Engine Dynamometer Testing
Engine
testing was done over four run sessions.
First couple of sessions provided some data, but much time was spent
resolving issues with the dyno setup, operation and
instrumentation.
The
basic test plan had the following objectives;
w
Initial
run-in
w
Set timing and injector flow
rates, generate MAP correction table, check staging and timing retard MAP set
points; balance EGT’s
w
Get coolant flow (with and without
thermostat) and oil flow data
w
Generate max torque and power
curves vs RPM
w
Generate fuel burn
data
w
Measure power and noise level
with/without secondary muffler.
Initial
problems with engine operation encountered were:
1.)
Problems with mixture adjustment, and engine running better with secondary
injectors turned off led to discovering that the staging relay was connected
incorrectly; secondaries were on at low MAP and vice versa.
2.)
Engine ran better on trailing plugs than on the leading plugs. After much double checking, head
scratching, and some further testing we finally concluded that the only
explanation could be that the timing was actually very much early. And the only way that could be true was
if the timing marks and the pointer did not relate to the position of the
rotors.
By
viewing the position of the apex seal through the two spark plug holes,
measuring angles, etc. we made a new TDC mark on the pulley, and a new 20 Degree
BTDC mark. Then reset the static
timing on the crank angle sensor, and fired up. Ran well, and disabling leading had a
bigger effect than disabling trailing, one would expect. To verify timing we put a pressure
transducer into the trailing plug hole, ran an oscilloscope trace triggered by
an inductor on the leading plug wire. One step back on the EC2 timing adjustment
gave us the trace we wanted. (Fortunately, operators know what the right trace
looks like based on their prior rotary development work). We then ran up to
The
original timing mark was about 14 degrees too early.
ENGINE
ADJUSTMENTS
The
preset staging point was found to be at about a 16.5” MAP, and the point for the
4 degree spark retard was very close to 22” MAP. After setting the timing to 20 BTDC at
5000 and MAP > 22”, it was found that slightly better power was achieved with
a couple of steps (1.375 degrees each) further advancing.
Setting
up the MAP mixture correction table on the dyno was a bit of a challenge. For a given MAP, the required mixture
correction was different for different RPMs. We then made up a typical propeller load
curve as a function of RPM, and tried to do the mapping by stepping MAP with RPM
and power values that fell on the curve.
Not easy, and didn’t really work because for given MAP there are two
places on the curve, one at higher RPM and lower load, and one at lower RPM and
higher load. Finally we kind of roughed it out, and then set at WOT for RPM of
about 5000 and up. The mapping
process will have to be repeated on the airplane with the actual prop load in
place.
The
other difficulty was that the mixture was much too rich at low MAP values. Setting the correction at steady state
conditions did not get the MAP values as low as occur when throttle is pulled
back from higher power; resulting in a very rich condition which cause the
engine to stall. Two factors
contribute to this; one is that the primary fuel injectors (which are the only
ones operating below 15” MAP) are high flow (570 cc/min), and the other is the
use of a fixed set fuel pressure regulator rather than one that is modulated by
manifold pressure.
It was
interesting to note that at one point 
we had the engine very seriously
flooded. It cleared relatively easily, probably due to the use of un-shrouded
plugs.
FLOW
MEASUREMENTS
After
re-calibration of the flow meters we got good data for oil, and coolant flow
with and without thermostat. No
data is available for the pressure drop in the dyno cooling loop, but my guess
is that it is lower than will be experienced in the airplane. Although the run is longer, the pipe
sizes are 1 ˝” and 2”, and the radiator is large.
The oil
cooling loop is likely more typical of pressure drops in the plane, with AN10
lines and an oil/water heat exchanger.
The oil pressure to the gallery was about 100 psig.
POWER AND
TORQUE
Wide
open throttle power and torque curves were run after adjusting the mixture
correction table and the timing.
Mixture setting was approximately for maximum power. Data was initially gathered up to about
6000 RPM, but looking at the data showed that the torque curve was still flat,
and hp increasing, so subsequent runs were made to 7000 RPM. Still no break in the hp curve. Given that the 2.17 redrive ratio limits
engine RPM to about 6300 (Prop @ 2900), there was no point in pushing things
further.
The
torque curve is distinguished by it’s flatness; varying only by 20 ft-lbs from
210 ft-lbs at 3000 rpm to a peak of 225 ft-lbs at 5000 - nearly constant from
4800 to 6800. The 225 ft-lbs is not
extraordinary, and could probably be improved by 9.7 compression rotors vs the
stock 9.0 rotors (for turbo applications) that are in the engine. Mazda’s data indicates about a 4%
increase in power with 9.7 vs 9.0 at 5000 rpm. It may also improve some as the engine
wears in - this is an engine with only a couple of hours of running since
overhaul.
