Cone error

[glowingturnip]

Just a quick question to hopefully confirm my understanding.  I'm just trying to figure out exactly what effects cone error (ie axis of the scope not being exactly parallel to the axis of the equatorial mount) is.  I'm thinking:

- polar alignment - none, since depends on the mount being aligned only.  Perfect polar alignment should mean no field rotation even with a significant cone error, it just means the scope is pointing in a different place than it's supposed to, but is still tracking correctly in RA

- guiding - I'm thinking also no effect if there is a perfect polar alignment.  Decent guiding can be had with the guidescope looking in a slightly different place than the main scope (that's why we have adjustable guidescope rings after all), and indeed I suspect that if you had perfect polar alignment then you could point the guidescope anywhere and it would still work, right ?  Thus cone error no effect on guiding since is similar to the guidescope being off-axis to the main scope which doesn't have much/any effect

- Goto alignment - definitely does cause an effect, which is why we have 3-star align.

Does that sound right ?

[kalasinman]

Sorry, no answer, just a question. I can't understand how a 3 star alignment could fix cone error. The example illustrations I've see show perfect as a half circle resting on the horizon. If the scope is high or low to the mount, the half circle becomes stretched, or compressed out of round (hence cone error). This would mean the apparent distance between the stars would be compressed or stretched relative to the mounts tracking. Yes? I can see how a star align could fix up the goto of a star, or even a region.--Jack

[glowingturnip]

was thinking about that - actually it's amazing that goto works at all, it has to surmount cone error, polar-align error, time error, latitude and longitude error and altitude error, sounds like a lot of unknowns for a single triangulation. 

[IanL]

Just a quick question to hopefully confirm my understanding.  I'm just trying to figure out exactly what effects cone error (ie axis of the scope not being exactly parallel to the axis of the equatorial mount) is.  I'm thinking:

- polar alignment - none, since depends on the mount being aligned only.  Perfect polar alignment should mean no field rotation even with a significant cone error, it just means the scope is pointing in a different place than it's supposed to, but is still tracking correctly in RA

- guiding - I'm thinking also no effect if there is a perfect polar alignment.  Decent guiding can be had with the guidescope looking in a slightly different place than the main scope (that's why we have adjustable guidescope rings after all), and indeed I suspect that if you had perfect polar alignment then you could point the guidescope anywhere and it would still work, right ?  Thus cone error no effect on guiding since is similar to the guidescope being off-axis to the main scope which doesn't have much/any effect

- Goto alignment - definitely does cause an effect, which is why we have 3-star align.

Does that sound right ?

All absolutely correct.

Perhaps to clarify a bit on guiding, all that matters at the outset is polar alignment - if you have perfect polar alignment then cone error has no effect on guiding accuracy.  By definition, any offset in pointing between the main OTA and the guide scope is cone error in one of the two, or indeed in both of them, so it follows that such offset has no effect on guiding accuracy.

Where the pointing offset does matter is where you have less than perfect polar alignment.  If your main OTA and guide scope are pointing to the same place in the sky, field rotation will take longer to become apparent in your image (i.e. you can expose for longer and get round stars).  If there is an offset, then field rotation will become apparent sooner.  How soon depends on the offset, how badly out your PA is, etc.

I also have a set of guidescope rings, which are firmly installed in the bottom of my astro-junk box.  They can cause significant flexure between the main OTA and the guidescope and I found they cause far more problems than they solve (and I have an offset of about a degree and a half between the imaging scope and the guidescope).

I do a basic polar alignment through the polarscope with EQMOD and don't do any drift alignment or software alignment.  As noted I have a degree and a half of offset between guider and imager and at 510mm focal length with my Canon 500D I can happily take 30 minute guided exposures with no field rotation problems, so really it's only something to worry about if you actually see the problem in your images and setup!

One scope being off-axis from another has no bearing whatsoever - assuming both scopes have the same star bang in the centre of the field of view they will be as close to parallel as makes no difference.  In theory the two scopes should be pointing in towards each other by a small angle, but the distance to your target is close enough to infinity that the angle is so small as to be immeasurable in practice.

Sorry, no answer, just a question. I can't understand how a 3 star alignment could fix cone error. The example illustrations I've see show perfect as a half circle resting on the horizon. If the scope is high or low to the mount, the half circle becomes stretched, or compressed out of round (hence cone error). This would mean the apparent distance between the stars would be compressed or stretched relative to the mounts tracking. Yes? I can see how a star align could fix up the goto of a star, or even a region.--Jack

I don't think you've got that right.  Cone error does not compress or stretch anything.  Assuming you have perfect polar alignment, the mount's RA and Dec axes will always describe a circles that follow the circles of RA and Dec on the sky.  The OTA must also follow those circles regardless of how the OTA itself is aligned relative to the mount's two axes.

