SCT - Tracking analysis in different quadrants of the sky

[g__day]

I decided to do my last play with MaxPoint before I swapped to Tpoint. I improved the backlash in my DEC gears then did a three star alignment on the SkySensor2000-PC (Fomalhaut, Dipha and Rigel Kentarus) - which I re-tested and found very accurate gotos.

So I ended up performing a 162 star alignment and saw very interesting behaviours. I slowly swept around the sky - starting in the South East and heading North.

By 30 stars the pointing error (corrected) was an impressive 35 arc seconds, but as I slowly swept around to different quadrants of sky the tracking each new quadrant seemed off by a consistent amount in a new direction. By 100 stars the pointing error (corrected) was 1.8 arc minutes. It stayed like this up to 162 stars covering the full night sky.

The tracking errors seem quadrant based to me (forming almost a H pattern). So I am pondering if this is most likely a slight mirror shift (using a C9.25 SCT) in each new quadrant of the sky?

The raw pointing in each quadrant of the sky seems very consistent. I am hoping that Tpoint is smart in how in addresses point errors - by adjusting for pointing by deciding the correction required using those points closest to the target. Is this the case anyone?

Thanks, Matt

[frolinmod]

You can go direct to the Tpoint author and ask him. Just register over at the Software Bisque support forums at www.bisque.com and post your question in the TheSkyX Pro forum. Patrick Wallace will likely answer in person.

[gary]

Hi Matt,

TPOINT doesn't work in that way.

Firstly, one should separate the function of TPOINT, which is used for
data analysis and model creation, from a "pointing kernel", that can accept
TPOINT model terms and apply corrections to where the telescope points.
In other words, TPOINT does not have a pointing kernel, that is the job
of some other piece of software, specifically the piece of software that
is responsible for pointing the telescope.

Once you put a pointing model in place, a pointing kernel that accepts a TPOINT
pointing model makes no reference to the original pointing data used to create
that model.

To appreciate the mechanism, the pointing data is analyzed to create the
terms for various equations that are used to model some specific systematic
mechanical "fabrication error" of the mount/OTA. Such "fabrication errors"
also include polar misalignment. Once the terms are created, the job of the
original pointing data is done and in order to apply pointing corrections, the
pointing kernel has in-built into it essentially the same set of equations that
TPOINT used.

The TPOINT philosophy is that each of the equations models the behaviour
of some particular mechanical fabrication error of the telescope. Some of these
errors will be constant in magnitude and direction across the sky. Others, such
as Non-Perpendicular Axis Error (NP), are not. So if a term such as NP is used, then
the corrections applied might be larger in some part of the sky compared
to others. So the pointing kernel can apply a correction specific to the area
of the sky that you are currently pointing, but the magnitude and direction of that
correction is a result of the sum of all the equations in the pointing model rather
than by referencing some earlier pointing data.

TPOINT and for that matter any other pointing analysis software can only
analyze and correct for systematic errors. A sudden mirror flop is an example of
a non-systematic random error.

By the way, the two clumps of data to the East and West in your scatter diagram
hint most likely to something such as a non-perpendicularity between your optical axis
and Dec axis - what TPOINT refers to as CH - a common fabrication error in most GEMs.

Hope this is helpful.

[hotspur]

Matt.

Try page 67 in your SS2k guide,re motor parameters,there is a setting 1 to 3 re motor shudders,I am trying to get my SS2K auto guiding,but having a job getting things to run smoothly.Have noticed the dec 'shudders' periodically-its apparently caused by normal variations in the motor characteristics-so the manual says.

We did extensive testing and tweeking last night,and although the PHD graph was very good,occasionaly there was a dip in dec,and we are wondering if this was a tiny motor 'shudder' that slightly throws guiding and ending up with 'eggs'

I am going to adjust the motor parameter (dec)from 1 up.and see if that makes any difference.So seeing that your using a SS2K and its motors-might be something to look at.

Look forward to seeing how this goes for you.

cheers Chris

[g__day]

Ernie,

Yep - I have had many conversation with the Bisque boys over the years!

Gary, I'm sure you are correct!

Thanks for that - I abreviated because I use the Sky6 PE now with Tpoint add on - and Tpoint states (pg 19 of the user manual) "You must map at least six points before Tpoint is able to perform a "fit" and begin to improve the pointing of the telescope". I took the author's own words that Tpoint improves the pointing as understanding it does it directly, but I really don't mind if its doing it indirectly - such so long as it's done!

Even if Tpoint only shows me how pointing errors vary across the sky - that'll give me leads to what challenges I face. So if it is mirror shift I can either try and mod the scope to remove it, change scopes, live with it or try and model it out if that is possible.

So tonight (weather permitting) I'll leave MaxPoint in the loop - but turn off its Apply Correction Model check box - and see what Tpoint has to tell me about the pointing I see happening!

Chris,

I always run the SS2K with the motors set to 3 (as I'm pushing 3 scopes and Losmandy bars). I set backlash in RA and DEC to zero and up PHDs Hysterisis to about 25% and guiding aggressiveness down to 65% I'd rather under correct than chase the seeing!

As a final SS2K related tip - when you select 3 approriate guide stars:

1. I normally try to avoid anything near the SCP
2. Avoid a meridan flip
3. Repeat the process twice - on the exact same stars - before you leave the process and check your alignment parameters.

That last point was a chance discovery of mine - it seems to improve the pointing discernibly for me.

Oh - and the DEC shudder - if you have placed the mount in Polar aligned mode (or maybe) if you have simply successfully aligned on 3 guide stars - the SS2K adjusts tracking on both motors simultaneously to adjust for polar mis-alignment (and refraction and non sideral movement rates). So if you are out of polar alignment the SS2K may be issuing periodic corrections in DEC (giving you a saw tooth pattern). I hope it doesn't - i.e. that it does small corrections every few milliseconds - not a significant correction every few seconds - love to know the answer to that one!


* * * * * *



All suggestions greatly appreciated BTW!

Matt

Finally

[gary]

Thanks Matt,

In the case of TPOINT with TheSky6, the pointing corrections will be performed
by the pointing kernel within TheSky. But as a user you don't necessarily need
to know this as TheSky integrates TPOINT nicely and what piece of software
is doing what is therefore opaque.

Where the "at least six points" requirement comes from is that by default,
TPOINT with TheSky automatically applies the terms ID, IH, MA, ME, CH and NP.
Just as you require six unique equations to solve six unknown variables, you need
at least six data points to make an estimate of solving these six terms. Strictly
speaking, you could get away with four if you turned off NP and CH, which
I don't recommend you do. ID, IH and MA and ME should always be used.

Recommend you try the "Suggest" button feature for adding additional terms if you have
not done so already. There can be subtle things such as gear eccentricities that
TPOINT's richer set of terms can often pick up that aren't available in the
package you were using previously.

[frolinmod]

I recommend you upgrade to TheSkyX and the Tpoint Add On if you're still using TheSky6 and Tpoint for Windows. The Tpoint Add On is a newer version of Tpoint and TheSkyX uses it better than TheSky6 did.

[g__day]

Ernie,

Why specifically - given my situation what is the delta you see worth spending all that $ and moving to a subscription based payment system? Are you suggesting pointing and tracking is far better under the latest version - because that is my focus at the moment?

I expect to find that there is a segment in the sky where my pointing suddenly steps down a notch - figuring out why will be an important part in my learning curve. Resolving this may give me all sky pointing around 30 arc seconds - which I'd be quite happy with on this gear.