Guiding accuracy

[Pitch Black Skies]

Hi guys,

What RMS should I be hoping to achieve and what is unacceptable?

How are my guiding requirements calculated?

I'm probably not explaining this right.

I am referring to part 3:15-3:50 in this video.

I don't understand how that's calculated.

Any help is greatly appreciated.

I use 533MC, 130PDS & HEQ5.

 

Edited March 30, 2022 by Pitch Black Skies

[The Lazy Astronomer]

Much as I enjoy Rory's videos, I think that was poorly explained. Basically he's saying for widefield imaging, you'd generally be shooting at a lower resolution and mount tracking errors would be less noticeable as each pixel captures a larger portion of the sky. When using longer focal lengths, generally you'd be at a higher resolution and tracking errors become more apparent.

There's a bit more to it than that, but someone like @vlaiv will explain it much better 😁

Generally, you'd want your total RMS to be less than or equal to 0.5x your imaging scale (in arc seconds), and ideally, a roughly equal error in RA and DEC - if one is significantly larger than the other, you may see star elongation even if the total RMS is good. 

Your image scale is ~1.2"/pixel (although it seems the general consensus now is that the sampling rate for colour cameras should be doubled from what the calculation suggests, so your actual scale is ~2.4"/pixel). Honestly, I am unsure whether this means you should be aiming for an RMS of 0.6" or 1.2" (someone more knowledgeable than me will confirm, I'm sure).

If the latter, that should be relatively easily achievable with your setup; the former will likey be more difficult.

[ONIKKINEN]

Like above, i think less than half of imaging resolution and similar errors in both axis is good guiding. No obvious excursions in declination and a periodic error that is in check is key here. I use the predictive PEC algorithm in PHD2 and it seems to be doing a decent job of keeping my somewhat poorly figured RA worm gear in check.

But looking at the guide graph can be misleading, especially with a newtonian and you may have good guiding without round stars. That is, i believe, due to mechanical issues with most off the shelf newtonians including: Focuser slop, tube deforming, mirror cell stability issues, secondary holder issues and more. All of these issues will lead to effectively taking your scope out of collimation randomly and during the session. Your guide scope (if using a guide scope) can still report decent guiding because its not aware of the other issues, unless the guide scope attachment method is also sub par (also common with newtonians). Experts here recommend OAGs regularly for newtonians for these reasons, and im beginning to see why since some issues still persist for me even though i dont think they are related to guiding exactly.

I will probably make the jump from a guide-scope to an OAG some time in the near future because i see "good guiding" but still varying decrees of issues with the actual subs.

[Pitch Black Skies]

5 hours ago, The Lazy Astronomer said:

Much as I enjoy Rory's videos, I think that was poorly explained. Basically he's saying for widefield imaging, you'd generally be shooting at a lower resolution and mount tracking errors would be less noticeable as each pixel captures a larger portion of the sky. When using longer focal lengths, generally you'd be at a higher resolution and tracking errors become more apparent.

There's a bit more to it than that, but someone like @vlaiv will explain it much better 😁

Generally, you'd want your total RMS to be less than or equal to 0.5x your imaging scale (in arc seconds), and ideally, a roughly equal error in RA and DEC - if one is significantly larger than the other, you may see star elongation even if the total RMS is good. 

Your image scale is ~1.2"/pixel (although it seems the general consensus now is that the sampling rate for colour cameras should be doubled from what the calculation suggests, so your actual scale is ~2.4"/pixel). Honestly, I am unsure whether this means you should be aiming for an RMS of 0.6" or 1.2" (someone more knowledgeable than me will confirm, I'm sure).

If the latter, that should be relatively easily achievable with your setup; the former will likey be more difficult.

Thank you, lots of useful info there. You did a great job of simplifying it for me.

Sometimes my RMS is above 1.2". Does that mean my image is potentially blurring, provided the seeing is below 1.2"?

