PHD2 Guide Instability

[Clayton Yendrey]

Guide logs for 10/10 and 10/14
https://openphdguiding.org/logs/dl/PHD2_logs_Mxpk.zip

I know that there are camera issues:  10/10 with a QHY guide camera, driver issues with PHD2
10/14, physical sensor issues with the ZWO guide camera.

I hope to have the guide camera issue finally resolved in a week so that doesn't obscure the issue.  
I have, within TSX,  run a TrainPEC, and two TSX guide sessions.  One session, approximately two hours, was with the OAG/primary imaging camera doing the guiding.  This was followed directly by one hour on the OAG guiding.  This was followed directly by a 6 hour PHD2 guide session on the OAG (while also imaging).

There is a consistent result between what TSX indicates and what occurs in PHD2.  In straight tracking for the PEC, tracking is smooth and the guide/PEC star has minimal movement (ProTrack TSX feature enabled) and the star stays well centered with a 5.5 a-s PEC, P-P.  No measurable PEC after applying PE correction (which in TSX speak means the PEC was less than 1a-s P-P).

The TSX guide sessions exhibited no axis related excursions, and while aggression was not optomized, at close to default settings for DirectGuide, the OAG over the two hours averaged a Total RMS error of approximately.5 to .7 a-s.  During the hour run under OAG guiding on TSX, Total RMS guide error stayed in the .25 to .38 a-s range.  No odd axis related excursions or spikes.

That cannot be said of PHD2 guiding which exhibited both RA and DEC spikes of 2 a-s or greater mixed with periods of 'smooth' guiding which got shorter as the session continued.  This story was repeated last night.

In particular, the comments in my earlier posting concerning the poor tracking performance of the mount came the day after Software Bisque noted the extremely good performance and stability of the mount during the PEC training session.

I have noticed in the previous TrainPEC sessions at the first of the year and late last year, that since the replacement of the failed worm drive, I never see exhibited in TSX, the degree of axis instability (tendency to spikes) that are shown in PHD2.  I find the difference in unguided tracking performance (as recorded in both applications) concerning since those should be close to identical.

I'm at a loss to explain this growing amount of evidence that PHD2 might be involved in some manner with the mount tracking and guide "issues" shown in PHD2 but not in TSX.  Based on the more obvious problems found with the QHY drivers and USB traffic, it seems it is worth asking if a similar but more subtle issue(s) could be present with ZWO guide cameras or with USB connections in general for PHD2?

FYI - the ZWO cameras always operate under the ZWO native windows drivers, not ASCOM.

[Bruce Waddington]

It looks to me like you have static resistance (stiction) on the RA axis.  This shows up in both guide logs and is demonstrated by the occasions when the mount is slow to respond to guide pulses in RA then abruptly over-shoots.  I don't know how the SB drive system works but on other mounts this can be caused by an over-tight mesh of the worm gear with the worm wheel.   Obviously, none of this is going to show up during unguided or PEC measurement sessions when the RA drive is simply moving at a constant sidereal rate in the same direction.  You'll notice the PHD2 GA run doesn't show them either. Since you have identified the replacement of the worm drive as a pivotal point in time, this seems consistent with a meshing problem of some kind.  There is also some evidence that large RA guide pulses and the subsequent slowness of the RA axis to respond are causing the Dec axis to move.  We have seen this before on some other SB systems but have no way to diagnose the details.  On other mounts, this can be caused by a combination of an overly tight mesh in RA and a bearing weakness in Dec.

If you want to rule out other unlikely possibilities, you could do some guided runs without imaging or dithering to be sure there's nothing happening in the imaging train that would cause vibration or other unwanted movement (e.g. filter changes).  You could also use the Diagnostic Image logging function in PHD2 (Advanced Settings/Global tab) and capture guide camera frames when these larger excursions occur - it looks like movements of greater than 2px would be a suitable trigger.  If you look at the images, you should see that the guide stars really have shifted in the frame and that the frame is not truncated or otherwise corrupted.  Each event will cause a capture of 5 guide camera frames - 2 before the event, 1 associated with the event, and 2 more frames after the event.

Good luck,

Bruce

[Clayton Yendrey]

Bruce,
Thanks for the feed back.  I have not gone through the SB plunger adjustment procedure yet (it controls the worm/ring gear engagement tension).  However, I wanted to see how tight the RA plungers were (it is the DEC plungers that usually get adjusted to deal with mechanical cross talk).  In the SB mounts there are two screws that press down on a plate that presses down on springs that push down on plungers, etc.  They are supposed to be equal and the adjustment to OEM spec process calls for 'zeroing' the plungers by screwing them down (not hard) but to the point of no further adjustment possible then using that point to count off turns for the recommended starting point with further 'refinement' as indicated by mount performance.

I found that the left RA plunger screw / plunger is hard against the bottom/pressing on the plunger, no further adjustment possible.  The right hand is not that way.  I haven't made any corrections yet (cool front just blew through and some 20mph+ winds right now).  I also notified SB of this discovery to see if there is a rationale for it being something other than an "oops" when they serviced the mount.  I intend to go through the SB pdf on the plunger adjustment (starting from scratch) to get to the normal starting point.  Depending on where the tension winds up, there will be a need for a new PEC I think, so a chain of follow on 'stuff' once this process gets started.  Of course, it looks like a week or longer before clear skies are forecast again.

I also want the replacement guide camera in the OAG before jumping off on this, so it will be a bit before I can follow up.

Thanks again for the analysis, help.  I would not have thought to check the RA plunger tension otherwise since it rarely requires adjustment (unlike the DEC), at least on the Gen 5 MyT mounts.

Clayton

[Clayton Yendrey]

Just re-reading you reply and noticed the mention of DEC bearing weakness.  There is a possibility that this is a design "issue" with the MyT; since its original release it has been followed by a consistent (if small) percentage of the mounts exhibiting mechanical cross talk.  The 'fix' that has evolved from user experience has been to back off the plunger adjustments on the DEC axis up to the point that hysteresis is noted then tightening just sufficient to eliminate that.  The new Gen6 SB mounts got a slight pay load increase (50# to 70#) and larger axis shafts and bearings.  The back-channel chatter was that beside competing directly against the AP Mach2GTO, the Gen 6  up-spec of internal components was to address the mechanical cross talk problem.  It will take some time to determine whether that goal was completed; if the cross talk issue is no more, simply having the through mount cabling will give it an advantage over the Mach2GTO now that payloads are equal (I know of a few folks that went to the 1100 rather than the Mach2GTO because of the lack of through mount cabling capability in the Mach2GTO).  I don't know how accurate the chatter is; SB up-spec'd load capacity (bearings/shafts) in the entire Gen6 line so it may be the chatter concerning crosstalk being addressed in the Gen6 is wistful thinking.  ;-)

[Bruce Waddington]

I think the two problems are probably correlated.  By that, I mean that eliminating the static resistance on the RA axis is likely to eliminate any evidence of this "cross-talk".  If there is resistance on the RA axis, the energy associated with a large guide pulse may not get the axis moving but it "goes somewhere".  That's why these cross-talk events are usually associated with larger than normal guide pulses. With some mounts, the energy may just be absorbed and stored without causing any immediate effect but in other cases it may cause movement somewhere else.  That might mean deflection of a worm mounting block or, perhaps, movement of a bearing on the Dec axis.  The energy will always be released at some point, often causing an over-shoot, and of course there could be both effects - an over-shoot and an unwanted movement of the Dec axis.  I think the thing to focus on first is getting the RA axis behavior sorted out.