66 replies to this topic

[Bob4BVM]

Yes, the cell really needs a reboot, and I would also open up the back of the mirror box to improve ventilation.  Equilibration of the temperature of the mirror is a huge factor in telescope performance.

 

I have refigured a number of those mirrors.  Sometimes the glass itself does not have the best anneal, but hopefully your particular piece is OK and it has been ground flat on the back.  The wavy-backed Pyrex blanks can catch on mirror cell supports and act strangely in some cases.

 

Here is my mirror support article, which is based on testing optics under the sky and in my shop.  I think it's the most modern treatment of mirror cells: 

   http://www.loptics.c...rorsupport.html

 

Additionally, here is a link to my article about tracking down telescope issues, which you might find useful after your cell is rebuilt:

  http://www.loptics.c.../starshape.html
 

Good luck.

 

Thanks for those articles Mike. Very useful !

I am in the slow process of a complete redesign of my 17.5, and most everything will change. The mirror box will go away completely, along the lines of Oberon's Merope scope.

The only reason I showed the old cell was to illustrate that it is possible to put together something pretty low-tech and still get a scope up and running pretty quickly, & for very little cash.  I know people who will agonize endlessly over every decision and end up never getting started or never finishing, when they could have been out under the stars enjoying the scope; there is no better place to decide what improvements they want to make. That's the beauty of ATM, none of my scopes are ever really 'finished'.

Bob

[JimV]

Mike, I had read your well written article before, and found and reread it earlier today.  Thanks for posting the link.

Clearly I have problems with my seat belt sling pulling on the mirror, and likely torque from the cell itself when tilted.

It's easy to misdiagnose these tilt problems as bad seeing in low atmosphere.  The giveaway was the perfectly round diffraction patterns with mask in place.

 

One option is to drop out my tailgate frame and replace it with a tilt frame, while keeping the rest of the scope unchanged.  I already have delrin, as I had emailed JP a year ago about mirror cell advice.  I'm excited about fixing the problem because my short eyepieces will perform.

 

Why 45* for the Whiffletree?  Equal spacing?

[JimV]

I derived the equation for force on edge of mirror for Whiffletree holder, dependent on angles and lever arm ratio choices.

[Mike Lockwood]

The 45 degree separation has been used for some time, I'm not sure where it originated.

 

John Pratte and I tried it, and we have found that it works extremely well on mirrors from 12" to 32" in aperture, and I hope to test a 34" in the future.

[JimV]

It's a great design.

Thought you would be interested in the numbers.

Using 45, 45, 45 degree spread results in 0.38 of mirror weight pressure on each roller.

Using 30, 30, 30 degree spread results in 0.30.

If all 4 rollers sat at bottom of mirror, it would of course be 0.25.

 

Where the lower rollers are positioned is less critical than the uppers.

The higher the upper rollers, the more pressure on the mirror edge for all rollers.

Using 50, 50, 50 degrees gets you 0.43 mirror weight on each roller.

[Michael Miles]

... they have 2 or 3 8 inch dobs available for roughly 500 dollars or less. ...

 

Here's one that is more bang for your buck - out of stock at this site, but worth tracking down:

 

http://www.telescope...ector-telescope

[Nils Olof Carlin]

.

 

Why 45* for the Whiffletree?  Equal spacing?

I like to think of it like this:

If the mirror were supported only at the lowest edge, it would tend to bend forward and "potato chip", causing (low order) astigmatism. To compensate for this, you need to apply as much balancing force side-to-side. There are essentially 3 practical ways. The simplest is with two points, each at 45 deg to horizontal (or vertical if you prefer). Next level of sophistication is the 45 deg whiffletree - it is also balanced in the same way. After this comes the 180 deg sling. While all these cancel primary astigmatism, the most important mode of deformation, there are other modes that are more or less suppressed. With a really well made sling, the improvement over a good whiffletree support is small - a poorly done sling (as most are or have been - like seat belts!) is no doubt worse.

Other possible whiffletree angles (like 30 deg) do not cancel astigmatism and do not work as intended.

 Do check http://www.cruxis.co...ecalculator.htm before you decide - I suspect a 2-point support with rollers is fine for your 17.5", and the easiest imaginable to make.

 

Nils Olof

[prfesser]

Diving into this thread...Nils Olof, I was planning a 180 degree sling for my 14.75 x 0.75" quartz mirror.  What makes a sling well-done, what makes it poorly done?

 

Thanks!

Terry

[JimV]

Sorry, I'll interject.  I'm an expert on seat belt slings not working.

Mirror slides side to side, so I need to recollimate if I roll the scope to where a tree isn't blocking my view.

The sling has a habit of slipping off the mirror edge.  You can use double back sticky tape, but what a mess when that tears up.

The seatbelt stretches affecting collimation.

The worst problem is when pointing the scope less than vertical.  Sling isn't centered on mirror edge so it either lifts or pulls down on mirror deforming it.  I have triangular diffraction patterns on my 1/19 wave mirror.  If you don't have a mirror cell that tilts with the mirror during collimation, the sling will never be centered.

I supose I could use a wire sling, but that won't fix the lateral sliding problem.

I'm retrofitting whiffletree.

[Nils Olof Carlin]

Diving into this thread...Nils Olof, I was planning a 180 degree sling for my 14.75 x 0.75" quartz mirror.  What makes a sling well-done, what makes it poorly done?

 

Thanks!

Terry

Terry, Jim,

 

This old idea brings up the factors I believe are essential:

http://web.telia.com...g/threepart.htm

  Using two 90 deg slings for weight-bearing keeps the side position well secured without swinging.

  Cable slings can be positioned at or near the known plane of the COG (see the Cruxis calculator)

  Using the teflon sheat (or possibly beads) keeps friction low and force distribution close to ideal

  Clamps at the diagonal diameters minimizes the sensitivity to having the free ends slightly off height

  Long free ends makes height adjustment easier and also less critical

  The "upside down" 180 deg sling holds the mirror securely from moving much during transport, without adding more than a minimum force, and one that is completely symmetric

 

This is my personal thinking - I know of only a few who have tried, and no details about their actual results. My own 13.1" is too small to show problems, if any - two rollers would have been quite enough. Check with the Cruxis calculator first! But if my slings or a 45 deg whiffletree solution is more practical, you must decide. BTW I don't believe a more complex multiple whiffletree would be a meaningful improvement.

Any factors missing?

 

Nils Olof

[JimV]

A guy at TSP had the dual 90 degree wire sling well made.

I have an old style fixed cell so I'm ready for wheels.  It's something new to try with a different set of problems.

Maybe I'll just do two at 90* like you suggest.

[Mike Lockwood]

he worst problem is when pointing the scope less than vertical.  Sling isn't centered on mirror edge so it either lifts or pulls down on mirror deforming it.  I have triangular diffraction patterns on my 1/19 wave mirror.  If you don't have a mirror cell that tilts with the mirror during collimation, the sling will never be centered.

 

The sling should be placed on the center of gravity, which will not be centered on the miror edge, it will be a bit towards the rear surface.  The same is true for the contact point of rollers on a whiffletree.

 

Triangular distortion is pretty common in these old Coulter primaries.  Rotate the primary in the cell to see if the triangle shape rotates with the mirror, and if it does it is likely polished in.  If it doesn't it is likely a pinched optic.

[JimV]

I'm still using cork pads, and will soon move to delrin.  Stiction isn't helping the image.

Using cruxis calculator, 1/16+" (1/20 mirror 1.625" thickness) off COG for 90 degree rollers triples surface deformation to 4.5nm.

4.5nm / 500 nm = 1/100 wavelength of light.  Double that to 1/50 wavefront because mirror reflects.

If I'm 1/8+" off COG, I imagine error is non-linear, and will be > 15nm, which is 30/500 = 1/16 wavefront.

If mirror surface error and deformation are uncorrelated, then these errors look like RMS numbers - you can square them, add them, then sqrt the result.  When noise is uncorrelated, variances add.  Somehow I don't think that works here.  Deformations and mirror figure likely are not fine featured random noise.  Likely they influence large sweeping areas of the mirror.  If they accumulate, you're at 1/8 wavefront.  Forgive me if I'm cavalier with combining two number types, but I think the theme is correct.  Rollers need to be on COG.  My frame doesn't move with mirror.

I may be stuck with trying a wire cable since mirror posts are at 120* intervals anyway.  Cruxis site and Nils Olof tell me that 60* angle causes astigmatism for rollers.

 

[Nils Olof Carlin]

If mirror surface error and deformation are uncorrelated, then these errors look like RMS numbers - you can square them, add them, then sqrt the result.  When noise is uncorrelated, variances add.  Somehow I don't think that works here.  Deformations and mirror figure likely are not fine featured random noise.  Likely they influence large sweeping areas of the mirror.  If they accumulate, you're at 1/8 wavefront.  Forgive me if I'm cavalier with combining two number types, but I think the theme is correct.  Rollers need to be on COG.  My frame doesn't move with mirror.

I may be stuck with trying a wire cable since mirror posts are at 120* intervals anyway.  Cruxis site and Nils Olof tell me that 60* angle causes astigmatism for rollers.

To add RMS errors as you describe, they need not be random nor fine featured - they just need to be orthogonal, in the sense that Zernike polynomials are (if you disregard the central obstruction!). With PLOP, you can find the zernike polynomials for deformation - not so with the Cruxis calculator, unfortunately. But, unlike you, I see little reason to believe errors will add linearly- to do so, they would be identical in shape, or at least similar.

Can't you simply add new posts at 90 deg?

 

Nils Olof

[JimV]

Can't you simply add new posts at 90 deg?

I'm at edge of tailgate frame.

 

[JimV]

I think I'm going to use a single cable with a slotted clip clamped to it at bottom of mirror.

I'll have an adjustable pin fastened to the frame going through the slot.  This allows

the mirror to move in height and top to bottom, but not side to side.

 

[JimV]

I moved from cork to delrin mirror cell pads this weekend.

My planetary images greatly improved.  It's hard for me to understand how lateral grabbing causes the problem.

My thanks to Mike Lockwood and John Pratte for the mirror cell page and doing the research.