SBIG AOX versus SX AO

[gregbradley]

I have been wanting to implement an AO unit in my imaging for some time.

I wonder if the SBIG AOX is a superior AO unit to the SX version or do they both have similar performance?

Its a hard question I know as I suppose those who have an AO unit are likely to only have had one or the other.

Perhaps someone though has or has had another SBIG AO unit they can compare.

Are all AO units the same in performance or is it more some are better than others like Astro cameras?

Greg.

[Paul Haese]

I don't have any comparative evidence on which is better. What I do know is that the AOX is a multi generation technology now. It appears to have improved over each generation to its current point. It integrates very well with SBIG STXL camera. There are some minor glitches with the software but overall I can say I will never go back to conventional guiding again, such is the power of this technology. I am assuming that the Starlight Express AO has undergone similar development. That is about all I can offer.

[gregbradley]

Originally Posted by Paul Haese

I don't have any comparative evidence on which is better. What I do know is that the AOX is a multi generation technology now. It appears to have improved over each generation to its current point. It integrates very well with SBIG STXL camera. There are some minor glitches with the software but overall I can say I will never go back to conventional guiding again, such is the power of this technology. I am assuming that the Starlight Express AO has undergone similar development. That is about all I can offer.

Thanks Paul,

Its your images in particular that have made me want to implement an AO.

I have been finding out more about it. It seems an AOX can be made to work with a MMOAG and an ST7XME or ST8XME guide chip as the guider in the MMOAG. I wonder about the weight and the space requirements of that setup though. An ST7XME might be pretty heavy as a guider. I wonder if it can be stripped down to bare necessity?

Greg.

[Slawomir (Suavi)]

I have been also thinking of adding an AO unit to my setup, as I expect it should allow for collecting better quality data, in particular when used with a mass-produced mount.

[codemonkey (Lee)]

Can the AOX be used with non-SBIG cameras?

[rally]

No.

The interface is proprietory - the software and firmware to drive it is integrated with the controller which is part of the SBIG camera electronics and uses the onboard SBIG guider chip (second chip next to the main imaging chip) to provide the feedback info.

But I believe as has Greg stated, that others have discovered you can use your own choice of main camera (Non SBIG) and use an older ST series SBIG camera as the guide camera with the AO

Greg - the weight of an FLI plus SBIG ST series camera is probably less than the weight of a STL11000m with 8 filters !
The thing I dont like about the MMOAGs is they only use 4 grub screws in a circular dovetail on either end, it wont fall off, but its hardly the best design for concentrically holding a heavy payload aligned with the optical axis - all the load is being taken at 4 sharp points on either end and they permanently indent the dovetail making fine adjustment difficult later if you only want to rotate it say 1/2 a degree.

Rally

Originally Posted by codemonkey

Can the AOX be used with non-SBIG cameras?
. . .
Originally Posted by Greg
I wonder about the weight and the space requirements of that setup though. An ST7XME might be pretty heavy as a guider.

[codemonkey (Lee)]

Thanks Rally

[gregbradley]

Yes the 4 grub screws on the MMOAG is probably not the best solution. The proprietary FLI zero tilt adapter works much better.

So Rally you think it would be better to use an STL11 with the older AO units than a 16803 with an AOX and a ST7 guide cam?

Does the new AOX work better than the older SBIG AO units I wonder?

Greg.

[rally]

Not really sure Greg,

Its not really my area of expertise

It depends a lot on what you are trying to achieve and what camera you prefer to image with.
The AOL and AOX both operate slower than the AO7
I think the AO8 was in between

The AO7 used a tilt/tip mirror and thats why the camera sits perpendicular to the optical path, whereas the later AOs used a prismatic tilt/tip arrangement - its just flat parallel glass - so its a straight through light path.
As the design changed from mirror to prism and the size of the glass has progressively increased to cater for bigger CCDs, the increased mass has meant that the energy to tilt/tip has also increased - somewhere along the line the constraints of effective operation has meant that the maximum correction frequency has gone down.

So the AO7 can correct at higher frequencies than the others - but obviously only with a smaller aperture - aka smaller CCD chip

Is this good or bad ? - I'll leave that to others to argue about !
Some will say its a big disadvantage as you cannot correct for higher frequency wavefront aberrations, some will say it doesnt matter as this type of adaptive optics is so limited by its low overall frequency that any gain is better than nothing, and that the best we can hope to correct is only the slow stuff anyway !
Truth is all of these AOs are barely able to correct 50% of the movement

The Pros would like to be able to correct at high speeds up to 1000Hz which is difficult for mortals, but I think typical speeds are maybe 100Hz
But they are also trying to truly use adaptive optics for correction for the shape of the wavefront aberration which is much different to a simple tip/tilt correction.

The AO7 could go to about 20+Hz I think - the internal guide camera is only doing at most about 40 frames per sec anyway and it took about 10ms for the mirror to fully react to any move
The later AO models were much slower - 10Hz or less - I think they took about 50ms to react to a full move

So this makes them very good for high speed guiding correction which is useful - but really for true AO I think they are quite limited

In answer to your specific question If the 16803 is your chip of choice being a square STL11000 then that is really the constraint.
The underlying software will be the same, the speed of the AOX and AOL is probably about the same - I havent checked and havent owned an AOX so Im not sure about it - havent even read a spec sheet !
But you would still only be able to use an AOL (or AOX) with the STL11000m - the AO7 and AO8 are too small to cover the chip and the AO7 has some rotational effects occurring that would make things even worse on a big chip.

I think you need to assume that what you are going to get is much better guiding, so it keeps your stars centred without bloating due to continual wobbling between guide intervals - any magic for true adaptive optic correction isnt likely.

