More testing, Mesu2/ODK14

[NickK]

We need to take this as another subject, because what you propose is extremely interesting. Because essentially you will have mapped the optic system using this technique.

I have often wondered whether or not it is possible to correct an optical system mathematically rather than optically with field flatteners etc...

Simply put, if you know the field curvature and its repeatable, why not apply and inverse function to the image to correct it?

Very very interesting... topic for discussion, please start a new thread.

Agreed - this is very interesting on idea I had was to use the image analysis of the guide scope frames (ie 1-3 second exposures) and then produce a corrective vector space for the final image where the exposure is in the order of 10 minutes.

Naturally the maximum pixel depth is a limiting factor, as the difference in aberation for guide scopes etc. However I thought the idea had merrit for me being a OAG user (even with a partial image being monitored).

[ollypenrice]

Agreed - this is very interesting on idea I had was to use the image analysis of the guide scope frames (ie 1-3 second exposures) and then produce a corrective vector space for the final image where the exposure is in the order of 10 minutes.

Naturally the maximum pixel depth is a limiting factor, as the difference in aberation for guide scopes etc. However I thought the idea had merrit for me being a OAG user (even with a partial image being monitored).

Is there not a danger of the prism deforming the image in the guidescope, so making it unrepresentative of the primary image?

Olly

[michael.h.f.wilkinson]

Is there not a danger of the prism deforming the image in the guidescope, so making it unrepresentative of the primary image?

Olly

Agreed. Anyway, the off-axis guider image is at the edge of the field of the scope, so typically will have the worst problems. I do not know whether this can be done on live images (it might not be fast enough) but computing a quality map of the image plane should be possible at modest frame rates.

[Tom OD]

Just a thought, if you were imaging in Bin 2x2, won't that help beat any bad to mediocre seeing? Seeing as it would be a lower resolution.

T.

[ollypenrice]

Just a thought, if you were imaging in Bin 2x2, won't that help beat any bad to mediocre seeing? Seeing as it would be a lower resolution.

T.

Well, to go unbinned would be pointless at this FL because you'd be spending ages getting the data and losong the extra resolution to the seeing. It does seem weird going into bin 2 for Ha and luminance but that's 2.4 metres for you.

When we are all sorted and we have a night of great seeing we'll see if there is any future in unbinned. However, while I feel bin2 at short FL doesn't give you aything like 4x the signal, with 350mm aperture it seems to do so. There's a lot of discussion about this which I've yet to read. However, the seat of the pants reaction is to say that on this setup binning is super-fast compared with non-binning.

Olly

[NickK]

Olly.. succumb to active optics..

[ollypenrice]

Olly.. succumb to active optics..

I'm in favour in principle. I've seen both the SX and SBIG systems working here and work is what they did, though one owner did say the thing was 'a pain...'

On my own setups with FLs ending at a metre I don't think it applies.

Olly

[Darth Takahashi]

I'm not sure about the adaptive optics either but partly my concern is achieving enough space to put the thing in the image train, this is more of a concern when using the focal reducer, I think from memory it is 56mm for the Mewlon, not sure how that would work on the ODK???

How much back focus do you have Olly???

I have accidentally imaged unbinned (1x1) with the Mewlon (Ha 7nm) and it wasn't a disaster, in fact I definitely didn't see the full 4x gain in binning(2x2) with the ST10!!!

[vdb]

AO only work and add resolution in my opinion if

a) your mount has fast errors, ours doesn't

you have a guide star that allows several exposures per second so that many corrections can be made.

Yves.

[Darth Takahashi]

AO only work and add resolution in my opinion if

a) your mount has fast errors, ours doesn't

you have a guide star that allows several exposures per second so that many corrections can be made.

Yves.

Not sure that I'm understanding point ( please explain further?

[ollypenrice]

When I saw these things working they were flickering away at about 8 corrections per second from memory, but does that mean they were dealing in eigth of a second exposures? Quite honestly I've never given it a thought.

Olly

[Merlin66]

I'd say the binning would have to match the seeing and the focal length.

for a 2" FWHM star image, it would be 23.3 micron, and a 3" seeing would be 35 micron.

For good sampling is would infer a final pixel size of 11 micron ( for 2") or 16 micron (for 3")

[NickK]

Agreed. Anyway, the off-axis guider image is at the edge of the field of the scope, so typically will have the worst problems. I do not know whether this can be done on live images (it might not be fast enough) but computing a quality map of the image plane should be possible at modest frame rates.

If each frame is stored then you have concrete points in the time domain that could be analysed later. It doesn't need to be realtime. Applying the changes at the same time as combining would allow corrective wieghting to be done too to aid image resolution enhancement too.

To olly's point - if the guide scope aberations can be defined as a set of "constant" (a function) then you can also have the main scope set defined as quazi-constant too. When I mean constant, I mean you can have a set of parameters which can describe the behaviour in each situation of the input parameters.

It was just an idea to be investigated. Another option is to create a neural network that redefines itself on the quality of the image (as a function of the network itself). Big network given the size of the input data! A great excuse to abuse your GPU

[NickK]

When I saw these things working they were flickering away at about 8 corrections per second from memory, but does that mean they were dealing in eigth of a second exposures? Quite honestly I've never given it a thought.

Olly

Yes.

The optics systems available to us normals are based around watching the motion of the stars between each exposure (ie 8 exp per second).

The control system then has a loop that then moves a lens to correct.. it does that 8 times a second.

Big bad active optics use the light wavefront and deform the mirror using voice calls hundreds or thousands of times per second.. I think you're looking at 20K for a small cude deforming mirror for the secondary in a AO system.. then you have the cost of the sensors, prisms etc..