[Pitch Black Skies]

4 hours ago, ONIKKINEN said:

Like above, i think less than half of imaging resolution and similar errors in both axis is good guiding. No obvious excursions in declination and a periodic error that is in check is key here. I use the predictive PEC algorithm in PHD2 and it seems to be doing a decent job of keeping my somewhat poorly figured RA worm gear in check.

But looking at the guide graph can be misleading, especially with a newtonian and you may have good guiding without round stars. That is, i believe, due to mechanical issues with most off the shelf newtonians including: Focuser slop, tube deforming, mirror cell stability issues, secondary holder issues and more. All of these issues will lead to effectively taking your scope out of collimation randomly and during the session. Your guide scope (if using a guide scope) can still report decent guiding because its not aware of the other issues, unless the guide scope attachment method is also sub par (also common with newtonians). Experts here recommend OAGs regularly for newtonians for these reasons, and im beginning to see why since some issues still persist for me even though i dont think they are related to guiding exactly.

I will probably make the jump from a guide-scope to an OAG some time in the near future because i see "good guiding" but still varying decrees of issues with the actual subs.

Another really useful reply. I was considering upgrading my 130pds to maybe a 200pds but now realise that the difficulties you mentioned might even be more noticeable at the longer focal length and added weight.

I have noticed that the RA corrections are much greater than the DEC ones. I might have to adjust aggressiveness. I use an ASIair and not sure if there is PEC training capability with it.

I might invest in a longer dovetail to reduce flexure down the line. I ensured not to use the finderscope shoe for the guidescope, it didn't seem sturdy enough. Instead I attached a second dovetail on top of the rings and bolts the guidescope to that.

It makes sense to use the OAG. I don't really understand them, they seem a bit more complicated.

[Clarkey]

If you have an HEQ5 it is worth doing the belt modification to improve tracking. You should be able to get below 1” RMS with some fettling. I think mine is typically around 0.7. There is a predictive PEC algorithm within PHD2 in the advanced guiding settings. Change to this and you should see some improvement.

With regards to the 200pds this would be pushing an HEQ5. Although possible, any breeze will ruin your subs.

WRT OAG I am not sure it is needed on a 130. To be honest I have gone back to a guide scope on my RCI at 1600mm because of the issues I had with OAGing and difficulties finding good guide stars.

[Pitch Black Skies]

9 hours ago, Clarkey said:

If you have an HEQ5 it is worth doing the belt modification to improve tracking. You should be able to get below 1” RMS with some fettling. I think mine is typically around 0.7. There is a predictive PEC algorithm within PHD2 in the advanced guiding settings. Change to this and you should see some improvement.

With regards to the 200pds this would be pushing an HEQ5. Although possible, any breeze will ruin your subs.

WRT OAG I am not sure it is needed on a 130. To be honest I have gone back to a guide scope on my RCI at 1600mm because of the issues I had with OAGing and difficulties finding good guide stars.

Ordered the belt kit yesterday, hopefully it will make a difference.

I use an ASIair Pro, I don't think there is any PEC training capability with it yet unfortunately.

I should probably try to spend more time with the 130. It's a very capable scope.

Edited March 31, 2022 by Pitch Black Skies

[vlaiv]

15 hours ago, Pitch Black Skies said:

How are my guiding requirements calculated?

4 hours ago, Pitch Black Skies said:

Sometimes my RMS is above 1.2". Does that mean my image is potentially blurring, provided the seeing is below 1.2"?

While pixel size and guide RMS are often connected with some sort of rule of the thumb - I'd rather try to explain it a bit differently.

For the moment, forget pixel size. Forget pixels completely.

Let's just look what happens at focal plane of the telescope as the light comes in. There are three different major contributor to the blur that happens to the image.

- seeing

- guiding/tracking precision

- aperture size

Each one of these produces some level of blur on its own. In perfect conditions with perfect guiding - there is maximum magnification of the telescope that can be used. After that image just gets bigger without detail - it is blurred. Seeing of course blurs the detail and tracking precision does the same.

When all three are present - they combine and produce larger blur then each of individual components. Thing is - they don't combine in trivial way - just by simply adding some numbers. They combine a bit more complicated than that (in fact - quite a bit more complicated by process called convolution). However, we can simply things by using some approximations that will help us understand things.

