I finally got into the gear box today and found another issue.  In addition
to the sun gear pin failure inside the gear box, the flex plate came apart.
I found a piece of it in my cowl as a litte hint of what I was about to
find..

Here are some pictures.  The plate was pretty badly broken up and was only
being held together by Tracys aluminum damper plate and the steel bolt ring.
Pin in sun gear was also broken.  No apparent serious damage, just what I
assume is cosmetic galling of the input shaft, adapter plate, and bell
housing from the loose pin floating around.

Thinking back, I had an early sign of it at the Reno Pylon School.  After
one of the races I noticed that my home made cover over the flywheel had
been marred by the flywheel but was not in contact with it.  I attributed it
to the high air speeds (205 Kts Indicated coming down "The Chute") somehow
pushing my cover down onto the flywheel.  I put some creases into the cover
to make it stiffer which seemed to solve the problem, but perhaps the flex
plate had broken at that time.  With a mis-tracking flex plate it is easy to
see that sun gear pins would be undergoing additional stress.  So I am
calling the underlying cause of both issues the flex plate failure.

The large washers that covers the rubber dampers between the dampers and the
flex plate have made a significant groove into the aluminum of the
attachment plate.  I have never noticed that before, possibly was a function
of the twisting flex plate.  Question is, does it need to be replaced?

My rubber dampers are noticeably worn and soft and need to be replaced.  I
am getting quite a significant amount of oil coming past either my rear oil
seal or the oil seal on the gear drive.  That oil may have done in the
rubber dampers, that got too sloppy and stressed the flex plate that
eventually fatigued and broke, which further lead to the pin failure.

So I am undecided on what to do about the flex plate replacement.  I agree
that the iorn flywheel is not the answer.  The machine work would have to be
quite extensive, because the aluminum plate rides inside the ring.  After
all that machining, it would have to be re-balanced but that would be
difficult because it has the rear counter weight built in.

Using one of the aftermarket aluminum flywheels might work.  It would still
have to be machined to make space for the damper plate, but at least it
would be aluminum (easy to machine) and could then be balanced symmetrically
(perhaps even with the damper plate installed so it is all nicely balanced.
That would not be a low cost solution, however!

Paul, if I were to get an aluminum flywheel like this one:
http://www.racingbeat.com/RX7-1986-1992/Flywheels/11465.html
do you think that you could put it on your milling machine and mill in the
resesses for tracys damper plate?  Then would you have any suggestion on
where I could get it balanced?

I am also wondering if I have any business trying to make that much HP.  I
was about to install a water/methanol injection system to be able to use
more boost at Reno.  But it seems I may not be able to safely use the HP
that I already have...

-
--
David Leonard

Turbo Rotary RV-6 N4VY
http://N4VY.RotaryRoster.net
http://RotaryRoster.net

Wow Dave that is a game changer.

I have a guy up here in SZP with a really big lathe that can make
a new "flex plate" . As I recall the subject of flex plate run
out came up recently. Perhaps your plate was warped. Everybody
should check their plates for run out. Any run out puts a bending
stress on the plate due to dynamic imbalance. All parts of the
plate want to revolve in the same plane. If the plate does
not track it will introduce a bending stress in the material.

Also the alignment of the input shaft to the box must be
perfectly in line with the e-shaft axis. This is not easy to
check. Perhaps a low cost laser can be inserted into an empty
gear box starting at the prop bearing end and see where the light
spot falls on the counter weight. A holder for the laser can be made
and checked in a lathe for an aligned axis. If you UPS your empty box,
spacers and adapter plate  up I'll put Jim Poole on the job.

One quick and dirty way is to assemble Tracy's damper on the
input shaft with out bolting it to the flex plate. In fact
leave the flex plate out and off for the moment. Use the engine block
as an anchor for a dial indicator and check for damper plate run out.

Then measure the distance from the block to all four quadrants
of the damper plate.

Paul Lamar


It's been a while but I am still trying to find a reason for the
failure. At this point, I tend to think of gyroscopic forces at work
here.

We know that the setup is subject to high g turns, we also know that
the engine is spinning at high rpm. We also know that the flex plate
can, "flex", we also know that there are scratch marks prior to
failure.

My reading points to a fast spinning flywheel acting like a gyro. That
means the wheel resists to turns. When David makes high G turns, the
wheel flexes during the turn according to precession force, and
therefore the plate got flexed one time for each revolution of the
fly wheel.

The magnitude of the precession force changes according to the G force
of the turn, (or torgue applied to the flywheel), causing the flywheel
to spin at a different plane. (Thus toughing the housing and scratches
the housing)

Therefore, David's failure may be resolved by a smaller, lighter
flywheel, a stronger flex plate, or combination of the two.

Let me know what you think.