I always like people to imagine the most extreme example of cone error - take the OTA and stick it end on on to the dovetail saddle so it is pointing along the axis of rotation of the Dec axis.  What will the OTA do when the RA axis is rotated?  It will still follow a circle across the sky (in this case the celestial equator).  The downside is that you can't point anywhere else with the OTA arranged like that (as rotating the Dec axis will just rotate the field of view), but the point stands - the OTA always describes correct circles in any part of the sky where you can point it, but for extreme / larger amounts of cone error there will be increasingly large areas of the sky where the OTA cannot be pointed.

It is possible to model and compensate for the error in pointing once you have a three star alignment so the mount/handset needs can deal with such errors when slewing. Of course if you have really bad PA you will also have other slewing errors to deal with as the mount's axes won't follow the lines of RA / Dec but be at an angle to them.  In practice the pointing errors when slewing can dealt with by lower end mounts using algorithms to get close to the target based on the nearest one, two or three alignment points you have set rather than trying to create a complex geometry model of the mount's various errors. (See this for details of how EQMOD does it for example: http://eq-mod.sourceforge.net/tutindex.html).  Of course if you can do it the hard way and model everything (e.g. see TPoint: http://www.tpsoft.demon.co.uk/ ).

Once you've got your go-to sorted out, the mount's mechanical components take care of the rest (assuming you have reasonable polar alignment of course).

[rwg]

A one star alignment should be perfect if you are (a) perfectly polar aligned and ( no cone error - basically pinning one point on the sky to the direction the mount is pointing tells the handset all it needs to know to get the right RA/DEC for all other points.

If your polar alignment is off, but you have no cone error, a two star alignment should allow the handset to compensate - basically it can assume that the first star alignment is perfect and then when the second star doesn't quite turn out to be where expected it can calculate the offset from correct polar alignment that would lead to the observed error, and hey-presto it knows both the correct RA/DEC for the mount encoders and the polar misalignment. However, there's not enough information to work out whether the error is from polar misalignment, cone error or a mixture of the two.

Now, a three star alignment gives the mount two more pieces of data to play with in its calculations (basically how far off the 3rd alignment star is from its expected position - in DEC and RA). That gives it 6 pieces of information about the mount (3 times RA and 3 times DEC) and 5 unknowns (RA encoder position, DEC encoder position, Alt Polar mis-alignment, Az polar misalignment and cone error). So, in this last case there is enough information to work out both the polar misalignment and cone error - even if you have a mixture of both. Once the handset knows these errors it can correct for them on future gotos, etc.

At least, that's the theory...

In practice, while I was still using the handset (before going EQMOD), I found that I could do a 3 star align a couple of times in a row and get quite different values for the calculated polar alignment at the end - that was in spite of being very careful over each alignment, using an illuminated reticle eyepiece, etc. Sometimes I'd get good gotos, sometimes not, and I think it was all down to the way it calculates the cone error and polar misalignment being very sensitive to small errors in finding the alignment stars.

At this point I decided to get rid of my cone error, and it really made a big difference - accurate gotos every time. It also made aligning with tools like Alignmaster etc work better (in my opinion, anyway). I found that the procedure to measure and fix cone error was a bit of a pain, so I wrote a program to guide you through the adjustment process - ConeSharp.

So, in my opinion getting rid of cone error is definitely worth the time and trouble - everything just works so much better once it's eliminated.

cheers,

Robin

[glowingturnip]

Cheers Ian, exactly as I thought.  That helps.

I do have guide-rings now, but they're a darn sight better than before where I just had my guidescope just screwed into a single camera piggyback screw !  (should I leave the forum in shame ?) 

[JamesF]

I'm afraid I've not given this much thought myself so I'm going to just throw it out as a couple of random questions and have a think about it more this afternoon...

1) Does cone error prevent accurate drift alignment?

2) Can cone error be corrected with the OTA on a mount that is not properly polar aligned?

James

[glowingturnip]

Thanks Robin, I'll check out your program and let you know how I go.

To be honest, for my set-up, I find the 3-star so frustrating - it invariably asks me to pick stars I don't know, which turn out to be behind trees or almost at the horizon, and then fails after all that anyway.  I just do a 2-star align, which seem to let me pick much brighter well-placed stars, and then refine my pointing accuracy afterwards with stars I know near my target - works pretty well most of the time, and if not, it's usually a poor PA anyway.

[glowingturnip]

I'm afraid I've not given this much thought myself so I'm going to just throw it out as a couple of random questions and have a think about it more this afternoon...

