TPAS (Argo Navis): Who's used it?

[Rodstar]

After being encouraged by another observer for some time to start using the TPAS function with my Argo Navis to improve the pointing accuracy of the Argo+Mary Rose, I spent the last few days studying the relevant pages of the Argo Navis manual, and finally worked out how to use it.

I used the less than ideal conditions tonight to do a pointing run of about 20 stars using the Mode Identify function, then let Argo do some computations.

Bottom line? A HUGE improvement in pointing accuracy, such that every object I sampled was in the FOV of the Ethos. That is much better than previously, when I would often find myself ranging around the sky trying to find an object, having "located" it with Argo.

What have your experiences been with TPAS?

[Starkler]

I had a couple of half hearted goes to see if it would help with the pointing on my gso dob. I found that I could reduce the pointing error to about half of previous, but I suspect that something wasnt quite right mechanically in my setup. TPAS can only account for variables that are constant and can be predicted.

On my SDM dob, TPAS isnt required at all as the scope points superbly without it, which goes to show what a great job Peter Read did with the construction geometry of my scope :D

[gary]

Hi Rod,

Great to hear you managed to get out and perform a TPAS run.
Do you happen to recollect what your RAW and FITTED RMS were
on the 20 stars? These are interesting figures to know and you
can review them by scrolling to the end of the REVIEW DATA submenu.

Now you are familar with TPAS, you could actually now hold down your
end of the conversation with the engineers that are responsible for the
pointing of scopes such as the Kecks and Geminis. For example, if you
mention your Collimation in Azimuth (CA) term was so and so, or
your fitted RMS or PSD was such and such, they would know what you
are talking about because the same techniques and language is used
by the professionals.


Best Regards

Gary Kopff
Managing Director
Wildcard Innovations Pty. Ltd.
20 Kilmory Place, Mount Kuring-Gai
NSW. 2080. Australia
Phone +61-2-9457-9049
Fax +61-2-9457-9593
sales@wildcard-innovations.com.au
http://www.wildcard-innovations.com.au

[ausastronomer]

Hi Rod,

Good to hear you got your head around it. It can be a little daunting at your first attempt.

With any "hand made" telescope, regardless of how well it is made, there will always be "fabrication errors". Particularly, when it is made from wood. As you have noted, TPAS improves the pointing accuracy dramatically.

It is also worth mentioning if you are limited for time, you can do a short pointing run of about 5 stars, carefully selected, just use the Index Error (IE) term only and pointing accuracy will usually be improved to put everything into the FOV of a medium power eyepiece. Clearly, the larger the group of stars sampled the better the pointing accuracy attainable, but it is comforting to know that this shortcut offers a good improvement, if you don't have much time.

Cheers,
John B

[astroron]

I find the accuracy of the ARGO NAVIS is as good as I require.
Most objects are in the field of my 13mm nagler and most times in the field of view of my 9mm plossl.

[Rodstar]

Thanks for all of those comments, everyone.

It is incredible that the pointing accuracy of various scopes varies so greatly. Geoff and Ron, you must have very precisely manufactured gear not to need TPAS! I have been getting quite frustrated with pointing errors, particularly near to zenith.

Gary, the raw RMS was 9.1 and computed RMS was 3.1. All of the various terms that can be used with an alt/az set-up showed a significant improvement, so I employed all of them. As for individual stars, all were improved, some quite dramatically. The widest raw figure for a star was about 36 arc minutes, and the widest computed figure with the new terms was about 15 arc minutes, from memory. The closest computed figure for a star was a matter of about 24 arc seconds!

Gary, I have a question which others might benefit from the answer to. Because the terms were so helpful, I saved them all, ie chose "Use and Save". Next time I fire up the scope, and I do another star sample, what will Argo be comparing my new sample to? Will it be the saved terms, or zero terms? Do I change the terms from the saved ones from last time to the new terms, in the same way as I did last time by overriding zero terms with the computed terms? What if I want to return to a zero term - do I just opt not to use the term at all?

One other question, while I am peppering you with Argo questions....It appears that my fixed alt position is about 90.41 degrees. I assume if I plug that in, I also switch the auto alt ref function off?

[Rodstar]

Thanks to John (Ausastronomer) for giving me the low down re TPAS on the phone this morning.

It may well be that when my scope fell over shortly after I first got it, the alt encoder (which came off - attached the arm) was not put back on properly. This is no doubt adversely affecting the natural pointing accuracy of the scope. John is going to drop in during the week to have a look, and see if we can't get the numbers down even further.

