POLL - Guiding

[Rodd]

I have been discussing guiding with a friend and we disagree.  I wish not to give away my (and his) position by stating my opinion, so a poll is in order.  Please read the following and answer the poll question....which is a two part yes/no response.  However, if one wishes to elaborate, assign conditions, or make stipulations, one may do so.  

Conditions

1) Maximum guiding accuracy of mount is 1.0 total rms (in arcsec) 

2) Capacity of mount is unlimited for sake of discussion.

3) Seeing is not the limiting factor  

4) Camera is a CCD and the exposure lengths must be at least 10 min for the long focal length F8 system and 3 min  for the wider field F4 system (i.e no short ungided subs!)

5) Binning data unnecssary - seeing equals imaging resolution

Questions

A - Would a guide resolution of 1.0 rms  be sufficient for imaging at a focal length of 5,000 mm and a resolution of 0.2 arsec/pix?

B - Would a guide resolition of 1.0 rms be sufficient for imaging at a focal lenth of 500 mm and a resiolution of 3.5 arcsec/pix?

 

EDIT:  After realizing that being "sufficient" is a subjective concept, I wish to inform that what I am after here is, will guiding errors of 1.0 rms be visible in the high resolution image?  will they but visible in the low resolution image?  Some may think that an image is satisfactory while others may not.  So I am not talking about image aesthetics, or whether the image is "good enough" for you.  I am asking about the visibilty of guiding errors in the image and the general concept of whether this guiding accuracy is sufficient for the imaging systems above.  Would it normally be your recommendation to a beginner.

  

Looking forward to seeing your answers

Edited June 12, 2022 by Rodd

[ONIKKINEN]

A= Not sufficient. Not even close really, and if someone guides like this they will definitely lose resolution to poor guiding and so shooting at this resolution is a pointless waste of SNR.

B= Definitely good enough and i doubt there would be a very noticeable difference with better guiding at this resolution.

I think a guiding error of 50% of your imaging scale is where things start going to the nitpick and diminishing returns category.

[CraigT82]

A - nope

B- yep

(Now I’m half expecting you to post a perfect looking image captured with the parameters outline in A!)

[Pitch Black Skies]

Did you mean to write 500mm for A? Either way, I suppose the focal length is irrelevant as you have given the sampling rate.

A = 0.2"/pixel

B = 3.5"/pixel

If seeing is not the limiting factor, I'm confused as to how guiding errors wouldn't be visible at 1" total RMS if the sampling rate is 0.2"/pixel.

Edited June 12, 2022 by Pitch Black Skies

[DaveS]

A, nope, not a chance. you need a mount capable of 0.1", and put it on top of a mountain in the Atacama with laser adaptive optics.

b, yep.

[Craig a]

A: no 

b:yes

[newbie alert]

A) nope

yes

With A  using a 5000mm FL scope and imaging at .2 then I'd be fairly confident using it for planetary or solar system objects but not for deepsky and 10 min subs, massively over sampled and need to guide under .2 to break even( problem is alot of people use pixels to quote their guiding and not arc mins)

Fairly confident that you knew the answer..

 

Edited June 12, 2022 by newbie alert
Changed

[Rodd]

4 hours ago, Pitch Black Skies said:

Did you mean to write 500mm for A? Either way, I suppose the focal length is irrelevant as you have given the sampling rate.

A = 0.2"/pixel

B = 3.5"/pixel

If seeing is not the limiting factor, I'm confused as to how guiding errors wouldn't be visible at 1" total RMS if the sampling rate is 0.2"/pixel.

No. I meant 5,000.  The idea is long focal length vs short. And you are correct—guiding errors would be visible a 1” rms and a sampling rate of .2”. So your answer would be “NO” for A.  

[Pitch Black Skies]

2 hours ago, Rodd said:

And you are correct—guiding errors would be visible a 1” rms and a sampling rate of .2”

What's the person's reasoning as to why A would be sufficient/not show guiding errors?

[Rodd]

1 hour ago, Pitch Black Skies said:

What's the person's reasoning as to why A would be sufficient/not show guiding errors?

At first I did not understand your question in the framework of the example.  Then I realized that you assume my position and that of my friend.  I am not quite ready to "break Blind'" as they say.  Suffice it to say that I put conditions on the question for the poll that were not necessarilly applied to our discusion.  I would rather wait and see if my friend replies to the poll.  Then we shall see.

Edited June 12, 2022 by Rodd

[Pitch Black Skies]

4 hours ago, Rodd said:

And you are correct—guiding errors would be visible a 1” rms and a sampling rate of .2”.

Apologies, I thought you were telling me your position on it when you stated the above. I figured that if you believe guiding errors would be visible with (A), surely that would suggest you believe (A) to be insufficient.

[Rodd]

Just now, Pitch Black Skies said:

Apologies, I thought you were telling me your position on it when you stated the above. I figured that if you believe guiding errors would be visible with (A), surely that would suggest you believe (A) to be insufficient.