This is only relevant for longer focal lengths, atmospheric effects will mean that what is happening on one corner of your CCD is not necessarily the same as what is happening on the other ! - ie there is enough physical distance in the lower/middle atmosphere between these stars across say an angular patch of sky across 30-60 arc secs.
So the gains may or may not be there across the complete image, but your guiding will be better !! and that will improve your FWHM - better stellar profiles.

You can use a ST7 for AO control by using a Remote Guide Head - so the camera can be positioned on top of the OTA rings for arguements sake and just plug the RGH into the MOAG - makes life easier if you have the spare $500+ worth of RGH laying around.

The other thing is that the latest SBIG cameras and AOs mean that the guide camera is in front of the filters not behind with the main chip on the ST, STL series. This makes guiding especially high speed better - especially if you are using NarrowBand, but if you are using an RGH and ST7 then that has already been achieved anyway. So that makes good sense as a cheaper option too.

Bit of a ramble - hope it helps

[Shiraz (Ray)]

don't know if any of this will help Greg - it's almost impossible to glean anything useful from the published data.

The SX device uses 4 stepper motor drives and has 5 ms steps - but they don't tell you how big the step is, so there is no way to work out the maximum slew rate. You should be able to make the SX unit work with any set of cameras. The device is sold as a corrector for mount errors, with no mention of seeing correction. I have used an earlier model SX. It worked as specified and integrated seamlessly with phd2. It was capable of correcting errors (whatever the source) of up to about 5hz for normal excursions at 1"/pixel.


The SBIG website is similarly uninformative, but the AOX uses what appears to be a voice-coil-like magnetic drive and it can work at up to a 10hz update rate (for an unspecified correction amount etc.). Claims are made for it being able to correct seeing errors, but the supporting data provided on the website are very old simulation results, based on unspecified assumptions plus some (old?) imagery, again taken under unspecified conditions - not tests from the sort of system that you would use, so not of much benefit to your decision making.

FWIW, both devices probably do basically the same thing, so it looks to me like the SX is the best option for your setup, purely on the basis of ease of getting it to work with your cameras. If you had an SBIG camera, the AOX would be the clear choice. The SBIG device likely has an edge in correction speed, but that is conjecture and not based on any real data. In any event, the ability to run an AO at high speed will depend on the availability of a bright guidestar and it would be interesting to hear from others what rates they can actually guide at in the real world. My understanding is that others see benefit from an AO at a fairly low guide rates in good seeing. You can find an informative discussion on guide rate considerations in the old AO8 manual on the SBIG website - the general ideas for stepper drive (like the SX) and direct drive (like the AOX) are presented.
"The AO-8 is somewhat slower than an AO-7, since the motor and
gearbox can only tip the motor at 18.75 degrees per second, or 75
pixels per second. In this design moves are slew-rate limited. What
this means to the user is that the user should use a focal length that
does not produce too large of a star image. For example, if the seeing
is bad and the star is 9 pixels across, it is probably jumping around by
4.5 pixels or so with each look, which would take 60 milliseconds to
move. If under such conditions you reduced your focal length to
where the star was only three pixels across, only 20 milliseconds are
spent making the move, and you can achieve a higher update rate with
the AO-8. The total overhead in the software for an AO-8 move is 45
milliseconds. If, for example, you are taking 50 millisecond
exposures with an optimized focal length (~3 pixel Full Width Half
Maximum stars), the total cycle time is 45 + 50 + 20 ms, or 115 ms (8
updates per second). The AO-7 advantage was that a move of any
length is 20 milliseconds, but the exposure and overhead times were
similar. Based on our experience and knowledge, the corrections
made by the AO-8 help with reducing ground layer atmospheric
turbulence, and are quite effective in reducing guide errors and wind
buffeting. The smallest increment of move for the AO-8 is 0.075
degrees per axis, or 2.7 microns, sbout a third of a pixel."

[gregbradley]

Rally and Ray, thanks for the informative posts. Very helpful. Its hard to find real world experience on these units.

Greg.

[Paul Haese]

Just for some info on my use.

I have found that most stars of around mag 9 will produce speeds of 1.7Hz. The fastest I have found so far was using mag 1.7 star and I tried speeds of 10Hz. I could have gone faster but did not try. I am told that the AOX can operate at 30Hz. Which is very fast but not sure if this is really practical for long focal length imaging. There are not that many bright stars that close to lots of objects.

I have found that guiding at 1.7 Hz is still better than native guiding at 1-2 second exposures. Good seeing makes for very tight star profiles and sharp detail. Poor seeing produces sharp detail but not as sharp as in good seeing but still better than not using this type of guiding. So most of the correction is likely mount related and some seeing correction. A word of warning though. You really need to have your guiding really nailed for this to work. I spent several nights working on honing the guide settings both min and max moves. Not only that PEC and PA should be fully corrected and as close to the pole as possible. Mount corrections should be minimised to produce good results.

I can confirm that the AOX is a voice coil magnet Drive x 4 (two in each axis). I have seen inside of these units and have seen one working in exercise mode.

[gregbradley]

Thanks Paul. It's primarily your sharp images that has brought it forward in my planning.
Greg

[Ken]

Not sure I can add much to this but I recently added a SXao to my homemade Fork mounted Newt. hoping to improve guiding. The latest version has moved from serial to usb and more power is available to the motors according to SX. If the mount has problems with backlash while calibration is in operation the unit will move beyond its normal limits and 1 of the motors will jam. Only way to unjam is to turn that motor shaft with a tiny screw driver usually having to take it off the scope,not fun in a imaging session.
While I don't have a lot of experience yet my results show that if the seeing is bad it will not help much but if the seeing improves it will keep things very tight with PHD plotting mostly in 1.5 arc sec circle at 2 m fl.
Clear skies Ken.