Jump to content


Photo

EQMOD Polar Alignment vs Celestron ASPA precision

  • Please log in to reply
24 replies to this topic

#1 Moromete

Moromete

    Viking 1

  • *****
  • Posts: 665
  • Joined: 15 Feb 2012
  • Loc: Romania

Posted 28 August 2013 - 02:04 AM

How precise is EQMOD Polar Alignment routine compared to Celestron's ASPA (done after a good 2+4 alignment)?



I'm refering to this EQMOD routine http://www.youtube.c...h?v=FZgPZnC7zKc

I would like to answer someone who has experience with both alignment methods.

#2 rmollise

rmollise

    Hubble

  • *****
  • Posts: 15558
  • Joined: 06 Jul 2007

Posted 28 August 2013 - 08:03 AM

It's in the ballpark, but the ability to do repeated iterations with the AllStar firmware (or the similar routine now in the SynScan HCs) puts it on top. That said, I can easily do three-four minute exposures with the EQMOD routine. Might be able to go longer, but my limit is usually 3-minute subs.

#3 ghataa

ghataa

    Viking 1

  • -----
  • Posts: 625
  • Joined: 20 Jun 2011
  • Loc: Central, NJ

Posted 28 August 2013 - 10:14 AM

My experience is the same as Rod's. Both are good but the ability to iterate with the ASPA enables refinement. That said, the EQMOD PA is quite good and can get several minute exposures (ie 7-8 min at 600 mm).

George

#4 Moromete

Moromete

    Viking 1

  • *****
  • Posts: 665
  • Joined: 15 Feb 2012
  • Loc: Romania

Posted 28 August 2013 - 10:34 AM

Thanks Rod and George for helping and sharing your experiences!


I thought that Celestron's ASPA is more accurate because of 3 reasons:
- you use it at much higher powers (when a reticle eyepiece or CCD is implied) than EQMOD Polar Alignment done through the lower power polar scope => lower polar alignment error
- with ASPA polar alignment error is computed better after an alignment on 2+4=6 stars, not only on 3 stars like with Skywatcher HC which seems to have still a broken firmware 3.35 in repesct of ASPA and GOTOs in AZ mode
- the model printed on Skywatcher/Orin polar scope lens I thought is not very precise, as is the case with Takahashi/Vixen/Losmandy etc. polar scopes.




Question: has EQMOD a software routine similar to ASPA but based on its N-Star alignment and not on mount's polar scope?

#5 tjugo

tjugo

    Viking 1

  • -----
  • Posts: 956
  • Joined: 06 Nov 2007

Posted 28 August 2013 - 10:44 AM

Hi,

I completely agree with Uncle Rod.

The EQMOD routine 'just' eliminate the guess work. For me is good enough for short focal lenght guided imaging.

ASPA is better.

There is no EQMOD routine equivalent to ASPA, however there are some cost effective solutions to polar align your GOTO mount, for instance alingmaster is very good, not expensive and as easy as ASPA.

Cheers,

Jose

#6 Stew57

Stew57

    Mercury-Atlas

  • *****
  • Posts: 2537
  • Joined: 03 May 2009
  • Loc: Silsbee Texas

Posted 28 August 2013 - 10:44 AM

I know some will argue but here is the question I just asked the celestron engineer that writes the software.

"Is the DEC offset correction computed from the 2 alignment stars, or does the calibration stars correct for both DEC offset and cone error? Are the calibration stars weighted as an average of the 4 calibration points, or are they factored in individually?"

His reply

"I would need to check but I think the DC offset is calculated after just the 2 stars. If you don't fancy perusing the equations directly, you can find out by clearing the DEC offset (factory reset), and doing a 2+0 align and then check to see if the DEC offset is non-zero.
The calibration stars are an average. They "calibrate" the cone error. That is all. Whether you do 2+1 or 2+4 you effectively have a 3 point alignment model."

Maybe there is a trade secret that he is not revealing but every time he always answers that the 4 calibration stars are averaged as 1. Maybe I am misunderstanding but it does not look like a 2+4 = 6 star alignment.

