Paul,

I saw where at least one other member of our group tried using the
calculators at www.wallaceracing.com/runnertorque.php and found that the
results didn't make sense. Here's some formulas for our 13B engines our
group may wish to test experimentally. These are based on the underlying
principles used for the Wallace Racing calculators but adjusted to fit the
results you obtained on the Mazdatrix dyno.

Runner length (in inches) for peak torque = 157,000/rpm.

Runner area (in sq. in.) for peak torque = rpm/2281

Example - How to use:

From your testing, the peak torque occurs at about 88% of the rpm for
peak
h.p.

H.P. Peak desired at 6,000 rpm.

Peak torque occurs at .88 x 6,000 rpm = 5,280 rpm

Runner length = 157,000/5,280 = 29.7 inches

Runner area = 5280/2281 = 2.315 sq. in.

Runner diameter = 2 x sq. root of 2.315/pi = 1.72 inches

I used parametric variation (which is exactly the same thing as
cut-and-try in the lab, but it sounds more sophisticated) to deduce
their equations then calculated new constants to fit the data from your
Mazdatrix dyno runs. To see if the values made sense I checked against
the following knowns for the rotary:

I recall that the diameters and lengths of the runners on the peripheral
port you tested at Mazdatrix were intended to duplicate the LeMans engine,
and that engine had been optimized about as far as possible for road
racing.

The formulas used in the calculators at wallaceracing appear to be based
on the following premises:

1.The optimum runner length is directly inverse to the rpm.
2.There is an intake velocity that works well for a broad range of engine
configurations and rpm.

We know that the intake event on a reciprocating 4-stroke engine is driven
by 180 degrees of crankshaft travel. Racing literature points out that as
a result of valve overlap, the intake event is actually somewhat longer.

The rotary's intake event lasts a full 360 degrees. There's a period of
overlap, around 45 degrees, where the intake event for one chamber is
ending and and the event for the next chamber is starting. But to the
intake runner, the events occur at 360 degree intervals.

I examined the runner length results to see how they compared. Lengths
resulting from the above formula were 1.87 times longer than those
recommended by the Wallace calculator for a 4 cylinder 160 cu. in. engine.
Considering how the intake event of a rotary compares with a
reciprocating engine, the ratio seems reasonable to me.

I found the resultant formula for runner area a little surprising. Part
of the difference is due to the improved volumetric efficiency of the
rotary.

I often make a quick-and-dirty estimate of v.e. by looking at the peak
of the torque curve. Over the years I've observed that 1 cu. in. of
displacement results in 1 ft. lb. of torque at the peak for a lot of
engines. Your engine produced 1.25 ft. lb. per cu. in. suggesting a v.e.
of 125% compared with typical reciprocating engines.

To sound more erudite, this could all be calculated in terms of BMEP. The
fundamental units (force & length) are the same - but BMEP is expressed
more elegantly (and opaquely). Converted, 1 ft. lb. per cu. in.
corresponds to a BMEP of 151 psi. Reciprocating engines typically run
between 120 to 150 psi.

As a check of the other equation, I visited with Steve Beckham to compare
with the final engine he and Everett Hatch built. The intake velocity
they arrived at by independent experiment was only about 8% higher than
that resulting from the above runner area formula.

It would be desirable to design experiments to test the rpm range where
these equations apply, and to test the premises themselves.

I look forward to your comments.

Philip Groelz

Here is the Power Sport dyno power curve with the
smaller runner compared to the larger runner Lemans engine.
I am not sure what the exact length was on the power sport engine.

Paul Lamar


Philip,

Excellent post. I was under the impression though that intake length and
diameter can change the torque peak point radically. With the Lemans
engine with its variable length manifold and two inch ID tubes at
maximum extension (plus 175 mm more than the immobile section) the
torque curve peak occurred at 6,000.

Doug in Japan

Doug,

I didn't answer right away because I've been pondering your question. I
suspect you might be thinking about the almost explosive interchange of
energy many systems exhibit at resonance. Riders of highly tuned 2-stroke
dirt bikes refer to it as getting on the pipe. There has been and still
is a lot of effort expended in this direction with the result that the
volumetric efficiency is about 20 25% above a typical recip.

I feel there is still a lot to be learned about this little engine. And
the more we learn, the more we'll be able to achieve with it. We are
trying to apply what's been learned about both the two stroke and four
stroke engines and are having some success.

The basic breathing is substantially different, however. It's natural
to look at the rapid unporting and think aha, that's like a two
stroker. But then look at the four stroke engines and observe that this
engine has the same four events. I haven't seen much recognition of how
much slower the intake and exhaust events are with the Wankel design.
With a two-stroke, the breathing takes place in about one-third of a
revolution. With a four stroke engine it takes place in little over half
a revolution. With either engine there is no net flow in the pipe for two
to three times as long as the valve (or port) was open. But with the
Wankel, both the intake and exhaust events last the entire revolution.
Without the interruption, resonance effects should be stronger.

The duration for an individual chamber is actually even longer because
adjacent ports are connected to the conditions in the runner while the
edge seal is passing across the opening. About a year ago, one of our
members submitted a chart showing he measured the intake event for one
chamber at 440 degrees, and the exhaust event at 400 degrees of e-shaft
rotation. From studying my own engine the numbers look about right.

The equations in my previous post reflect this aspect of the Mazda
13B/Wankel breathing process.

What are the implications Pipes are longer and smaller diameter, which
have both good and bad aspects. That should result in lower friction
losses and higher resonant peaks. Longer also means heavier and more
trouble fitting into an engine cowling.

There's a substantial amount of time that the intake pipe is connected to
the exhaust pipe, somewhat like a highly tuned four cycle race engine or
even a two stroke engine. How can this event be better managed

Philip

The intake cycle in any one chamber is 270 E-shaft degrees. Each chamber
follows in sequence with no delays. The engine is as close to
a continuous combustion engine that one can find next to a gas turbine.

"How can this event be better managed" Do a p-port RX8 engine
with side exhaust.

Paul Lamar

Paul,
It seems that many newcomers to the rotary are
still confused about the fact that the rotary
is a4 cycle engine.On the exhaust side it seems
whatever scavenges the best works best. The Mazda
racing exhaust in the un muffled era was a continuous
expanding taper from the port to the exit. The two
tubes were blended into 1 about 14" from the port.The
megaphone exited under the passenger dooron the RX-2.
The outlet was3-3/4" diameter. This pipe dynoed well,
and was the loudest exhaust I have ever heard outside
of a AA-FD dragster. The design was intended to allow
the pressure wave to expand all the way to the exit. My
suggestion, since we must run a muffler is to have the
muffler as far downstream as possible. The exitcan be
pretty small once the flow breaks up. Lynn might
comment here. Don't the new rotary cars put the muffler
right at the rear bumper Intake tuning seems more
important to making power.

Bill Jepson

The RX8 muffler is huge. It is about 30 inches long and 8 or
ten inches in diameter,


Paul Lamar