Each of components can be represented by RMS value. Guide RMS comes as RMS value already. Seeing comes as FWHM, but can be converted to RMS by simply dividing it with 2.355. Telescope aperture or more precisely Airy disk it produces can also be represented as RMS (although somewhat more complicated calculation) - but for the time being, we won't pay much attention to it - other than to say that for most apertures - like 80mm+ - it is the smallest component.

You will notice that I emphasizes the smallest component in last sentence - that is for a reason.

Simplified formula for calculating total blur goes like this: square_root(first_rms_squared + second_rms_squared + third_rms_squared).

So we have square root of sum of squares. If this reminds you of Pythagorean theorem - then great, because we want to look at it in that way:

why? Well, because of this case:

If shorter leg is significantly smaller than longer leg then hypotenuse is almost the same length as longer leg. I'll reiterate in a bit different way - if one component of the sum is much smaller then the others - then it is contributing much less.

You will say - hold on, that is true for ordinary sum as well: 10 + 1 = 11, 10 and 11 are not that far away. Yes, but look what happens when we add 10 and 1 in quadrature - square_root(100 + 1) = ~10.05

Look how much smaller the difference gets when things are added this way.

What does it all mean - how big your guiding error is then? Well - that depends on largest factor, or what your guide RMS is compared to other big RMS out there that is seeing.

You mention seeing of 1.2" FWHM - well, year, that does not happen , or rather happens once a year on average site if at all.

Usual value is 2" FWHM or 3" FWHM if seeing is average to poor. 1.5" FWHM is excellent seeing for most sites.

Let's translate that into RMS:

1.5" FWHM = 0.637" RMS

2.0" FWHM = 0.85" RMS

3.0" FWHM = 1.274" RMS

4.0" FWHM = 1.7" RMS

In order not to contribute much - we need guide RMS to be quite a bit smaller than seeing RMS, and if your guide RMS is 1.2" - it is never significantly smaller than seeing RMS.

Guide RMS needs to be as small as you can make it. Simple as that. Only when you reach 0.2-0.3" RMS levels - you can say, ok, so I made it small enough compared to average seeing conditions (~x4 smaller) so I don't have to worry too much about it.

What about that - versus pixel size thing? Making your RMS half of your imaging scale is good rule of the thumb that works for most common cases. Here is an example:

Say you are imaging in 3" FWHM and you have 1.2" RMS guide error. Rule of the thumb says you should have 2.4"/pixel - as that is twice your guide RMS.

Let's see if that is true.

3" FWHM is 1.274" RMS and combined with 1.2" RMS that gives: sqrt(1.274^2 + 1.2^2) = 1.75" RMS or ~4.12" FWHM (when we multiply back with 2.355).

Optimum sampling rate for that level of blur is 4.12 / 1.6 = 2.575 - which is very close to 2.4

Even if you add aperture size in the mix - you still get very close results in common range - that is (1.5"/px - to 2.5"/px, 2" FWHM - 3" FWHM seeing and 0.7"-1.2" guide RMS, 4"-8" aperture).

However, if you want accurate results - there are complex formulae that will calculate effective resolution of your system and what pixel size to use (but these contain some approximation - like perfect optics, which is not always the case and so on ...)

Bottom line - make your guiding the best you can (lowest RMS value) always. Don't "settle" for it until you reach 0.2-0.3" RMS. Mind you - that low numbers are not always possible with mass produced mounts, so do research of what can be done and at what cost.

[iantaylor2uk]

Good explanation vlaiv - in terms of seeing RMS though, how does the widespread use of multi-star imaging impact this factor - it seems to me you can go to 0.5 secs or 1 secs guide camera exposure, and the fact you are guiding based on 8 or 10 stars must surely reduce the effect of seeing (compared to guiding on a single star). 

[vlaiv]

2 minutes ago, iantaylor2uk said:

Good explanation vlaiv - in terms of seeing RMS though, how does the widespread use of multi-star imaging impact this factor - it seems to me you can go to 0.5 secs or 1 secs guide camera exposure, and the fact you are guiding based on 8 or 10 stars must surely reduce the effect of seeing (compared to guiding on a single star). 

