Re: Guiding AP mounts with PHD


Michael Fulbright <mike.fulbright@...>
 

I use Hysteresis on RA and ResistSwitch on DEC and have very good
guiding results on my Mach1.

I have some backlash in DEC I really ought to adjust but instead I just
turn on the auto backlash compensation in PHD2 and it works great so
I've put off getting out the hex wrenches.  :)

I can't image any use case for ResistSwitch on RA.  The gear will always
be loaded and if you are getting too much chasing in RA it is probably
because your MinMove isn't set high enough.

The "RA OSC" measure should be around 0.5 - you can use this as a guide
to as if you are sending corrections in RA too often or not enough.

(From the PHD2 User Guide: http://openphdguiding.org/PHD2_User_Guide.pdf)

The 'RA Osc'  value shows the odds that the current RA move is in the
opposite direction as the last RA move. If you are too aggressive in
your guiding and over-shooting the mark each time, this number will
trend toward 1.0.  If you were perfect and not over- or under-shooting
and your mount had no periodic error, the score would be 0.5  Taking
periodic error into account, the ideal value would be closer to 0.3 or
0.4.  If this score gets very low (e.g. 0.1), you may want to increase
the RA aggressivness or decrease the hysteresis.  If it gets quite
high (e.g. 0.8), you may want adjust aggressivness/hysteresis in the
opposite direction.
Michael Fulbright

On 2/15/2018 2:51 PM, Ron Kramer ronkramer1957@gmail.com [ap-gto] wrote:
Once I get more experience with it working well - I'll start to switch
options to compare.

On Thu, Feb 15, 2018 at 1:09 PM, Bill Long bill@outlook.com
<mailto:bill@outlook.com> [ap-gto] <ap-gto@yahoogroups.com
<mailto:ap-gto@yahoogroups.com>> wrote:

Interesting. I just installed PHD2's latest build on my imaging
machine prior to my last run, and the default for RA was
Hysteresis. In looking at their documentation they mention:


The ResistSwitch algorithm behaves much as its name implies.  Like
the hysteresis algorithms, it also maintains a history of past
guide corrections, and any change of direction must be
"compelling" in order to issue a reversing guide command.  This
is appropriate for declination guiding, where reversals in
direction are both suspect and likely to trigger backlash in the
gears.  For that reason, ResistSwitch is the default algorithm for
declination but not for right ascension, where valid direction
reversals are expected.


I wonder if other folks are using Resist Switch on both axis as well?


------------------------------------------------------------------------
*From:* ap-gto@yahoogroups.com <mailto:ap-gto@yahoogroups.com>
<ap-gto@yahoogroups.com <mailto:ap-gto@yahoogroups.com>> on behalf
of Ron Kramer ronkramer1957@gmail.com
<mailto:ronkramer1957@gmail.com> [ap-gto] <ap-gto@yahoogroups.com
<mailto:ap-gto@yahoogroups.com>>
*Sent:* Thursday, February 15, 2018 7:00 AM
*To:* ap-gto@yahoogroups.com <mailto:ap-gto@yahoogroups.com>
*Subject:* Re: [ap-gto] Re: Guiding AP mounts with PHD
Never thought I'd be sharing settings. I had nearly a year of
problems (Skywatcher mount) but they continued after getting the
mach1.  I did several things at once and am not sure what solved
the problem.
I remeshed the gears (easy) and removed slop out of both.  (maybe
that was it?)
but I also installed PHD on a new computer (clean) with new
defaults.  And maybe that was it? or both?)

The old defaults had hysteresis for RA. The new defaults are
"resist switch" on both. I left it at that, latest installed
defaults..
I also changed from 80mm to side by side 80 and 127mm scopes - so
there is more weight and complete rebalance. ( I don't think that
was it ) but my point is I changed so much that
it just started working well - and I have no idea what made the
change..

Oh - I also swapped from a ASI290mm camera to a lodestar.  The
initial test did not show that was improved.  BUT I eliminated all
the star lost problems I had been having. It didn't hurt going to
the lodestar guide camera that's for sure.
It's also imperative that the pixel size be correct.


On Wed, Feb 14, 2018 at 2:45 PM, marfig1970@yahoo.com
<mailto:marfig1970@yahoo.com> [ap-gto] <ap-gto@yahoogroups.com
<mailto:ap-gto@yahoogroups.com>> wrote:

I am also interested in the best PHD2 settings for the Mach1.
What guidance algorithms are you using?