#7 SkipW

SkipW

    Ranger 4

  • -----
  • Posts: 337
  • Joined: 03 Feb 2011
  • Loc: Oklahoma, USA

Posted 28 August 2013 - 09:35 PM

After reading an earlier discussion, I factory reset my ASGT and ran a two-star alignment with the dec index marks intentionally off so there is a non-zero dec index error. Dec index and cone were both reported as zero. Repeat, but this time adding one cal star. Dec index and cone both had non-zero values after adding only one calibration star.

I don't think there's enough information available to solve for polar misalignment and cone or index errors after only two alignment stars. I was a bit surprised that both errors could be calculated with only one additional data point, but apparently they can.

#8 Stew57

Stew57

    Mercury-Atlas

  • *****
  • Posts: 2537
  • Joined: 03 May 2009
  • Loc: Silsbee Texas

Posted 28 August 2013 - 09:53 PM

Thanks that is interesting skip. I wonder how the mount can point with a 2 star align without compensating for the dec offset. Did you add additional calibration stars seeing how the values change?

I am planning on exploring this some time but am away from my mount for a good while due to family medical problems. Anyway thanks for the info.

#9 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3923
  • Joined: 12 Oct 2007

Posted 29 August 2013 - 01:21 AM

Yep. And after each added cal. star the dec. offset and cone will change. And if you power off and restart they will still be there and help with your pointing if you then do a 2+0. And if you go in and change the dec. offset a bit, your pointing will then be worse after a 2+0 - even if the cone stays the same. And the pointing error won't be a simple shift in dec. - it will be complicated.

Many things like this may not be in a manual or on a web page, but can be deduced from the underlying principles, and then confirmed by simple tests like these.

Frank

#10 Moromete

Moromete

    Viking 1

  • *****
  • Posts: 665
  • Joined: 15 Feb 2012
  • Loc: Romania

Posted 29 August 2013 - 02:35 AM

It's such a pity EQMOD doesn't have a software routine similar to ASPA but based on its N-Star alignment considering how much powerful a PC is compared to the HC.

I think such a routine could give a lower polar alignment error than Celestron ASPA considering EQMOD N-STAR alignment can calculate the position of much more than 2+4 stars, as Celestron HC does.


Can someone ask the EQMOD developer if he can/wants to develop a polar alignment routine based on his N-STAR alignment?

#11 cn register 5

cn register 5

    Viking 1

  • -----
  • Posts: 760
  • Joined: 26 Dec 2012

Posted 29 August 2013 - 02:42 AM

After reading an earlier discussion, I factory reset my ASGT and ran a two-star alignment with the dec index marks intentionally off so there is a non-zero dec index error. Dec index and cone were both reported as zero. Repeat, but this time adding one cal star. Dec index and cone both had non-zero values after adding only one calibration star.

I don't think there's enough information available to solve for polar misalignment and cone or index errors after only two alignment stars. I was a bit surprised that both errors could be calculated with only one additional data point, but apparently they can.

If you look at the alignment maths there is enough information to get the polar axis misalignment after two stars but it has to assume a value for the Dec axis and cone errors. After all the fork mounts can get a polar axis misalignment and they only have the option to use two stars. The same experiment will show this, after two stars you get a polar axis error.

If you want to do an ASPA after two stars use a star on the same side of the meridian as the two align stars.

Chris

#12 cn register 5

cn register 5

    Viking 1

  • -----
  • Posts: 760
  • Joined: 26 Dec 2012

Posted 29 August 2013 - 02:44 AM

Can someone ask the EQMOD developer if he can/wants to develop a polar alignment routine based on his N-STAR alignment?

YOU can ask the EQMOD developers, they have a Yahoo group for this sort of thing.

Chris

#13 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3923
  • Joined: 12 Oct 2007

Posted 29 August 2013 - 03:02 AM

If you want to do an ASPA after two stars use a star on the same side of the meridian as the two align stars.