No. I was just saying that if you felt they would be visible then you were correct and that your answer to A would then be no.    Maybe I let a cow out of the barn. My apologies.  It’s silly anyway.  No harm done 

[Pitch Black Skies]

18 minutes ago, Rodd said:

No. I was just saying that if you felt they would be visible then you were correct and that your answer to A would then be no.

So you believe A is sufficient so?!

 

 

 

 

 

 

 

 

 

 

I'm only joking 😁👍

[Rodd]

3 hours ago, Pitch Black Skies said:

So you believe A is sufficient so?!

 

 

 

 

 

 

 

 

 

 

I'm only joking 😁👍

You got me. I am on my phone and your whole message is not visible on the screen.   I didn’t see the last part until I Scrolled down after visiting page. 🤭

[vlaiv]

My personal view is that there is no such thing as "sufficient" guiding.

Lower RMS is always better, regardless of sampling rate applied.

It is true that at some point we enter domain of diminishing returns, and it is good to know where this domain is, but it is not related to focal length nor to sampling rate.

It is related to other contributors to final FWHM - namely seeing and aperture size.

To answer the poll questions:

For both A and B - question does not make sense as guide RMS and if we deem it "sufficient" is not related to either sampling rate or focal length.

In order for Guide RMS to be "inconsequential" to total FWHM - it should be at least x3-x5 smaller than largest component that contributes to FWHM.

This is usually the seeing. If we convert all values to RMS, then we can see that

2" FWHM seeing ~= ~0.85" RMS

3" FWHM seeing ~= ~1.275" RMS

80mm aperture Airy disk ~= 0.6" RMS

From this we can see that even for small scope of 80mm in poor seeing of 3" we need to guide at 0.25" - 0.425" for guiding change to make minimal difference.

As for rule of thumb that guiding RMS should be half of pixel scale - well, that is just looking things "in reverse". It is ok to give some peace of mind, but if one wishes to get the best out of their gear - then it is not sufficient.

It is sampling rate that should be adjusted to match achieved FWHM and not the other way around (we can't really impact FWHM, but we can match pixel scale to FWHM, either by using reducers, or binning or other means ...).

[Rodd]

1 hour ago, vlaiv said:

As for rule of thumb that guiding RMS should be half of pixel scale - well, that is just looking things "in reverse". It is ok to give some peace of mind, but if one wishes to get the best out of their gear - then it is not sufficient.

It is sampling rate that should be adjusted to match achieved FWHM and not the other way around (we can't really impact FWHM, but we can match pixel scale to FWHM, either by using reducers, or binning or other means ...).

Violation of the [poll conditions...NO BINNING allowed.  Show me allot of experienced imagers using a mount that can only guide at 1" rms with 17" scope shooting at 5,000 mm.  If they bin their data x 10, then wthey wouldn't be using the long focal length.  It makes no sense.  

One can't realistically equate a sampling raye of 0.2 with 3.5.  If you do--you are no longer really shooting at 5,000mm--you are shooting at wide field pixel scales but without the widefuield FOV.

One can always change the data so the guide error can't be seen (changing sampling rate).  But that is not what the poll question is about.  

 

Edited June 13, 2022 by Rodd

[Rodd]

Poll is effectively broken.  Vlaiv is my friend and both of our positions are now known.  For teh record, What he says may be true--but it misses the point.  The experiment is using a 1" rms mount to shoot with a long focal length scope without altering the data (binning).  We are definitely hitting different nails......the mystery to me is how it is believed that we are hitting teh same nail.

[dan_adi]

On 12/06/2022 at 15:56, Rodd said:

I have been discussing guiding with a friend and we disagree.  I wish not to give away my (and his) position by stating my opinion, so a poll is in order.  Please read the following and answer the poll question....which is a two part yes/no response.  However, if one wishes to elaborate, assign conditions, or make stipulations, one may do so.  

Conditions

1) Maximum guiding accuracy of mount is 1.0 total rms (in arcsec) 

2) Capacity of mount is unlimited for sake of discussion.

3) Seeing is not the limiting factor  

4) Camera is a CCD and the exposure lengths must be at least 10 min for the long focal length F8 system and 3 min  for the wider field F4 system (i.e no short ungided subs!)

5) Binning data unnecssary - seeing equals imaging resolution

Questions

A - Would a guide resolution of 1.0 rms  be sufficient for imaging at a focal length of 5,000 mm and a resolution of 0.2 arsec/pix?

B - Would a guide resolition of 1.0 rms be sufficient for imaging at a focal lenth of 500 mm and a resiolution of 3.5 arcsec/pix?

 

EDIT:  After realizing that being "sufficient" is a subjective concept, I wish to inform that what I am after here is, will guiding errors of 1.0 rms be visible in the high resolution image?  will they but visible in the low resolution image?  Some may think that an image is satisfactory while others may not.  So I am not talking about image aesthetics, or whether the image is "good enough" for you.  I am asking about the visibilty of guiding errors in the image and the general concept of whether this guiding accuracy is sufficient for the imaging systems above.  Would it normally be your recommendation to a beginner.

