Why is there no mount with auto-polar-align?

[rogerg (Roger)]

If a digital compass or an IPhone can find north/south, why can’t an EQ mount with GPS, current date/time, current lat & long, and other computation, be made to have an integrated azmith adjustment to automatically determine north/south and rotate the azmith of the mount to perfect alignment?

Similarly: If a digital angle gauge/angle meter can find 31 degrees elevation, and an iPhone can tell us if our tripod is perfectly flat or not, why can’t an EQ mount be made to have an integrated altitude adjustment and automatically determine the correct altitude (in conjunction with GPS and current date/time data) and move elevate the mount to the correct azmith?

Is this just a case of cost?

I was surprised when the first GPS telescopes came out, as i thought this was what the “you beaut fantastic GPS!” feature would have given us. But now many years later I’m still surprised there’s no mount on the market that will auto-align its self once plonked down on the ground.

Is it just much harder than I’m thinking it is???

Roger.

[mithrandir (Andrew)]

It's another two motors, would rely on the tripod/pier head being perfectly level, known local magnetic variation, and a CCD to confirm the alignment stars were where it though they should be.

Any magnetic compass is only as accurate as the variation you feed it. A bit of reo in the concrete or a nearby steel fence throws my Silva compass off about 5 degrees. Polar alignment needs better values than "S is over thataway".

With a bit of practice it doesn't take that long to polar align, at least for visual.

[rogerg (Roger)]

Originally Posted by mithrandir

It's another two motors, would rely on the tripod/pier head being perfectly level, known local magnetic variation, and a CCD to confirm the alignment stars were where it though they should be.

I don't see that it should rely on tripod/pier level at all - the fact is you could have a level in the tripod knowing how "off level" it is, this could be taken into consideration quite easily I think.

Originally Posted by mithrandir

Any magnetic compass is only as accurate as the variation you feed it. A bit of reo in the concrete or a nearby steel fence throws my Silva compass off about 5 degrees. Polar alignment needs better values than "S is over thataway".

Interesting. I don't have an iPHone to test it, but I thought if magnetic interference would be a problem then the electronics, battery, metal in a device like an iPhone would have the same effect as reo, if it did affect the type of compass used in such devices?

Originally Posted by mithrandir

With a bit of practice it doesn't take that long to polar align, at least for visual.

Of course that's true, but the same could be said for Goto: It doesn't take much to learn star hopping to the point it's extremely efficient to find almost any NGC object, but people still like Goto

Thanks for your reply, all interesting food for thought

[bartman (Bart)]

Hi Roger,
There is a scope that auto aligns with a ccd built in.
It's a Meade ETX-LS. According to the article in Sky at Night Magazine - November 2009 - it takes about 10 mins to align and keeps objects in the 'eyepiece' for up to an hour.
I'm not sure if it would be good for astrophotography. I'd buy one if I had some spare cash!
Bartman

[mithrandir (Andrew)]

Originally Posted by rogerg

I don't see that it should rely on tripod/pier level at all - the fact is you could have a level in the tripod knowing how "off level" it is, this could be taken into consideration quite easily I think.

Mine has a level built in. If I level the tripod both e/w and n/s with a bubble level, including rotating the level 180 deg to even out any error in it, and then look at the built in level, the bubble is outside the centering circle. That can be corrected simply because the error is on the n/s line. I don't know what would need to be done if it was ne/sw.

Interesting. I don't have an iPHone to test it, but I thought if magnetic interference would be a problem then the electronics, battery, metal in a device like an iPhone would have the same effect as reo, if it did affect the type of compass used in such devices?

How accurate do you think the iPhone compass is? I would expect it would do for finding your way around when you are in effect using it as a GPS. My GPSs will all tell me which way is north, but only if they are moving and can work out the course I am on.

The fields induced by the electric currents are small are unlikely to matter. Any magnetic compass is affected by any magnetic materials around it. A fixed compass on a ship can be compensated for the metal in the ship. A portable one can't. You can never tell what will be close enough to affect it.

The iPhone level function has more chance because gravitational variation is much less, assuming a non-accelerated frame of reference.

Originally Posted by bartman

It's a Meade ETX-LS. According to the article in Sky at Night Magazine - November 2009 - it takes about 10 mins to align and keeps objects in the 'eyepiece' for up to an hour.

It is also an alt/az mount, so it is the same set of motors to align and track. The GPS tells the software where the mount is. The CCD lets the software work out where it is pointing.

All in all I don't see it as economic for an equatorial mount.

[renormalised (Carl)]

Originally Posted by mithrandir

It's another two motors, would rely on the tripod/pier head being perfectly level, known local magnetic variation, and a CCD to confirm the alignment stars were where it though they should be.

Any magnetic compass is only as accurate as the variation you feed it. A bit of reo in the concrete or a nearby steel fence throws my Silva compass off about 5 degrees. Polar alignment needs better values than "S is over thataway".

With a bit of practice it doesn't take that long to polar align, at least for visual.

Or, you could design a system much like what they have in fighter aircraft, using a gyrocompass and INS. But that would probably become too expensive to do.

