There are at least four critical issues with the belt type PSRUs.  I'll
tackle them in the
order that  I have seen the most problems with the myriad of units out

there.

First and foremost:   Many belt type PSRUs ignore the very high (~1000

lbs)

side loads
that the belt (ANY belt) imposes on both of the shafts.  Each shaft

needs a

pair of bearings
to pick up these side loads.  A key design requirement of any PSRU is

that

it must keep the
loads in the engine the same as the car application, i.e.  ONLY torque
coming out
of the crank - side and thrust loads must be very small.  This implies a
crankshaft pulley
that has two bearings (one on each side of the pulley) and a
misalignement-capable
coupling.  There are various commercial couplings available, such as the
double row
roller chain wrapped around two closely located sprockets, that permit

SMALL

angular
and lateral misalignment.

Second - Rolling element bearings have absolutely finite life and fail
dramatically.   They
MUST be properly designed for the load-speed combination and then

replaced at

regular intervals (hours of use, not condition) and inspected at perhaps
1/4th of expected
life intervals.  So, if the design manual says B10 life is 3000 hours, I
would want to look
at the rolling elements and race every 750 hours or less.  BUT - since

we

all make errors,
I would want to look at the first one ever made every 250 hours.

Sealed

bearings
eliminate the requirement for separate seals and a lot of effort to make
them work,but
preclude inspection of the bearing.   I don't like this. An open bearing
with removable
seals can be opened up, degreased, inspected, regreased and resealed.

More work

and design effort, but much safer.

Third -  The output shaft and flange should look just like the end of

the

crank on
a similar HP Lyc or Continental.  Many look like a 1/2 scale replica.

Since

the
life of these is proportional to the DIAM to the FOURTH POWER !!!!!   A

little

too small is WAY too weak.   This is critical as the small shaft will
fatigue and break.
Many under sized PSRUs survive because they only fly 10-20 hrs per year

and are

crashed in a couple of years.  Not enough time to fail, YET.  Also, most
aircraft
conversion make nowhere near the claimed HP, which also saves a bunch of

butts.

Fourth - The belt system must be properly engineered TO THE

MANUFACTURE'S

RECOMMENDED HORSEPOWER AND TORQUE AND SPEED design factors.
This is often ignored.  They will work for a while, but are probably
overheating and
ovewrstressing the rubber and fabric.   Bad news.

This is not a condemnation or recomendation of any specific belt-type

PSRU.

This is
just how this particular cowardly mech eng would evaluate one if my

soft, pink,

irreplacable butt is going to ride around in the air behind it.

Bill Freeman
Long EZ builder & pilot
BSME, MSME

Northwest Aero products and Belted Air Power both make Belt style reduction units that appear to be substantial
in design. One makes a rotary specific unit. Find them with the Google search engine. Lynn E. Hanover

The AirCraft Rotary Engine NewsLetter.              Powered by Linux.
ACRE NL web site.
http://home.earthlink.net/~rotaryeng/            http://www.linux.org