---In ap-gto@yahoogroups.com <mailto:ap-gto@yahoogroups.com>,
<ronkramer1957@...> wrote :

Bill - 90% aggression seems excessive. I had my initial
problems, but the one thing that always helped was little
aggression else overcorrection was the result.
I let things run - and every 3-4 minutes I tweaked settings
depending on the graph details.  I ended up here.
I also think that one of the major factors in the improvement
of my new mach1 mount might have been the re-meshing of both
axes. There was a good bit of play in both axes upon receipt
of the mount.
Very easy to adjust as the manual explains. I've always read -
longer exposures to reduce poor seeing, I'm now going to
adjust this with shorter exposures as Roland suggests.

My latest results and settings.

https://scontent-ort2-1.xx.fbcdn.net/v/t31.0-8/26758301_10213070253583834_4228601485580199512_o.jpg?oh=2dbea085eb7e6ad2655587db5382c924&oe=5B081855
<https://scontent-ort2-1.xx.fbcdn.net/v/t31.0-8/26758301_10213070253583834_4228601485580199512_o.jpg?oh=2dbea085eb7e6ad2655587db5382c924&oe=5B081855>

On Tue, Feb 13, 2018 at 3:36 PM, Bill Long bill@...
<mailto:bill@....> [ap-gto] <ap-gto@yahoogroups.com
<mailto:ap-gto@yahoogroups.com>> wrote:

Yeah during the summer, I am usually in the 1.5-1.75"
range for my images, using the defaults in PHD2 and my
FSQ106ED scope + OAG, and not the method discussed
here. So I am somewhat eager to see what occurs using
this same method in the Summer months and using my
TOA130 as well.


I just upgraded to PEMPro 3 via the free upgrade
promotion, and plan to set aside the next good night
to work on PE Curves. Mine is a few years old, so I
suspect that it needs to be refreshed. I will do the
unguided test you suggested, with PE on and off and
see what the results are before and after the PEMPro
3 work. That is a good suggestion.


But back to the method here -- I do think this test at
least shows that this method is much better than just
accepting the defaults, and much better than taking
the suggestions from the Guiding Assistant itself (it
told me to set MinMo to 0.10, and 0.15 in the two
tests I did, but with 3-4 second exposures).


I'll try and set the RA Aggression to 90% and work
from there and see how RMS looks over time.



------------------------------------------------------------------------
*From:* ap-gto@yahoogroups.com
<mailto:ap-gto@yahoogroups.com>
<ap-gto@yahoogroups.com
<mailto:ap-gto@yahoogroups.com>> on behalf of
chris1011@... <mailto:chris1011@...> [ap-gto]
<ap-gto@yahoogroups.com <mailto:ap-gto@yahoogroups.com>>
*Sent:* Tuesday, February 13, 2018 12:24 PM
*To:* ap-gto@yahoogroups.com
<mailto:ap-gto@yahoogroups.com>

*Subject:* Re: [ap-gto] Guiding AP mounts with PHD

I took some images and noted that the FWHM (at
0.93"/px) was about 2.5-3" over the course of the
night.

It is not unusual to have 2 - 3 arc sec stars this
time of year. I get that here with any of my scopes,
even the mighty 17" during these winter months. Once
summer gets here my results get better and better
until late summer - early fall when the nights are
pristine and very calm. That's when I achieve my best
seeing and stars can be as small as 1.2 arc sec with
the 17". A smaller aperture, shorter scope cannot
achieve that, of course but sub 2" can be gotten with
a 5" refractor.

As far as RA, you will not get any useable data or
know what to do by looking at guiding data. If you
want to know the health of the RA axis, you MUST run
an unguided test to see what your periodic error is
doing. PE correction must be off. Once you run a few
worm cycles you can judge how well the basic mount is
tracking in RA. Then do the same unguided test with PE
correction turned on. If the periodic error does not
improve significantly, then it's time to do a PEMPro
run (or any software that can create a PE curve) and
load it into the mount. Then do another run with PE
correction turned on to verify that the mount is now
tracking more accurately.

One thing I would do is to set the RA aggressiveness a
bit lower than Dec unless the seeing is pristine.

Rolando


-----Original Message-----
From: Bill Long bill@... <mailto:bill@...> [ap-gto]
<ap-gto@yahoogroups.com <mailto:ap-gto@yahoogroups.com>>
To: ap-gto <ap-gto@yahoogroups.com
<mailto:ap-gto@yahoogroups.com>>
Sent: Tue, Feb 13, 2018 1:47 pm
Subject: Re: [ap-gto] Guiding AP mounts with PHD



So I finally had a chance to test this out last night.
Here are the steps I took:

1. Calibrate near the equator and meridian.
2. Allow PHD to guide for a short period using 1 sec
exposures.
3. Turn on Guiding Assistant (which turns off guiding
and gives a readout of error parameters)
4. Allow guiding assistant to run for 300 seconds.
5. Note the Peak error for Declination (in pixels).
6. Set MinMo to the peak error for Declination on
both axis (in my case 0..50), set aggression to
100% for both axis.
7. Make no changes to Hysteresis (default = 10)
8. Allow PHD to guide and watch the graph.


