Bill Freeman wrote:

Paul,

Just a suggestion for someone with a lot of curiosity and time
on his hands.  Seems like the whole coolant pump issue has
been causing problems, so a test rig seems to be needed.
If you are willing to test at room temp. you can eliminate the
engine, provided you can drive the pump without it being
on the engine. Otherwise a running engine with external
plumbng to a radiator and some valves with multiple pressure
pickoffs would be useful.

The key seems to be the backpressure in the system, basically
the resistance to flow which is the sum of the individual resistances.
Apparently Mazda intended for there to be a substantial outlet
restriction, based on your pix and description of the stock outlet geometry.
This wasn't an accident, I suspect. My guess is that with inadequate
outlet resistance (restriction) the whole system pressure is lowered
which may cause the inlet pressure (NPSH) to be too low and
cause cavitation (closed system).  Since a centrifugal pump is
designed for a particular NPSH, it appears that this will come
from the outlet restriction (minus the rest of the restrictions down
stream) raising the pressure in the system.  I imagine that with
inadequate outlet restriction that the outlet pressure is some lower
value, so the pump inlet pressure is that lower value minus the
losses, and may be too low to avoid cavitation.

For example - assume that with no intentional outlet restriction
the pump outlet pressure is 5 psi.  The piping and radiatior cause a pressure
drop of 3 psi, leaving an inlet pressure of 2 psi at the pump.   If an
outlet restriction is introduced the pressure between the impeller and
the restriction jumps to 20 psi, is lowered to 15 psi at the outlet
restriction, and is lowered to 12 psi by the piping and radiator.
Now the pump inlet pressure is 12 psi, which should substantially
improve the caviatation resistance of the pump.

A test rig with multiple pressure pickoffs and a variable restriction
(gate valve or ball valve) downstream of the pump would
answer a lot of questions.  A flow meter would be nice but is a
expensive.   You can make a simple venturi flowmeter and
calibrate it with flow into a bucket on a scale, timed.  You
would want a delta P guage from just upstream the venturi to
the min diameter point.  You could rig this up laying on the
top of the engine with a car by just building up the unit on
a piece of plywood and connecting to the pump with
radiator hoses.

Slowly close the outlet valve and record the pressures in the system
and (if available) the flow rates.  My guess is that there will be
an optimum point of flow which has a significant outlet restriction to
force the pump up onto the operating curve.

Please point out any errors in reasoning -  don't claim to be a pump
expert.

Bill

 
The Aircraft Rotary Engine Newsletter.              Powered by Linux.

http://home.earthlink.net/~rotaryeng/            http://www.linux.org