How to get your Fast f/2.8 Newtonian to work like it should (TS Hypergraph 6 and Sharpstar 15028 HNT)

[GalaxyGael]

How to get your Fast f/2.8 Newtonian to work like it should (TS Hypergraph 6 and Sharpstar 15028 HNT)

 

Abstract

Fast Newtonians with hyperboloidal mirrors offer fast optics with more stringent requirements on focus, tilt, collimation and other settings. What goes hand-in-hand with requirements from the user, is accuracy and manufacturing precision expected of the maker, and this is lacking in many cases. I thought it would be useful to summarise some efforts to get my TS Hypergraph 6 (aka Sharpstar 15028 HNT) fine-tuned, and hope that this might be useful to anyone who has, or is thinking of getting, one of these fast hyperboloidal mirrored newtonians. Note: the title of this post assumes that collimation is good and there are no underlying issues with camera or optical train tilt, corrector spacing and related effects.

1. Some issues with the 1^st generation 15028 HNT

The 2019/2020 models of the Sharpstar 15028 HNT (hyperboloidal primary mirror f/2.8 fast newt) were known to have issues that were well documented here and elsewhere. Loose mirrors, moving during slewing causing miscollimation, issues with secondary, lose screws and other problems. Many of these were solved, but with intervention from owners that went beyond (IMO) expectations from a 2k+ EUR scope.

Apparently, Sharpstar mention that the 13028HNT has a redesigned or improved primary cell, which I assume means they found a middle ground. Their engineering approach to fixing the looseness in the earlier models was crude and as you will see, overcompensated, causing just as much of a problem but of a different kind. I did not have any loose elements, loose screws or any issues detailed by @Davey-T and others on SGL and elsewhere recently. Any user can search and refer to those useful threads, as I did, to deal with issues depending on the version of scope you have. If you have the TS, Sharpstar, Omegon models from late 2020, or 2021, these are likely the later generation and you can tell by taking out the primary and trying to (re)move it. If it is solidly stuck in place, you have the primary cell ‘designed’ to fix looseness. You will likely need to do a few simple mods to get it singing.

2. Sherlock Holmes-ing the cause: Overcompensation to the primary mirror cell

Sharpstar/TS etc. overcompensated the primary mirror cell looseness issue to such an extent that I had to undo much of the modifications. It was very tightly inserted in the cell, with no play in any direction - too tight, basically.

The first inkling was after a conversation with Kostantin Firsov (thank you Kostya), who also uses the Sharpstar 15028HNT to great effect. He found 3 axis astigmatism on his primary mirror and linked it to pressure from the sidewalls and from the brass pins under the mirror. You can see these in Fig. 1(a,b). Second, there is vinyl tape over the felt edging of the three vertical sidewalls that contain the primary mirror. If you have the original batch, the mirror is quite loose. If you have one like mine, the mirror is very tight. Remove these tapes and leave the felt as-is. The mirror slips gently back into position, about 0.5 cm/s meaning it does contact the felt ever so slightly. It had to be pushed back in using the as-received setup.

 

Fig. 1. (a) The base of the primary mirror cell containing 3 friction fit brass mushroom-shaped pins, one of which was lower in height compared to the other two. (b) The three vertical sidewalls that house the mirror are lined with felt, and in this example, a crude vinyl tape addition to increase thickness.

 

I removed these pins as Konstantin did, because one of them was lower in height that the other, and it tilted the mirror. Worse, any tightness on the retaining ring would cause asymmetric down-pressure where the mirror touched the brass points. They are small and cause unnecessary pressure on a sensitive optical element as-received, since the retaining rings with new felt pads to keep the mirror from moving, was bolted down fully. I removed the brass pins and replaced with silicone glue to give perfect flatness and adhesion but in areas that can easily be accessed with an xacto/stanley knife if it needed removing. This mod in final form is shown in Fig. 2(a,b).

Fig. 2. (a) Application of No-nails silicone glue to three regions, causing 15-20 mm circular adhesions areas when the mirror is placed. Their location allows easy removal when needed. The mirror is now exactly flat with respect to the mirror cell and nothing presses on the mirror side, bottom or top.

 

The retaining ring is reapplied but gently so, with felt only toughing the edge and no pressure, as it should be. Now, all pressure is removed from the sides, from the top and from the bottom, eliminating any pinching or astigmatism in the hyperboloidal primary. (Side note: interesting to see that the centre doughnut is just a shadow masked region from metallization and dielectric coating, such that you can see right through it when held up to the light – no sticker adhesive on this mirror, see Fig. 3).

Fig. 3. Image of the primary mirror showing the shadow masked region during metallization and dielectric coating that defined the central doughnut. The doughnut has no metal coating. The shiny edge of the mirror in this image was always accessible to incoming photons and not covered by the retaining ring.