[Tandum (Robin)]

I think maxpoint does something like this. It runs around taking images and plate solves them against a database. Sounds a bit like Barts Meade ETX except you need a camera and PC.

[AndrewJ]

Gday Roger

Automated "polar" aligning is relatively simple, but would be costly for a std EQ mount due to the power required for the extra motors, and possible redesign of the adjustment mechanisms to support automated work.
An IPhone etc will get you close, but for a proper align, ( do reduce DEC drift etc , you need the ability to do very fine adjusments.
Just for fun, i have started to automate my wedge, piccies attached.
I have the DEC working well, but had to redesign the AZ to reduce the power reqd to move a loaded wedge.
The plan is to write a simple application to watch a webby feed whilst drift aligning, and have it adjust the wedge.
One day

Andrew

[rogerg (Roger)]

Looks like a good step forward Andrew! I hope you'll let us know how it all goes when it's finished

[g__day (Matthew)]

Roger,

To answer your initial question - its the required precision, versus all the mechanical errors / tolerances of the interacting gear - that makes this economically unachievable.

To get a rough alignment - isn't too challenging. But to get within say 30 arc seconds of your celestial pole, and account for all the flexures and non perfect orthonogality of your OTA components, and account for seeing variations - requires human intervention and a fuzzy logic system.

Simple example - how level can you get your mount. Can you think of a way to get your mount level to within 20 arc seconds. Now also ask is the base of your mount perfectly level in all directions to within a micro metre? I can't think of how one might get the base of a mount level to better than 5 arc minutes of true level.

Next is your OTA perfectly flat on your mount, or is its light path only within 2-3 arc minutes of aligned to the mount holding it?

Do any of the components of your rig flex by more than a few arc minutes or tens of arc seconds - and does this vary with elevation of the target?

Close to isn't too hard to achieve, but the high precision and rigidity of all the interacting components required for long duration, long focal length imaging is very, very exacting.

Look at Tpoint - it may model hundred of stars to account for pointing, tracking, flexure and non-orthonogality of your rigs' components - and it has tens of interacting variables and fitting methods.

Bottom line - the precision required exceeds the economic capability / cost point of the gear amateurs consume.

Matthew

[Steffen]

Originally Posted by mithrandir

How accurate do you think the iPhone compass is? I would expect it would do for finding your way around when you are in effect using it as a GPS. My GPSs will all tell me which way is north, but only if they are moving and can work out the course I am on.

The fields induced by the electric currents are small are unlikely to matter. Any magnetic compass is affected by any magnetic materials around it. A fixed compass on a ship can be compensated for the metal in the ship. A portable one can't. You can never tell what will be close enough to affect it.

The iPhone level function has more chance because gravitational variation is much less, assuming a non-accelerated frame of reference.

The iPhone's compass will correct for magnetic declination at your location, but being a magnetometer it is still at the mercy of nearby magnetic materials distorting the field.

Therefore - compass, shmompass! What would be interesting to see is whether the gyroscopic device in the upcoming iPhone 4 will be sensitive enough to detect the Earth's rotation and hence the orientation of the Earth's axis in a reasonably short amount time. That would be the bee's knees for polar alignment. However, I don't think a gyroscope sensitive enough to pull this off with sufficient accuracy would fit inside an iPhone…

Cheers
Steffen.

[Paul Haese]

If PEMPro can get your mount polar aligned to ridiculous accuracy then it must be a simple matter of writing the code to tell several motors to adjust the azimuth and altitude on the mount. It should do this in an iterative fashion. Afterall the accuracy to guide is already present, why wouldn't it be possible to get auto alignment.

Good question Roger. I think this is possible and achievable.

[rogerg (Roger)]

Originally Posted by Paul Haese

If PEMPro can get your mount polar aligned to ridiculous accuracy then it must be a simple matter of writing the code to tell several motors to adjust the azimuth and altitude on the mount. It should do this in an iterative fashion. Afterall the accuracy to guide is already present, why wouldn't it be possible to get auto alignment.

True, and not only that - systems like the Gemini have built in modelling that can tell you (assuming they work correctly for your hemisphere ) exactly what correction to make to your Alt and Az adjustment in order to achieve perfect alignment. So using something like Gemini you in theory don't need the PC.

This would be a little more time consuming and requiring more user intervention than my initial "plonk it down and it works it out" though But close.

Rog.

[Bassnut (Fred)]

Originally Posted by Paul Haese

If PEMPro can get your mount polar aligned to ridiculous accuracy then it must be a simple matter of writing the code to tell several motors to adjust the azimuth and altitude on the mount. It should do this in an iterative fashion. Afterall the accuracy to guide is already present, why wouldn't it be possible to get auto alignment.

Good question Roger. I think this is possible and achievable.

Ive discussed this with Andrew before (yours is taking too long to make Andrew ), yes its certainly achievable and infact a fairly simple automation project with an iterative drift algorithm or probably in one hit with Tpoint.

Initially pointing south and leveling only has to be approximate if all 3 axis were contolable.

