David's Astronomy - Notes 2022-09-14 (S1037)

Issue
Poor guiding on some targets due to the selected Guide Star being either too dim (SNR only 15) or being probably just a random patch of warm pixels (SNR only 6 or 7), leading to the guide star being lost and causing trailed star images. The guide star is selected even though there are clearly better (higher SNR) candiate stars in the image (and within the specified ROI).

Description
This particular issue seemed to appear a few weeks ago, and is seen in the 3 different versions of PHD2 tried in this period : 2.6.10dev3, 2.6.11 & 2.6.11dev2.

Session is run by automated program control. For each target guiding is initially via PHD2's OpenGuiding Server using the "guide" method. A ROI region is specified (but this has only relatively recently added)

{"method": "guide", "params": [{"pixels": 2.0, "time": 10, "timeout": 30}, false, [341, 229, 2415, 1622] ], "id": 151}

The method starts looping exposures, performs 'autofind' to select a guide star, and begins guiding. The method returns a 'SettleDone' event either when guiding has settled or the when the method timesout without guiding being settled.

PHD2 Debug Log shows that the

02:14:53.962 00.004 11704 Star::AutoFind called with edgeAllowance = 0 searchRegion = 30 roi = 2415x1622@341,229 02:14:53.963 00.001 11704 AutoFind: using ROI 2415x1622@341,229

Event Monitoring wiki page on PHD2's GitHub site gives the following the following example Guide Method call:

{"method": "guide", "params": {"settle": {"pixels": 1.5, "time": 8, "timeout": 40}, "recalibrate": false, "roi": [20, -50, 1280, 960]}, "id": 42}

I had problems with the method being accepted if I include "roi": so I exclude the roi keyword from the call. The ROI specification is nevertheless accepted by PHD2 as can be seen by PHD2 Debug Log which shows :

02:14:53.962 00.004 11704 Star::AutoFind called with edgeAllowance = 0 searchRegion = 30 roi = 2415x1622@341,229 02:14:53.963 00.001 11704 AutoFind: using ROI 2415x1622@341,229

If I use the "settle", "recalibrate" and "roi" keywords like the wiki example

{"method": "guide", "params": ["settle": {"pixels": 2.0, "time": 10, "timeout": 30}, "recalibrate": false, "roi": [341, 229, 2415, 1622] ], "id": 43}

PHD2 returns an error

"error":{"code":-32700,"message":"invalid JSON request: Unexpected character on line 1 at \": {\"pixels\":...\""}

But PHD2 does accept the following ok "[ ]" brackets for params keyword replaced by "{ }" bracket.

{"method": "guide", "params": {"settle": {"pixels": 2.0, "time": 10, "timeout": 30}, "recalibrate": false, "roi": [341, 229, 2415, 1622] }, "id": 87}

When the issue with 'selected guide stars having only SNR 6 or 7' was first noticed, the Minimum SNR was set to higher and higher values in an attempt to 'discourage PHD2 from choosing a guide star that is known to be too faint for sustained guiding'. Value was eventually raised to 50. This appeared to work for some targets but not others. Hypothesis was that if the exposure contained a star with SNR > 50 it was picked ok, but if it didn't then PHD2 would typically ignore stars with SNR 20-49 and just choose one with a SNR of only 6 or 7 instead. Based on this hypothesis Minimum SNR was then reduced down to a value of 13 for sessions S1036 and S1037. However this didn't stop the problem from sometime occuring, with either a star with SNR only 6 or 7 being picked (clearly lower than 13) or one being picked with SNR of 16 that was higher than the Minimum SNR of 13 but not higher enough that the star was readily lost after 2 or 3 minutes. In all these case better candidate (higher SNR, non-saturated) guide stars are clearly present in the exposures.

Whilst changing the Minimum SNR to a new value of 20 was considered it is noticed that the position of guide star auto-selected is nearly always situated on or very close to the border of specified ROI and is rarely chosen within the central part of the ROI where the better candidate stars are often present

Looking at the position of the guide star that PHD2 has auto-selected (green box) for each of the guide runs during the last observing session it can be seen that the guide star is located on or very close to the ROI border (white rectangle) in 12 out of the 14 guide runs. A more centrally positioned guide star has been selected in just 2 guide runs (For Info, The yellow rectangle shows the FOV of the Main Imaging Camera, the Blue Circle is the target object)

Guide Star SNR 56

Guide Star SNR 633

Guide Star SNR 442

Guide Star SNR 6

Guide Star SNR 17

Guide Star SNR 35

Guide Star SNR 33 (guide star of SNR 6 initially selected)

Guide Star SNR 13

Guide Star SNR 148

Guide Star SNR 18

Guide Star SNR 197

Guide Star SNR 6


	Session Aims & Highlights
	- Observing Result - Night Summary Plot - Session Event Log

	Operational Issues
	- Critical Issues (0), Major Issues (1), Minor Issues (3), Small Defects (2), Continuous Improvement (1)

	Images from 2022-09-14 >> [ Local Files >> ]

	PHD2 Update 2.6.11dev2
	Investigation - PHD2 AutoFind Anomalies

This Web Page:	Notes - Session 1037 (2022-09-14)
Last Updated :	2024-09-25
Site Owner :	David Richards
Home Page :	David's Astronomy Web Site

David's Astronomy Pages
Notes - Session 1037 (2022-09-14)

Session Aims & Highlights (2022-09-14)

Operational Issues (2022-09-14, S1037)

PHD2 Update 2.6.11dev2

Investigation - PHD2 AutoFind Anomalies

David's Astronomy Pages Notes - Session 1037 (2022-09-14)

Session Aims & Highlights (2022-09-14)

Operational Issues (2022-09-14, S1037)

PHD2 Update 2.6.11dev2

Investigation - PHD2 AutoFind Anomalies

David's Astronomy Pages
Notes - Session 1037 (2022-09-14)