Skywave (SKW) Pro Beta Test for Collimation and Wavefront Analysis

Here is after using SKW for secondary adjustments. This was all near real-time under my seeing limited conditions.


t

Here are a few other screenshots using the Pro version. These are the Wavefront Analysis and MTF of my scope under seeing limited conditions. It would be nice to use this as a bench test with an artificial star....but way too much work for me.

Interesting. Where do we find that piece of software?

Ah, ok! Interesting. Maybe I'm wrong, but it seems similar to the Roddier test.

SkyWave-Collimator version is now available, it can be downloaded from here:

https://www.innovationsforesight.com/support/skg_download/

The SKW-Collimator license is free, the user buys a mathematical model for his telescope.
SKW-Collimator is also standard in the new SkyGuide and SkyGuard (SKG) versions.

Here is a link to some documentation and tutorial:

https://www.innovationsforesight.com/support/skg_skw_documentation_tutorial/

The SKW Pro version is scheduled to be released at the AIC 2022,  May 20th and 21st 2022 at San-Jose CA.

SKW is using our patent pending AI based wavefront sensing (AIWFS) technology. For more technical details here are two peer reviewed SPIE proceeding references related to the two conference lectures I gave on this topic, the last one was a joint work with Dr. John B. Hayes adjunct research professor at Wyant college of optical science, university of Arizona.    

[1]  Gaston Baudat, “Low-cost wavefront sensing using artificial intelligence (AI) with synthetic data”, Proc. SPIE 11354, Photonics Europe, Strasbourg, 4-1-2020
[2] Gaston Baudat, John B. Hayes, “A star-test wavefront sensor using neural network analysis”, Proc. SPIE 11490, Optical Engineering + Applications, San-Diego, 8-21-2020
[3] Paul Hickson & Greg Burley, “Single-image wavefront curvature sensing”, SPIE Adaptive Optics Astronomy, vol. 2201, 549-553 (1994)ohn

John is an active member of astrobin, here is a link to some of his experiences with SKW:

https://www.astrobin.com/txk760/H/?q=john%20hayes

Following Roddier and Roddier's work on curvature sensing (CS) used with success in the context of AO  (1993),  Hickson & Burley (1994) (see [3]) have shown, using the same irradiance transfer equation,  that a single defocused star image is enough to retrieve the wavefront phase without ambiguity with proper defocus and care.

SKW levearges this work and others, it is designed for telescope alignment in this context, but not only. It uses long exposures under the Kolomogorov's turbulence model for retrieving the wavefront, hence the telescope aberrations, either native or misalignment induced, the main task for the SKW-Collimator version.

The AIWFS technology it self can be used for optical metrology, in the lab. Experiments (see [2])  done by John comparing with a Twyman-Green interferometer (PhaseCam 6000 from 4D) have shown very good agreement in an optical bench. Sensitivity around 1/1000 of the wave, accuracy around 1/100 of a wave at the time.

The latest tests done on the lab comparing with a Shack-Hartmann sensor (SH) brings now the accuracy with our new mathematical models (Neural Networks) around few 1/1000 of a wave, below are some results:



Of course one does not need such performance for telescope alignment. One should remember that a Strehl ratio of 80% is about 0.075 wave rms. Therefore accuracy in the range of few 1/100 of a wave is enough, especially under seeing limited conditions at which the telescope operates.

Below a test done on the sky with an actual star between SKW and a SH sensor on the same telescope, at the same time, at Patrick space force base (Florida) in November 2021:



The methods are in very good agreement, despite a below average seeing (FWHM was around 2.3", Fried's parameter r0 = 4.5cm) at the time.
The mathematical model was trained with seeing limited synthetic (simulated) images (long exposures).
Obviously this telescope needs some alignment.
As a matter of fact the model reports/outputs the seeing, acting as a seeing monitor too.

One interesting aspect of the AIWFS technology is that one does not need any dedicated wavefront sensor hardware, a simple camera and a mean to defocus is enough.

Also when using an imaging camera one can, at once, access several stars which offers the unique opportunity to measure field dependent wavefronts and aberrations (on and off axis at once) with a single frame and sensor, no need of several wavefront sensors or any mechanical scanning.

In order to leverage this capability  we are actively working with the DAG team in Switzerland and in Turkey for the implementation of our AIWFS  in the context of active optics (for keeping the telescope optics aligned and the primary mirror M1 figure on specifications, M1 is a deformable mirror pretty much like the VLT).
This is a 4m meters telescope for the East Anatolia Observatory:

https://www.innovationnewsnetwork.com/dag-turkeys-21st-century-grade-international-astronomical-observatory/17050/

We are still accepting some beta testers for SKW Pro, anybody interested please feel free to contact me directly.
One of the goal is to implement the ESO technique applied to the VLT active optics using field dependent astigmatism from AIWFS for retrieving mirrors misalignment values at once.  We are in contact with Stéphane Guisard (ESA/ESO) who is helping us for providing the amateur astronomers with a similar technology and methodology.
For information here is the basic paper on the topic:

https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.507.8063&rep=rep1&type=pdf

Gaston Baudat:
In essence one traded coma for a nice symmetric star pattern. Without any optical quantitative measurement it will be hard to spot this, beside below average in focus performance and likely more than normal off-axis aberrations (here the star was on axis during the alignment process).

Hi Gaston,

After reading your description I got a bit „confused“.

I understand your point that a symmetric defocused star image may leave residual optical aberrations when in focus (I observed this myself in some optics).
However, I would claim that with more experience, an astronomer would know that the final judgment on alignment is done in focus. Of course, this requires a good seeing condition.
Nevertheless, one would see if the optics is misaligned, as in your example to the right, if one has a bit of experience in star testing a telescope.

One should say of course, that many instruction manuals which come with OTAs are stopping short of illustrating a well aligned optics. Usually it ends with having the donut symmetrical. This might explain why „at least 70%“ of scopes are looking like in your example .

CS,
Björn

Gaston Baudat:
Hi  Björn,

All valid points.

I concur that when doing a star test it makes sense to check the collimation result in-focus (on and off axis) as well, however it may be challenging to judge because seeing.

On the right example with a near perfectly symmetrical "donut" one may easily come to the conclusion that the scope is aligned and the job is done, especially under seeing conditions. Scintillation makes it hard to detect a subtle intensity gradient in the diffraction pattern due to coma.

I guess my point was that even under good conditions one should often expect that a well aligned telescope (reflector) should exhibit a asymmetrical defocused star.
Aiming for symmetric one, like suggested in most telescope user manual collimation procedure, may often lead to a sub-optimal scope alignment, a failed collimation, while having a sense of successes.
I have not doubt that with experience one can eventually overcomes this, but it is not straight forward for most.

The long term (mean) wavefront is a plane-wave, seeing averages out (seeing is a phase noise with a zero mean).
There wavefront senors, SH or SKW,  allow measurement with accuracy in the order of 1/100 of wave even under below average seeing conditions.

Having access to the wavefront, hence quantitative optical collimation, removes any guess from the equation, IMO.

Gaston,

I fully agree with your statement. For the classical procedure to work one needs experience and sky conditions which permit a fine-tuning of the optical alignment. That's a reason for me to pursue some different techniques to align or simply confirm alignment of my optics.

I guess, the ultimate access to the wavefront would be an interferometer in auto collimation but that's a bit too extreme for most. 

Björn