While it was new to use PHD2 with this short of
exposures, I could see that the bobble that Roland
refers to was ignored and that the mount was corrected
for legitimate reasons only. I took some images and
noted that the FWHM (at 0.93"/px) was about 2.5-3"
over the course of the night. The majority of all
corrections sent to the mount were RA corrections,
with a few Dec changes here and there. Not sure what
to make of this? I did play with some of the values to
see if I could improve RMS, but none of the changes I
made seemed to make a positive difference.

Anyhow, I changed back to my previous settings
(basically using defaults in PHD2 and 3-4 second
exposures) as a test and noted that the stars in my
images were worse off (4-4.5" FWHM) so I believe there
is good reason to use this method. I just hope I did
it right! 😊 My system does currently have an issue
with tilt (20%) that I need to resolve, so I believe
that the improvement in star quality would have been
much better, had the train not had the tilt issue in
it. Overall I was pleased though.

This still brings up the question of whether or not
Hysteresis should be used for either of the axis, or
if Resist Switch or other algorithms would work better
with the AP mounts? Also, should I be concerned at all
that the majority of corrections were made in RA (and
some of them were pretty significant corrections)?

------------------------------------------------------------------------
*From:* ap-gto@yahoogroups.com
<mailto:gto@yahoogroups.com> <ap-gto@yahoogroups.com
<mailto:gto@yahoogroups.com>> on behalf of
chris1011@... <mailto:chris1011@...> [ap-gto]
<ap-gto@yahoogroups.com <mailto:gto@yahoogroups.com>>
*Sent:* Wednesday, January 31, 2018 4:51 PM
*To:* ap-gto@yahoogroups.com
<mailto:gto@yahoogroups.com>; ap-ug@yahoogroups.com
<mailto:ug@yahoogroups.com>
*Subject:* [ap-gto] Guiding AP mounts with PHD
Howdy,

Howard and I had a chance to log in to a customer's
Mach1 mount last night to help him with some guiding
issues using PHD. He had sent his calibration graph
and it looked quite good, but he was having some
problems guiding. After doing some mechanical tests we
set up parameters in PHD to get the mount to track and
respond accurately. Since PHD is somewhat different
from MaximDL, which I use, we had to approach the
settings a bit differently.

The first step is to do a quick 2 - 3 minute Unguided
run using 1 second guide exposures, and look at the
guider graph to see what the maximum excursions are in
Declination. The guide star will bounce around a
certain amount and this peak error will be the seeing
that you cannot guide out with normal guide software.
This P-V value, which in our case was between +-0.5
and +-0.8 arc seconds is what I call the Guide Star
"Bobble" limit. Trying to correct for that with guide
moves to the mount is impossible (only a fast acting
AO system can chase that seeing error).

So, knowing that the minimum seeing error is +-0.8 arc
seconds, you will want to set the initial Min Move
setting in PHD to be approximately this value as a
starting value. We set the Min Move to 0.8 arc sec,
the guide rate at 1x and the aggressiveness to 100%. 
This means that no correction pulses are sent to the
mount while the guide star is bobbling within that
envelope, but once it exceeds even slightly, the mount
gets a full correction command (-0.8 arc sec) to bring
it back toward the zero position. We turned guiding on
and the result was that the mount responded quickly
whenever the error exceeded the bobble limits and
overall guiding was tight, accurate and almost the
same RMS value, ~0.35 arc sec, for both axes. An
exposure with the main camera showed tight round stars.

So, as a starting point:
Determine the amount of guide star bobble above and
below the axis
Set the Min Move to that value
Set Guide rate to 1x sidereal
Set the guide star exposure rate to 1 second
Set Aggressiveness to 100%
Begin guiding and note the RMS value for both axes.

You can then change any of the parameters to see if
you can tune the RMS to a lower value. We tried longer
guide star exposures, but found that the response to
tracking errors became more and more sluggish and less
accurate. So for these seeing conditions a faster
guide rate of 1 per second resulted in the lowest RMS
error on both axes. In pristine seeing it might allow
longer guide exposures, at least that is what we found
using PHD Guiding.

Rolando





Join main@ap-gto.groups.io to automatically receive all group messages.