700 replies to this topic

[Oberon]

#2
 

Full size image here.

 

[Oberon]

#3

 

Full size image here

 

Edited by Oberon, 05 July 2019 - 10:34 AM.

[Oberon]

#4

 

Full size image here

 

[Starman1]

Do I read this correctly that the stiffest designs are #1 and #2 with a slight edge to #2 on the roll component?

[Oberon]

And here is an early sneak preview of a curved spider for comparison...

Full sized image here.

[Oberon]

Do I read this correctly that the stiffest designs are #1 and #2 with a slight edge to #2 on the roll component?

Absolutely.

The surprising bit was that the roll component on #1 proved less than expected, less than it feels when handling. Grab it and it is very touchy and unable to resist small movements, but due to the significant radius of the centre section movement is soon arrested.

All those spiders performed excellently with the bump test, the differences are only marginal. The curved spider was a very different story and the graph doesn't do its chaotic oscillations justice.

Edited by Oberon, 05 July 2019 - 10:57 AM.

[ckh]

Johnathan,

 

What do you mean by "central support"?

 

In #2 and #4 the vertices of the spiders are off center from the CG of the diagonal and support. If you turned the square support tube 45 degrees and attached the vanes to opposite corners, the support would be thru the center of gravity and possibly a little stiffer due to the reduced roll moment. I suppose the approach you chose simplifies construction.

 

The poor performance of the circular vanes is not surprising, even though they are almost 1mm thick.

 

Super job!  We will all learn a lot from your efforts.

 

Carl

[Oberon]

Carl, you are right to observe that the off centre nature of my mounts in #2 and #4 has a slight weakness that would be improved by centering the vanes; I'd noticed the effect myself (most obviously the curved blue line in #4) and intended to talk about it. The flip side is that the off-centre design enables a much shorter moment for supporting the mirror; you can snug the secondary right up inside the vanes. However, doing that would have compromised the "like for like" intention of the experiment where each spider was supporting the secondary mass at a moment of 100mm. Therefore I need to make a short version of #2 and compare it with #2. When I do that I will also make a version with the vanes on center and basically compare all three.

That said, so far, and as expected, #2 performs superbly and is the superior design among these 4. Yet there remains room for improvement. I'm happy to play along here and creep up on the optimum design experimentally.

So far as the curved vanes are concerned, for consistency I made a curved vane spider using the same 0.5mm aluminium sheet hoping to at least get a direct comparison. I knew it would perform poorly, but...well lets just say it was so hopeless I couldn't measure it, the laser quickly left the entire target. And it wobbled.

 

[Starman1]

Personal experience in the field has told me the best curved spider design is )o( 

In exactly that orientation.

[Oberon]

Then you will be interested in these plans...

 

I'm predicting that #8 will be at least as good as #7, but has the added benefit of spreading the diffraction around evenly over the field of view.

Of the designs shown here, only #5 and #8 spread the diffraction evenly around 360 degrees; if you consider it carefully you will notice that in both #6 and #7 the opposing vanes track exactly the same angles, and thus only spread the diffractive energy over half the view, and at double the strength.

Edited by Oberon, 05 July 2019 - 06:41 PM.

[ckh]

Thanks for the response Jonathan.  I suppose a compromise would be to bring one pair of vanes up to the part of the back of the diagonal that is closest to the primary, with the other two vanes diametrically opposite (even through the position or width of the vanes would not be symmetrical.) Then the moment of the diagonal along the eyepiece axis is minimized and the distance between the vane pairs that resist pitch and yaw is maximized. The same strategy might work with wires as well.

 

 

Did you tension the vanes?

 

Looking forward to more results.

 

Carl

[MitchAlsup]

It seems to me that if the point where the two vanes touch were coincident with the CoG of the secondary, that there would be an even lower rotational moment.

 

[tommm]

That's basically the Albert Highe secondary holder configuration. It has about zero moment arm for roll, but longer moment arm for pitch compared to the offset design (note Oberon's comment that you can "tuck the secondary up between the vanes).  Totvos says this is very rigid, even with the relatively long moment for pitch he has on his holder - post #668 here. Also post #151 in this thread.

 

Very nice compact design with the adjustment screws down at the mirror.

