Page 4 of 4

Re: Lhires III fine tuning

PostPosted: Wed Apr 06, 2016 3:48 am
by Paul Luckas
Further to Andy's most recent results, I've been investigating the residual "spectral image doubling" on my Lhires. I've found a large variation in the doubling effect depending on where my target star is located along the slit - with the effect worsening as I move the target along the length of the slit from one end to the other.

I'm trying to visualise what may be the cause of this, and finding it hard to wrap my head around the '3D' light path in the Lhires.

Images attached show the quality of a typical spectral image with the star positioned at one end of the slit (as viewed in the guider) compared with the star positioned on the centre of the slit (as viewed in the guider). Moving the star to the other end of the slit results in an even worse spectral image (not attached).

In both of the above cases, the slit and target are in sharp focus as viewed in the guider and the collimator has been focused well on calibration lines. The main mirror is adjusted as per the manual - i.e., seeking a balance between astigmatism and vignetting - and the grating is not pinched, and shows pretty good alignment across the length of the micrometer adjustment.

Of note, my Ne/Ar lines are very asymmetrical (thin at the top, and broad at the bottom). The 'thin' end of my calibration lines corresponds to the top 'good' spectral image - perhaps a clue as to what is going on.

Comments most welcome.


spect.PNG (312.74 KiB) Viewed 3265 times

spect2.PNG (319.54 KiB) Viewed 3265 times

Re: Lhires III fine tuning

PostPosted: Wed Apr 06, 2016 1:56 pm
by Benjamin Mauclaire
Hello Paul,

With all the elements given, it seems that your main mirror needs to be adjusted a bit different with star feed back.
It's not an easy task but when it's done Lhires stars producing spectra.

When seting mirror, put your mount in sleep mode, then recednter the bright star, take 2-10 s spectrum in bin 2*2 or 3*3 and repeat it.
The best star's position on slit isn't allways slit's center even for acq ccd spectrum image or autoguiding camera.

Curious to see your calibration lines.


Re: Lhires III fine tuning

PostPosted: Thu Apr 07, 2016 11:39 am
by Paul Luckas
neonHD54879-1.jpg (12.16 KiB) Viewed 3229 times

Re: Lhires III fine tuning

PostPosted: Thu Apr 07, 2016 3:27 pm
by Robin Leadbeater
Hi Paul,

Just for comparison. The sharpness of the lamp lines varies only slightly along the slit in my LHIRES, being marginally less sharp at the top and bottom of the field. see attached (35um slit. 2.6 pixels = 33.5um FWHM)

If in your case the focus, both in the horizontal and vertical, changes along the slit (presumably you can adjust the doublet to bring any position along the slit into focus?) I would be looking for something out of square in the slit direction relative to the optical axis. (camera, grating, slit ?)


Re: Lhires III fine tuning

PostPosted: Fri Apr 08, 2016 1:42 am
by Paul Luckas
Thanks Robin,

My neon lines used to resemble that (thin at the top and bottom) but with a little more smile, before I upgraded to the new calibration lamp assembly. I wonder if things have moved out of alignment as part of the disassembly and upgrade process. I'm also talking to Andy off line, and will revisit grating tension in its holder too

It's interesting to see so much variation in these units (having also had long conversations with Bernard off line) - though Andy and I seem to have almost identical spectral images.


Re: Lhires III fine tuning

PostPosted: Fri Apr 08, 2016 11:11 am
by Robin Leadbeater
Hi Paul,

These are with a 1200 grating which probably explains the smaller smile

My lamp lines do not actually change in width along the slit, it just looks that way because of the way the intensity of the lamp varies. The checks of FWHM on the right which were measured top, centre and bottom of the field and confirm they are actually ~the same thickness (within ~+-10%). Are your lines actually wider at the bottom?


Re: Lhires III fine tuning

PostPosted: Sat Apr 09, 2016 2:25 am
by Paul Luckas
Hi Robin,

My lines, particularly at longer wavelengths - do not show consistent FWHM across their length (probably up to +/- 20% difference) - one of the reasons I'm revisiting all of this. Annoyingly, the 'thin' end has the lowest FWHM but the poorest throughput. The 'fat' end is the area of best throughput but larger FWHM. It's possible to focus on the 'fat end' but I don't quite reach the best FWHM, and it seems much harder to home in on best collimator focus at the 'fat end' (at least visually).

I spent quite some time yesterday experimenting with grating positioning in the holder, clamp tightness, slit positioning (albeit there's not much room for adjustment with the slit) and calibration lamp servo position tweaking. All of this has me no closer to understanding the relationship between all of the components, but I have noted some interesting things during the process. For example, small movements of the grating in its holder change the 'vertical' position of the spectral image quite markedly - so much so that I've been able to adjust my mirror back to a more satisfactory angle. The Ar/Ne lines are now considerably more symmetrical and also quite vertical too, at least at shorter wavelengths.

My experiments with tension on the grating clamps yielded little - at least with calibration lines. Any effect of tightening and loosening the clamps on the quality of calibration lines was surprisingly absent.