Sounds like I should not be using all of the terms, too. For example, I undertand now that in most circumstances one does not use ECES and ECEC at the same time, and that IE, CA and NPAE vary between sessions and should not be saved.

In relation to the Fix Alt Ref Auto function, I understand that the TPAS IE function renders this entirely otiose, so I will set this to OFF from now on.

Hope I got that right?!:shrug::)

[ausastronomer]

Hi Rod,

You got it "almost" right.

In regard to this quote, "I undertand now that in most circumstances one does not use ECES and ECEC at the same time, and that IE, CA and NPAE vary between sessions and should not be saved."

What i said was this:-

Irrespective of your telescope type or mount type, never save IE (Index error in elevation). It is not repeatable from one session to another. It is like a "super" fix alt ref and renders the altitude reference angle term redundant, if you use a TPAS model.

In terms of the other errors, I said whether you saved them or not was entirely dependant on your mount type and what was causing those errors. I said that with a "truss" dob, with a "sling" type mirror cell it was not a good idea to save NPAE (non perpendicular axes) or CA (collimation error) because each time the scope was set up the truss poles can bed slightly differently from session to session. In addition the mirror can settle differently in the sling making these errors "non repeatable" from session to session. I mentioned that ECES and ECEC "could" be saved from session to session, again dependant on what was causing those errors. For instance if ECEC (Elevation Centering Error Cosine) was your most significant error term and caused by something like your altitude encoder being improperly centered on the altitude bearing axis of rotation, it would be a repeatable error and the term could be saved. I also mentioned that I didn't save any terms with my truss dob.

With a scope like a CNC machine made tubed dobsonian, or an equatorially mounted scope, things are somewhat different as there is a lot more repeatability in the setup, in regard to "some" of the terms.

Cheers,
John B

[gary]

Hi John,

Thanks for the post,

A little background as to what is going on here might be helpful.

As you know, the encoders used are incremental encoders and have no
zero reference point. As it turns out, one needs to establish the zero reference
point for at least the Alt encoder. This is the job of the FIX ALT REF step.

FIX ALT REF has a powerful feature called AUTO ADJUST, whereby after
performing a two-star alignment, the ALT REF point is automatically adjusted
based on the information obtained from aligning on those two stars.

When one uses TPAS and the mount type is set to ALT/AZ DOBSONIAN,
something interesting happens when at least four stars have been sampled
and the IE term is set to COMPUTE. The IE term is the Index error in Elevation
term and is analogous to what AUTO ADJUST would normally do. What happens
is that rather than adjusting the Alt encoder zero reference point based on
the position of just two stars, one is now using all four stars. As the number
of sampled stars increases, all can be used to contributing to the establishment
of the Alt encoder zero reference point. In other words, TPAS effectively
provides a "super auto adjust" function.

In practical terms, as you point out, sampling just a small number of stars,
say four or five and just fitting IE can make a dramatic improvement to the
pointing performance of some scopes.


Best Regards

Gary Kopff
Managing Director
Wildcard Innovations Pty. Ltd.
20 Kilmory Place, Mount Kuring-Gai
NSW. 2080. Australia
Phone +61-2-9457-9049
Fax +61-2-9457-9593
sales@wildcard-innovations.com.au
http://www.wildcard-innovations.com.au

[h0ughy]

I watched Gary and Ed Roca with his 18" go through this at ISSAC in 2006 - the more they did the more accurate it got. Really amazing stuff - very clever man that one who sits in the stream!!

[gbeal]

A great thread, and one which I am sure others could benefit from, me especially.
Any chance of a law-mans tutorial for those who haven't managed to make it all as simple as those that have seem to suggest, please.

[Rodstar]

Thanks to John for coming over last night and checking my altitude encoder.

John set up a laser pointer directly at the centre point of the encoder, and then moved the scope up and down. We noticed that the light point did not remain fixed on the centre of the encoder, and in fact it moved about 5mm to the right when moving the scope tupe from the upright position to the horizontal position. That being the case, the encoder was obviously not square. John tweeked the position of the encoder until the laser point was constant through full motion.

Unfortunately the skies were heavily clouded so I was not able to test the results of John's handiwork. It will be interesting to do another TPAS run in the next few days and see how the pointing accuracy is improved.

[Rodstar]

Gary,

I have held back waiting for someone else to jump in who is more proficient with TPAS than I.

Perhaps it would help to know what specifically you need to know, or if you just want a general run through the procedure? The Argo Navis manual has full instructions, and I was able to mostly work it out just by following them. My difficulty is with working out the best combination of terms, and I will no doubt benefit from further input about that as I make more attempts with TPAS pointing runs.

