A few minor problems for the 13B. One hole was ten thou
  off or so but it could be drilled out with no problems.
  One cup tab was a bit too long so that could be hacksawed off.
  We had a few typical bolt circle problems but easily fixed by
  drilling the holes out to 27/64th instead of 10 MM.

  Other than that the 13B was usable. The starter hole
  was a bit on the small side but I can set it up in my
  cheap Chinese mill and bore it out about .020.
  The starter hole should be 2.640.

  I was working from Tracy's dwg and I should have
  checked the dwg against Tracy motor plate which I also have.
  The drawings had a few errors and I probably made one or
  two my self. The dwg does not match Tracy's plate :)

  The RX8 extra bell house bolt was way off by about a half inch so
  that is going to require a whole new dwg which I have about done.
  I want to make the plates fit both the 13B and the RX8 motor.

  Jeff I'll get the  fixed dwg to you in a day or so.


  Paul Lamar

  Paul
  So this is usable. If  you want I can open up the starter hole
  here. It
    measures 2.620 on the solid model. I depend on the solid
   model for
    geometry for expediency. If you need to digitize the bolt
  pattern on the
    housing we can do that. We need absolute numbers. I do not
    know the
    difference between the RX7 and 8.  the half inch hole problem,
  does that
    make it unusable??? No worries, we just get it right for the
  folks that
    need them...
  jeff
  I'll ship it back if you want it. I only have one RX8 end housing
    and it is on my RX8 P-port test engine which I don't want to
  take apart.
    I think I got it measured fairly accurately.
  If you have a 13B end housing that would be a good one to
    digitize.
    The solid model is now fine. I fixed it. Jarrett has it now and
    he is converting it to a Solid Works file for you. I made a
  couple of
    other changes while I was at it. After you do it three or four
  timesyou get better and faster. The new version is much more
  accurate as
    I had been there and done that :) All ribs are now the same
  width and
    parallel. We need to drill all the 10 mm holes out to 27/64th
  as that
    will ease the bolt circle problem.
  No change in the back pattern.
  We should look into a spun or pressed bell housing and use
   the two
   hollow
    dowel pins on the end housing to locate the unit more
  accurately. Along
    with a lighter weight gear box of course :) The bolt and spacer
  idea is
    less than ideal.
  Paul Lamar

   Paul

   I think when you send me the new geometry I will make a
  template, send
   it to yo then you can check for fit. I am not in favor of
    making the
   holes larger. I can use my machine as a CMM and get the number
  exact.  I
   did that with the rotor housing and it is perfect.... We save
   these master coordinates for all future parts. This should be
   done all
  around  the block.
   I have machined bell housings before. Lots of aluminum chips but
  can be
   done fairly quick as opposed to finding and getting someone quote,
   setup, bid,etc. Lots of time and money..
   My guess a billet block of 6061 for a bell housing would be at
  least 60
   lbs..Al plate is a little over $3.00/lb.  Figure $200 for
  material.  It
   will machine down to about close to 8  lbs or less. If the
  housing is
   not too complicated with lots of ribs, this could be machine in
  1 day.
   More,  obviously for the first article sample, but for any
  production  at  all,  $600 to $700 is doable. Eventually get a
  cast part , but this
   way we could get something out cheaper and sooner...
   I agree the bolt and spacer arrangement needs to be fixed....

   jeff

   No doubt about it that is the ultimate solution. I rejected it in
   the past because of the cost. Prices have come way down on NC mill
   stuff as a lot of people have the machines payed off.

   The problem with sand casting is weight. The foundry has
  problems with
   anything thinner than .2 inches. Also the cost of sand
   castings has
   gone way up. NC milling is now more competitive.

   Try these until Jarrett gets the SW version out.

   Paul Lamar

  Paul

  Tell Jarrett no need for file conversions anymore. I have a new
    SW2009  seat so all will work directly......

  jeff

  Great. Thanks Jeff.
  I'll bet it will load Rhino native directly :)

  Paul Lamar

  Paul, I did this over lunch so it's the quick version. I didn't
    put the cups
    on as we know they are strong enough. I just fixed one end
 where the
    cup goes [the farthest from CL] and the second one I fixed it such
    thatit was still moveable from the first yet fixed in the vertical
    plane [similar
    to how flexible an engine mount 'could' be. The trick here is
 if I
    make both
    ends fixed [ie cast in concrete] then there is some degree to
    stress that
    is carried via tension between the two fixing points. When I put
    one fixed
    and one fixed in the vertical only ?I think we get a better
    representationof the loads w/ a mount that has midle
 deflection at
    load. It looks ok still
    to me.. but I haven't spent too much time going over the results.
    I can do
    the 'real deal'  test later tonight but thought this might
 help in
    the mean time.
  Let me know..
  Jarrett
  That ia much better.
  Jeff this is the new design fixed with the RX8
    hole in the right place. I also added a rib on the highly loaded
    left side. The stress level is now very conservative which is what
    I was shooting for.
  Thanks Jarett.
  Paul Lamar
  No problem, do you want me to do one w/ the cups on it.. or are we
   good at this point?

   Jarrett  [geeze I've got to slow down when I'm typing.. and then
   proof read.. sorry about that...]

   NO cups needed. We are good. Jeff might need a solid works file.

   Paul Lamar


  Ok, Jeff, let me know if you do.  Btw those PSRU holes where put
 in on
  a 5.875" B.C.
  Two holed on the top x 12 holes [30deg OC] Each being 0.25dia.  That
  sound correct?


  Jarrett


  Jarrett

 I got the iges...... there where no holes for the psru and no register
 in the back.  All the geom comes in as splines. That make it difficult
 to separate out for machining..do you have in SW a standard set
 made in
 lines and arcs. It's easier to manipulate and takes a lot less memory.
 If you don't' I will make a version here that's cleaner but we
 need to
 get the design absolutely established. send straight up SW from
 now on,
 no more conversions.

 jeff

 Did you try the step file? Jarrett has it in SW and he will send it.

 Paul Lamar


Here's the SW file.. unfortunatly this too has splines [I've yet to see
any that don't when they come over this way..]
On this part [due to it being a basic 2D geometry for machining [2.5D]]
one might as well just switch
it to a dxf/dwg. The file is attached... I don't have the rear register
in there [didn't have time to go back
and look up the dim's for it]. I put the PSRU B.C in but need Paul to
confirm the dim's [about 3 postings
above this one]..

Sorry not a huge help..  :/

Jarrett

I think the splines are when I use the "fillet curve" option.
It ask for a radius but I think it "blends" the curves. One of
the really neat features of Rhino. That is why it does cowling's
and other compound curved shapes so well. It will also blend
surfaces and loft.

Fortunately sldprt is not a text file and does not need to be zipped.

Paul Lamar