I had thought that re-programming the position of the relco lamp over the slit (possible using the new calibration unit's electronics) might see some shift in slit illumination to a more centralised maximum, with thinner top and bottom, but this was not the case. For the record, the only affect of changing the relco's position over the slit was throughput and line length. It's probably best left at factory position IMO. Due to the onset of stormy weather (after months of clears skies I might add) I have been unable to test the effect of these adjustments on a stellar spectral image to see if any of the 'doubling' effect has dissipated, but I'm keen to see if there's any change.



Re: Lhires III fine tuning

PostPosted: Sat Apr 09, 2016 1:35 pm
by Robin Leadbeater
Hi Paul,

My example above was pre upgrade. Attached is an example post upgrade but with a 600 grating (and this time unbinned) so not exactly like for like. Focus is good top and centre, with the line profiles clearly showing the shape of the 35um slit with flat top and steep sides but deteriorates a bit towards the bottom of the field (green line profile)


Re: Lhires III fine tuning

PostPosted: Wed Apr 13, 2016 7:32 pm
by AndyWilson
I thought I'd add a further post to this topic as I've done a bit more fine tuning following offline discussions with Paul Luckas and François Cochard.

First, just to emphasize that in my earlier posts my main issue was due to the grating holder pinching the grating. This was easily fixed by reducing the tension in the screws of the grating holder, while making sure the grating was still held firmly.

Since then I noticed that the maximum throughput and so signal did not coincide with the thinnest spectrum. When setting up the spectrograph I use the graph feature in Maxim DL to create a plot of the average pixel values in horizontal rows. This is a bit of an art as well as a science, but what I try to do is maximise the area under the graph, rather than simply making the spectrum width as narrow as possible. Note this isn't a graph of the spectrum looking at wavelength, rather a graph looking across the spectrum height.

I did some experimentation last night by adjusting both the Lhires III mirror position (using the screw at the base of the Lhires III) and telescope focus to try to maximise the throughput, while trying to minimise the spectrum width. I used the bright star Regulus with 60 second integration times to smooth out the atmospheric seeing. Below are the images and graph plots of average pixel values across the height of the spectrum using Maxim DL. Though it is a bit subjective I believe adjusting the mirror position has increased the throughput by up to 30%. A significant success! On both occasions the Lhires III mirror was fairly central, and I have just moved it from one side of the exact centre position to the other.

Spectrum image before mirror adjustment.
Before image
Regulus_Before_20160412_small.jpg (44.89 KiB) Viewed 3087 times

Spectrum image after mirror adjustment.
After iamge
Regulus_After_20160412_small.jpg (45.13 KiB) Viewed 3087 times

Spectrum average pixel count before mirror adjustment.
Before graph
Regulus_Before_20160412_graph.jpg (51.48 KiB) Viewed 3087 times

Spectrum average pixel count after mirror adjustment.
After graph
Regulus_After_20160412_graph.jpg (52.51 KiB) Viewed 3087 times

I may be able to improve things even further, though I am happy with my results and want to spend time acquiring data :-)

Best wishes,

Re: Lhires III fine tuning

PostPosted: Thu Apr 21, 2016 2:03 am
by Paul Luckas
Thanks Andy,

As per François' request - here's my update for the record.

Alignment of the grating to the grating holder, while providing for minimal 'drift' up and down the CCD during wavelength changes, does not guarantee grating alignment with the slit. I've found it necessary to adjust my slit holder significantly (see attached photo) in order to 'straighten' my calibration lines (i.e., better align slit to grating 'grooves'). In conversations and comparisons with Andy, clearly my particular Lhires is not as 'square' as one might like.

The result has been dramatic - with an increase in R (as measured by ISIS) of about 10-15% at 5411Å. Moreover, it's now much easier (quicker) to ascertain best focus of the collimator during a session at the spectroscope.

Caveat - this adjustment is by no means absolutely necessary (my spectral images and profiles were fine in most respects). Only for those with cloudy skies and time on their hands.

I've also been adjusting calibration lamp positioning over the slit noting only very minor improvements.

On the subject of spectral image "doubling", while still present, it is not as bad as it was. It certainly does not appear to adversely affect the resulting throughput or profile. Bottom line, it's probably best to start with a focused star on the slit and work back from there. Of note, my mirror is adjusted quite far over towards the slit carrier side (i.e., minimum Littrow angle / maximum vignetting). I'm not sure why, but I seem to need to do this in order to get my spectral image up into the CCD frame.

Finally, I still have a nagging feeling that collimation of my RC optics may also be at play here - with both Andy and I using RC's and (possibly) the only reporters of the spectral doubling feature).


FullSizeRender.jpeg (54.75 KiB) Viewed 3013 times

slitangleafter.jpeg (45.25 KiB) Viewed 3013 times

FWHM1.PNG (260.19 KiB) Viewed 3013 times

spectra1.PNG (252.5 KiB) Viewed 3013 times