Henry Nee

I think you are on to something Henry. Dave is fond of doing  aerobatics
from time to time. We are constrained somewhat by the diameter of the
ring gear. If this is the primary cause the holes in the stock flex plate
near the hub are stress concentrations and should be done away with. The
RX8 flex plate does not have them. Unfortunately it does not fit Tracy;s
damper. Tracy's damper is also rather large in diameter and that aggravates the
problem. In other words a smaller diameter damper would be a big improvement.
The mass on the flex plate out near the ring gear could be reduced by cutting
a bunch of large holes. That too would help reduce the gyroscopic forces.

I think a further reduction in diameter of the damper in this design is called
for. Perhaps I can use 1/2 by 1 4130 steel tubing instead of 3/4 by 1.5.
I'll make a few stress calculations. Smaller rubber blocks can be used and
the overall weight reduced.

Paul Lamar


Paul

I am not a mechanical engineer so I don't know what is necessary in
the flex plate. But I am worried that RX-8 flex plate may not be
strong enough, even though it look much better. This is because RX-8
don't make turns over 1G, but many home build aircrafts can regularly
turn with multi-Gs.

As for Tracy's damper, If you don't mind, I want to know more about
its usefulness. So correct me if I am wrong. My reading suggests that
torsional dampers are necessary to dampen the torque pulses on the
crank shaft. Harmonic balancer is actually a kind of torsional
vibration damper. Torsional Vibration requires a torsional damper to
dampen it, and long drive shaft, due to the torsional flex inherent in
the long shaft, makes it a much bigger problem. I would think that a
torsional damper is actually a torsional shock absorber.

I suspect that torsional vibration is a much bigger problem for piston
based engine, as torque reversal happens on the compression stroke.
The slowing piston acting on the crank shaft through the connection
rod created an amplified reverse torque force on the crank shaft,
resulting in torsional vibration.

The rotary does not have connection rod connected to a typical crank
shaft. The rotor rotates and drives the shaft directly. This may be
the reason why it has lower torque at low rpm comparing to a regular 4
stroke, piston engine. Therefore I am guessing that the damping
requirement on a rotary is much less than that of a similar piston
engine.

Tracy's damper is really not a torsional vibration damper, but more of
a impact reduction device when engaging the clutch. Since we do not
have a clutch, may be we don't need it as much as we think? Do you
know if anyone had experimented with a direct connection on a rotary
without the damper?

Also, do you know anything about the stock Mazda torsional damper
design and placement? Does a torsional damper exist between the engine
and the transmission? Or it is simply an engagement damping device? If
it is simply an engagement damping device, then I would speculate that
we don't need such a damper as we have no clutch (props can turn
freely when we crank the engine, wheels can't turn without moving the
vehicle at the same time when cranking, thus a clutch and all related
parts is necessary in a car, but may not be necessary in an aircraft.)

I hope someone can help to proof or disproof my thoughts. If we can
remove an unnecessary damper, it is some weight saved.

Henry Nee

Tracy's damper should be called a rubber coupler. I smooths the torque pulses
from the engine applied  to the gears. E-shaft torsional resonance is not a
problem in the 2 rotor rotary. Gone. Does not exist.
The e-shaft is extremely stiff in torsion.


Paul Lamar


Paul,

From looking at your diagram of the rotary torque curve vs degree of
crankshaft rotation, if one could design a damper that was able to
absorb between 250 and 600 foot pounds of torque during the torque
peaks, then release it fairly smoothly and completely just as the next
torque peak reaches about 250 ft-lbs, with maximum recovery of power
and minimum loss to heat, AND also be able to adapt the rate at which
it did this, depending on the engine rpm, then wouldn't this be the
theoretically smoothest damper?

If so, I am picturing rubber bushings to absorb the power, and valved
dashpot cylinders [mounted in parallel with the rubber bushings]  that
control the release of the stored energy, with the rpm dependent timing
controlled by the dashpot valve, which is able to change the size of the
orifice according to the rpm.  This dashpot valve could be a modified
one way 'check valve' that can compress as quickly as needed, but
release more slowly.  I am also reminded of the old mechanical centri-
fugal ignition timing advance mechanism not that my dashpot valve
would necessarily resemble it, but simply that simple mechanical
devices can be designed to successfully play such roles.

The advantage of this approach over a silicone-filled viscous damper
would be less conversion of power to heat, and recover more as
usable power although compressing the rubber bushings would
necessarily convert some power to heat.

Is the reasoning behind this approach to damper design fundamentally
sound?

Gene Kahn

It sounds workable but costly to implement. It also sounds like a
damper and spring arrangement.  Rubber already has a considerable
amount of damping or hysterias.  The typical sprung clutch plate
found on all manual transmissions was used for years in Ross gear boxes
on rotaries but proved to be prone to catastrophic failure.
In theory it looked ideal by storing energy and giving it back
with little increase in temperature.

Tracy's rubber bushing version seems to hold
up better. I just think it can be improved while cutting it's polar
moment of inertia which will reduce the gyro reaction stress on
the flex plate.

Paul Lamar