Username:

Password:

Altair AstroDIO DehumidifiersAtik CamerasModern AstronomyDavid HindsNe3 Filters

Author Topic: PEC  (Read 7027 times)

0 Members and 1 Guest are viewing this topic.

Offline Kellys_eye

  • -----------------------
  • Poster God
  • *****
  • Posts: 1772
PEC
« on: 11:46:32, 16 October, 2006 »
I need someones experience in these matters to sort something out in my mind.

Would it not be possible to develop a PEC system that tracked the WHOLE of a worm/wheel for errors and kept the record (electronically) within itself and applied it to the tracking system each and every time the setup was used?

Why, if PEC is a constant, should we have to rely on external tracking to maintain accuracy?  I'm thinking that any given tracking mount should be able to have these errors permanently programmed out.

Something also tells me I'm missing something in my thought on this... what?

TIA

Dave Gill

250/f4.5 newt, 150/f5 refractor, Mak 127, lots of bits'n'pieces mostly homebrewed.

Offline DaveS

  • Poster God
  • *****
  • Posts: 1949
  • The SkyShed Observatory, South Suffolk, UK
Re: PEC
« Reply #1 on: 12:22:57, 16 October, 2006 »
The problem with correcting PE totally, is that it's pretty much an impossibilty, even for large professional installations.

All we can do, with our amateur kit, is to minimise it best we can.

The worm-wheel, and 'worm' itself, are are only part of the problem. The drive motor gear box has a very large step-down ratio, via a series of tiny gears, these, in the case of Meade kit, connect the motor shaft to the 'worm', via a pair of transfer gears. The 'worm' itself then of course connects the 'drive chain' to the large RA drive wheel.

Each and every component in the chain, contributes to Periodic Error. Any flexure of the mount, as a whole will also add to the PE.

The ony real part of the PE that we have some control of, is that related to a 360 deg rotation of the 'worm-shaft' itself. If PE arising from this, is truely periodic, which can compensate for this with PEC software/ firmware. If the errors are random, then we have no chance.

We can sometimes improve the PE, by adjustment of the accessible gears in the 'drive-chain', to improve the 'mesh', and thus reducing backlash.

To produce a PE curve, for the enttire 360 deg rotation of the large RA drive wheel, and it's own multitude of period errors, relating to machining tolerancee would, IMO, be a huge undertaking. The amount of data required would be enormous.

This issue apart, you would need to track a single star throughout a 24 hour continuous period (one complete rotation of the large RA drive wheel, in order to collect the data from which the PE could be calculated.

Any deviation, from this, such as doing it in sections of time, at a different DEC, or very far away from the optimum 'seeing' position near the meridian, would have a significant effect on the accuracy of the data collected. This would of course then be transferred to the Fourier analysis of this data, thus producing an inaccurate PE curve, from which in turn the PEC data would be derived.

A nice idea, but IMHO, totally un-realistic, unfortunately.

Dave
Dave

Offline Kellys_eye

  • -----------------------
  • Poster God
  • *****
  • Posts: 1772
Re: PEC
« Reply #2 on: 17:02:37, 16 October, 2006 »
Dave,

Thanks for that lengthy and informative reply.  Just what I needed!

So, what's needed first of all is some means of reducing (or removing entirely) the gearing?  Direct drive huh? Hmmm....

I'm not sure your statement
Quote
The amount of data required would be enormous.
would put anyone off though.  If corrections are made at even one second intervals it only amounts to 86,400 over 24 hours.  Considering even basic EPROM capacities of Megabytes this wouldn't (shouldn't) pose an issue.

Quote
Any deviation, from this, such as doing it in sections of time, at a different DEC, or very far away from the optimum 'seeing' position near the meridian, would have a significant effect on the accuracy of the data collected.

How does variation in DEC affect the PEC?  The RA wheel/worm turns at the same speed and over the same teeth regardless of the DEC setting don't they?

I'm alluding to the possibilty of the PEC processor knowing exactly where it was on the worm/wheel cycle at all times regardless of whether you loosen the clutch and slip the mount around a few degrees.



Dave Gill

250/f4.5 newt, 150/f5 refractor, Mak 127, lots of bits'n'pieces mostly homebrewed.

Offline tomhow

  • Tom How
  • Galactic Poster
  • ******
  • Posts: 3121
  • Making equipment is far more fun than buying it.
    • The Curdridge Observatory
Re: PEC
« Reply #3 on: 20:19:18, 16 October, 2006 »

How does variation in DEC affect the PEC?  The RA wheel/worm turns at the same speed and over the same teeth regardless of the DEC setting don't they?


There are many things to take into account. Although most scopes RA turns at a constant rate, the sky appears not to - the tracking rate has to be adjusted to take into account atmospheric refraction (known as the "king rate" of tracking on some scopes)

Another factor - if you measure your medium period periodic error (ie that due to the worm shaft, ie c. 4-10mins) using a star on the celestial equator, it might give perhaps 40 seconds peak to peak... make the same measurement of a star halfway up to the pole it gives less because the sky appears to move past more slowly the higher you go.

Measuring the whole 24 hour cycle would probably not help because of the other errors such as variations in tracking speed with voltage, temprature, load, balance,  refraction (itself dependant on many factors) etc and the effects of flexure and slippage would still require some form of feeback guiding over long exposures. These factors compound and conspire to effect each other to make an unsolvable pickle.

In summary, mechanical imperfections in the telescope drive are only one part of the range of problems that make guiding a requirement.



Tak Sky 90, Atik 490, Homemade Mount, OAG, Lodestar

The Curdridge Observatory

Offline Kellys_eye

  • -----------------------
  • Poster God
  • *****
  • Posts: 1772
Re: PEC
« Reply #4 on: 20:39:32, 16 October, 2006 »
Thanks Tom,

If the PEC error is (for the sake of simplicity) a function of the worm/wheel error then, as I've stated, the error remains the same regardless of whether you are pointing at the hoizon or zenith.  True, the motion of the star is less but the worm/wheel error will always be the same.

Whilst it would be ideal to remove PEC for a whole 24 hour cycle this would be impractical since you'd only ever need to track for the length of a night (say 12 hours) and, even then, for most applications, a much shorter individual period.  That said, if the error could be removed (or at least substantially reduced) then it should be possible to make exposures of 30-60 minutes without guiding.  I realise I may well be short of the mark in that statement but what I'm getting at is that out-of-the-box tracking errors should be reducible to more acceptable levels by built-in PEC - I'm sure there are many instances where short-comings in mechanical construction are offset by electronic feedback/correction?

But, as you say, there are many more errors introducible to the whole PEC issue so perhaps I am chasing an impossible dream?

Can anyone further enlighten me on the lowest practical PEC figure for non-guiding applications based on a typical CCD and, say, a 30 minute exposure?

Dave Gill

250/f4.5 newt, 150/f5 refractor, Mak 127, lots of bits'n'pieces mostly homebrewed.

Offline tomhow

  • Tom How
  • Galactic Poster
  • ******
  • Posts: 3121
  • Making equipment is far more fun than buying it.
    • The Curdridge Observatory
Re: PEC
« Reply #5 on: 20:51:23, 16 October, 2006 »

Can anyone further enlighten me on the lowest practical PEC figure for non-guiding applications based on a typical CCD and, say, a 30 minute exposure?


Assuming spot on polar alignment, acceptable periodic error dependson the camera (pixel size) and scope (focal length).

For example, a sony icx098 (webcam chip) on a 10inch LX200 would give an image scale of 0.5 arc sec per pixel.
Something like a KAI-11000M (9um pixels) with a 300mm short refractor/camera lens would give an image scale 6.2"/p

The magnitude of tracking error you are prepared to put up with is largely a function of image scale.

Tracking is much less of an issue at shorter focal lengths, which is part reason behind the current fashoion for short refractors.

You see a lot more good wide angle images than you see good close up images.
Tak Sky 90, Atik 490, Homemade Mount, OAG, Lodestar

The Curdridge Observatory

Offline DaveS

  • Poster God
  • *****
  • Posts: 1949
  • The SkyShed Observatory, South Suffolk, UK
Re: PEC
« Reply #6 on: 23:23:04, 16 October, 2006 »
Dave

I don't think I can add any more to what Tom has said, I think he has summarised the issue nicely.

Dave
Dave

Offline Chris Graham

  • Poster God
  • *****
  • Posts: 1081
  • dum de dum de dum
    • http://www.backyard.8m.net
Re: PEC
« Reply #7 on: 08:31:12, 17 October, 2006 »
I always wondered about this topic too, cheers for the great info ;)
-Skywatcher 8" Reflector on HEQ5 with Skyscan
-Orion ED80 Refractor
-70mm Grab and Go Guide Scope
-Canon EOS 350D, Toucam Pro 2


Astronomy & Veggies

Offline MikeMS

  • The Regulars
  • Poster God
  • ***
  • Posts: 1503
Re: PEC
« Reply #8 on: 08:56:32, 17 October, 2006 »
Being a bit out of my depth here, but I would imaging it would not be possible to get your mount to be so level to undertake this sort of exercise. The level bubble in most mounts would be far too inaccurate.

Out of interest, how do you get professional mounts to be dead level? Is there a very long spirit level for people to use.

Thank you to the above posters for a very interesting read.

Mike MS

Offline coatesg

  • Senior Member
  • ****
  • Posts: 606
  • Graeme Coates
    • Chromosphere
Re: PEC
« Reply #9 on: 09:26:23, 17 October, 2006 »
Out of interest, how do you get professional mounts to be dead level? Is there a very long spirit level for people to use.

Well - actually they don't have to be. All that has to be done is to align the polar axis of an equatorial mount with the axis of the earth. However, having a level base just makes this easier to achieve (so azimuth changes don't impact on altitude and vice-versa).

I personally use the small spirit bubble levels on my mount to get it level, and then run through the Gemini routines for alignment (which just involve moving the alt/az on the mount to point it at various stars and then performing a final correction). These can be done regardless of the level nature of the base. Drift alignment is a bit trickier with a non-level mount, though a few iterations alternating between south and E/W will reduce any "cross-adjustments".

Cheers,

GC
Graeme Coates
http://www.chromosphere.co.uk
Titan + Gemini L4v1.04 :: 14" f4.6 Newt :: WO FLT110 :: FC-60NZ :: ST2000XM + CFW8 :: DMK 41AF02.AS

Offline mcrossley

  • Poster God
  • *****
  • Posts: 2721
  • Wilmslow, Cheshire
    • Wilmslow Astro
Re: PEC
« Reply #10 on: 11:53:35, 17 October, 2006 »
Nice thread techy :)

You also have to ask how much error the RA gear actually introduces. 

For example, I trained my PEC the other night, after training I measured the error again.  PemPro gave the residual PE as +/- 0.6 arc secs (1.2 arc secs peak to peak).  The graph showed that the total RA error was about 3 arc secs peak to peak.  The error in the Y axis (due to seeing) was about 1.75 arc secs.  So add the seeing noise to the remain PE gives 2.95 arc secs, which is all the observed error (give or take a nadge).  It appears that the RA gear does not introduce significant 'high' frequency errors. 

I did observe though that there was a slow drift back and forth in RA over the monitoring period of about 15 mins that may be due to the RA gear. But I also think may be (possibly more likely) due to the needle roller bearings on the RA shaft.  My RA shaft is very slightly undersized for the needle bearings and I suspect the drift may be due to the RA shaft "climbing over" each roller.

There was no drift in Dec.

Mark
Mark
C14 SCT + Hyperstar » SV4 Apo » MI-250 + Gemini » QHY8 » QHY5 » ToUcam
http://www.wilmslowastro.com/

Offline coatesg

  • Senior Member
  • ****
  • Posts: 606
  • Graeme Coates
    • Chromosphere
Re: PEC
« Reply #11 on: 12:11:49, 17 October, 2006 »
It appears that the RA gear does not introduce significant 'high' frequency errors. 

This is the key - if the error in RA due to PEC is "slow" (ie no sudden jumps), then as long as you are not using a really long focal length, then any autoguider that's setup well should be able to correct the error without trailing, even if it is as large as +/-20arcsec. Of course having said that, the less work it has to do, the better the results are likely to be, and that's why we all get hung up over PEC errors  :-\

When you're down in the few-arcsec range, as long as there are no jumps due to drive errors, then seeing is far more important anyway. A benefit of matching ccd chips to focal length of telescope (making pixels resolve about 2 arcsec for example - planetary imaging excluded) is that the tracking appears better (a 5 arcsec wobble is only 5 pixels Peak-to-peak, rather than 10 if the scale is say 1arcsec/pixel) - and you don't lose too much because the seeing usually destroys the detail at the 1 arcsec level anyway (or at least it does for me here in the UK!)

Cheers,

GC
Graeme Coates
http://www.chromosphere.co.uk
Titan + Gemini L4v1.04 :: 14" f4.6 Newt :: WO FLT110 :: FC-60NZ :: ST2000XM + CFW8 :: DMK 41AF02.AS

Offline tomhow

  • Tom How
  • Galactic Poster
  • ******
  • Posts: 3121
  • Making equipment is far more fun than buying it.
    • The Curdridge Observatory
Re: PEC
« Reply #12 on: 12:15:24, 17 October, 2006 »

I did observe though that there was a slow drift back and forth in RA over the monitoring period of about 15 mins that may be due to the RA gear.



Surely that statement is the pretty much the definition of telescope mount periodic error?

The magnitude of the periodic error, is, to a certain extent, less important than its character. If the curve is smooth with no sudden jerks, then autoguiding will cope happily. If the tracking is unpredicableand jerking, the autoguiding will not help.

When working towards autoguiding it is advised that you "check" to make sure the mount tracking is smooth before you start mucking about with extra telescopes, guidecameras and the like.

Any more, and i think i will be repeating myself...

http://astro.neutral.org/articles/autoguiding/tracking_autoguiding.html
Tak Sky 90, Atik 490, Homemade Mount, OAG, Lodestar

The Curdridge Observatory

Offline mcrossley

  • Poster God
  • *****
  • Posts: 2721
  • Wilmslow, Cheshire
    • Wilmslow Astro
Re: PEC
« Reply #13 on: 12:24:47, 17 October, 2006 »

I did observe though that there was a slow drift back and forth in RA over the monitoring period of about 15 mins that may be due to the RA gear.


Surely that statement is the pretty much the definition of telescope mount periodic error?


Well yes and no, normally when we talk about mount periodic error the period we refer to is for one revolution of the worm (4 minutes with my G11).  The slow drift back and forth in RA over 15 minutes may or may not be periodic, I'd need to record it over an hour or two to see if was periodic or random.  If my hypothesis that it is due to the RA needle bearings is correct, then I would also need to calculate their rate of roll against the RA shaft to see if it matches any observed longer term PE.

There is probably a 24 hour periodic error as well due to the RA wheel not being absolutely centred on the RA shaft.

Now and go and read that link over lunch, thanks Tom... :)
Mark
C14 SCT + Hyperstar » SV4 Apo » MI-250 + Gemini » QHY8 » QHY5 » ToUcam
http://www.wilmslowastro.com/

Offline Kellys_eye

  • -----------------------
  • Poster God
  • *****
  • Posts: 1772
Re: PEC
« Reply #14 on: 12:39:03, 17 October, 2006 »
Thanks for all the contributions in this post - it has certainly helped my understanding of the problem of PEC.

The link is also very informative - thanks everyone.

Dave Gill

250/f4.5 newt, 150/f5 refractor, Mak 127, lots of bits'n'pieces mostly homebrewed.

 

ukbuysellRemote Imaging from AustraliaSharpSkyblank APTUKAI on Facebook
Powered by SMF 2.0.15 | SMF © 2006, Simple Machines LLC
DarkBreak by DzinerStudio. Theme modified by The UKAI Team

Page created in 0.585 seconds with 37 queries.
TinyPortal © 2005-2012