Jump to content


Photo

AVX Calibration Stars - Function?

  • Please log in to reply
31 replies to this topic

#1 cpl43uk

cpl43uk

    Messenger

  • -----
  • Posts: 412
  • Joined: 09 Feb 2008
  • Loc: Bristol, UK

Posted 08 July 2013 - 05:12 PM

I have the AVX and so use the 2 star Alignment. This asks for 4 stars for accurate GOTOS and it works fine. However if I find later that the alignment is not quite right I know I can re-align on a new star and eliminate one of the original alignments. My question. Can one of these new stars be a previously used "calibration star" in the original procedure? Secondly, if I do re-align (using any star), what happens to the calibration stars? Do I need to re-do them?
Basically "what are the calibration stars used for and what do I do with them if I re-align?"

Thanks

#2 dr.who

dr.who

    Apollo

  • *****
  • Moderators
  • Posts: 1299
  • Joined: 05 Jan 2012

Posted 11 July 2013 - 10:50 PM

Yes you can use the same star. What they do is help build a more detailed aka precise model of your night sky. The more data points for reference the better assuming you center them up right. A 12ish mm illuminated reticle eyepiece helps with precise centering

#3 frito

frito

    Apollo

  • -----
  • Posts: 1183
  • Joined: 05 Oct 2012
  • Loc: Fremont, CA

Posted 12 July 2013 - 12:08 AM

actually how the nexstar software works is this

the first two stars are its actual reference points that it builds the sky model on, you screw one of them up and your done it will not work very well.

calibration stars are there to correct cone error, an error inherent to mounts esp German EQ mounts.

http://stargazerslou...ach&section=...

thats cone error.

in a perfect world if you had a perfect mount with a perfectly aligned OTA on the mount cone error would be negligible. calibration stars attempt correct for this and any other errors that may have been caused by many other things, poor centering of previous alignment stars, slop in the mechanics. you name it.

if you take the time to accurately align your mount there should be no reason to realign it in one night. if your tracking is poor its probably because your polar alignment is poor, the mount compensates for polar alignment errors when slewing but while tracking it does not so a poor polar alignment will result in an object slowly moving out of the field of view. this is completely normal.

#4 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3920
  • Joined: 12 Oct 2007

Posted 12 July 2013 - 05:39 AM

It is somewhat a lost cause, but cone is not the only thing found by the calibration stars - and a key value that is important to know accurately is the declination index error - which you can see listed along with the cone error. Your mount could be perfect and have no cone or other physical errors - but if the encoder value for dec. = 0 is even slightly off, it will throw off the pointing across the sky.

I just checked the AVX manual and it just says the calibration stars will improve pointing across the sky - and I think that is good and doesn't specifically point to cone as the critical value. Both Cone and Dec. Index are calculated in the calibration process, and both are important.

Frank

#5 Stew57

Stew57

    Mercury-Atlas

  • *****
  • Posts: 2535
  • Joined: 03 May 2009
  • Loc: Silsbee Texas

Posted 12 July 2013 - 08:14 AM

Just an added note. The calibration stars are all averaged into 1. The 2+4 alignment is actually a 2+1, as the 4 calibration stars are weighted as one.

#6 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3920
  • Joined: 12 Oct 2007

Posted 12 July 2013 - 08:27 AM

I disagree with that. I do not think it works that way. Each star contributes information about the error in its part of the sky, and that error is different in different parts of the sky, and averaging all the star positions to one would lose that information.

Frank

#7 cn register 5

cn register 5

    Viking 1

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

Posted 12 July 2013 - 08:55 AM

I think Mark's opinion can be discounted. He's presented no evidence for it at all.

Far more likely is that all the alignment and calibration stars are fitted to the pointing matrix using something like a least squares fit to minimise the error.

Chris

#8 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3920
  • Joined: 12 Oct 2007

Posted 12 July 2013 - 09:13 AM

My question. Can one of these new stars be a previously used "calibration star" in the original procedure?