 

The rest of the scope is excellent, no issues. I blackened many elements, but they do not need to be changed. Focuser is solid, stable, shift-free (even during inward and outward focus), and it is remarkably thermally stable because of the thick carbon tube. No issues with the secondary, although I might update the collimation screws and the main spring behind the mirror down the line, but only if it causes an issue with age. No astigmatism in the secondary found.

 

3. Redirecting energy back to the star – flaring, clipping and a turned down edge on the primary

The issue I noted all the time was a shiny edge to the mirror when looking down the tube. I had linked this originally to a reflection of the inner edge of the retaining ring, and that was true. But what is not obvious is that these scopes have retaining rings that are not wide enough to cover what is clearly a turned down edge that redirects photon energy to the point where star cores are ill-defined (read: horrible), with asymmetric flaring that follows the stars location in the frame, segment clipping of flaring etc. Refer to my astrobin recent images (https://www.astrobin.com/users/codwyer/), with fair warning that if you like nice stars as I do (in the main), brace yourself. I was determined to remove this effect and improve the star profile better than sharpstar did with the original configuration.

I spent weeks taking images that were amazing (IMO) in terms of signal (jaw-dropping is at the edge of overstating, f/2.8 is a different world), but too horrible to add additional imaging time, or even bother processing properly. You can see a few on my astrobin where they serve as placeholders for re-imaging soon. What was important was Kostya Firsov’s statement that he saw the exact same issue. That was important and confirmed that my collimation was near perfect, as was backfocus etc., which I had assumed all along was my fault and a limitation of my capability to tame f/2.8. It turns out that I can collimation this thing from completely miscollimation very effectively, which was a big (but long winded) plus from this process. 

Kostya Firsov used medical x-ray film as an edge mask, which reminded me of what I used to do for astrometric purposes on RCs, sacrificing some aperture for contrast and preventing unwanted flaring. Planetary imagers do this too for contrast/sharpness. I did not expect it to be necessary for a circular retaining ring (i.e. not 3 mirror slips where masking is known to be effective). The edge of the primary is defacto the worst part of the mirror, so I 3D printed this mask (Fig. 4(a)), attached it and looked down the barrel. Pitch black, no reflective edge around the mirror (Fig. 4(b)).

Fig. 4. (a) Image of a 3D printed mask obtained from the Thingyverse, originally designed for the SW 150PDS which has a similarly sized mirror. It has an inner diameter of 144 mm, perfectly covering the 3 mm of mirror edge, which I assume accommodate the turned down edge region. The screws holes or the mirror slips, which are ignored in my case. (b) Image looking down the tube with mirror mask in place. That mirror will need to be caked in pollens/organics before I take it out again!

 

I had about 80 minutes of clear sky last night, got collimated, setup and took a single 4 minute exposure of pacman nebula (NGC 281), high in the sky near zenith, lots of stars of various sizes and a good test. The background-extracted single sub in shown in Fig. 5 with some crops of two regions.

Fig. 5. A single 240s exposure of NGC 281 using Hypergraph 6 and ASI2600 MC Proc, -10C, Gain 100, no filters. Background corrected and HLVG only. (Bottom) Selected regions showing star profiles, completely fixed from astigmatism and turned down edge effects.

 

Well, I now have a scope performing almost ideally. When I captured previous images, two things usually happened. I saw the clipping the subs, packed up and went inside completely demotivated. If they came through after integration, processing was halfhearted as stars just got worse, were poor candidates for effective masking, etc. The nebulosity, galaxies etc. were fantastic from Bortle 6 without filters at f/2.8, and that was enjoyable to see in the images, but the stars ruined every target for me since getting this scope.

 

4. Conclusions and outlook

For owners of this scope, it can perform really very well and it seems that the current generation versions require untightening of the mirror and masking of the turned down edge. That will give nice round stars, with a reduction even of the squareness of the PSF usually seen in these fast newts. With these modifications, you will have a ‘photon hoover’, that is collimation and temperature stable. The tightness of the primary indicated which ‘generation’ you have, but mirror masking is the singular action that improves acquisition data considerably.

Addendum: primary masking may also be useful in general for fast scopes with primary mirrors (or indeed any Newtonian, especially those with mirror clips). It can improve contrast and reduce or eliminate any weak diffraction flaring and is a very effective method to do this.

I will update this thread with the first full integration using this mildly modified scope, probably M45 since I imaged that for the first time ever recently (yes, I have never imaged anything from M45 to all targets around Orion ever!!), as in my astrobin you can see how bad the stars look.

