wrote in message
...
On Aug 27, 3:32 pm, "Peter Dohm" wrote:
"Jim Logajan" wrote in message

.. .

Ron Wanttaja wrote:
On Tue, 26 Aug 2008 21:07:46 -0700 (PDT), "
wrote:

everyone says "ooh -- auto...dangerous" but no
one can explain exactly why.

1. Ignition systems with insufficient redundancy.
2. PSRU failures.
3. Difficulty in implementing efficient liquid cooling systems.

But doesn't the Rotax 912 have reduction gearing and liquid cooling? It
is
getting put into an awful lot of aircraft models - particularly LSAs.

That's true, and the biggest annoyance (of which I am aware) is that they
have increased the recommended "idle" speed to increase the service life
of
the PSRU--which is of the spur gear type. I don't know whether any of
the
belt or chain type PSRU installations have a similar requirement.

As to cooling: there were a lot of liquid cooled aircraft engines in
WWII,
but the the aircraft they in which they were installed looked a lot
different from their air cooled counterparts.

Peter

Even belted PSRUs have vibration nodes. The Glastar in which
we put a PSRU'd Soob didn't like 1400 engine RPM; it semed to be an
argument between the flywheel's inertia and the prop's. Running it at
that RPM for long would have torn the teeth off the belt. I didn't
notice if there were further nodes at 2800 and 5600. Adjusting belt
tension didn't change anything.
I've read about (and encountered) cases of cooling problems in
auto conversions. Many builders underestimate the amount of heat that
needs discarding, and also make mistakes in radiator installation and
baffling. I've seen rads mounted out in the breeze where they not only
slow the airplane but suffer from airflow problems created by the
vortices generated around the rad. I've seen a couple of small rads
mounted behine the front cowl openings, where they're supposed to get
ram air, but without proper baffling to separate the incoming air from
the air behind the rads the pressure differential is minimal, causing
low flow, and air eddying around the rad further interferes with
flow.
In the Glastar I mounted the big, full-size rad (from the same
car as the engine) behind the engine, at an angle so that the top edge
was at the firewall and the bottom was forward about 8". Baffling
around the rad made sure that ALL air leaving the cowl (except for a
bit leaving around the hot exhaust pipes) had to go through the rad,
so I had maximum flow. A lip on the cowl outlet to accelerate air away
from the opening lowered the pressure further so that max differential
was maintained between the front and rear of the rad. And even with
all this the engine's coolant temp reached max in an extended full-
power climb on a summer day.
The P-51 had an underbelly scoop and a variable-geometry outlet
behind it. The rad was in this housing. Inlet and outlet shape and
size were critical, and I've heard that the designers were so clever
that they even got a little thrust as the cooling air expanded and was
accelerated a little when it left the outlet. OWT, maybe, but there's
lots to learn from their design anyway. It's worth noting that the
inlet was much smaller than the rad's area; Mr. Bernoulli tells us
that pressure increases as airflow slows and decreases as it
accelerates, so the divergent duct between the inlet and rad face
slowed the air and increased its pressure. Same principle used in
numerous places in a jet engine.

Dan
I was only thinking of the exact ratios that place the same teeth in use on
each successive rotation of the belt.

Torsional resonance can be extremely difficult to monitor andI am glad that
you were able to identify it before it became a dissaster. For the moment,
my own project and the decision to build around a PSRU or use a direct drive
aircraft engine has been pushed further into the future. But I have
wondered whether the elimination of critical speeds might be the true
purpose of those little springs in the driven plate of a manual clutch.