This will help, but another fun fact overlooked, that applies to all mounts and is evident in the math: cone affects pointing everywhere all the time. It isn't just something that shows when you do the meridian flip. It's a smaller effect if you stay on one side, but it affects gem's, forks, alt-az - all of them. All with a slight cone error anyway.

Frank

#14 cn register 5

cn register 5

    Viking 1

  • -----
  • Posts: 760
  • Joined: 26 Dec 2012

Posted 29 August 2013 - 04:13 AM

The sync that's done at the start of the ASPA routine will help because it should eliminate errors in pointing to the star used for ASPA but cone may make a difference to the determination of the polar error.

It's complicated because the axis positions for both align stars are affected by the mount errors as well. Celestron seem to have assumed that cone and Dec offset do not matter for a fork mount because their current alignments for these mounts are two stars only. Similar reasoning should apply for a GEM used on one side of the meridian only.

But certainly, for best precision use multiple stars on both sides of the meridian.

Chris

#15 telfish

telfish

    Messenger

  • *****
  • Posts: 474
  • Joined: 17 Nov 2010
  • Loc: Adirondack Mountains NY

Posted 29 August 2013 - 08:41 AM

I use Alignmaster and can get my Polar alignment down to under 1 arc min. Usually around 37 arc seconds. That is doing 4 iterations.

However I am concerned how accurate these numbers are due to any cone error and focuser slop/ adapter tolerances.

Would using stars on either side of the meridian even these things out?


I use a 12 mm illuminated reticle eyepiece or my DSLR to center the stars.

Any thoughts on possible inaccuracy introduced by these factors into any scope based routine?

Does the type of scope and the focal length alter the accuracy of any ASPA routine due to mirror flop/enhanced accuracy with greater magnification?

I have tried the ASPA in the Synta 3.35 version and just can't get it to be repeatable.

#16 rmollise

rmollise

    Hubble

  • *****
  • Posts: 15558
  • Joined: 06 Jul 2007

Posted 29 August 2013 - 09:30 AM

Thanks Rod and George for helping and sharing your experiences!


I thought that Celestron's ASPA is more accurate because of 3 reasons:


I can only relate what I have experienced. The EQMOD routine is more than good enough for the kind of imaging I (and most folks outside the Jason Ware/Bob Gendler league) do. It's easy, user friendly, and just makes me like EQMOD all the more. :cool:

Some years back, work was started on an ASPA style routine for EQMOD, but it was never completed for several reasons. One being that given the good EQ-6 polar scope and the effectiveness of EQMOD's "helper," there just wasn't much further demand for one. ;)

#17 Alph

Alph

    Surveyor 1

  • -----
  • Posts: 1759
  • Joined: 23 Nov 2006
  • Loc: Melmac

Posted 29 August 2013 - 02:46 PM

It's such a pity EQMOD doesn't have a software routine similar to ASPA but based on its N-Star alignment considering how much powerful a PC is compared to the HC.



EQMOD utilizes a mount modeling algorithm that does not allow for easy separation of the polar axis alignment error from other pointing errors. In a nutshell, EQMOD employs a fancy geometrical interpolation method that has no understanding of underlying sources of pointing errors.

#18 cn register 5

cn register 5

    Viking 1

  • -----
  • Posts: 760
  • Joined: 26 Dec 2012

Posted 29 August 2013 - 04:13 PM

It's such a pity EQMOD doesn't have a software routine similar to ASPA but based on its N-Star alignment considering how much powerful a PC is compared to the HC.



EQMOD utilizes a mount modeling algorithm that does not allow for easy separation of the polar axis alignment error from other pointing errors. In a nutshell, EQMOD employs a fancy geometrical interpolation method that has no understanding of underlying sources of pointing errors.

Please give us a detailed technical explanation of what you are saying. You must have done this to be able to make such a comprehensive denunciation of what they have done.

Chris

#19 Alph

Alph

    Surveyor 1

  • -----
  • Posts: 1759
  • Joined: 23 Nov 2006
  • Loc: Melmac

Posted 30 August 2013 - 12:24 AM

Please give us a detailed technical explanation of what you are saying.


