Re: (ATM) secondary-holder/spiders

[Chuck Grant <grant@aretha.llnl.gov>]

> From: PFAFF2@delphi.com
> 
> > > > > In affixing the diagonal mirror to the holder you must balance
> > > > leaving it loose enough so that there are no visible effects
> > > > from stress, with or without thermal expansion, and mounting
> > > > it firm enough that the collimation does not change as you
> > > > orient the telescope in different directions.
> > > > 
> > > > If it is firm enough not to move, you don't have to worry
> > > > about it falling off (under normal conditions).
> > > > 
> > > >
> > > The conditions will be ment if you glue the (diagonal) to the holder
> > > in 3 places with RTV with 1/16" or more in thickness.
> > > Bob Pfaff 
> > > 
> > 
> > Not quite.  If you make the blobs too thick, it will flex too much.
> > 
> > Chuck
> > 
> I think you are nit picking here. I did not mean 1/4" blobs of glue.

Well yes I am kind of picking nits so that what is posted is 
completely correct without any external assumptions.  I know you
know what you are talking about.  I just think you didn't write
exactly what you meant.

> The RTV must be thick enough to stretch and not transmit to much
> of the distortion of the holder to the diagonal. Some ATM's think
> 1/16" thick blobs are on the thin side. Anyway if the diagonal does
> sag it will be parallel to the holder and should not change the optical
> path within limits.

Unfortunately this is just not true. The center of gravity of the
diagonal is not in the plane of the three support points (that plane
is the surface of the diagonal mount).  So unless the diagonal is oriented
exactly horizontal, there will be some twisting moment applied to the
interface between the diagonal and the diagonal mount, and the forces 
on the three blobs will not be the same.  I don't know how significant
this effect is, but it exists, and it clearly gets worse as the blobs 
get thicker.  Considering that for diffraction limited viewing an f/4 
Newtonian has to hold collimation on the order of less than a millimeter 
of optical axis shift at the image plane, I would estimate that making
your blobs too thick could be a real problem.  Especially for big heavy
diagonals.  Any twisting of the diagonal position, of course, causes 
much more optical axis shift at the image plane than the same amount
of parallel sag.

> 
> Has anybody ever done or heard of an analysis on what the correct
> thickness of RTV should be? It seemes to be just a gut feel now and what
> has worked before.

I wish I knew of such an analysis, or quanitative experiment.

> Bob Pfaff 
>