One of the approximation made in above discussion is that seeing and guiding/tracking accuracy are independent variables (linearly independent vectors add in quadrature).

This means that seeing FWHM is one that would be measured with perfect mount and that tracking / guiding accuracy is one that would be measured in completely still atmosphere.

There is usually some correlation between the two - but it can be reduced with using longer guide exposures - often 2s is given as enough time for atmosphere seeing to average out and tracking issues not to show (but sometimes 4s and longer is needed if seeing is particularly poor). Mount must have smooth error and be mechanically sound for this.

Multi star guiding helps to further separate the two, so it is beneficial thing. It won't improve above results - but it will make them more accurate. It shortens exposure needed to isolate seeing effects. With multi star setup even 1s is enough to remove seeing from guiding equation.

[Pitch Black Skies]

6 hours ago, vlaiv said:

While pixel size and guide RMS are often connected with some sort of rule of the thumb - I'd rather try to explain it a bit differently.

For the moment, forget pixel size. Forget pixels completely.

Let's just look what happens at focal plane of the telescope as the light comes in. There are three different major contributor to the blur that happens to the image.

- seeing

- guiding/tracking precision

- aperture size

Each one of these produces some level of blur on its own. In perfect conditions with perfect guiding - there is maximum magnification of the telescope that can be used. After that image just gets bigger without detail - it is blurred. Seeing of course blurs the detail and tracking precision does the same.

When all three are present - they combine and produce larger blur then each of individual components. Thing is - they don't combine in trivial way - just by simply adding some numbers. They combine a bit more complicated than that (in fact - quite a bit more complicated by process called convolution). However, we can simply things by using some approximations that will help us understand things.

Each of components can be represented by RMS value. Guide RMS comes as RMS value already. Seeing comes as FWHM, but can be converted to RMS by simply dividing it with 2.355. Telescope aperture or more precisely Airy disk it produces can also be represented as RMS (although somewhat more complicated calculation) - but for the time being, we won't pay much attention to it - other than to say that for most apertures - like 80mm+ - it is the smallest component.

You will notice that I emphasizes the smallest component in last sentence - that is for a reason.

Simplified formula for calculating total blur goes like this: square_root(first_rms_squared + second_rms_squared + third_rms_squared).

So we have square root of sum of squares. If this reminds you of Pythagorean theorem - then great, because we want to look at it in that way:

why? Well, because of this case:

If shorter leg is significantly smaller than longer leg then hypotenuse is almost the same length as longer leg. I'll reiterate in a bit different way - if one component of the sum is much smaller then the others - then it is contributing much less.

You will say - hold on, that is true for ordinary sum as well: 10 + 1 = 11, 10 and 11 are not that far away. Yes, but look what happens when we add 10 and 1 in quadrature - square_root(100 + 1) = ~10.05

Look how much smaller the difference gets when things are added this way.

What does it all mean - how big your guiding error is then? Well - that depends on largest factor, or what your guide RMS is compared to other big RMS out there that is seeing.

You mention seeing of 1.2" FWHM - well, year, that does not happen , or rather happens once a year on average site if at all.

Usual value is 2" FWHM or 3" FWHM if seeing is average to poor. 1.5" FWHM is excellent seeing for most sites.

Let's translate that into RMS:

1.5" FWHM = 0.637" RMS

2.0" FWHM = 0.85" RMS

3.0" FWHM = 1.274" RMS

4.0" FWHM = 1.7" RMS

In order not to contribute much - we need guide RMS to be quite a bit smaller than seeing RMS, and if your guide RMS is 1.2" - it is never significantly smaller than seeing RMS.

Guide RMS needs to be as small as you can make it. Simple as that. Only when you reach 0.2-0.3" RMS levels - you can say, ok, so I made it small enough compared to average seeing conditions (~x4 smaller) so I don't have to worry too much about it.