Sorry, that's only on a need-to-know basis :grin:

#20 cn register 5

cn register 5

    Viking 1

  • -----
  • Posts: 760
  • Joined: 26 Dec 2012

Posted 30 August 2013 - 04:11 AM

Please give us a detailed technical explanation of what you are saying.


Sorry, that's only on a need-to-know basis :grin:

So just your usual wind and *BLEEP*, I thought so.

Chris

#21 Phil Sherman

Phil Sherman

    Surveyor 1

  • -----
  • Posts: 1517
  • Joined: 07 Dec 2010
  • Loc: Cleveland, Ohio

Posted 30 August 2013 - 11:09 AM

Please give us a detailed technical explanation of what you are saying.


Sorry, that's only on a need-to-know basis :grin:

So just your usual wind and *BLEEP*, I thought so.

Chris


EQMOD's pointing model is very easy to explain. Every time you do a sync on an object, EQMOD records the mount's altitude, azimuth, and the motor encoder values. This allows an exact repositioning to that point in the sky, which never changes as long as you don't move your mount or release the clutches. Once you have a number of sync points defined, all goto operations are computed using one of two methods.

If the target is contained within a triangle formed by three sync points, then the three points are used, with the equation of time, to compute the encoder values for the target. The mount is then instructed to move to that position. If the target lies outside of a triangle of sync points, the closest sync point is used by itself. In this case, the equation of time and the target's RA and DEC are again converted to an altitude and azimuth but a displacement from the closest sync point is used instead of a surrounding triangle of sync points.

One interesting effect of this is that you can sync on the same star multiple times during an evening. Each sync will give a new altitude, azimuth, encoder value which is stored in a sync point table. The table can be saved and restored which is ideal for a mount on a fixed pier. There's also a recovery routine that can resync the scope to the sky model if you've had a power failure, which causes loss of the motor encoder values. This can also be used for a multi night portable session to bring the mount back into sync with the saved sky model.

EQMOD also runs the mount using the ASCOM platform, which allows multiple programs to concurrently control the mount. The mounts that EQMOD controls also support pulse guide, which means that any camera that PhD will control can be used as a guide camera, without requiring a connection to the ST4 guide port. One less wire and one less piece of hardware to carry and have fail in the field.

Phil

#22 rmollise

rmollise

    Hubble

  • *****
  • Posts: 15558
  • Joined: 06 Jul 2007

Posted 30 August 2013 - 12:22 PM

Actually, none of the mounts supported by EQMOD have encoders except for the new alt-az EQ6, and EQMOD doesn't use them.

#23 Moromete

Moromete

    Viking 1

  • *****
  • Posts: 665
  • Joined: 15 Feb 2012
  • Loc: Romania

Posted 30 August 2013 - 12:31 PM

I propose that all EQMOD users should ask kindly the EQMOD developer to create an ASPA routine based on his N-Star alignment. This routine should give better accuracy than Celestron's. I can't believe such opportunity has been missed.

What do you say guys?

#24 rmollise

rmollise

    Hubble

  • *****
  • Posts: 15558
  • Joined: 06 Jul 2007

Posted 30 August 2013 - 12:38 PM

As I said earlier, there were plans for one, and there is some preliminary code for it in some of the earlier releases. But not much was done thereafter. It doesn't hurt to go over to the Yahoogroup and ask Chris about it, though, I suppose. It's something that hasn't been mentioned in a long time.

#25 Alph

Alph

    Surveyor 1

  • -----
  • Posts: 1759
  • Joined: 23 Nov 2006
  • Loc: Melmac

Posted 30 August 2013 - 03:08 PM

I propose that all EQMOD users should ask kindly the EQMOD developer to create an ASPA routine based on his N-Star alignment. This routine should give better accuracy than Celestron's. I can't believe such opportunity has been missed.

What do you say guys?

As I said, not doable. They tried and failed for the reasons mentioned earlier. The EQMOD does not model the geometrical mount errors/terms.






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics