PHD2 : Auto-select star too early after meridian flip

Hello,

I noticed a (small) issue with SGP and PHD2 in my last 3 sessions.

It seems that, after the meridian flip, SGP tells PHD2 to auto-select a star BEFORE resuming frame looping.
As a result, PHD2 autoselects the star on a frame corresponding to the “Before meridian flip situation”.

Of course, when PHD2 resumes looping a few seconds later, this star is lost (since meridian flip has occured !) but PHD2 always finds a (false) star in the search box and reports “guiding” to SGP, even if this guiding is of course very poor.

In my opinion, SGP should wait until PHD2 reports it is looping (then we are sure that PHD2 continuously refreshes a field AFTER meridian flip), and only then, should send the “auto-select star command” (which works very well when done manually). Is there a simple (automatic) way to do that ? (I didn’t find it in SGP or PHD2 and I don’t want to wake up to do this manually…)

Versions : SGP 2.6.0.25, PHD2 2.6.4.

thanks !

If you have log files for these events, I’m sure it would help. an SGP log for sure and a PHD2 log file if possible.
The latter is located in documents\Phd2

Actually, SGP does not tell PHD2 when to select the star, PHD2 determines that on its own. We’d need to see the logs to determine what actually happened.

yes, we’ll need to see both the PHD2 Debug Log and the Guide Log (PHD2 log instructions) in addition to the SGP log.

Hi,

here are these logs :

https://www.dropbox.com/s/o53a9kb0tmqc1x0/PHD2_DebugLog_2018-02-21_192141.txt?dl=0

The meridian flip began at around 21:23.
After the MF, I distinctly saw an auto-select star being done on the “old” (before MF) PHD image. Then, loss of this star after frame looping resume and finally choice of “nothing” in the 25 pixels searchbox (there was, however, an HFD measured but it was definitely not a star, rather a pixel cluster of low-frequency noise, resultinf in awful guiding).
At about 21:30, I manually stopped PHD2 to issue a manual auto-select star command, which picked a good star. Everything was OK after that.

Thanks for the logs. According to the logs the ASCOM driver returned a stale camera frame from before the meridian flip. I think you will have to bring the issue to the attention of the ASCOM driver developer (Atik). I’d be glad to talk about it with Atik if you want to put them in touch with me (andy.galasso@gmail.com)

Here are the details from the logs to support the conclusion:

the last complete exposure before the meridian flip, note guide star position = (460,164)

21:23:59.872 00.000 6820 ScheduleExposure(3000,3,1) exposurePending=0
21:24:03.252 03.166 3956 Exposure complete
21:24:03.286 00.000 6820 Star::Find(25, 460, 163, 0, (0,0,0,0), 2.0, 0) frame 1786
21:24:03.286 00.000 6820 Star::Find returns 1 (0), X=460.45, Y=163.81, Mass=25859, SNR=97.1, Peak=4520 HFD=2.8

next 3s exposure begins

21:24:03.308 00.000 6820 ScheduleExposure(3000,3,1) exposurePending=0
21:24:03.399 00.000 3956 worker thread servicing REQUEST_EXPOSE 3000

SGP tells phd2 to stop guiding for the meridian flip:

21:24:05.474 00.102 6820 evsrv: cli 04D3BC28 request: {"method":"stop_capture","id":1004}
21:24:05.474 00.000 6820 StopCapturing CaptureActive=1 continueCapturing=1 exposurePending=1

PHD2 tells the ASCOM driver to abort the exposure (2.1 seconds into the 3s exposure), and the ASCOM driver reports that the exposure was successfully aborted

21:24:05.528 00.003 3956 ASCOM_AbortExposure returns err = 0
21:24:05.528 00.000 6820 OnExposeComplete: Capture Error reported

PHD2 has now stopped; SGP finishes the meridian flip

[02/21/18 21:26:27.959][DEBUG] [Telescope Thread] Telescope: Syncing to J2000 RA: 6,30289361587789 Dec: 22,6248376659468^
[02/21/18 21:26:28.043][DEBUG] [Center Scope Thread] Automatic Auto Center Success - Total Error <= Allowable error: 1,4 <= 10,0

PHD2 starts looping

21:26:32.725 00.000 6820 StartCapturing CaptureActive=0 continueCapturing=0 exposurePending=0
21:26:32.730 00.005 6820 ScheduleExposure(3000,3,0) exposurePending=0

camera frame arrives, it is a stale frame from before the flip. We know this because it finds the same guide star @(460,164) from before the flip

21:26:35.825 03.074 3956 Exposure complete
21:26:35.882 00.000 6820 Star::AutoFind called with edgeAllowance = 0 searchRegion = 25
21:26:35.958 00.000 6820 Autofind returns star at [460, 164] 76.7 Mass 26275 SNR 97.5

next exposure – a valid camera frame, phd2 finds a weak star nearby in the search region @(476, 141) and locks on to it

21:26:36.010 00.000 6820 ScheduleExposure(3000,3,1) exposurePending=0
21:26:39.104 03.094 3956 Exposure complete
21:26:39.139 00.001 6820 Star::Find(25, 460, 163, 0, (0,0,0,0), 2.0, 0) frame 2
21:26:39.139 00.000 6820 Star::Find returns 1 (0), X=476.29, Y=140.92, Mass=1660, SNR=16.4, Peak=2128 HFD=2.4

Thanks a lot Andy.
You described EXACTLY what I directly saw in the PHD window.
I contact ATIK with that issue, forwarding them your conclusions and email.
We will see what happens !

Many thanks again,

Olivier.

1 Like