[vdb]

for AO to be effective you want multiple corrections per second, for multiple corrections per second you need a bright star, otherwise it will not be detectable at lets say 0.2 sec exposure time and hence AO will not obtain better results then a good mount ...

[rfdesigner]

When I saw these things working they were flickering away at about 8 corrections per second from memory, but does that mean they were dealing in eigth of a second exposures? Quite honestly I've never given it a thought.

Olly

Olly

I see your initial question has been more or less answered.

I've been reading round this whole thing for a while now and have discussed it with some people invloved with professional systems.

There are several things I've found of note to amateurs:

The low order abberations (tip / tilt) account for a large percentage of the total system error.

Whilst the best systems run at many 10s of Hz, even running as low as 5Hz can be of beneft, depending on accuracy, clearly delay between imaging and moveing the image is bad news, I think this is part of the reason for the lodestar being defaulted to 2x2 binning, it downloads in approx a quarter of the time.

Seeing is not generally dominated by any one source. I found a wonderful paper from observations and tests in the paranal mountains where they established that their seeing came from three sources...

A: Jet stream located at about 7km up, in a 500m or so thick band.

B: Gound effects, below a couple of hundred meters.. the lower the level the more impact it had.

C: residual all hight abberations.

each source accounted for about 1/3rd the total seeing.

Thus using an tip-tilt system in our scopes we can see that high level abberations will possibly account for 2/3rds the seeing and ground level for 1/3rd. When you look at the light paths down from the sky, the ground effects should be common between the guide camera and the imaging camera, but the high level effects will not. Thus it becomes useful to use multiple guide stars and average them oterwise you add their errors to the imaged object. In fact one really needs two guide cameras, and a double OAG.... to get that working some slightly cleverer software might need writing.

Anyway. if you don't have guide stars bright enough then it can't help you much.

I feel a discussion on lazers comming on

just mulling possibilities

Derek

[Darth Takahashi]

OK got it, I'm glad I never bought one then.

There was one going secondhand on TE-LES-KOOP not so long ago, AO8 for about 400 Euros.

[Darth Takahashi]

Olly

Seeing is not generally dominated by any one source. I found a wonderful paper from observations and tests in the paranal mountains where they established that their seeing came from three sources...

A: Jet stream located at about 7km up, in a 500m or so thick band.

B: Gound effects, below a couple of hundred meters.. the lower the level the more impact it had.

C: residual all hight abberations.

each source accounted for about 1/3rd the total seeing.

Derek

From all of us in the Netherlands... thank you very much Derek!!!:( Someone pass me the Valium:eek: I can feel a bout of depression come on...

All joking aside, I'm laughing inside when I here others complain about the seeing. I'm lucky if I can get 4" seeing maybe 10% of the time here in Holland and probably I'm being to optimistic. I had thought about measuring it but decided against in because it might make me suicidal.

[rfdesigner]

From all of us in the Netherlands... thank you very much Derek!!!:( Someone pass me the Valium:eek: I can feel a bout of depression come on...

All joking aside, I'm laughing inside when I here other complain about the seeing. I'm lucky if I can get 4" seeing maybe 10% of the time here in Holland and probably I'm being to optimistic. Id had thought about measuring it but decided against in because it might make me suicidal.

Perhaps your seeing is so bad becasue you're that bit lower, the more your bad seeing is due the gound level effects the better.. you can do something about those.

[Darth Takahashi]

Perhaps your seeing is so bad becasue you're that bit lower, the more your bad seeing is due the gound level effects the better.. you can do something about those.

Tell me more, what can I do about the ground effects??? There is actually a guy here in Holland who wants to make a mountain, I kid you not:D:D:D

[NickK]

I'd have thought that NL's main problem is the density of the air, being low altitude. Then add humidity with the sea..

You could use a laser - possibly powerful enough to get through the first 100-300m of air. Use a single camera with a filter at the same wavelength then look at how the beam is reflected. I have a sneaking thought that you'd need a very high focal length to detect any variations though given the pixel size of amateur sensors.

[Darth Takahashi]

I'd have thought that NL's main problem is the density of the air, being low altitude. Then add humidity with the sea..

You could use a laser - possibly powerful enough to get through the first 100-300m of air. Use a single camera with a filter at the same wavelength then look at how the beam is reflected. I have a sneaking thought that you'd need a very high focal length to detect any variations though given the pixel size of amateur sensors.

Sounds like I might have everything I need.

1. Green lazer pointer bought in Thailand 300mW power:D

2. Mewlon 3000mm native focal length which can be extended using either the ExtQ(1.5x) or a WO Barlow(2x).

Whats next? Simply measuring the deviation? Then what?

[ollypenrice]

Go up on stilts like Russel Croman or the Tivoli place in Namibia? They know about stilts in the Netherlands!!

Olly

[michael.h.f.wilkinson]

Go up on stilts like Russel Croman or the Tivoli place in Namibia? They know about stilts in the Netherlands!!

Olly

We use them. They are called piles, and most of our houses would sink if we did not have them

[rfdesigner]

Sounds like I might have everything I need.

1. Green lazer pointer bought in Thailand 300mW power:D

2. Mewlon 3000mm native focal length which can be extended using either the ExtQ(1.5x) or a WO Barlow(2x).

Whats next? Simply measuring the deviation? Then what?

Correct for it.. fast. SX-LFAO or similar.

By the way, the pro's use a sodium Laser (well some do) as this will activate high altitude sodium atoms.

If you use a fast camera and a pulse lazer then you can look at the appropriate height by timing the camera and having a powerful pulsed lazer.

Star Wars anyone?

Derek