It is
likely that a slight increase in torque in the 5000-6000 rpm range could be
achieved with longer ‘tuned’ induction runners. This would provide a bit more power in
the typical operating range; but without further real data comparison there is
no good way to estimate the effect. This manifold is 4 ˝” flange to flange,
engine to throttle body, and the throttle body barrels are about 5”, a design
that is well suited to my ‘packaging’ design.
For a
‘first shot’ custom design system, the performance is great. 285 hp from this package is impressive,
and a slightly higher reduction drive ratio, and a variable pitch prop would
take full advantage.
FUEL
CONSUMPTION
The
fuel consumption data was very limited because of a high degree of scatter
caused by poor flow meter performance.
The computer logged data for this parameter was essentially
unusable. Some good data was
measured by doing timed runs, i.e.; set up a steady state engine operation and
take weight readings of a fuel supply container at one minute intervals for five
or more minutes. The number of
these runs was limited to 4 as these were the last tests we were doing and ran
out of time. The results are as
follows:
|
RPM |
Mixture |
MAP (“Hg) |
Avg EGT
(F) |
HP |
BSFC |
mpg |
|
5,100
WOT |
Max
pow |
26.6 |
1,570 |
220 |
0.52 |
12.5 |
|
5,100 |
Leaned |
23.8 |
1,556 |
155 |
0.49 |
16.9 |
|
5,000 |
Max
pow |
24.8 |
1,544 |
200 |
0.54 |
12.9 |
|
4,550 |
Max
pow |
21.8 |
1,488 |
150 |
0.55 |
15.6 |
The max
power mixture setting was approximate, and was typically with a lambda of about
0.94, and maybe 100 - 150 degrees rich of peak EGT. The “leaned” run was done at about the
same amount lean of peak. The
leaned run may have no relevance in actual operation with a prop because rpm was
held as mixture was reduced by varying load while trying to maintain about 24”
MAP. Power dropped significantly, but estimated fuel economy is clearly
improved.
The
BSFC values seem good; pretty much in line with expectations for a rotary
engine. The miles/gal values were
computed using the Velocity factory values of airspeed as a reference and
adjusting it by the cube root of HP.
EXHAUST GAS
TEMPERATURE
Before any adjustment, the exhaust
gas temps were within 80 F of one another at power, with #1 rotor being the
hottest. Using the Mode 4
adjustment on the EC2, one step of adjustment brought rotors 1 and 2 to nearly
identical temps, with #3 being about 30F higher.
Max
power at 5000 rpm showed EGTs about 1500F, and about 1600F at 6000 and
255hp.
It was
interesting how they varied with mixture adjustment. As the mixture was leaned (all rotors)
they reversed with #3 being the lowest and #1 being the highest; and the spread
increased to about 100F. This
indicates that #3 rotor was running the leanest, and as the mixture was leaned
to the lean-of peak side of the curve, it’s EGT dropped the furthest. There were also some changes noted from one run to another as we did
different runs with the spread having increased at the same conditions; for
reasons undetermined.
EFFECTS OF THE
MUFFLERS
The
engine has a tangential manifold/muffler with the exit out the rear end (2
3/4”). Just a few inches downstream
is a secondary custom muffler that has an inner perforated tube (1/4” holes)
surrounded by an outer tube. The
inner tube has a slight restricting orifice in the center to encourage some of
the flow to go through the holes.
The system was connected to the dyno facility exhaust system which also
contains a large muffler.
Power
runs were made with the dyno muffler disconnected, and with both the secondary
muffler and dyno muffler disconnected.
There was essentially no change in the power output from the baseline
case.
Sound
level measurements about 8 ft away at about a 45 degree angle from the exhaust
exit were taken with and without the secondary muffler. The readings at full power were 126 dB
without the muffler, and 120 with the muffler. Both are high, but the good news is that
the secondary muffler provides a 6 dB reduction without affecting the
power. NO doubt about it; it was a
powerful sound. Low power/RPM sound
was reasonably gentle; probably similar to a Lycoming.
MANIFOLD
PRESSURE
One of
the reasons that the hp continues upward at the high RPM is the ability of the
engine to “breath” reasonably easily.
Even in this case, the measured absolute pressure at the exit from the
throttle body (the MAP) is seen to drop off significantly at high RPM. More power could likely be achieved at
the higher RPMs with larger throttle body bore. The three barrel (one per rotor), 44mm
(1.73”) diameter bore was selected as adequate for up to 6000 or so, and still
provide good control at the low end.
Some of
the drop noted is likely to be occurring in the air box; the manifolding of the
air into the throttle body. The
flow area in the air box is always larger than the ports downstream, but is
restricted by my cowling design from being larger.