Wavefront Estimator (finally!)

[pfile]

i recently fitted my TMB92SS with an orion flattener based on the experience of another user. the field is pretty flat - much flatter than i was ever able to achieve with the TRF-2008 0.8x reducer.

i thought it would be a good time to try Mike Schuster's WavefrontEstimator script. wow - this script is amazing. really nice work.

the analysis might be garbage in/garbage out, because as you can see from the defocused star image, there are some diffraction patterns that don't belong there. the dimmer one closer to the center turned out to be some crap suspended from a spiderweb crossing the aperture, but even though everything is all cleaned up, the stronger of the two diffraction pattern remains. the only thing i can think of that could be impinging on the light cone is the OAG in my filter wheel, but somehow i didn't expect that to manifest itself as such a "point" diffaction source. looking down the OTA i don't see anything sticking into the path of light. anyone have any ideas what this might be?

i wish i had realized the script handled scopes with a central obstruction - i would have tested my AT6RC+TRF-2008 before i replaced it with the TMB92.

Code: [Select]


[2016-06-27 21:57:55] Point spread function estimation:
[2016-06-27 21:58:00] Strehl ratio: 0.815
[2016-06-27 21:58:00] Strehl diameter: 4.07 ?m, 1.66 arcsec

...

[2016-06-27 21:57:50] Aberration estimation:
[2016-06-27 21:57:50] Z5 (2, -2) Primary astigmatism oblique: 13.2 nm RMS
[2016-06-27 21:57:50] Z6 (2, 2) Primary astigmatism vertical: -8.8 nm RMS
[2016-06-27 21:57:50] Z7 (3, -1) Primary coma vertical: 2.4 nm RMS
[2016-06-27 21:57:50] Z8 (3, 1) Primary coma horizontal: -3.4 nm RMS
[2016-06-27 21:57:50] Z9 (3, -3) Primary trefoil vertical: -1.2 nm RMS
[2016-06-27 21:57:50] Z10 (3, 3) Primary trefoil oblique: -1.1 nm RMS
[2016-06-27 21:57:50] Z11 (4, 0) Primary spherical: -30.9 nm RMS
[2016-06-27 21:57:50] Z12 (4, 2) Secondary astigmatism vertical: 0.4 nm RMS
[2016-06-27 21:57:50] Z13 (4, -2) Secondary astigmatism oblique: -3.4 nm RMS
[2016-06-27 21:57:50] Z14 (4, 4) Primary tetrafoil vertical: -3.6 nm RMS
[2016-06-27 21:57:50] Z15 (4, -4) Primary tetrafoil oblique: -1.7 nm RMS
[2016-06-27 21:57:50] Z16 (5, 1) Secondary coma horizontal: -0.6 nm RMS
[2016-06-27 21:57:50] Z17 (5, -1) Secondary coma vertical: 0.1 nm RMS
[2016-06-27 21:57:50] Z18 (5, 3) Secondary trefoil oblique: 0.2 nm RMS
[2016-06-27 21:57:50] Z19 (5, -3) Secondary trefoil vertical: 0.9 nm RMS
[2016-06-27 21:57:50] Z20 (5, 5) Primary pentafoil oblique: -0.6 nm RMS
[2016-06-27 21:57:50] Z21 (5, -5) Primary pentafoil vertical: -1.2 nm RMS
[2016-06-27 21:57:50] Z22 (6, 0) Secondary spherical: -13.2 nm RMS


FSQ it is not, but given the modest price i suppose it's not horrible.



rob

[RickS]

Looks very interesting, Rob.  I should try running it on my AP140.  Have you tried rotating the OAG to see if the diffraction artifact moves?

[mschuster]

Hi Rob, thank you for posting.

Not sure the diffraction source, on my FSQ/QSI the OAG pickup is far enough off-axis to not be visible.

The estimated on-axis Strehl ratio of ~0.81 is just better than the conventional so-called diffraction-limited level of 0.8, the maximum acceptable level of wavefront aberration for general observing.

The primary aberrations present are spherical and astigmatism, with spherical dominating and easy to see in the defocused images as radially symmetric exposure differences in the Fresnel rings. On-axis spherical aberration with a flattener would not be a surprise IMO.

Best,
Mike

[pfile]

Looks very interesting, Rob.  I should try running it on my AP140.  Have you tried rotating the OAG to see if the diffraction artifact moves?

the OAG itself can not be rotated but i have that optec rotator from SRO1 in the imaging train, so i can change the angle of the OAG with respect to the OTA. i will try - last night i re-did the tests after cleaning the junk off, but right when i finished i realized that i had accidentally used the L filter instead of G. so i need to re-do the test anyway. it would also be interesting to see what things look like thru R and B.

the flats from this setup seem very slightly asymmetrical; i suppose this could be caused by the pickoff prism - perhaps at f/5 the shadow would not be as crisp as it is on my f/8 reflectors.

Hi Rob, thank you for posting.

Not sure the diffraction source, on my FSQ/QSI the OAG pickup is far enough off-axis to not be visible.

The estimated on-axis Strehl ratio of ~0.81 is just better than the conventional so-called diffraction-limited level of 0.8, the maximum acceptable level of wavefront aberration for general observing.

The primary aberrations present are spherical and astigmatism, with spherical dominating and easy to see in the defocused images as radially symmetric exposures differences in the Fresnel rings. On-axis spherical aberration with a flattener would not be a surprise IMO.

Best,
Mike

do you think that changing the flattener spacing could improve things? based on eyeballing star shapes, the spacing seems to be very close to correct, however, the results from FWHMEccentricity definitely show some weirdness. taken together with WE it may mean that the spacing is not correct. on the other hand the flattener is an off-the-shelf part, so it could simply be mismatched.

it's possible that the seeing here obviates the need for higher-quality glass, but i don't have an apples-to-apples comparison with something like an FSQ.

rob

[mschuster]

do you think that changing the flattener spacing could improve things?

Maybe. Guessing: To flatten a field with relatively simple optics may involve a tradeoff and an increase in on-axis aberrations. Also, according to an S&T review, an earlier version of the telescope (with no flattener) exhibited some spherical aberration.

Thanks,
Mike

[pfile]

maybe i will not let the perfect be the enemy of the good. i know that there's enough sag in the train such that when the OTA is more horizontal i start seeing corner aberrations, so there may not be much point in polishing this turd any further...

rob

[pfile]

maybe i was on a double star or something... i'm not seeing either artifact now. not sure if i'm looking at a bug in SGP or not, but even the frames that should have been taken thru the G filter show "L" in the fits header. so this may have been GIGO... going to try again.

rob

[pfile]

interestingly, the results on the "clean" image are very much the same (strehl=0.796), with very similar primary/secondary spherical aberration, but the oblique astigmatism is much lower at 0.6nm RMS. so the deformed image must have fooled the script a little bit. also i made sure i was using the G filter this time, and the FITS headers do show G now, so the analysis of the L images yielded reasonable results when interpreted as G.

rob

[mschuster]

the oblique astigmatism is much lower at 0.6nm RMS. so the deformed image must have fooled the script a little bit.

On the original deformed images, I noticed that the stronger of the two diffractions is located in different places. I would expect this position difference to affect estimates of astigmatism more than spherical due to asymmetry.

so the analysis of the L images yielded reasonable results when interpreted as G.

Partly due to the fact that the script uses central wavelengths for L- and G-bands that are nearly the same.

On my FSQ aberrations do seem to vary a bit with band, with RGB Strehls estimated as 0.90, 0.93, and 0.94.

Best,
Mike