I am happy to give a summary of how to commence a pointing run, collecting the data as a starting point if you think that would be helpful? :shrug:

[gbeal]

Rod,
any help is appreciated. I am not a complete dummy, but the whole TPAS thing has eluded me, and in essence I guess it would be easier to be "shown" rather than read it and try. The Argo is all but in operation on the new 16" truss, and once it is up and running I will likely run a trial to see if I can help the pointing. Any improvement in accuracy is beneficial.
Given the wide range of Argo converts on this forum a simple "do this, then this" tutorial could be good perhaps?
Gary

[Rodstar]

Gary, it is really interesting to see how many very experienced and capable amateurs have baulked at using the TPAS technology. Based on the very limited survey to date, it seems only about 1/7 of those with Argos are actively using TPAS. That may be food for thought for Gary and Mai, who I am sure would love to see more people using the TPAS function, and using it with confidence. To my mind, it is a huge technological advance, once understood.

I am happy to prepare a basic intro, but I am going away tomorrow for 5 days, so it will not be posted until mid next week. If anyone pips me to the post, I will not mind!!!

[gary]

Thanks for the posts.

The most important function of any telescope pointing device is
to assist the user in achieving their own stated pointed
performance goal.

For example, for many visual observers, the pointing goal might
be articulated to be to drop the object within the FOV of the
eyepiece of choice across the whole sky. For imagers, the goal
might be to blind point the scope so as to deliver the object
onto a small CCD array.

When going from the cataloged co-ordinates of some object, such
as a star or galaxy, to where the telescope computer 'thinks' it
is, requires a complex series of transformations. For example, the
effects of precession and nutation of the Earth, atmospheric
refraction and so on result in the apparent place of an object
being different to its cataloged position. Another potential
source of error is the various geometric, gravitational flexure
and eccentric bearing errors within the mount/OTA. Devices such
as mechanical setting circles and Digital Setting Circles (DSC)
do not take effects such as precession, nutation and refraction
into account and any geometric error within the mount compounds
the problem so that the pointing error residual might become
significant enough that the system does not achieve the
end-user's pointing goal.

Argo Navis is a Digital Telescope Computer (DTC) rather than a
DSC. This is an important technical distinction rather than a
marketing one.

Unlike a DSC, the Argo Navis DTC takes into account effects such
as precession, nutation and refraction and through its
Telescope Pointing Analysis System (TPAS) can assist the user in
helping to diagnose and potentially compensate for some of the
more common systematic errors that occur within a mount/OTA.

All mounts, even the Kecks and Gemini telescopes, have geometric
fabrication errors to some extent. For many amateur enthusiasts, the
errors are sufficiently small on their telescopes that they achieve their
pointing goals without needing TPAS. Therefore, like a cure for
some ailment, TPAS is certainly does not need to be prescribed
to everyone. However, for those with chronic systematic pointing
problems, TPAS can often be a great benefit.

Over the coming weeks I will attempt to make some posts walking
one through the various steps of TPAS.

Also be sure to attend my talk at the forthcoming South Pacific
Star Party where you can learn more about telescope pointing.

Best Regards

Gary Kopff
Managing Director
Wildcard Innovations Pty. Ltd.
20 Kilmory Place, Mount Kuring-Gai
NSW. 2080. Australia
Phone +61-2-9457-9049
Fax +61-2-9457-9593
sales@wildcard-innovations.com.au
http://www.wildcard-innovations.com.au

[ausastronomer]

Hi Rod/Gary Beal,

I will write something up over the next few days in respect of dobsonian mounts.

Cheers,
John B

[gbeal]

Thanks guys. I had to ask, as I am sure there are many who also don't fully understand it. Look forward to whatever you can offer up.
Gary

[gary]

Whenever one installs encoders on a mount for the first time, it is
highly advisable to perform what we call the Daytime Encoder Test.

If you are considering performing a TPAS run for the first time and have
never performed a Daytime Encoder Test, now is the time to consider doing it.

Power the Argo Navis unit OFF.

Using a high-powered eyepiece or reticle eyepiece, sight and center a distant
terrestrial object. For example, I often like to use power pole insulators.

Now power ON the unit.

DIAL up MODE ENCODER and press ENTER.
Spin the DIAL until it shows AZ/ALT ENC STEPS.
The two values shown in the lower line of the display should both be zero.