My answer agrees with the manual - you should keep "alignment" stars on one side of the meridian, and "calibration" stars on the other - and not mix them.

But - not to complicate things too much I hope - if a calibration star was in the east originally, but hours have passed so that it is now in the west, I believe you should be fine using it as an alignment star. What matters is where it is (which side of the meridian) when you "use" it.

Frank

#9 Stew57

Stew57

    Mercury-Atlas

  • *****
  • Posts: 2535
  • Joined: 03 May 2009
  • Loc: Silsbee Texas

Posted 12 July 2013 - 09:20 AM



Check teamcelestron search for "calibration stars".

Just a snippet from the thread written by the engineer that codes the HC and MC firmware.

"That said, the last 4 points in a 2+4 alignment each measure cone error against the "backdrop" of the two star alignment. In that sense, this is really just a 3 point alignment with the last point being an average of 4 measurements. Why would the 4 measurements of the cone error vary? One cause is mount flexure as the OP eluded to. In principle you could use some of the extra data points in the 2+4 to back out the mount flexure terms. The legacy handcontrol did not have the computational prowess, the code space, or even sufficient RAM to do that kind of computation."


I wish it worked like T-point also, but it doesn't.

#10 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3920
  • Joined: 12 Oct 2007

Posted 12 July 2013 - 09:34 AM

That reads as an over-simplified description of the method, and I stand by my interpretation - which is consistent with Chris's.

Furthermore, it focuses on cone error and ignores dec. index error - which I claim is quite important to get right.

I would go further and say that - since the calibration stars cannot distinguish flexure from refraction or whatever - in effect they do help compensate for all those effects - even if they don't explicitly model those terms like t-point does.

Frank

#11 frito

frito

    Apollo

  • -----
  • Posts: 1183
  • Joined: 05 Oct 2012
  • Loc: Fremont, CA

Posted 12 July 2013 - 09:36 AM


Check teamcelestron search for "calibration stars".

Just a snippet from the thread written by the engineer that codes the HC and MC firmware.

"That said, the last 4 points in a 2+4 alignment each measure cone error against the "backdrop" of the two star alignment. In that sense, this is really just a 3 point alignment with the last point being an average of 4 measurements. Why would the 4 measurements of the cone error vary? One cause is mount flexure as the OP eluded to. In principle you could use some of the extra data points in the 2+4 to back out the mount flexure terms. The legacy handcontrol did not have the computational prowess, the code space, or even sufficient RAM to do that kind of computation."


I wish it worked like T-point also, but it doesn't.


i was trying to find that info from him on there a little bit ago lol.

explains why when i align my mount because im using such a small and solidly mounted OTA after the first calibration star the last three are barely off being perfectly centered.

about the hemisphere thing with the alignment, if you follow the suggested stars the mount's firmware usually won't give you alignment stars that are not good to use in my experience. that is not the case with synscan firmware's 3 star align and probably one of the biggest reasons why people tend to say the nexstar system has more accurate goto's but the added cone error averaging could also have an effect with larger scopes etc.

#12 frito

frito

    Apollo

  • -----
  • Posts: 1183
  • Joined: 05 Oct 2012
  • Loc: Fremont, CA

Posted 12 July 2013 - 09:39 AM

That reads as an over-simplified description of the method, and I stand by my interpretation - which is consistent with Chris's.

Furthermore, it focuses on cone error and ignores dec. index error - which I claim is quite important to get right.

I would go further and say that - since the calibration stars cannot distinguish flexure from refraction or whatever - in effect they do help compensate for all those effects - even if they don't explicitly model those terms like t-point does.

Frank


i doubt nexstar has dec index error correction because the firmware is the same across all the mounts and the CG-5 AGT for starters has no DEC or RA indexes to check against. how can you tell they have the same firmware? easy PEC tool appears to work with the AGT even though the mount has no worm indexes and due to that its impossible for it to save the data and have a reliable result.

#13 Stew57

Stew57

    Mercury-Atlas

  • *****
  • Posts: 2535
  • Joined: 03 May 2009
  • Loc: Silsbee Texas

Posted 12 July 2013 - 09:57 AM

Frank,

I asked the engineer he said they were averaged to account for cone error and mount flexure. It makes no difference to me as it is what it is, but it is not more than that. :grin: I did not specifically asked about dec index error but will if I get a chance.

It seems to me that all error even mirror flop will be included as the controller has no way to differentiate between mirror flop or mount flexure. The question really is the 4 calibration stars averaged as a single data point, or are they 4 separate data points. He stated that the HC did not have space for 4 separate data points but that may change as the + has more room. The explanation may be completely wrong as I am in no position to check him but I tend to believe what he says.

My brain (according to the national geographic channel) thinks it is more of an expert than it really is. I am probably completely wrong in my understanding of the calibration stars function. You may be also though, so I would suggest joining teamcelestron.com and asking Derik directly. He is a really nice guy and go into depth if you really are interested.

#14 cn register 5

cn register 5

    Viking 1

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

Posted 12 July 2013 - 10:38 AM

The Dec offset is one of the things that is calculated as part of the alignment process, along with Ra offset, cone error and the two polar align errors.

They are all shown in the HC and it is the same for all the GEM scopes. The switches only provide a consistent starting position, if you don't have these then the alignment marks do the same job.

As for how the calibration stars are averaged I think that there's more to this than taking an average of the axis positions and the hour angle and dec positions of the four stars. I read the same info and didn't interpret it as that. I expect the exact process is proprietary. Whatever it is it does a good job.

Chris

#15 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3920
  • Joined: 12 Oct 2007

Posted 12 July 2013 - 10:45 AM

doubt nexstar has dec index error correction because the firmware is the same across all the mounts and the CG-5 AGT for starters has no DEC or RA indexes to check against.


I haven't actually used those mounts without index switches - but the principle is the same: even if the mount is mechanically perfect, it needs to know which dec. encoder reading is exactly 0 degrees - and it does that by looking at a bunch of stars. I thought that all the mounts displayed a dec. index error in the hc but that may not be true. Even if they don't display it, they still need it internally - and they get that information from the stars.

Frank

#16 frito

frito

    Apollo

  • -----
  • Posts: 1183
  • Joined: 05 Oct 2012
  • Loc: Fremont, CA

Posted 12 July 2013 - 10:53 AM

i'm probably getting worm gear index and error index confused.

the CG-5 has visual index marks you have to align before aligning the mount otherwise it will think its at them when you start it and will slew past limits etc. so yeah it has a software index generated at each alignment procedure.

#17 Stew57

Stew57

    Mercury-Atlas

  • *****
  • Posts: 2535
  • Joined: 03 May 2009
  • Loc: Silsbee Texas

Posted 12 July 2013 - 11:08 AM

I thought that was done via the alignment stars. How close the alignment stars land toward center depends on how accurate I line up the alignment marks and are polar aligned. The first calibration star seems off the same amount no matter how I start the alignment with regards to the index tabs. Am I missing something?

One experiment I will try when I get back into town is;
Do a 2 star alignment with one calibration star. Slew around a bit and then choose to go back to the calibration star. Note how close to center it lands.

Then do a 2 star alignment and add 2 widely spaced calibration stars one of which should be the one chosen above. Slew around a bit and the slew back to the calibration star.

If the calibration stars are a single data point the accuracy of the first run will be greater than the second by a small margin. If the calibration stars are separate data points then the pointing accuracy to the star will be unaffected.
Does this sound like a reasonable test?
I

#18 rmollise

rmollise

    Hubble

  • *****
  • Posts: 15554
  • Joined: 06 Jul 2007

Posted 12 July 2013 - 11:13 AM

I disagree with that. I do not think it works that way. Each star contributes information about the error in its part of the sky, and that error is different in different parts of the sky, and averaging all the star positions to one would lose that information.

