On Thursday, January 8, 2015 at 12:22:04 PM UTC-5, Phil Hobbs wrote:   
   > On 01/08/2015 11:35 AM, Phil Hobbs wrote:   
   > > On 01/08/2015 11:05 AM, ggherold@gmail.com wrote:   
   > >> On Wednesday, January 7, 2015 11:23:10 PM UTC-5, Phil Hobbs wrote:   
   > >>> On 1/6/2015 8:54 PM, ggherold@gmail.com wrote:   
   > >>>> On Monday, December 29, 2014 9:06:55 AM UTC-5, Phil Hobbs wrote:   
   > >>>>> I'll probably have to get a couple to play with.  The   
   > >>>>> nonpolarizing thing and the asymmetry make them potentially   
   > >>>>> harder to use than PBSes for anything at all fancy--for   
   > >>>>> instance if the phase shift between p and s is more than a   
   > >>>>> degree or two, the output polarization will be a mess.   
   > >>>>>   
   > >>>>> Cheers   
   > >>>>>   
   > >>>>> Phil Hobbs   
   > >>>>   
   > >>>> Hi Phil, (and I thought SO was dead.) If there is any absorption   
   > >>>> you get a phase shift. (But you know that.) We've used the phase   
   > >>>> shift in metallic beam splitters.   
   > >>>   
   > >>> Interesting--what do you use it for?  IME it's usually a nuisance.   
   > >>> (You folks have some pretty cool ideas.)   
   > >>>   
   > >>> Cheers   
   > >>>   
   > >>> Phil Hobbs   
   > >   
   > >>   
   > >> This was my colleagues idea... we bought a bunch of inconel (I   
   > >> think) metallic beam splitters from thor labs, and found the one that   
   > >> gives ~90 degree phase shift.  This was stuck into a Michelson   
   > >> interferometer and then one looks at both the output beams.  (You   
   > >> need a second beam splitter to pick off the beam that goes back to   
   > >> the light source.)  Then with two signals in quadrature you can get   
   > >> direction information as you move one of the mirrors.  Pretty neat.   
   > >> You also get much better noise immunity when counting fringes... it's   
   > >> almost all phase noise and no amplitude noise. (If you understand   
   > >> what I mean by that.... the quadrature signal looked at x-y on a   
   > >> 'scope is a circle and noise (bumping the table) moves around the   
   > >> perimeter of the circle, but doesn't change the radius, that was not   
   > >> for you Phil, but for those reading along at home.)   
   > >>   
   > >> George H.   
   > >   
   > > Cute, and much cheaper than using waveplates, if you can afford the   
   > > power loss.  (BTW I never feel like you're talking down to me, don't   
   > > worry.  There's all sorts of stuff I don't know, or have forgotten, or   
   > > haven't thought about.)   
   >   
   > I was just running over the details of this in my head.  (I'm working on   
   > my third edition, and this might make an interesting subsection, with   
   > your permission.)  The intensities of the two output beams in a regular   
   > Michelson have to sum to a constant, because there's nowhere else for   
   > the power to go.   
   >   
   > With a lossy coating, that isn't so, of course, but the details are a   
   > little bit subtle, ISTM.   The side port interferes two beams that have   
   > each been reflected once and transmitted once, whereas the back port   
   > (towards the laser) interferes a twice-reflected beam and a   
   > twice-transmitted beam.   
      
   Yup, I can do this in my head now, but it's easier the first time to draw the   
   picture.   
   >   
   > The phase shift on reflection and transmission are different, so the   
   > intensities coming out the two ports don't sum to a constant.  One might   
   > think that the beamsplitter is a linear element, so the reflection and   
   > transmission coefficients don't depend on what the other beam is doing,   
   > but interestingly I don't think that's the case: the E field inside the   
   > metal will depend on the phase shift, so the total absorbed power will   
   > depend on the phase shift.   
      
   I'm not quite following you here.  When I try and think about the details   
   of what's going on inside the metal I get a little confused,   
   I've got refelcted 'bits' coming from both deeper and shallower   
   in the metal.  It's hard to see how that all works out in detail.   
      
   We use a slab beam splitter for the quadrature trick and there is   
   some (input) polarization dependence... I think this is mostly due to S and P   
   reflection/ refraction difference at the air-glass interface.   
   And that a cube beamsplitter wouldn't show this.   
   >   
   > That makes the beamsplitter function as a bilinear system, in the same   
   > sort of way as a Pockels cell.  It ought to be possible to produce   
   > cross-modulation in one beam by modulating the other one.  That would be   
   > a fun addition to your interferometry kit!   
      
   So wiggle the path length in one arm... But what am I detecting?   
   Both output ports have the parts of the beams from both arms.   
   Are you saying the cross modulation with give a signal at double   
   the wiggle frequency?   
   And how fast do I need to wiggle?   
      
   George H.   
      
   >   
   > Cheers   
   >   
   > Phil Hobbs   
   >   
   >   
   >   
   > --   
   > Dr Philip C D Hobbs   
   > Principal Consultant   
   > ElectroOptical Innovations LLC   
   > Optics, Electro-optics, Photonics, Analog Electronics   
   >   
   > 160 North State Road #203   
   > Briarcliff Manor NY 10510   
   >   
   > hobbs at electrooptical dot net   
   > http://electrooptical.net   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)