Its the cost of a 3 axis system thats the problem, not whether its possible or not.

Well, it could actually be "plonk down and it works it out", with only rough level and south pointing.

[bojan]

Why not the processor which controls the movement of AR and DEC axis simply take measurements (via camera) and apply required corrections?
Some dobsonian driver software essentially does this (by performing 3 star alignment procedure) and the whole mount (including ortogonality of the shafts) is analysed and parameters are stored.
Apart from small amount of field rotation, this should be enough.. no need for second set of motors, really.

[Bassnut (Fred)]

Originally Posted by bojan

Why not the processor which controls the movement of AR and DEC axis simply take measurements (via camera) and apply required corrections?
Some dobsonian driver software essentially does this (by performing 3 star alignment procedure) and the whole mount (including ortogonality of the shafts) is analysed and parameters are stored.
Apart from small amount of field rotation, this should e enough.. no need for second set of motors, really.

Ive been wondering about that. Tpoint makes a map and compensates for all alignment errors for pointing, but I dont think it compensates RA/DEC tracking, would seem a logical extention and not that hard. Perhaps it does, havent tested drift with an out-of-alignment mount after a Tpoint map.

[rogerg (Roger)]

Originally Posted by Bassnut

Ive been wondering about that. Tpoint makes a map and compensates for all alignment errors for pointing, but I dont think it compensates RA/DEC tracking, would seem a logical extention and not that hard. Perhaps it does, havent tested drift with an out-of-alignment mount after a Tpoint map.

Gemini does this when in "Adaptive King" tracking mode, it will use the modelling to track accurately using both RA and Dec. In practice I don't use it because I need to get rid of the field rotation anyhow.

[bojan]

Then there is really no problem for processor to report something like this:
"Move your polar axis by such-and-such to the east, this-and-that up"..
Screws with verniers on them are really not that expensive to make.

Originally Posted by rogerg

Gemini does this when in "Adaptive King" tracking mode, it will use the modelling to track accurately using both RA and Dec. In practice I don't use it because I need to get rid of the field rotation anyhow.

[gary]

Originally Posted by Bassnut

Ive been wondering about that. Tpoint makes a map and compensates for all alignment errors for pointing, but I dont think it compensates RA/DEC tracking, would seem a logical extention and not that hard. Perhaps it does, havent tested drift with an out-of-alignment mount after a Tpoint map.

Hi Fred,

I can answer that for you.

TPoint, on its own, is a pointing analysis system. As you are aware, it provides
the parameters for a pointing model.

In order to get the most benefit from a system such as TPoint, the same pointing
models need to be implemented in the telescope's control system kernel. The
control system kernel is then responsible for pointing and tracking in real time.
Apart from the corrections provided by the pointing model, the control system is
also factoring in effects such as Earth rotation, refraction, etc.

Since tracking is the differential of pointing, good pointing performance is a prerequisite
for good tracking performance if the system is tracking "blind", that is, without the
benefit of guiding on a guide star.

So for telescope's such as the AAT and the Gemini's the operators perform a TPoint
pointing analysis and the parameters extracted from that analysis are then fed into
the scope's control system kernel. Software Bisque with their TheSky and TPoint
offerings also provide a system called ProTrack which they use with their ME mounts
that provides these type of real-time tracking corrections. Our own Argo Navis
has an integrated pointing analysis system called TPAS, the outputted parameters
of which are then used in integrated models in its own pointing and tracking kernel.
So, for example, if you interface a slew and track system to it like a ServoCAT
or SiTech, it feeds them tracking rate data that is being continually corrected
in real time.

To give you an example of the 'static' behavior of a system such as TPoint
and the 'dynamic' real-time behavior of the control system kernel, consider
for one moment refraction. When inputting data into TPoint, the user has
entered their latitude as well as the current atmospheric pressure, temperature
and the computer's system has to have its time accurately set to compute the
current Hour Angle of stars. These parameters are then used to correct the apparent
position of each star owing to the effects of refraction. The control system
kernel, on the other hand, needs to be making corrections owing to refraction
to both the pointing and tracking on a continuous basis. If the temperature or
pressure are different on that night compared to the night of the TPoint run, then
the correction made for refraction at any given elevation will also be different.

As you are aware, polar alignment error in systems such as TPoint and our
own TPAS are treated just like any other source of systematic mount error.
The advantage of this approach, is that unlike a drift test, polar misalignment is
considered simultaneously with all the other errors in the system. Since the
errors are entwined in a complex and knotty way, unraveling the polar misalignment
parameters requires reasonably complex analysis techniques.

Of course, there is also no magic point in the sky to which one can align the
mount's RA axis and achieve a 'perfect polar alignment' that is then valid for all
points in the sky to which the scope can point. In other words, the best place
to align the RA axis of the scope depends on where the scope is tracking at
any one instant. A good compromise is to align the RA axis with the refracted
pole.

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

[Bassnut (Fred)]

Thanks for that Gary, and the time for such a comprehensive explanation. So thats what ProTrack does ;-).