>Paul,
 >
 >I found the KSE site and the pumps aren't that expensive but yes they
 >are still overkill, unless a bypass valve can be milled into them and as
 >such will allow us to adjust the flow we want.
 >Could you explain the advantages of the way Jim routed the second outlet
 >into the front housing?
 >I believe both inlets go into a common cavity in the front housing.
 >Could the pump inlet be connected to the outlet of the radiators?
 >
 >Bill I had no luck on finding any company website close to the name you
 >mentioned. Could you send the URL to us?
 >
 >Doug in Japan
 >
 >I can't read Jim Medere's mind but perhaps this drawing will give us
 >some clues.
 >
 >Also this is an excerpt from Max Bentele's SAE 1991 book Engine
 >Revolutions.
 >
 >Max was working for Curtiss Wright in 1958 during the early days of the
 >rotary.
 >
 >"Chapter 8 - The Rotary Engine Era"
 >
 >"Cooling of the trochoid housing presented the greatest challenge. In a
 >reciprocating engine, cylinder head and piston dome are exposed to all
 >phases of the working cycle. In contrast, the individual portions of the
 >engine outer housing are exposed only to their respective phase.
 >Temperature and flow velocity of the working medium, and thus the heat
 >flux to the housing wall, vary along the trochoid circumference. Two
 >extremes exist: In the combustion zone, the very high heat flux will
 >induce nucleate boiling of the cooling water; in the intake area, the
 >inflowing fuel/air mixture is, due to fuel evaporation, colder than the
 >ambient atmosphere and much colder than the cooling water. The design
 >task was therefore twofold: (1) to prevent overheating and resulting
 >thermal cracking of the housing in the combustion zone and (2) to
 >minimize thermal distortion of the housing and to assure effective gas
 >seal operation."
 >
 >"We soon realized that NSU's circumferential cooling flow was incapable
 >of fulfilling all requirements of high-performance engines. We selected
 >axial flow, back and forth between the side housings, with the flow
 >velocities in the various rotor housing sections adjusted to their
 >respective heat input. In the combustion zone, the high-velocity cooling
 >flow removes the steam bubbles generated by nucleate boiling, and thus
 >permits extremely high heat fluxes to the housing without detrimental
 >effects. This cooling scheme was also beneficial to the structural
 >integrity of the rotor housing. Engine test runs confirmed its
 >effectiveness in all respects. It was patented and became the standard
 >for all high-performance Wankel engines."
 >
 >What Mazda did not do is adjust the flow velocity between front and rear
 >rotor housings let alone the 3 rotor. Rotor housings are interchangeable.
 >
 >BTW nucleate boiling is good boiling.
 >
 >Paul Lamar
 >
 >Paul & Doug,
 >The MEZIERE site is www.meziere.com <http://www.meziere.com/>
<http://www.meziere.com <http://www.meziere.com/>>. They have
 >changed their site since my last visit and no longer have the pump HP
 >charts available. They probably got too many questions about the
 >effectiveness of the electric water pumps. The KSE pump was the other
 >pump I first checked out. KSE is supposed to have a mount to run the
 >pump directly off the back of a dry sump pump. They designed it for the
 >sprint cars that run the dry sump pump off the cam cover.
 >Bill Jepson
 >
 >--

Bill

Thanks. Yea I did find that site but the amperage draw verses the
performance looked dubious to say the least. Here is a good company
that offers many pumps albeit ones we would have to adapt. They seem to
have experience supplying the racing community.

http://www.stewartcomponents.com/index.html

Perhaps we can modify one.

The tech notes they offer are especially pertinent as pump performance is
directly related to the head they have to pump against. 55 gallons a
minute at zero head is unrealistic.

http://www.stewartcomponents.com/tech_tips/Tech_Tips_6.htm

As quoted:

"Proper bench testing of accessories is the only proper method of
development. The accessories affect so many functions of the engine that
testing them on a running engine on the dyno is a total waste of effort.
The coolant pump is a great example of an accessory that must be tested
and developed off the engine. To bench test the coolant pump you must
know pressure drops at a given flow for all the components of the
cooling system.

Following is a typical Winston Cup engine at 100 GPM:

Lower radiator hose = 1.5 PSI
Block and cylinder head - each (at 50 GPM) = 8.5
Outlet manifolding = 1.25
Top radiator hose = 2.25
Radiator = 1.5
Total = 15.00 PSI

In addition to having the proper flow restriction as expressed in GPM @
PSI, the cooling system pressure and temperature must be known. All
these conditions are duplicated for the bench test.

Energy losses due to driving the coolant pump can only be calculated when
all conditions are duplicated and torque and RPM measured. Amp draw of
the drive motor is not an accurate measure of the energy required to
drive the pump. Torque must be measured with a load cell and horsepower
calculated from there.

Doug in Japan

Good work Doug.
Excellent web site. I have always been impressed with Stewart Warner.
They always present excellent data and good information.

Paul Lamar

Doug,
I agree that for tests to yield realistic results they have to include
as true-to-life conditions as possible. I have been on the Stewart site
before myself and they are obviously an actual supplier to industry, as
opposed to selling vapor ware. The pump from Meziere I was talking about
it a separate engine driven pump. They still sell it, although it seems
like an afterthought on the current site. We would probably be better
off if we used the KSE pump or adapted a Stewart unit if we actually
need anything but a stock unit.
Bill Jepson

Bill, Paul

I would like to hear your design thoughts about having a housing machined
for this BMW underdriven water pump. I believe would  basically be a
cylinderical shape with water coming into a flooded chamber around the
impeller and a tangentical outlet above it.    Flows look adequate.

I am waiting on more detail flow verses head graphs from Stewart
Components and I'll let you know.


Doug

Doug, I don't know what that file is. Not jpg and not a zip file. How did you
come up with that?

Paul Lamar

Paul,


I dunno.  I just went to this Bimmer catalog website, enlarged it, and
copied it into my files.
I have been able to find this pump for $172.

http://tinyurl.com/5xm5ar

Here is another website with a catalog containing brackets and many many
examples of serpentine belt set-ups.  I found this very valuable to
understand what is possible on the accessory side of our Mazda engines.
Some of the V-8 engines pictured look like they have been influenced by
Japanese 'anime'  - animation.   Alright guys "Transform"

http://www.marchperf.com/catalog.html

Doug in Japan