you refer to the “Calibration Distance” parameter in the Guiding tab of the Brain?
- Is the value calculated by the brain automatically picked up and used in PhD2?
Any comments on the calibration plot not being orthogonal? – ever seen this?
Thanks for the help.
Ed
From: Lenny Shaffer <le...@lenny-shaffer.com>
Sent: Wednesday, December 15, 2021 9:13 AM
To: edk <flae...@gmail.com>
Subject: Re: [open-phd-guiding] Calibration Plots are not orthogonal and often Calibrations fail
Have you increased the size of the selection box or run the setup through The Brain to find the correct step size for your setup?
--
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/8952259b-878f-430e-9ef3-973c61944949n%40googlegroups.com.
--
Murphy's Law of the Week: "Murphy invented round holes and square pegs."
https://github.com/OpenPHDGuiding/phd2/wiki/EQASCOM-Settings
Then you need to create a new PHD2 configuration profile and enter all the appropriate parameters. Then follow this procedure while trying to get a workable calibration:
If the calibration is still whacky, then something is wrong with the mount or the mount control software. In that case, use the Star-Cross tool and/or the Manual Guide tool to see if you can get the mount to move consistently and repeatably in all 4 directions. Until that happens, the mount is un-guidable.
I don't know why you couldn't find any explanations in the manual, they exist - here's the section discussing calibration problems and how to understand them:
https://openphdguiding.org/man/Trouble_shooting.htm#Problems-_Mount_Control
As I said, your problems seem to be so severe that something very basic is wrong with the setup - either the guide speed or the mount.
Good luck,
Bruce
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/fbc4f6a5-58be-4142-9ae5-e5897f3c265an%40googlegroups.com.
Yes, my first attempts were at the wrong rates. All new attempts and associated correspondence, which still fail in DEC control, are at between. O.6 and 0.9 rates.
Per correspondence, I will post the Guide Logs.
Your patience, and contributions, are very much appreciated.
Ed
From: open-phd...@googlegroups.com <open-phd...@googlegroups.com> On Behalf Of mj.w...@gmail.com
Sent: Saturday, December 18, 2021 4:25 AM
To: Open PHD Guiding <open-phd...@googlegroups.com>
Subject: Re: [open-phd-guiding] Calibration Plots are not orthogonal and often Calibrations fail
Hi Ed
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/087d7310-53dd-453e-bbb0-897548b30dc9n%40googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/752c4753-91ff-43e0-a45b-f5c7882ad5c3n%40googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/06166b59-988d-4266-8cf2-dd2b9cce1564n%40googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/9d252d83-92da-4d23-8049-9ac827d23d7fn%40googlegroups.com.
Actually I am guiding with PHD@ and EQMOD only.
The USB Root Hub was turned off per earlier note to another person who commented on it.
As seen in the below data, PHD2 is requested to have an RA rate of 9, which shows up in the mount calibration data below, but the mount is responding at 92.57 a-s/sec!!! How the command of 9 is converted to 92 is the problem. What is the reason/mechanism that is taking a command from PHD2 of 9 a-s/sec and turning it into 92.57 a-s/sec. What setting/other input could generate this is the insight I am hoping to get help on. Yes, it is clear to me as well what the issue is, but what is the cause and how to fix it.
I will try tomorrow to generate the star figure called for by PhD2 as a diagnostic, then go from there.
Also, others have noted that NINA does not have the option/function to generate Video so all photos are in long exposure mode.
Ed
From: open-phd...@googlegroups.com <open-phd...@googlegroups.com> On Behalf Of Brian Valente
Sent: Monday, December 20, 2021 9:43 AM
To: Open PHD Guiding <open-phd...@googlegroups.com>
Subject: Re: [open-phd-guiding] Calibration Plots are not orthogonal and often Calibrations fail
Uncheck the computer's ability to power off the usb device, as shown
On Mon, Dec 20, 2021 at 6:42 AM edk <flae...@gmail.com> wrote:
What then are the correct USB Power Management and Power Settings to use?
On Monday, December 20, 2021 at 8:41:09 AM UTC-5 mj.w...@gmail.com wrote:
Hi Ed
Your Palm Tree test sounds like you may have had the camera configured in Video Mode, outputing high frame-rate video, as used for Planetary Imaging.
The Long Exposure Mode is the one outputting the pinstripes.
Guide Rate has gone from too small, 1.5arcsecs/sec, to way too big, 92.57 and 105.981 arcsecs/sec:
Until you better understand your software, try guiding with just PHD2 and EQMOD.
USB Power Management and Power Settings are in Windows Device Manager:
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/CAJa45i4Kg3RQgqDhNKdiQM1Qpw%2BrgQ3Ag_a4WvPP0KSWvwo%2Bqg%40mail.gmail.com.
Hello Michael,
I apologize for the empty file, I do not understand how that happened. In the future I will strip out non-active viewing times to make it more manageable.
PhD2 does not show the log for the last run I performed in its list of upload logs, even if the “show logs with no guiding” box is checked. The correct logs do appear in the Help/Open Log Folder tab. The correct log is attached, and has been stripped of all but that evenings active viewing times, including the calibration attempts. The original log was 35M and may have not uploaded due to size.
The instructions for the Star-Cross app explicitly require taking an extended photograph using the main imaging camera. I run the main imaging camera through NINA and so, YES, I was using NINA. Is there an issue with using NINA for this requirement?? If not NINA – Then how take the picture? Should the star-cross images shown up on the NINA photograph taken while PHD was executing the star-cross app?
When I attempt to calibrate using some instructions provided in the thread, including having the weights horizontal and the telescope perpendicular to the RA axis, PhD2 gives a warning that the calibration is not reliable because it is being attempted too far from celestial DEC=0. I am at 26deg N which points very high in the sky.
I have now contacted Sky-Watcher, and High-Point, to ask about issues with the EQ6-R mount interface to PHD, including providing them with a writeup and representative logs. I am also looking to reload EQMOD and the mount drivers.
Leaving no stone unturned, I will go back and attempt to run a calibration at 0.9 a-s/s
I do very much appreciate your help and expertise.
Wishing you a fine holiday season.
Best regards
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/87d418a3-1388-4dae-9240-5fe66aefc558n%40googlegroups.com.
https://openphdguiding.org/logs/dl/PHD2_logs_uJE8.zip
Very best regards, and again, Thank You.
Clear Skies to all
Ed
FL USA