What about that - versus pixel size thing? Making your RMS half of your imaging scale is good rule of the thumb that works for most common cases. Here is an example:

Say you are imaging in 3" FWHM and you have 1.2" RMS guide error. Rule of the thumb says you should have 2.4"/pixel - as that is twice your guide RMS.

Let's see if that is true.

3" FWHM is 1.274" RMS and combined with 1.2" RMS that gives: sqrt(1.274^2 + 1.2^2) = 1.75" RMS or ~4.12" FWHM (when we multiply back with 2.355).

Optimum sampling rate for that level of blur is 4.12 / 1.6 = 2.575 - which is very close to 2.4

Even if you add aperture size in the mix - you still get very close results in common range - that is (1.5"/px - to 2.5"/px, 2" FWHM - 3" FWHM seeing and 0.7"-1.2" guide RMS, 4"-8" aperture).

However, if you want accurate results - there are complex formulae that will calculate effective resolution of your system and what pixel size to use (but these contain some approximation - like perfect optics, which is not always the case and so on ...)

Bottom line - make your guiding the best you can (lowest RMS value) always. Don't "settle" for it until you reach 0.2-0.3" RMS. Mind you - that low numbers are not always possible with mass produced mounts, so do research of what can be done and at what cost.

Awesome, will have to give this a proper read later. Legend.

[Pitch Black Skies]

8 hours ago, vlaiv said:

While pixel size and guide RMS are often connected with some sort of rule of the thumb - I'd rather try to explain it a bit differently.

For the moment, forget pixel size. Forget pixels completely.

Let's just look what happens at focal plane of the telescope as the light comes in. There are three different major contributor to the blur that happens to the image.

- seeing

- guiding/tracking precision

- aperture size

Each one of these produces some level of blur on its own. In perfect conditions with perfect guiding - there is maximum magnification of the telescope that can be used. After that image just gets bigger without detail - it is blurred. Seeing of course blurs the detail and tracking precision does the same.

When all three are present - they combine and produce larger blur then each of individual components. Thing is - they don't combine in trivial way - just by simply adding some numbers. They combine a bit more complicated than that (in fact - quite a bit more complicated by process called convolution). However, we can simply things by using some approximations that will help us understand things.

Each of components can be represented by RMS value. Guide RMS comes as RMS value already. Seeing comes as FWHM, but can be converted to RMS by simply dividing it with 2.355. Telescope aperture or more precisely Airy disk it produces can also be represented as RMS (although somewhat more complicated calculation) - but for the time being, we won't pay much attention to it - other than to say that for most apertures - like 80mm+ - it is the smallest component.

You will notice that I emphasizes the smallest component in last sentence - that is for a reason.

Simplified formula for calculating total blur goes like this: square_root(first_rms_squared + second_rms_squared + third_rms_squared).

So we have square root of sum of squares. If this reminds you of Pythagorean theorem - then great, because we want to look at it in that way:

why? Well, because of this case:

If shorter leg is significantly smaller than longer leg then hypotenuse is almost the same length as longer leg. I'll reiterate in a bit different way - if one component of the sum is much smaller then the others - then it is contributing much less.

You will say - hold on, that is true for ordinary sum as well: 10 + 1 = 11, 10 and 11 are not that far away. Yes, but look what happens when we add 10 and 1 in quadrature - square_root(100 + 1) = ~10.05

Look how much smaller the difference gets when things are added this way.

What does it all mean - how big your guiding error is then? Well - that depends on largest factor, or what your guide RMS is compared to other big RMS out there that is seeing.

You mention seeing of 1.2" FWHM - well, year, that does not happen , or rather happens once a year on average site if at all.

Usual value is 2" FWHM or 3" FWHM if seeing is average to poor. 1.5" FWHM is excellent seeing for most sites.

Let's translate that into RMS:

1.5" FWHM = 0.637" RMS

2.0" FWHM = 0.85" RMS

3.0" FWHM = 1.274" RMS

4.0" FWHM = 1.7" RMS

In order not to contribute much - we need guide RMS to be quite a bit smaller than seeing RMS, and if your guide RMS is 1.2" - it is never significantly smaller than seeing RMS.