1) Does cone error prevent accurate drift alignment?

2) Can cone error be corrected with the OTA on a mount that is not properly polar aligned?

James

just my thoughts...

1) no, cone error shouldn't affect drift alignment at all - it doesn't really matter where the scope is pointed to drift align - use a star at the meridian where the drift is greatest, and using a star near zenith and star near horizon to try to isolate the errors down to alt and az errors separately, but theoretically you could do it on any star I suppose (though pretty hard using Polaris !)

2) again theoretically, if you could guarantee using your setting circles or whatever that your scope was pointed dead on 90 degrees Dec, then all you'd have to do to check cone error would be to point at a distant object, flip the scope through 180 degrees RA and see if you're still pointing at the same point or not, same way you align the polar scope itself.  In practice though, I'd imagine small errors in lining up to 90 Dec would invalidate this technique pretty quickly.

[rwg]

You can use ConeSharp to correct cone error as long as you are roughly polar aligned (within a few degrees should be fine). The way it works is to take two readings, one just east of the meridian and one just west. Polar misalignment in altitude will have practically no effect. Polar misalignment in azimuth cancels out between the two readings completely (but cone error adds up).

Even if there is an effect it's only of the nature of 'measured cone error is a few percent more/less than actual cone error' - repeat the procedure a couple of times and you will still be able to tune the cone error down to essentially zero.

cheers,

Robin

[themos]

Hello Robin, I just tried ConeSharp with CdC (planetarium) and EQMOD Simulator. I couldn't get past Step 3, ConeSharp threw up an error dialog saying "At this step,  the mount must be on the east side of the pier, pointing west". And that was after pointing at a star with Az 189:33:00 and Dec  00:44:48 and with EQMOD reporting "PierSide: East, pointing West"

[rwg]

Hi Themos,

That's odd - I developed it with the simulator and just re-tested it and it works fine - CdC 3.10, EQMOD ASCOM Simulator 1.27L. You did 'GOTO' the alignment star rather than manually slewing there? With a manual slew it's possible to cross the meridian, but even then the PierSide in ASCOM would show 'West, pointing East'. I will send the source code over to you and then you can experiment - it's easy to remove the check on Mount.SideOfPier, but I don't understand why you would be getting a different value for that property than I do when the UI looks the same...

cheers,

Robin

[themos]

It seems that SideOfPier handling was changed after EQMOD 1.27p ( I was running 1.27f) and even before that change, there is a special setting to select different meanings of it. Seeing as many people may be running drivers that get it wrong, I would just stick to checking the Azimuth of the star and not the SideOfPier property.

[rwg]

Thanks Themos, I will remove the check on 'SideOfPier' since it seems to do more harm than good

Robin

[rwg]

New version of ConeSharp uploaded which fixes the side-of-pier issue (and source is now in Github for those who like to tinker).

http://www.sharpcap.co.uk/conesharp

cheers,

Robin

[Wirral man]

On 05/09/2014 at 13:15, glowingturnip said:

Just a quick question to hopefully confirm my understanding.  I'm just trying to figure out exactly what effects cone error (ie axis of the scope not being exactly parallel to the axis of the equatorial mount) is.  I'm thinking:

- polar alignment - none, since depends on the mount being aligned only.  Perfect polar alignment should mean no field rotation even with a significant cone error, it just means the scope is pointing in a different place than it's supposed to, but is still tracking correctly in RA

- guiding - I'm thinking also no effect if there is a perfect polar alignment.  Decent guiding can be had with the guidescope looking in a slightly different place than the main scope (that's why we have adjustable guidescope rings after all), and indeed I suspect that if you had perfect polar alignment then you could point the guidescope anywhere and it would still work, right ?  Thus cone error no effect on guiding since is similar to the guidescope being off-axis to the main scope which doesn't have much/any effect

- Goto alignment - definitely does cause an effect, which is why we have 3-star align.

Does that sound right ?

ah right cone error im on it right now thanks thats

1 off the issue list cheers

[andrew s]

On 05/09/2014 at 12:15, glowingturnip said:

Perfect polar alignment should mean no field rotation even with a significant cone error, it just means the scope is pointing in a different place than it's supposed to, but is still tracking correctly in RA

Unfortunately, there is no such thing as perfect polar alignment. Due to atmospheric refraction the altitude of the effective pole depends on the air mass you are looking through. The best we can do is aim for the mean refracted pole.

On some wide field professional telescope they can adjust the altitude of the polar axis to account for this to avoid field rotation. Probably not a real issue unless you have a combination of wide field and large image scale.

Regards Andrew