Acknowledgements

In finding out what was not wrong with my scope, I learned a lot thanks to @Davey-T and others. Clearly, a newer generation of this scope was wrong in a different way, but a simple way. Thank you also to Konstantin Firsov who confirmed asymmetric flaring/clipping was an issue and for confirming astigmatism due to over-tightened elements. His input was vital to make decisions on what to do to remedy the situation.

 

Edited November 15, 2021 by GalaxyGael
fix typos

[vlaiv]

My only concern is different expansion coefficients of mirror cell and mirror itself.

There is a reason why mirrors are not glued to their cells - it can produce same effects as pressure on the sides.

What are your thoughts on silicone glue you used?

[GalaxyGael]

6 hours ago, vlaiv said:

My only concern is different expansion coefficients of mirror cell and mirror itself.

There is a reason why mirrors are not glued to their cells - it can produce same effects as pressure on the sides.

What are your thoughts on silicone glue you used?

You're right, but I had little choice since the pressure on the sides, uneven pressure from the top from the differing height brass pins underneath made a very bad situation. 

The best option I had available was an adhesive (not so much a glue) and we will see how it fares over the winter. Sharpstar mention a 'Pyrex-alike' material, that is a 'home-brew' borosilicate formulation or Duran glass probably. Its fractional expansion is in the range 3-4 x 10^-6 / K. Its not fused quartz, but it is reasonable. I used small spots of adhesive in three places near the edge. The sidewalls and retaining ring have sufficient room for expansion, and from cold contraction I assume that the amorphous substrate does so linearly all in directions to maintain optical quality (that is my assumption). I cannot tell yet if adhesive rigidity on the underneath surface will stress the rest of the mirror itself, but I am guessing now that it will be less worse than the as-received case. If there's a change long term or on very cold night, I will report it.

I had though of using an adhesive that has some 'wobble' like pure silicone but was concerned with tilt/sag in the mirror. Time will tell, and so will the images.

[GalaxyGael]

So we have had a run of clear night, although the moon is between 98% waning to 88% last night. 

Temperatures were down to -3, coldest yet this year, and the primary mirror mask is working quite well. 

A crop from a raw single sub of 120 s of M45 with 88% moon.

 

Edited November 23, 2021 by GalaxyGael

[ImNewHere]

Thanks for this guide here. I did all this, but there is a good 5mm between my mirror and the retaining ring now. Also the retaining ring on mine sticks over the edges of the turn down by about 5mm all around, but then again I do have the first one that was released in the US. On collimating mine I found the mirror flopping thing, which is why this was valuable for me. I used automotive gasket maker silicone for water pump housings due to it being non corrosive and having good expansion and stability properties with heating and cooling.

My big question is, should I be concerned about that 5mm space between the retaining ring and the mirror? If so what would you recommend for me to make the space up?

[GalaxyGael]

Are you referring to a vertical 5mmm gap away from the retaining ring? If so, that should be OK. Mine is separated vertically by roughly that amount as i used a double side padding normally used for putting towel hangers onto tiles in bathrooms - very sturdy, doesnt contract or expand and resistant to moisture/vapor. 

[ImNewHere]

Yep, that was what I am referring to. That makes me feel better. So did you put that in on the bottom of the ring where it is open and have it just touching the edges of the mirror?

[GalaxyGael]

No. Once the retaining ring was screwed back and the distance optimized with the grub screws to avoid downward pressure on the mirror, I added the printed ring aperture mask on top of the retaining ring, i.e. added on top of the fully assembled mirror cell. It masked the edge of everything essentially.

[ImNewHere]

Okay, sounds good to me. The retaining ring on mine is about 5mm from touching the mirror so there is no downward pressure, and it has a 5.75 inch ID so should catch the mirror if by some odd reason it were to come loose, though the silicone seems to be holding it well.

[CCD-Freak]

I have been working on my SS15028HNT to resolve the mirror cell issues and it does a great job with my ASI-533MCP. 

I just purchased an ASI-2600MCP with it's larger sensor so I may need some more tweaking. 

I have been having much fun with this little astrograph.  (^8

[Cheman]

Hi, and thanks to the OP for this thread.

I have a first version 15028HNT Sharpstar, with the very loose primary.  My original solution was I inserted shims of ptfe that I had that was the correct thickness to take up the slop between the mirror sides and the cell.  This seemed to work fairly well.  However, I never liked the mirror cell in the first place so I decided to go with a different, although more expensive and drastic solution.  I replaced the cell with one from https://www.aurorap.com/  It took a bit of modification to the scope end ring to get it to fit but the end result is a much more robust cell.  Pics attached. After reading your modifications, I will make an apature mask.  Then I think I will have a very nice end result

Che