Guide RMS needs to be as small as you can make it. Simple as that. Only when you reach 0.2-0.3" RMS levels - you can say, ok, so I made it small enough compared to average seeing conditions (~x4 smaller) so I don't have to worry too much about it.

What about that - versus pixel size thing? Making your RMS half of your imaging scale is good rule of the thumb that works for most common cases. Here is an example:

Say you are imaging in 3" FWHM and you have 1.2" RMS guide error. Rule of the thumb says you should have 2.4"/pixel - as that is twice your guide RMS.

Let's see if that is true.

3" FWHM is 1.274" RMS and combined with 1.2" RMS that gives: sqrt(1.274^2 + 1.2^2) = 1.75" RMS or ~4.12" FWHM (when we multiply back with 2.355).

Optimum sampling rate for that level of blur is 4.12 / 1.6 = 2.575 - which is very close to 2.4

Even if you add aperture size in the mix - you still get very close results in common range - that is (1.5"/px - to 2.5"/px, 2" FWHM - 3" FWHM seeing and 0.7"-1.2" guide RMS, 4"-8" aperture).

However, if you want accurate results - there are complex formulae that will calculate effective resolution of your system and what pixel size to use (but these contain some approximation - like perfect optics, which is not always the case and so on ...)

Bottom line - make your guiding the best you can (lowest RMS value) always. Don't "settle" for it until you reach 0.2-0.3" RMS. Mind you - that low numbers are not always possible with mass produced mounts, so do research of what can be done and at what cost.

So I've read your post and really enjoyed it, and it's not the first time you've answered one of my questions with such clarity and detail. I really think you should write an astronomy/astrophotography book, seriously. I've been wanting to say it for a while.

Now here's an utterly stupid question.

How is someone supposed to know what FWHM they have at a given location?

[vlaiv]

27 minutes ago, Pitch Black Skies said:

How is someone supposed to know what FWHM they have at a given location?

Well, you can't know it in advance simply because there is no way of telling what will the seeing be on particular night.

What you can do, however, is do a retrospective and form "a sense" or "a feeling" of what your gear is capable of delivering (and also test above formulae).

You can always measure FWHM of star profiles in your subs after you take them to see if they match your expectations.

It is best to record guide RMS and also note down seeing forecast for the night and see if it matches measured FWHM in your subs.

You can find seeing forecast for your location here:

https://www.meteoblue.com/en/weather/outdoorsports/seeing/london_united-kingdom_2643743

(just select proper location - I made a link to London as location).

There is column that describes seeing in arc seconds - that is FWHM value without impact of guiding and aperture size.

When using above formulae, do pay attention that two things can skew up results.

1. Local seeing. Above forecast is for global seeing, but local seeing effects can make things worse. Things like shooting over heated houses in winter or large bodies of water or pavement in summer can cause issues and worse final FWHM than calculations suggest

2. Type of scope and its optical performance.

Above I outlined how things work for perfect aperture (actually - I never did give airy disk RMS approximation so I'll do it here).

RMS approximation of aperture is given by 47.65/aperture_size where aperture size is in millimeters.

80mm scope will have RMS value of ~0.595625"

100mm scope will have RMS value of ~0.4765"

150mm scope will have RMS value of ~0.318"

200mm scope will have RMS value of  ~0.23825"

It falls quickly (as it depends on inverse of aperture size) and becomes insignificant to other sources as we have seen.

Above however holds only for diffraction limited scopes. It is often case that poor focusing, or using focal reducer or other optical corrector (like coma corrector) - reduces outer field aberrations, but makes scope less than diffraction limited over whole field (it's small tradeoff for good looking stars at the edges), and this can give something worse FWHM than predicted by above formulae.

Above formulae seem to give best case scenario - so that is something to keep in mind.

 

 

[edarter]

I was about to ask exactly the same question about how to confirm the seeing conditions on any given night. I use NINA, which shows a FWHM graph for stars in each image - I thought that was focus related, but is it effectively the seeing conditions for that image?

Thanks
Ed