"Adaptive Optics" - any experience and any good?

[gorann]

I have seen that both SBIG and Starlight Xpress sell "adaptive optics". The SBIG versions only seem to fit SBIG cameras but Starlight Express makes one that may fit many brands. They booth look fairly old in the electronic design with RS232 plugs and I have the impression that they have been around for a while and not evolved much. I cannot remember having seen any posts from people using them and wonder if they are not worth the extra expense. Still, the principle seems very smart: an optical unit in the light path that adjusts minor movements in a guide star on a millisecond basis, while the guide camera also sends the usual slow signals every odd second to the mount.

Here are the products I found:

http://diffractionlimited.com/products/cameras-accessories/adaptive-optics/

https://www.sxccd.com/sxv-ao-lf

Any comments anyone?

[Whirlwind]

A better term for these would "active" optics rather than adaptive optics.  They won't correct for seeing in the same way that scientific adaptive optics will do.  

However, they can be useful for fast guiding where that might be useful such as mirror movement, flexure etc or where you might have a mount that has rapid error changes that don't respond well to corrections (e.g. lots of backlash).  This is because you only have small amount of glass to move rather than the whole mount.  To get the ultrafast adjustments you do need a very bright guide star though which isn't always practical on longer focal telescopes.

The SX AO does have a USB connection though.  They tend not to update their images very frequently. 

[DaveS]

SX available from FLO

Here

I've considered this on and off (More off than on) for my ODK12. I don't need to guide, the DDM mounts do that very well indeed, but the AO could correct slow star wandering.

[andrew s]

2 minutes ago, DaveS said:

SX available from FLO

Here

I've considered this on and off (More off than on) for my ODK12. I don't need to guide, the DDM mounts do that very well indeed, but the AO could correct slow star wandering.

Somewhat puzzled by this Dave. What do you see as causing the slow wandering that the DDM fails to correct for?

Regards Andrew 

[DaveS]

Atmospheric turbulence "seeing". The DDMs guide on their encoders with a sky model so don't see a guide star. They are very good at correcting atmospheric refraction and wind gusts but the small scale wavering isn't seen.

[andrew s]

4 minutes ago, DaveS said:

Atmospheric turbulence "seeing". The DDMs guide on their encoders with a sky model so don't see a guide star. They are very good at correcting atmospheric refraction and wind gusts but the small scale wavering isn't seen.

Thanks that's what I thought. In the end for however good the feedforward control or feedback control of secondary variables ultimately you need to close the loop on the target variable (star) for ultimate precision.  Not that that is always required. 

Regards Andrew 

[vlaiv]

In my view - completely useless thing.

It won't correct for seeing for different reasons:

- it only corrects for first order aberration - tilt

- seeing aberrations, or specifically tilt is very local phenomena - seeing happens in higher in atmosphere at at least couple of kilometers of altitude. It is enough to move couple of arc seconds away from a guide star and that translates to significant length at that distance (for example 20 arc seconds at 5km is already about half a meter - seeing cells tend to be less than that). Any sort of wider field will be distorted by different amount of tilt - you can easily see this effect if you look at planetary recording or maybe Lunar recording - distortions (jumping around of the features - mostly due to tilt) is different at different places. Correcting for tilt at location of guide star will do nothing for rest of the image and probably cause more harm then help.

- it does corrections on time scales that are most of the time above time scales of seeing changes. It does something like 30hz max if I'm not mistaken, and seeing often goes as fast as 100hz or more.

It does have one use, and it is at best limited at that - correcting for a rough mount. If you want good guide performance but your mount is rough and it has small time scale error large enough (like 0.5"-1.0" or similar jitter) that can't be corrected with guiding because guide exposure is longer and mount is not as responsive enough - this will catch it and react provided that seeing is good enough that any guide star deviation from true position is due to mount and roughness rather than being "masked" by seeing.

On the other hand - why buy such expensive unit that might not perform as expected when you can instead sell current mount + invest that money into purchasing a new better mount that will be smoother and guide better.

 

 

[pete_l]

1 hour ago, vlaiv said:

On the other hand - why buy such expensive unit that might not perform as expected when you can instead sell current mount + invest that money into purchasing a new better mount that will be smoother and guide better.

Why? If a €1500 device can give a €4000 EQ8 the same level of tracking performance as a €13000 GM2000 then it would be worth looking at
I cannot say that it would, but the concept is there. The question is whether an Active Optics unit can actually deliver?
That is what nobody seems to be able to say definitively: either yes or no.

There are plenty of opinions, but very little practical experience.

[andrew s]

No withstanding @pete_l comment on theory the idea of the device makes good sense in control theory.