Now rotate the mount in Az (or if it is an equatorial, lock the Dec axis and
rotate the mount in RA) through 360 degrees and reacquire the target.

Note the left-hand displayed value. It should be ideally within a step
or two of the original zero reading. Keep in mind that if you have, say,
10,000 step encoders, when the value reaches 9999 it will wrap back to 0000.

Once satisfied, try rotating the mount again in Az but in the counter direction
to what you previously did. When you re-center the target, ensure that the value is
again within a couple of encoder steps.

Now try swinging the scope in Az back and forth a few times, to try and
induce an encoder slip if something mechanical is amiss, such as a loose
set screw. Again re-center the target and check that the Az encoder value
on the display is within a step or two of zero.

Now do the same procedure, but this time on the Alt (or Dec) axis.
If you have a Dob, obviously you will not be able to rotate the tube through
a full 360 degrees. Instead, rotate the tube all the way through its full travel
above and below the target before re-centering it, this time taking note
of the right-hand displayed value, which should be within a step or
two of zero.

Sometimes when we ask users if they performed a Daytime Encoder Test,
they respond, "Sure, I did it during the night". :doh:

The advantage of performing the test during the daytime is that it is
easier to find a fixed object. Stars won't do as they have apparent motion
due to the Earth's rotation. (Northern hemisphere observers could use Polaris
in a pinch, but ...) The other advantage of performing the test during the
daytime is that if something is amiss mechanically, you can literally get
down on hands and knees and inspect up close the encoder installation.
At night, it is dark, often cold, damp and sometimes you are tired.
During the day with the benefit of sufficient light and a clear head, it is
often easier to spot something that is going wrong.

If the mount does not pass the test, then you have some work to do
to find the cause. Often it is a loose set screw or an encoder not mounted
to coincide with the axis in question. Also make sure that the tangent arms,
which are the brackets that go from the encoder to some fixed part of the scope,
typically the side of the rocker box on a Dob, don't rotate around the axis
of rotation themselves. They don't need to be fixed rigid at the far end and
can be allowed to 'float', by cutting a well toleranced slot into the end of the
arm and sliding it onto a locating pin. Though they can be allowed to move
in or out radially from the axis or up and down away from the side of the
rocker box, they should never rotate or shift around the axis of rotation,
as in doing so they are providing for a positional uncertainty in the encoder system.

For portable scopes where removing the Alt encoder is required to transport the
telescope, it is worth repeating the Daytime Encoder Test now and then.

Once confident of the integrity and repeatability of the encoder installation,
one is then ready to perform one's first TPAS run.

The first part of a TPAS run is acquiring data and that will be covered in
the forthcoming Part 2 of this tutorial. :thumbsup:

Best Regards

Gary Kopff
Managing Director
Wildcard Innovations Pty. Ltd.
20 Kilmory Place
Mount Kuring-Gai NSW 2080
Australia
Phone +61-2-9457-9049
Fax +61-2-9457-9593
sales@wildcard-innovations.com.au
http://www.wildcard-innovations.com.au

[manxtriton]

Hi Gary,
Have you written the second part of this TPAS tutorial? I can't find any reference to it on the site,
kind regards
Clive,
Warwick,
England

[ausastronomer]

Hi Clive,

The term "forthcoming" means it will be available in the near future. I would suggest Gary will post it to this thread within the next 2 weeks or so.

Cheers,
John B

[gary]

Hi Clive,

Greetings to you there in the U.K.

Great to see you read Part 1 of the TPAS tutorial, for which I thank you.

Part 1 was only posted five hours and eight minutes prior to your question, so
you got to read it 'hot off the presses' as they say and I wouldn't be surprised if
one were to run one's finger across the computer screen, the ink probably still isn't dry. :)

I anticipate you will see Part 2 in the days and weeks ahead. This is a
work-in-progress being delivered online in next-to-real-time.

So thanks again for your interest and stay posted for the next installment!

Best regards

Gary Kopff
Wildcard Innovations Pty Ltd.

[goober]

I had a first play with TPAS tonight with my alt-az mount, sampling a dozen stars. The best combination of terms was IE (always use this, I assumed) and NPAE (-16.2 +/- 4.2). It gave me a fitted RMS of 4.5' and PSD of 5.9'. For 5000 step encoders, I think that's near the limit of resolution. Adding other terms, especially CA, blew out the results.

After the sample, I tried the model out with my 154x/32' eyepiece, and everything hit bang on, well within the field of view. Couldn't fault it.