  

Looking forward to seeing your answers

It's simple really, if one can guide at 1"  RMS error with an image scale of 0.2 "/pixel and get good results (round stars), then 10 micron, AP, ASA, Astelco, Planewave etc. would have never existed as mount manufactures. 

[vlaiv]

19 minutes ago, dan_adi said:

It's simple really, if one can guide at 1"  RMS error with an image scale of 0.2 "/pixel and get good results (round stars), then 10 micron, AP, ASA, Astelco, Planewave etc. would have never existed as mount manufactures. 

Round stars don't depend on total RMS error.

You can have RMS of 2" and still get round stars (at various image scales).

In order to get round stars - RA and DEC error must be roughly the same, independent and random in nature.

[dan_adi]

7 minutes ago, vlaiv said:

Round stars don't depend on total RMS error.

You can have RMS of 2" and still get round stars (at various image scales).

In order to get round stars - RA and DEC error must be roughly the same, independent and random in nature.

Then I am missing something. So you are saying if I get large guiding errors my eccentricity won't be affected but only FWHM?

[vlaiv]

19 minutes ago, dan_adi said:

Then I am missing something. So you are saying if I get large guiding errors my eccentricity won't be affected but only FWHM?

I'm saying that eccentricity is affected in specific cases - namely different RMS in RA and DEC, correlation between errors in RA and DEC and in general non random behavior of guide error.

If none of those are the case - you have same (or nearly the same) RMS in RA and DEC, they are independent and random - then yes, you will just get round stars with increased FWHM.

From above you can see that round stars don't depend on total RMS - there are (quite often) cases where you have large RMS but you still get round stars.

 

[Rodd]

1 hour ago, vlaiv said:

Round stars don't depend on total RMS error.

You can have RMS of 2" and still get round stars (at various image scales).

In order to get round stars - RA and DEC error must be roughly the same, independent and random in nature.

I think it goes without saying that what is desirable is round AND small. What good are stars the size of grapefruit?  

[Pitch Black Skies]

Yeah but if the grapefruit were the same distance as the stars they would actually be tiny.

[wimvb]

2 hours ago, vlaiv said:

Round stars don't depend on total RMS error.

You can have RMS of 2" and still get round stars (at various image scales).

In order to get round stars - RA and DEC error must be roughly the same, independent and random in nature.

 

2 hours ago, dan_adi said:

Then I am missing something. So you are saying if I get large guiding errors my eccentricity won't be affected but only FWHM?

The crux here is "random in nature". What would the cause of this randomness be, other than seeing and grit in the gears?

Random steps in X and random steps in Y will result in a pattern that fills a circle. The size of the steps determines the size of that circle. If the system is mechanical, it can't be tinkered with and optimized, simply because the deviations are random. In fact, any attempt to "fix" such a system by adding corrections, will only result in a wider circle (increased RMS).

Many years ago I was assigned the task to improve a process for film deposition. The aim of the process was to deposit a thin hard coating of 100 nm on a substrate. The process involved hands on interactions by the technicians, and one "fix" they came up with was to change the deposition time (in AP, we would call that the exposure time), depending on previous results. They adopted a scheme where they would increase the deposition time if the previous run had resulted in a film that was thinner, and decrease the deposition time when the previous film was thicker than the target. By simple statistical analysis I showed them that the deposition process had a built in randomness, and all  they did was increase the variation in film thickness by trying to manipulate it. Had they used a constant deposition time, the variation in thickness would have been less.

How does this translate in guiding practices? Quite simple acutally. If the guiding variations you see are not caused by the mechanics, but are truly random, then you shouldn't send any guide pulses. PHD tries to estimate the randomness in its "high frequency star motion" parameters and avoid chasing them through its MinMo and aggressiveness settings. Mechanical deviations (mount movements) are usually much slower than these high frequency motions, and are seldom random in nature, and it's them we try to guide out. The PHD manuals don't say this, but if seeing varies from one session to another, you should run the GA every session (takes only 2 minutes, and measuring backlash isn't needed). Afterwards you should set MinMo to the recommended value and keep an eye on aggressiveness. Keep it low, and exposure time long enough, if seeing is poor.

Edited June 18, 2022 by wimvb

[wimvb]

On 12/06/2022 at 14:56, Rodd said:

A - Would a guide resolution of 1.0 rms  be sufficient for imaging at a focal length of 5,000 mm and a resolution of 0.2 arsec/pix?

B - Would a guide resolition of 1.0 rms be sufficient for imaging at a focal lenth of 500 mm and a resiolution of 3.5 arcsec/pix?

Hmm, according to the premises, situation A has a camera with 5 um pixels, whereas situation B has a camera with 9 um pixels. I would start with swapping cameras:

A. 9 um @ 5000 mm fl: 0.35 "/p

B. 5 um @ 500 mm fl: 2 "/p

😉