If you are trying to accurately control say a flow of liquid and have a slow sticky valve you can add a small fast smooth valve in parallel to do this. It was a common practice when I was a control engineer. These units seek to do the same thing.

So probably best for lower class mounts. 

Regards Andrew 

[gorann]

9 hours ago, vlaiv said:

In my view - completely useless thing.

It won't correct for seeing for different reasons:

- it only corrects for first order aberration - tilt

- seeing aberrations, or specifically tilt is very local phenomena - seeing happens in higher in atmosphere at at least couple of kilometers of altitude. It is enough to move couple of arc seconds away from a guide star and that translates to significant length at that distance (for example 20 arc seconds at 5km is already about half a meter - seeing cells tend to be less than that). Any sort of wider field will be distorted by different amount of tilt - you can easily see this effect if you look at planetary recording or maybe Lunar recording - distortions (jumping around of the features - mostly due to tilt) is different at different places. Correcting for tilt at location of guide star will do nothing for rest of the image and probably cause more harm then help.

- it does corrections on time scales that are most of the time above time scales of seeing changes. It does something like 30hz max if I'm not mistaken, and seeing often goes as fast as 100hz or more.

It does have one use, and it is at best limited at that - correcting for a rough mount. If you want good guide performance but your mount is rough and it has small time scale error large enough (like 0.5"-1.0" or similar jitter) that can't be corrected with guiding because guide exposure is longer and mount is not as responsive enough - this will catch it and react provided that seeing is good enough that any guide star deviation from true position is due to mount and roughness rather than being "masked" by seeing.

On the other hand - why buy such expensive unit that might not perform as expected when you can instead sell current mount + invest that money into purchasing a new better mount that will be smoother and guide better.

 

 

Thanks, this is quite clarifying and I agree on everything you say,  but it seems like the manufacturers may agree with you and are selling this primarily to adjust for mount error and not for seeing.  Actually, I have realized that even my relatively cheap EQ8 is "seeing limited". Thus at good nights I get an RMS of 0.4" / pixel while on average seeing nights it is more like 0.6 - 0.8", so "adaptive optics" would not help at all, possibly the opposite as you say.

[gorann]

8 hours ago, pete_l said:

Why? If a €1500 device can give a €4000 EQ8 the same level of tracking performance as a €13000 GM2000 then it would be worth looking at
I cannot say that it would, but the concept is there. The question is whether an Active Optics unit can actually deliver?
That is what nobody seems to be able to say definitively: either yes or no.

There are plenty of opinions, but very little practical experience.

As you say there seems to be very little practical experience, as far as I can see NONE in this thread (so far). It probably means that this has not been a selling success and that is why no new models are coming out. People may have realized that it is better to trim the mount or upgrade than spending 1500 or more Euros on an active optics unit.

[vlaiv]

I did read once a review of this unit, and author did get tighter stars by using it, but can't remember what mount it was on.

Will try to find it again and post a link

12 hours ago, pete_l said:

Why? If a €1500 device can give a €4000 EQ8 the same level of tracking performance as a €13000 GM2000 then it would be worth looking at
I cannot say that it would, but the concept is there. The question is whether an Active Optics unit can actually deliver?
That is what nobody seems to be able to say definitively: either yes or no.

There are plenty of opinions, but very little practical experience.

Yes, you are right that some mounts are rather expensive, but why consider GM2000 and not something like Mesu 200?

EQ8 + AO device will place you in Mesu 200 price range, and I have sneaky suspicion that Mesu 200 will deliver at least as good performance if not better without complexity of running an AO unit.

[gorann]

Actually this autumn my EQ8 is behaving exceptionally well. Maybe he is trying to prove himself after I put him in the garage and replaced him with a new and shiny Mesu200 on the pier. But then I built a second obsy for him and a 14" Meade ACF that I got cheaply. And now the EQ8 and 14" SCT play along quite well, even getting down to RMS 0.4"/pixel on a good night. After some soul searching I realized that it was a bit unfair to put the EQ8 in the garage and his missbehaviour was most likely due to bad seeing some nights. He is not a Mesu but after I tightened up his backlash, he is really quite good.

Edited September 5, 2019 by gorann

[AKB]

8 hours ago, gorann said:

As you say there seems to be very little practical experience, as far as I can see NONE in this thread (so far).

IIRC, Sara @swag72 had one of these.  Perhaps she can provide some actual experience. 

[blinky]

There were members years ago that had these, think they have possibly left the forum nowadays but I seem to remember that there were not rave reviews! Something about it being tricky to get a bright enough star and I’m sure somebody else said it was variable whether it made a diff or not, the jist of it was ‘not with the hassle’

[swag72]

I had one of these for a while - I thought it would be the Shangri-la of great data, despite already having a pretty decent mount (Avalon Linear). At the time I was using a long focal length scope...... I think it was when I was playing with a C9.25.  At the time the only software that the SXAO used for guiding was Maxim and the recommendation was that you use very short exposures. I could never find a bright enough guide star to get the short exposures that were supposed to make the difference. Then I found that almost every night without fail it would run out of adjustment and so most of the night would be wasted. 