Of course, then I powered off the Argo and lost everything. Just to compare, I did a two star align and went around the objects again without TPAS, and the hit rate was just as good as with the TPAS model. Go figure :)

The other strange thing was I went through Gary's FIX ALT REF test to get an accurate fix on my docked scope. At rest my scope points slightly down, and I had set it an angle of -12.67 degrees (read it somewhere). Went through the alignment, and put the scope back at docking position, go into MODE ENCODERS. My angle now reads +7.63 degrees! Not sure what happened there, but I changed my FIX ALT REF to that and it seemed to work fine.

[gary]

Hi Doug,

Thanks for the post.

Great to hear you are out with the Tele Vue and Argo Navis.

It is always interesting to note what your RAW RMS point performance was
by way of comparison. If you have sampling data still in memory, If you then DIAL
up SETUP MNT ERR, REVIEW DATA and spin the DIAL clockwise until it reads
END OF DATA, the bottom line will show the RAW RMS, designated with
a capital Greek Delta symbol (shaped like a triangle). If you then press ENTER
and spin the DIAL one click, you will see DELTAS=FITTED. Press ENTER again
and the display will show the FITTED RMS for that data, designated by the
lower case Greek delta symbol (shaped like a stylized d). Using the REVIEW
DATA menu and the ability to change between RAW and FITTED display mode,
it is sometimes interesting to scroll back and look at the error residuals of
the individual sampled stars.

The bottom line is that the results TPAS reports are "as it really is" in a statistical
sense based on the data that you have collected by sampling.



I gather you were using AUTO ADJUST ON when you performed the FIX ALT
REF step and two-star alignment?

The ALT REF point is with respect the mount. For example, visualize the mount's
Az axis standing vertically. The zero degree ALT REF point is when the OTA
is 90 degrees away from this vertical Az axis. If the mount happened to be
level, this would correspond to when the tube were level and pointing
toward the horizon.

If the mount were level and the tube in the lowest 'rest' position looks below
the horizon, the ALT REF point at that 'rest' position should indeed be
a negative value.

The AUTO ADJUST ON feature is very tolerant to the initial guess of the
ALT REF point. It simply takes the initial value as an estimate and after the
two-star alignment, adjusts it on the basis of the expected separation of the two
alignment stars compared to the angular separation was based on the encoder motions.
When picking the two alignment stars, make sure the two not only have
differing Azimuths, but differing Alts as well.

Once the encoder installation has been verified by the Daytime Encoder
Test and ballpark pointing established when using AUTO ADJUST OFF,
we recommend to then use AUTO ADJUST ON on subsequent observing
sessions.

After performing the FIX ALT REF with AUTO ADJUST ON and the
two-star alignment, if one then brings the OTA back to the rest position
(commonly the vertical stop on a Dob) and DIALS up MODE ENCODER,
it is instructive to note the right-hand displayed ALT encoder value.
This should approximately correspond to the ALT REF point that AUTO ADJUST
has computed.

AUTO ADJUST ON can do an amazing job of refining the ALT REF point.
However, it only uses the two alignment stars as inputs into this process.

When one samples at least four stars and fits a model using the IE
term (which should always be used) and then puts that model in use,
TPAS now takes over the job of AUTO ADJUST ON. Rather than just use
two stars to determine the ALT REF point, TPAS uses as many stars
as you have sampled and will take into account any other error terms you have
defined into the model as well. In other words, the IE term, which stands
for Index Error in Elevation has an association with the ALT REF point.
TPAS with the IE term in place therefore becomes a 'super AUTO ADJUST'
feature.

Thanks again for the post.

Best Regards

Gary Kopff
Managing Director
Wildcard Innovations Pty. Ltd.
20 Kilmory Place, Mount Kuring-Gai
NSW. 2080. Australia
Phone +61-2-9457-9049
Fax +61-2-9457-9593
sales@wildcard-innovations.com.au
http://www.wildcard-innovations.com.au

[goober]

Thanks for your reply, Gary. Very interesting.

I did note down the raw and fitted RMS values, but they're at home. They certainly did change around a bit as I used different combinations of terms.

I did complete a successful daytime encoder test before the FIX ALT REF check. I went through the steps with AUTO ADJUST ON. That's why I was suprised to get a value returned from MODE ENCODERS of +7.6 degrees (plus???!!!) even though the scope is angled "down" below the horizon at rest. Perhaps my back yard is sloped!

I replaced my initial FIX ALT REF value with this +7.6 degrees in the MODE SETUP - SETUP ALT REF. Should I do this, or should I always leave AUTO ADJUST ON (which seems to work fine from my vast experience of three sessions).