[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

REFLECTOR: Liquid cooling auto conversions

Further comments on engine cooling.  These are just opinions, impressions,
and a few conclusions; informed by some compressible flow analysis in my
distant past, and observations of engine conversion projects of others.

Yes, the radiator has to be properly sized.  But I think that's the easy
part, because no matter what you do with radiator dimensions, if you don't
get the air flow through it, it's not going to work.  We all can get a
fairly good grasp of incompressible fluids, like water flowing through a
pipe, and a pump to push it; but compressible flow of air from free stream
dynamic pressure head is a whole different ball game.

Sizing the radiator goes hand in hand with ducting configuration.  A thicker
radiator core (longer air path) has more resistance to flow (more pressure
drop) and is only better if you slow the air flow sufficiently to go through
with low delta P.  Then, of course, there is the variable of fin density;
which affects cooling and pressure drop.

The pressure we have to work with is limited to the dynamic head.  And if
air isn't treated right in the ducting it will form back eddies and pressure
waves, and find lots of ways to give you less flow than you calculate from
your intake area.  Core thickness can be traded for x-sectional area only if
the ducting is designed to get the air slowed and through it.

Probably the two critical regimes are idling on the ground and full power
climb.  If these work it should work for cruise at any altitude your likely
to go.  Presumably, same indicated air speed, same cooling.  On the ground,
forget about prop 'suction'.  There isn't any.  Even with the engine speeded
up you'll find that if you are more than half a prop diameter away from the
prop the negative pressure is very small, and probably still about zero
toward the hub.  So plan on having an auxilliary fan or use exhaust
augmentation pumping.  If you're set up for natural convection (cool air
going in at the bottom, hot air out the top) that will help, but likely not
handle it.

Pay attention to duct inlet geometry; smoothly transition the air to lower
velocity (maybe 6 to 12 times the inlet area, depending on your core
thickness and if you expect the air to go around a corner).  Sudden
expansions or contractions cause pressure drops.  Transition the exit to an
area about 1.6 to 1.8 times the inlet area for minimum cooling air drag.

Assuming engine efficiencies are all roughly the same; Turbo Tom Wyatt's
survey shows the need for roughly 2 cu in. of radiator matrix volume per hp.
Typically; unless the ducting is well designed for pressure recovery
(conversion of dynamic head to static pressure) core thicknesses greater
than 2 - 2.5" for typical radiators may be risky in terms of pressure drop.
Given these rough guidelines, than a 220 hp engine needs somewhere around
180 to 220 sq. in. of radiator face area.  Add a little margin of safety;
mabe 220 to 250 sq in.  If you can get the air to flow through a 3.5" thick
core, maybe around 175 sq. in.

For a rotary engine, the radiator maybe can be a little smaller, but the oil
cooler has to be correspondingly bigger.

Well, this isn't meant to be definitive and probably raises as many
questions as are answered.  Some where in here my main point is that I think
most of the cooling problems with liquid cooled alternative engines has more
to do with air flow than radiator sizing.

Those of you doing conversions probably already know all this.  How do your
data points fit into the above guidelines?

Comments, disgreements and other info welcome.

Al Gietzen  RGE  755V   (planned Rotary, either 3 rotor normally aspirated,
or turbo 2 rotor).  I'll know what engine I'm using when I bolt it on.

>A Fluidyne engineer (Ontario, CA) designed my system.  He made two aluminum
>racing radiators designed for each of my wing roots.  I used Alan Shaw's
>(Wingco) design for wing root installation.
>Fluidyne insisted on two things:  First: Make the cores big and wide.  I
>have three cores (each almost one inch each) running horizontally for a
>total cooling area of 3 inches across.  Second: Capitalize on the cubes.
>Wayne's right about cubic inches.
>Basic engine/cooling specs for my 173 Elite:
>All aluminum Chevy V-6 (350 HP estimated)
>4.5 liter (Not same as 4.3 Vortec)
>Dry sump oil system
>Fuel injected
>Length: 18" from tank to tank
>Height: 6"
>Width: 3 1/2"
>Fin Area: 5 3/4" high X 14" long X 3" wide
>Cost: $300 for each radiator
>Fluidyne's phone:  888 FLUIDYNE
>Tim England, the Canadian V-6 Velocity builder told me that my system may
>be excessive.  My reply:  I want excess since most all - I take that back -
>all V-6 and V-8 auto conversions I've seen or read about have cooling
>problems.  Although well designed, Tim's plane boiled over during taxi at
>Oshkosh last summer.  I want excess cooling for excessively hot summer days
>and long taxi situations.  I also put a full size air conditioning core in
>the nose with 3/4" braided, teflon hoses with AN fittings.  This will be my
>heater.  It also provides more redundancy since a full-size fan will cool
>it during taxi and dump the hot air out in the summer.
>Good luck with your Mazda, and it sounds like you're doing some very good
>preliminary research.
>Dennis Martin