I found it far more trouble than it was worth. I did my homework and in theory I thought it should make a difference. I was sad to see that there appeared to be no discernible difference in my data quality. I guess it was something that I had to try, but I did not find it worked for me and so gave up with it and moved on.

[gorann]

19 minutes ago, swag72 said:

I had one of these for a while - I thought it would be the Shangri-la of great data, despite already having a pretty decent mount (Avalon Linear). At the time I was using a long focal length scope...... I think it was when I was playing with a C9.25.  At the time the only software that the SXAO used for guiding was Maxim and the recommendation was that you use very short exposures. I could never find a bright enough guide star to get the short exposures that were supposed to make the difference. Then I found that almost every night without fail it would run out of adjustment and so most of the night would be wasted. 

I found it far more trouble than it was worth. I did my homework and in theory I thought it should make a difference. I was sad to see that there appeared to be no discernible difference in my data quality. I guess it was something that I had to try, but I did not find it worked for me and so gave up with it and moved on.

A short but very informative review Sara that I think confirmed the general though of this thread - at present this is not the way to go for better guiding. Thanks a lot!

[michael8554]

4 hours ago, swag72 said:

would run out of adjustment

A long time ago I read the PHD2 instructions for AO, I seem to remember that PHD2 would allow the AO to reset when the end stop on the AO was reached. 

Michael 

[swag72]

6 hours ago, michael8554 said:

A long time ago I read the PHD2 instructions for AO, I seem to remember that PHD2 would allow the AO to reset when the end stop on the AO was reached. 

Michael 

I don't think when I was using it PHD2 was in existence......

Unfortunately my experience was a while ago , in the earlier days so to speak, and so I have nothing but brief anecdotal evidence as I didn't do any comparisons in those days with data with / without the AO.

[Area 52]

It's all about the atmosphere, period.  I've seen perfectly good scopes sold because the owner judged the optics visually ... or by imaging on bad nights.  No consideration was given to the atmosphere.   I used a CCD guided telescope assisted by a SBIG AO from 2001 to 2008 discovering asteroids.  Sub-arcsecond tracking.. or for a matter of fact... arcsecond tracking for any period of time is a true rarity at best... even with an AO.  The most steady seeing I had was during foggy weather.  The Minor Planet Center (MPC) recommends ~ 2 arcseconds per pixel.  This recommendation recognizes the rarity of arcsecond tracking/seeing and the effects on an image.  I recommend reading the IAU/MPC site on effects of atmospheric conditions on imaging.  Using UCAC4 or Nomad catalogs to reduce the raw data allows for astrometry accuracies of ~ .2 to .3 arcseconds through advanced algorithms, only.  My scope was set at 2.7 arcseconds per pixel.  Using UCAC 3, I reduced my residuals on average to~ .3 arcseconds.  Again, my scope with AO VERY rarely tracked to arcsecond accuracy for any period of time.  AO's are not for the week of heart.   My AO/scope combination was a pain in the rear to setup.  It took weeks to get the numbers working, but, without a doubt, I would never have achieved the tracking I did without it.  According to the manufactures, your bell curve can be reduced by approx. 25-30%.  It is my thought that adjusting a mirror is faster with less vibration than trying to reverse a motor with backlash adjustment.  A tip: Chasing turbulence/seeing at high rates is counterproductive.  I recommend corrections in the .5 to 3 - 4 second range.  You don't want to chase the seeing  but giving it time to allow for a small averaging effect.  Again, the IAU-MPC has a lot of information on this subject.

[Physopto]

I have the SX AO unit. It works fine. I have used it on my NEQ 6, on my EQ8 and on my GM1000. Once set up correctly it tilts the glass until a set point then bumps the mount, resets to centre position starts again. Stars are well controlled. But I have to admit  it can be a pain in the butt to get working until you understand it properly.

I have found by trial it is certainly not needed with the GM1000. In fact possibly makes things slightly more less focussed to the eye. I hasten to say this was tested with both my WO 132 and Tak 85mm on that mount.

It works as Sara has said with Maxim, but I found that if it even clouded overslightly and it lost sight of the guide star, it would send the mount off on a hunt, ruining the sub.  I did get PhD working and that was much better. But by that time I was moving on to the 10 Micron.

There are too many people who actually have not used an AO unit and who make uninformed judgements. I am not blaming anyone as discussions are a good method of discovering the truth. But one person's pain in the rear may not be another's.

