Perhaps its time for a new front cover. The best place for the water rad is
turning out to be on the spark plug side of the engine. Unfortunately that
complicates the pluming somewhat because the water pump inlet and outlet are
on the other side of the engine. Perhaps we should take a page out of the
RX8 design and make a new front cover incorporating the water pump housing
with the water coming out and going in  on the other side of the engine.

Another problem is the pump should be mounted as low as possible so any
drop
in water level will not cause the pump to cavitate. We built some remote
pumps a few years ago that used the stock Mazda pump cartridge. One other
point is the pump would be more eff. if the water came stright into the
impeller.  Here are some shots of that remote pump.

I am looking for feed back here. Does anybody else have an ideas they would
like to see incorporated.

Paul Lamar

Paul. The  aluminum end housings I designed for the four/six rotor upgrade
have the water pump moved to the lower plug side. The inlet and outlet are
both out the back of the housing as I am using a center housing with
parallel rather than series cooling. The water pump is an aluminum big block
chrysler cartridge which I bought new on ebay for $39. It has a cast high
flow impeller. The water outlet on the center housing is on the top for air
bleed, the inlet height is the same as the pump outlet. I am currently
working on tooling installation and wireing at the shop/hanger to cut some
mdf  casting patterns probably in february with the gantry router. I sent
some solid models to Jeff a month ago which are very close to  what the
final  pattern will be. The open coring lofts for the side ports are not on
these models which would not be used with p-ports anyway. I will do some
modified models with the ports incorporated into the patterns when I start
back on this part of my project (schedule says dec 18 to get back on the
engine design)

I have run several of the parts with meshcam and generated
G-code for pattern and machined versions. I have a glitch on three of the
fourteen models, one of which is the front cover with waterpump. The problem
is a .0015 gap between the layers making up the solid, not hard to correct
just not done yet. Also I have incorrect tool allocation on that part which
may be part of the problem. I just generated the G-code to test the software
before I bought it. It came bundled with Mach 3 cnc for $300 but the meshcam
part would be $185. I still can't get my render to work on the laptop but
the parts show up fine in meshcam so you might get Jeff to send you
something printable.

As far as I know everything in the solid models is to
scale including bolt and dowell locations and the e-shaft bearing machining.
There are also dynafocal mount iterations of the front cover which have the
tension bolts threaded into the  part that includes the motor mounts. The
stock rear iron could be used with the aluminum center and front housings
but there is also an aluminum rear housing with a 3" flange extension
eliminating the  spacers to Tracy Crook's  PSRU plate and it includes an
additional 6 mounting bolts to the plate.


        Feel free to use any or all  of these designs as you see fit, or
save some time by modifying the parts as the basis of another design. I am
using the RX8 rotor housings which have the water seal "O" rings in the
rotor housings rather than the side housings but with the addition of the
side housing seal grooves RX7 rotors and rotor housings would interchange.
You might look at the sealing of the center housing pair of parts. It is
tripple "O" ring sealed with a telescopic joint interface which you might
adapt to your two piece end housing with wear plate. I used the two piece
design, just made the facing part full  size and .811 thick. The end
housings don't need the telescopic sealing because the e-shaft bearing
flanges bolt into the facing section and hold the parts together.

george grimes