Drift in RA

[Les Dickens]

I am trying to pin down the cause of a steady drift in RA.  To this end it would be useful if I could determine from the guide log (attached) whether or not the guide star is actually drifting in the guide camera (I assume not as that's what PHD prevents!).  Is it possible to do that, as it would point to differential flexure as the source of the problem (even though the drift is in RA only, which seems a bit odd).  I loaded the log into PECPrep and that showed no linear regression, if that's relevant.

The drift rate is ~ 1.5 arc-secs / min.

Thanks,

Les 

[Brian Valente]

Hi Les

I don't have any comments on your drift rate (hopefully others will chime in) but I did notice these things in your guiding:

Mount = EQMOD HEQ5/6 (ASCOM), connected, guiding enabled, xAngle = 86.8, xRate = 1.419, yAngle = 173.6, yRate = 0.821, parity = +/+
Norm rates RA = 5.8"/s @ dec 0, Dec = 3.2"/s; ortho.err. = 3.2 deg
X guide algorithm = Hysteresis, Hysteresis = 0.200, Aggression = 0.490, Minimum move = 0.200
Y guide algorithm = Resist Switch, Minimum move = 0.300 Aggression = 30% FastSwitch = enabled
Backlash comp = enabled, pulse = 349 ms
Max RA duration = 3000, Max DEC duration = 2540, DEC guide mode = Auto
RA Guide Speed = 6.0 a-s/s, Dec Guide Speed = 3.0 a-s/s, Cal Dec = 16.6, Last Cal Issue = None, Timestamp = 28/02/2021 19:54:31

this looks to me like your guidespeed is too slow, and your aggression settings are quite low as well.

You should be at least 0.5x for guidespeeds on both axis

i would stick with the default values for aggression on both axis

Also the min moves you have are quite high relative to your guiding results. RA and DEC have a fair bit of room to wander about in there, so you might consider tightening that up as well

--
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/7349557d-6b02-47f9-9a5d-7a238bcdcea3n%40googlegroups.com.

--

Brian 

Brian Valente

portfolio brianvalentephotography.com 

[Les Dickens]

OK - clear night here tonight, so I will try your suggested changes.  

Thanks!

Les

[bw_msgboard]

Hi Les.  I would say the drift rate isn’t interesting, guiding can deal with that pretty effectively.  Your guiding results are limited by the uncorrected periodic error in the mount:

 

 

Most of this is coming at the primary worm period of 480 seconds.  PHD2 would do a better job of reducing this if you hadn’t shot yourself in the foot by reducing the aggressiveness.  As Brian said, you should restore the guide parameters to their default settings.  Beyond that, you can explore getting a high-quality PEC programmed into the mount or you can try using the PPEC algorithm in PHD2.  If you do that, give it an initial period of 480 seconds and let it run awhile to see how it goes.  Your elongated stars could be due to the 2x difference between your RA and Dec guiding rms or it could be due to differential flexure, or both.  If you look at this document, you can find a technique in the appendix for testing to see if it’s differential flexure:

 

https://openphdguiding.org/tutorial-analyzing-phd2-guiding-results/

 

 

Good luck,

Bruce

 

---

To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/e60a0f26-547f-494f-992c-81b657a5bfb0n%40googlegroups.com.

[Les Dickens]

OK - thanks for that.  Last night I tightened up everything  that can be tightened up and made the changes that Brian suggested: the tracking is hugely improved as a result.    

 I created the current PEC by running PHD with guide output disabled and a very short exposure time in order to generate a lot of data.  Looking at the Log Viewer analysis, I  clearly threw away some of that data along with the noise during the PECPrep processing.  I will have another go and in the meantime, take a look at the tutorial and the PPEC algorithm, as you  suggest.

 Thanks again,

Les