The AO unit is expensive and not everyone's cup of tea. To be brutally honest I don,t think I would buy it again, but only because I bought the 10 Micron mount.

Derek

 

[Area 52]

I see we're discussing two different situations.  Let me explain how an AO helped me.  I used a Meade 12" LX200, F:6.3 reducer and a self guiding SBIG-9E with exposures of at least 6 minutes each.  Focal length of the system was ~1750mm.   The full moon completely filled my image, so relatively small field of view.   All of my data is supported by actual equipment measurements, not on mfg ads.  At this magnification the atmosphere becomes an issue no matter the mount or optic quality.  The optics or mount quality has nothing to do with seeing or transparency.   They are two different subjects.  That is why the professional observatories us TRUE AO optics.  They have the best drives and optics money can buy.   All low magnification does is bring the point spread down due to the effects of not seeing the sky for what it is.   Meade and Starlight Express both use the vernacular AO loosely.  True AO optics uses lasers to measure atmospheric distortion; thereby, changing the shape of the mirrors to correct for the effects on a microsecond level.  Neither Meade nor Star Light express does this.  They are more or less medium speed drive correctors.... and this isn't a bad thing.  Meade makes inexpensive mounts with relatively so-so optics that is cost effective.  Periodic error can be +/- 10 arcseconds.   An AO was critical to getting my error down to an acceptable level at my F ratio.  $1500 for an AO sure beat spending $20,000 for a Mathis mount plus optics. 

Very short exposures can "beat the ceiling".  In other words, the effects of atmospheric distortion are minimized.  By using very short exposures stacked together using inexpensive equipment, I have seen planetary pictures that are absolutely incredible.  Results, high magnification with low atmosphere effects. But the effects of long exposures bring to light the demons of the sky.   My error was finally reduced to ~ +/- 1 arecsecond.  Not bad. This included atmospheric conditions .... on good nights.  Consider also the altitude of the object being imaged.  Looking over head, you looking through one atmosphere.  But go down 30-40 degrees and see what happens.  Overhead, I could capture 21st mag asteroids.  At 30 degrees down, I did good to catch 17 mag rocks.  How many atmospheres and how much mud was I looking through?  I guess what I'm saying,  If you're using low magnification, you have fairly good equipment shooting long exposures, an AO is not necessary.  But if the image shows effects either from the mount or atmosphere (it will not help bad optics) then it can be of benefit.   But be forewarned, expect to spend many hours extinguishing the dragon's breath.

[ollypenrice]

I'd forgotten that Sara had tried one of these. I know of only three people using them and the one who posted most recently was Peter Goodhew. Peter was finding it impossible to avoid effects of differential flexure in his dual 6 inch refractor rig so, if I have this right, he installed an AO unit on the 'slave' scope and now finds that this is the one which actually gives the lowest FWHM.

On 04/09/2019 at 23:44, DaveS said:

Atmospheric turbulence "seeing". The DDMs guide on their encoders with a sky model so don't see a guide star. They are very good at correcting atmospheric refraction and wind gusts but the small scale wavering isn't seen.

Autoguiders don't react fast enough to follow the seeing so, in theory, the advantage should be more or less equal whether guiding on encoders or on stars, I'd have thought? 

My 'digest' of what owners have told me about these units is that when they work they make a difference but getting them to work is often a pain. By far the most positive experience is Peter's. In effect it allows two parallel scopes to be independently guided. Our dual TEC rig does sometimes suffer slight trailing on the slave side but my low tech solution is to use that side for RGB.

Olly

Edited May 19, 2020 by ollypenrice

[DaveS]

OK Olly. As it happens I have given up on thinking about AO, too much hassle for dubious (In my case) rewards.

[Petergoodhew]

I'v been using the SX AO for a few months.  As Olly says, I use it to keep the second scope on my dual rig perfectly aligned and to compensate for differential flexure.  At 1200m focal length, and with minor differences between the scopes (tube, focuser, camera, and adapters are all slightly different) - and for 30 minute exposures - the flexure is enough to cause problems. The AO fixes it perfectly. I wouldn't recommend it as an adaptive optics solution (SX call it Active Optics, not Adaptive Optics). Yes the device is fast, but as Sara says, when I tried it my Lodestar X2 just wasn't able to pick up a bright enough star to refresh at the rate of the AO unit.  Maybe with a faster scope, shorter focal lent=gth, and wider field of view it should be better - but that's just a hunch, not experience on my part. Installing it, and getting it running with PHD2 was very easy.  If the drift exceeds to capacity of the AO then PHD2 sends a pulse to the mount to nudge a little to recentre the AO unit.  It's worth noting that the AO cannot function without a Lodestar - so this does get expensive as it's not just the cost of the AO unit.

 

Peter