Frank


As has been said, star 1 and 2 are for the basic model of the sky. Just as in two star alignments for fork mount scopes. Stars 3 - 6 are to deal with cone alignment and other problems likely in GEMs, no matter how well made. ;)

#19 rmollise

rmollise

    Hubble

  • *****
  • Posts: 15554
  • Joined: 06 Jul 2007

Posted 12 July 2013 - 11:16 AM


I haven't actually used those mounts without index switches - but the principle is the same: even if the mount is mechanically perfect, it needs to know which dec.


It does need to know where it starts from, but it does that with index marks. ;)

#20 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3920
  • Joined: 12 Oct 2007

Posted 12 July 2013 - 11:21 AM

Yes - the index marks correspond to the switch positions - and as such they are assumed to be slightly offset from the true 0/90 location and not perfect. That offset has to be calibrated very accurately - for a gem or fork mount - if the pointing is to be accurate. Both dec. index offset and cone apply to gem's and forks - whether they use switches or manually aligned index marks.

Frank

#21 jrcrilly

jrcrilly

    Refractor wienie no more

  • *****
  • Administrators
  • Posts: 33790
  • Joined: 30 Apr 2003
  • Loc: NE Ohio

Posted 12 July 2013 - 12:22 PM

Yes - the index marks correspond to the switch positions - and as such they are assumed to be slightly offset from the true 0/90 location and not perfect. That offset has to be calibrated very accurately - for a gem or fork mount - if the pointing is to be accurate.


That's true only if by "pointing" you mean the slew to the first alignment star. It has no effect at all on pointing accuracy after alignment; once aligned, pointing accuracy depends on the alignment. Any arbitrary initial orientation can be chosen. The only difference will be the amount of correction required to center the first alignment star because the mount bases that slew on the presumption that the initial position was on the marks (or switches).

#22 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3920
  • Joined: 12 Oct 2007

Posted 12 July 2013 - 12:26 PM

I completely disagree. The dec. offset is very different from the RA offset - which is what I assume you are referring to.

I think it is understandable to expect that the dec. offset is a trivial issue and somehow the cone value is critical - but in fact the dec. offset is a very important and elusive value that affects pointing across the sky - all the time - and not just for the initial stars.

Frank

#23 jrcrilly

jrcrilly

    Refractor wienie no more

  • *****
  • Administrators
  • Posts: 33790
  • Joined: 30 Apr 2003
  • Loc: NE Ohio

Posted 12 July 2013 - 12:55 PM

I completely disagree. The dec. offset is very different from the RA offset - which is what I assume you are referring to.


No need to assume. I was responding to your post, which specifically referred to the index marks (as seen in the portion I quoted). I specifically referred to them also.

#24 freestar8n

freestar8n

    Soyuz

  • *****
  • Posts: 3920
  • Joined: 12 Oct 2007

Posted 12 July 2013 - 01:08 PM

The "dec. index offset" I am referring to is the deviation in either the switch location or the index mark - from the true 0/90 degree dec. location. So whatever reference you use - it is assumed to be imperfect - and is corrected for by the calibration stars. For RA, any error doesn't matter too much - but for Dec., it matters a lot - because RA errors are independent of dec., while errors in Dec. are coupled to errors in RA. If that subtlety is not clear then none of this will make sense.

So - I am not referring to either the index marks or the switch positions - but the departure of them from the true 0/90 values. That is an issue that is trivial in RA and only affects initial pointing - but is elusive in Dec. - and affects all pointing - and not just to the initial stars.

Frank

#25 Alph

Alph

    Surveyor 1

  • -----
  • Posts: 1758
  • Joined: 23 Nov 2006
  • Loc: Melmac

Posted 12 July 2013 - 01:19 PM

The Dec offset is one of the things that is calculated as part of the alignment process, along with Ra offset, cone error and the two polar align errors


RA offset is not calculated as part of the 2+4 alignment process.






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics