I got all my autocad drawings redrawen pretty much from scratch
 into Rhino over the last three weeks. Basically the parts are a flat plate
 with an external shape for the part, from which small solids are
 subtracted
 for water manifolds, lightening, porting of intake and exhaust,stationary
 gear mounting, "O" ring lands and bolt holes.

  I have the pieces located within the plate part but am having trouble
 subtracting them from the main plate using the "solids difference"
 command.
 Some subtract as expected, however about half of them return an error
 message "boolian difference failed"

   I did a test run with the pieces as polysurfaces and had similar
 problems. I have since converted them to solids with the "create solid"
 command which appeared to work 100% of the time. I have eleven various
 parts for foundary patterns and another eleven pieces for finished parts.
 The engine requires six components, the rest are for differing motor
 mounts,
 water pumps, and a connector bellhousing to connect two engines end to
 end.

   I also have difficulty with the "solids union" command which
 returns "solids union failed"with the larger parts such as the sidewall
 of
 the bellhousing and the flange. Since some sections will subtract from
 these
 larger pieces and some similarly shaped sections will not, it appears that
 the larger parts may not be the problem. All the parts have a 3 degree
 taper
 on the edges for use as foundary patterns.

   The various pieces were constructed with a bottom surface a side
 wall and a top surface connected with the join command. The "cap planer
 holes" command doesn't seem to work on complex shapes, just circles and
 rectangles. The "surface from planer curves" command does work indicating
 the curves (outlines) are indeed in the same plane.

 Any help anyone can offer would be greatly appreciated particularly with
 the boolian commands. Also something which would indicate what error
 caused
 the error messages would be helpful. The manual shows how to apply the
 commands but lists very little help on what to do when they fail to work.
 Virtually the same documentation problems as Autocad.

 george grimes

 I want to see some renderings George :)

 Paul Lamar

 Paul, I would sure like to get your input on this engine and will send the
 JPG output just as soon as I can figure out why the difference command
 isn't
 working. I did a JPG of another part and know how to save a render to that
 format. Without the parts subtracted out all that would render is the
 external shape sort of like something with a Christmas cookie cutter
 shaped
 like a side housing.
   Yesterday my new desktop quit, something to do with windows system
 check deciding to shut everything down. I did the clean all the cable
 connections and t ran another few minutes. Now all I have is the BIOS
 screen. I don't have a lot of spares for this newer stuff so I will send
 it
 off -again-for the third time.
  If it stops raining I can get the electric service connected to the
 hanger which is pretty much wired up except for the larger machine tools.I
 should be getting the large format printer operating, but I am just about
 tired of computer problems and have no reason to expect that plugging it
 into the wall will get me airframe drawings in full scale.

 george grimes

 The difference command is very picky. You want to make sure both
 objects are truly solid. Choose "solid" and scroll down to "cap planer
 holes". That usually does the trick. Practice with simple shapes
 such as two blocks of different size. For complex shapes
 choose: analyzemass propertiesvolume and it will tell you where
 you are going wrong.


 Go to: http://www.frys.com/
 choose: electronic componentscomputer componentsmother boardsintel
 view all Intel boards. Scroll down to Item # 6033738

 This is a brand new ITX mother board that will fit in any old case.
 That INCLUDES the Atom processor for $80. << Some of the memory you now
 have may fit. If you buy a mother board, Fry's will sell you
 MS Xp for about half off. Around $100. You may need a SATA drive
 for about 500 gig for about another $100. Most old keyboards will work
 as well as VGA monitors you now have.

 They also have a complete Acer system using the Atom, a 160 G drive
 screen, 1 Gb memory, XP and keyboard for $200.

 This is a lot of bang for the buck in computers.

 Paul Lamar



 Paul, The computer is already all new parts. Two 500GB western digital
SATA
 drives and one maxtor 330 Gb One drive for Linux and the other 500 Gb for
 XP
 which I got for $102. I can put in a modum and microsoft will repair the
XP
 over the internet if I use the same mainboard it is regestered to. I had
to
 do that on the last crash. My ram is 2Gb Hitacchi and the moniter is a 19"
 1600 by 1400 resolution. I tossed the $300 video card when it burnt up and
 put in an older diamond viper which has been working fine for the 90 hours
 since the last crash. The first setup ran about 11 hours.. That failure
was
 the power supply that came with the case and I put in a 550 watt high end
 type with the extruded aluminum case.If the hard drive is crapped out My
 rhino licence is a one install (annual) student version. If it's that then
 I
 will probably not ante up again but just wait for the Rhino 5 which is due
 out this year.

 While I am stuck I will research the analize mass pr
 operties volume
 command options and see how to use it. I have used the cap planer holes on
 simple objects successfully, however these objects are much to complex for
 it to work. I have tried redrawing a few using only arc segments and
 increasing the minimum radius arcs to something larger. Still no joy. I
 tried a trick I use in Autocad when something wont extrude on a tapered
 solid. I use "curve offset" with an offset of .002 and just continue
 offsetting the curve until it fails to offset giving a return of "curve
 offset to two open curves" rather than "curve offset to one closed curve"

 Sometimes I can offset .250 and come back .232 to where I actually wanted
 to
 go and it works. If I offset this curve back to the original size it is
 exactly the same so I hadn't forced a shape change on the curve. I am
 familiar with the direction vector concept, just not why I should have to
 deal with it as the default operation of the software. If a solid is
 constructed from surfaces either the top or bottom normal has to be
 "flipped" before joining the surfaces. This is not really a problem just
 different. The supurb loft command in Rhino makes up for any small
 annoyances. And the accuracy and speed get a high rating as well.

 george grimes

 And it is still crashing? Thank God I use Linux for email :)

 If you are dual booting that could be the problem.
 I would not run two different O.S.'s on the same computer
 as there are all sorts of hardware registers that may get
 set up differently despite what some of the Linux advocates
 say.

 Learn to use "loft" as an option to "cap" complex shapes.

 Rhino is very picky some times on "difference" and "unions" on
 complex solids. Sometimes I need to "explode" them into
 a bunch of surfaces and "join" the surfaces one at a time.

 This also happens sometimes when you do a lot of editing. When that
 happens you should have a good idea of what the part is supposed
 to look like so start over without all the patches.

 One of the rules to keep things well organized is:
 make sure "snap" is always on. If necessary set snap
 to .001.

 I have been using it for close to ten years and I am still
 learning tricks from time to time.

 It is one powerful tool however. I am still using 3.

 Paul Lamar

Paul.  I made the solids originally from surfaces and joined the parts one
at a time. I noticed that some returned "2 surfaces joined into one open
surface" but hoped it wasn't a problem. I expected to see that until the
last surface was added. I haven't used the "cap planer holes" after the
join
command but can see how it might fill edge inconsistancies on the top and
bottom surfaces since they are planar. The sides are more like ribbon candy
and many failed the "loft" , "extrude tapered surface" and "edgesrf
"command
with either a jumble of cross connected lines or only a partial surface. I
generally used "planarsrf" on the top and bottom curves to make surfaces,
then deleted the curves. That was followed by using the boundary curve of
the surfaces which was segmented????? and the "loft" and "edgesrf" command
on these segments. The edge pieces were not continuous so I joined them
with
"join naked edges" including the closing edge until I had a joined
perimiter, then joined the top and bottom after checking sufrace normals
were facing out.

      I spent yesterday researching the analyze based commands and have 5
pages of notes on things to try. One interesting one is to fillet the edges
with the radius set to 0 to fill inconsistancies of the 87 degree corners.
I
was putting off filleting until I had complete parts, figuring if the
fillet
didn't work then that using ball end mills would fillet the square
corners.Also leaving out the fillets would make it more likely that anyone
I
give these designs to would have a better chance of importing into a
different CAD system.

    I looked up what the chevrolet vega block was made of a few daya ago.
They used A390 with 17% silicon and etched the cylinder bores after honing
to expose the silicon as the running surface. It is classafied as a high
strength alloy. An article on the WIKI site was of the opinion that the
early problems with the vega resulted from an undersized cooling system and
oil losses resulting in overheating and engine damage. The block was
apparently capable  of normal overhaul intervals. It seems to be the ideal
material for a one rotor end housing casting.I ran the same briggs and
stratton lawnmower for 22 years and 6 sets of wheels on the  same engine,
and that was an aluminum bore. I retired it when it needed the second set
of
rings and still have it in the shed.This engine was run about 3000RPM and
being air cooled ran hotter than I need in a water cooled rotary. It might
be interesting to research what alloy nthis engine used.

george grimes

If it were that easy I think Porsche and BMW would be using it instead
of Nicasil.

Paul Lamar



Paul. Well I  looked up the aluminum alloy that briggs and stratton used on
their lawnmower engine blocks dating back to 1953. They used ADC-12 which as
far as I can tell is only used in die casting. the analysis  is 85% aluminum
and 12% silicon with a minamum of 9 silicon, 1-1/2 copper and less than 1 of
the common addatives used in aluminum. Apparently the chevrolet vega and
briggs and stratton engines didn't use nikasil coatings. The current BMW
engine has an aluminum core inside a pressure cast  magnesium cylinder block
to prevent stud pullout if it was threaded into the magnesium. The type of
high end manufacturing they have a reputation for.

       I did a litle research on nikasil plating. It is electrically
deposited similar to chrome plating. The major differences seem to be a
powdered nickel silicon in the solution bath rather than solid anodes as
used in copper, nickel, and chrome plating. There is also a more complex
mixture in the solution bath which didn't sound like something to have in a
garage. A wetting agent is used to treat the powder before putting it in the
solution. There are several papers written by researchers in india listed on
the internet describing solutions, current density and part
cleaning/preparation.

       The major users of the process are single cylinder 2 stroke
motorcycle engine manufacturers in europe and Japan where it is used to
prevent the aluminum piston and cylinder from galling and more importantly
for friction reduction with the additional benefit of wear resisance.  The
surface also has some wick up of oil to improve lubrication. Several
articles mentioned running barrel faced chrome plated piston rings due to
sharp edged rings scraping off the coating. One manufacturer also uses
nikasil on the thrust faces of the piston skirts.

        I didn't find much on the cost of the process other than to see one
manufacturer mention it added $8 to the cost of the engine.If a running
surface for a rotary could be done for $50 it might be worthwhile. I would
have to locate a shop in the US to apply it and also determine if it would
work with the rubber rotor  oil seals. My two piece end housings will have a
$150 casting with no "O ring grooves on the rotor face side (RX-8), about 7
hours machining and 3 hours lapping. This could in my case make these parts
a throwaway at rebuild time since I have the machine tools sitting in the
hanger to do the machining in house.

        I still havent heard about the demise of my computer, however
McNeel is doing public trials of Rhino V5 with a free download to Rhino v4
owners. The student version of Rhino V4 is down to $159. Their monthly list
of bug fixes and new commands has come down exponentially over the last
several months so it might be time to try it as soon as I finish this
current current
project..

george grimes