Fringing ripple

[Bernard Heathcote]

Unfortunately fringing ripple isn't a new problem and has been discussed a few times on the forums during the last year, and I need to raise the topic again.

I've attached an image showing the ripple I have (not a new problem for me) with my SBIG ST-8ME (eShel) and ST-8XE (Lhires III). When the ST-8XE on the Lhires is replaced by an Atik 314L I do not get any noticable ripple. The same star and line was used for the three tests. I find that the eShel ripple is so bad that I have stopped using it until I can, hopefully, find some solution. Luckily I am able to use the Atik with the Lhires. I've read the information given my Benji, Schlatter, Lailly and Valerie, which is very useful but inconclusive. While flat fielding may reduce the effect it certainly does not remove it in my case.

I am most concerned about trying to get my eShel fully functioning and am wondering if there is any possible optical (mis)adjustment that could be the cause ... or is the KAF-1603ME 100% responsible? I ask this as I read that adjusting the main mirror of the Lhires can have some effect ... I have not tried this yet and would appreciated any comments first.

One possible, but expensive, 'cure' for the eShel is to replace the ST-8ME with the new SVXR-H694, or possible future Atik, with the large ICX694 CCD (thanks Thierry G for the info).

All comments and advice very welcome.

Cheers,
Bernard

[Robin Leadbeater]

Hi Bernard,

Interestingly I saw a significant reduction in ripples when I recently upgraded from an ATIK 16IC-S to a 314L+. However I had severe ripples and ripple free periods with the 16IC-S so I am not convinced this will be a long term solution.

Others, for example Jeff Hopkins have also seen significant variation with the same camera so there must be something subtle with the optical alignment/focus which affects this

Cheers
Robin

[Robin Leadbeater]

there must be something subtle with the optical alignment/focus which affects this

For a bizzare example of the variations in this affect a look at this post
viewtopic.php?f=8&t=176

I even saw this affect either side of a complete rebuild of the spectrograph and recollimation of the telescope following storm damage last February and also over the recent change from 16IC-S to 314L, though the overal intensity significantly reduced following the camera change

Robin

[Andrew Smith]

Bernard, Robin, I suspect these are interference fringes which are formed via multiple reflections within the spectrograph and or CCD camera. Robin’s example of variation with binning height tends to confirm this as this will change the optical path length.

It may be that deliberately de-collimating a component slightly could reduce or remove the problem without introducing too much additional aberrations.

It might be worth a try before getting a new camera that may not solve the problem.

Andrew

[Bernard Heathcote]

Robin,

Very interesting, and puzzling, your example of the ripple changes with binning position ... I have not noticed this. It does give hope that maybe the effect can be eliminated by some judicious optical alignment. I assume you've not yet had any feedback from Christian? I don't believe it is a processing problem as I see the ripple in the raw spectra. Clearly this ripple problem needs some high priority investigation ... but where to start? I need to go back through my records to see if I can find any variability in the ripple over the years and if there is, does it relate to any instrumental changes I may have made.

I agree with Andrew that another camera may not be the solution.

Cheers,
Bernard

[Robin Leadbeater]

Hi Andrew,

The source of this particular fringing in high resolution spectra has been traced to interference in the CCD cover glass. The spacing depends on the thickness of the glass. Apparently pros even specify cover glasses with a slight wedge to get around this potential problem. See here for a summary of knowledge so far
http://www.astrosurf.com/aras/fringing/ ... ripple.htm
The specific conditions that affects the severity though is not well understood. Small changes in collimation might well occur when changing cameras as the spacing in each case may not be exactly correct and refocussing the doublet will not bring things back exactly the same. I have heard comments that the better the spectrograph is collimated and focussed the worse the fringing, though it is not obvious to me why collimation affects it as the beam is converging through the cover glass. I found some time ago that the intensity can sometimes be affected by moving the adjustable mirror in the LHIRES


Bernard,

Christian confirmed that in this case it was not in the processing (though processing can produce fringes if for example the spectrum is tilted) The strange behaviour in this case though has us stumped. I have a feeling if we could understand what is going on in this case it might give a clue to solving it.
Re flats - I have not found conventional flats very successful for removing fringes either (The intensity and phase of fringes in normal flat just does not match - presumably because of the different light path and also due to shifts between taking the star and the flat) The most effective techinque I have found is to use a "star flat" ie divide by a bright hot star spectrum which has been divided by the library spectum for that star. This is similar to doing an instrument response correction but the response is not smoothed or only smoothed a small amount to reduce the noise. I have also sometimes specifically extracted the fringe pattern from a star flat taken on a different night and applied it to the star spectrum, shifting it as required to get the phase right.

Cheers
Robin

[Robin Leadbeater]

See here for a summary of knowledge so far
http://www.astrosurf.com/aras/fringing/ ... ripple.htm

Sorry that is a more specific link to Peter Schlatters study on the source of the fringes. The summary by Benji Mauclaire is here
http://www.astrosurf.com/aras/fringing/opus/index.html

cheers
Robin

[Andrew Smith]

Hi Andrew,

The source of this particular fringing in high resolution spectra has been traced to interference in the CCD cover glass. The spacing depends on the thickness of the glass.
Cheers
Robin

Hi Robin - That would have been my guess as to the cause but you never know. My MaxCam with E2V chip does not have a cover glass so I should not have a problem, but at the low resolution I am currently using it should not be an issue. Are the more commonly used CCD cameras in this field available with the cover glass removed from the CCD? I don't see why they are needed in a sealed unit?

I have come across one other cause which is due to the inter-pixel dead zones under exceptional seeing resulting in under sampling. Now that will never be an issue for me in Cheshire!

Andrew

[Bernard Heathcote]

Hi Andrew,

Kodak used to optionally sell many of their CCD chips without a glass cover, but the KAF-160X family wasn't one of them. I don't know what the new owner's plans are in this area, but if they don't improve their CCD technology they must surely lose the cheaper scientific market share to Sony.

Bernard

[Thierry Garrel]

Hello,
i have the feeling that interference problem and fringing ripple are of different nature. As described by professionals, fringes problem is due to non uniform sky background at small scale due to upper atmosphere continnum emmission. See Astronomy and CCD for a complete description. So even with a sky back gound sous traction, the correction is not perfect and has a strong dimensional sensitivity.
Interferences, show in Benji page is well known and are of different nature depending of optics setup.
So i think that what we have is a cross over the two phenomenon with mix results depending of upper atmospheric continuum glow and own optical setup.
So It could be a good idea to be in space and avoid any optics surfaces