Edited by tommm, 07 July 2019 - 11:24 AM.

[Oberon]

It seems to me that if the point where the two vanes touch were coincident with the CoG of the secondary, that there would be an even lower rotational moment.

 

SecondaryFrame.jpg

Yes. That was the essence of post #258 first paragraph. We will get to that.

[MitchAlsup]

Yes. That was the essence of post #258 first paragraph. We will get to that.

Thanks.

[Oberon]

#8

 

SS Crayfish

 

full size image here

 

Comments: a 4 vane curved spider, performance only very slightly improved over #5 the 3 vane curved, most noticeably the higher frequency oscillations which still took a disappointingly long time to settle. 

I had my fingers crossed that the opposing curves might prove move stable than typical curves; not that I had any good reason to think so, just wishful thinking. Now as a result of this test I doubt if there is any significant effect, and expect both designs #6 and #7 will produce very similar results.

Consequently my fearless prediction is that geometry is not the secret to a stiffer curved vane spider, only stiffer thicker vanes. Quite the reverse of a straight vane spider.

 

 

[mark cowan]

A spring is a spring...  What are the vanes made of?  That might be a place to seek improvements.

[BGRE]

To increase Young's modulus significantly beyond that of steel something more exotic like Tungsten heavy alloy is required, with an accompanying increase in density. Increasing the vane thickness by 25% achieves the same result at less cost and weight.

[Starman1]

Or changing to a material less prone to vibration, like CFRP or, simply, certain plastics.

Curved vanes are not really in tension, to speak of, so a thicker material could be as light.

There would be a limit to the design and secondary weight, of course.

[BGRE]

Yes, damping is also important epoxy composites can be good in this regard. Cast iron is better than steel , however cast iron sheet is somewhat rare and cast iron can have other undesirable properties. 

[Oberon]

#8

CF Crayfish

same geometry as post #266 but this time with 1.2mm thick Carbon Fiber vanes instead of 0.9mm Stainless Steel

significant improvement, especially in resistance to roll (blue), and damping of vibration.

Full size image here...

 

 

 

Edited by Oberon, 01 September 2019 - 07:41 AM.

[Oberon]

#7

SS Carafe

 

full size image here...

 

 

[Starman1]

The only commercial scope I've seen with the spider vane arrangement of post #272 is the Obsession 12.5",

and the secondary size is 54mm and the vanes are thicker than 0.9mm (I don't know what they are but it seems to be more than 1mm).

See image on RH side of page here:

https://www.obsessio.../12.5/index.php

The UTA I.D. is only about 13.5", too.

At least in that scope, with those dimensions, the vanes seemed quite stiff enough in the field.

 

That doesn't invalidate the comparison at all, but it does change the result for one scope, perhaps.

[Oberon]

Yes. So far the 3 SS vaned curved spiders have pretty similar performance, only the thicker CF curved vanes show a significant improvement. Which is probably what we should expect. I can well imagine that thicker vanes on a 12” scope with a smaller secondary would provide sufficiently stiff performance. I guess it would be nice if these tests went on long enough to provide some robust guidance here, like a calculator tool with inputs like:-

1. Count of vanes
2. Angle subtended by each curve
3. Internal diameter of UTA (unsupported area)
4. Mass of secondary and secondary support

5. Thickness of vane

6. Material (steel, aluminium, CF etc)

7. Depth of vane

8. Diameter of secondary support
9. Moment of secondary from vane

 

Thats quite a few variables to consider, and you’ll note I’ve ignored geometry. There is no way I’m going to be able to test each of these separately. This really wants to be done (by others) as an FEA analysis with a few built examples for verification.
 

Edited by Oberon, 01 September 2019 - 08:20 PM.

[gatorengineer]

Late to the thread, and you certainly have done a lot of work, but what you have built in post 271 is not close to being structurally the same as in 266.  If it were your results would be significantly different.  Not trying to be critical but structurally the are far from the same.  Also believe 8 is supposed to be a 120 and a 60 degree not 45 degree.  6 I believe is also supposed to be 90s....   

 

I'm interested as I have spider issues on a 36 f3.5 as well as having to design a spider for a 76" scope shortly.

 

Again not trying to be critical...