Paul,

 Thanks for your book it's an interesting read.
 I have put together a quick layout of how I would like to do my belly
 mounted radiator.  The attached pictures I hope explain what I am
trying to
 accomplish.
 You mention a separation of about 2" from the belly for boundary layer
 separation, does this have to extent the entire length of the
radiator or is
 the way I show it ok?
 I didn't see anything that gave me a frontal opening size
relationship to
 the area of the radiator, is there a percentage rule of thumb for this?
 Any other suggestions on what I'm showing to improve would be
appreciated.


 Randy

 Excellent illustrations Randy. Reduces the frontal area over the
 splitter plate idea and is far more cosmetic. A Canard P51.

 Is that Solid Works or Rhino 3D?

 You will still need the Kays & London wedge inlet diffuser page 36.
 Also you will need to take the bottom panel off as it will
 retard the hot air flow out from the front back side of the
 rad page 46. If you just lower the panel it will add frontal
 area. Not good. Just do away with it and let the hot air
 dump into that air stream.

 I would use Venetian blind cowl flaps page 51.

 I found the dims on the Neon rads I have. Here they are.

 The rule of thumb is 20% of the frontal area of the rad.
 With the P51 style scoop you are committed :) I dislike
 commitments all my old girlfriends used to tell me :)

 Paul Lamar




 Randy,


 I see you have shown two radiators.  Are you going to use
 both of them somehow?

 Doug in Japan.

 Yes. I was going to build a cross counter flow  rad (book  page 25 )
 for testing but that is now far down on the priority list as  Perry Mick
 has already tested it and it works well. If you are pressed for
 space it is worth it.

 Paul Lamar

 --

 Paul


 For years I was excited about that cross counter flow design you
introduced and actually did drawing to incorporate it under my engine.
That is until I read a comment last year on Stewart Component websites
claiming the such setup increase the back pressure  and therefore flow
significantly.

 http://www.stewartcomponents.com/tech_tips/Tech_Tips_5.htm

 I don't believe the longer effective water passages it will increase
pressure 16 times but it will be significantly more than double.  This
got me thinking about some to the water flow challenges I have had in
my line of work.  In essence, when the  water channels lengthen or
restrictions are increased,  the effect of back pressure on flow will
rise in a  non linear fashion.

 In a water heat exchanger, high velocity enhances turbulence, thus
wall contact time.  Velocity is upper limited by restriction and
internal pump fluid slippage.   Engine pumps are centrifugal  devices
hence produce highest flow when the system has the  least
restriction.  Said another way, maintain as high as water flow as
possible through the engine.  To get  300hp out of the 13B engine I'll
need,  according to Jim at Racing Beat, the man who's been there, at
least  40 gals a minute water flow. Choke the flow anywhere in the
system and pumping rate falls off as water pressure goes up. Knowingly
increasing the water pressure may not be a good thing for thin wall
items tending to corrode over time.

 For comparison purposes oil heat exchangers have positive
displacement pumps, have to squirt  a viscous liquid through
inherently small openings  and thus have a slower flowing, robust heat
exchanger in need of built-in turbulators and must be able to
withstand high pressures..  Many oil heat exchangers are designed to
be cross flow because the characteristics of wall contact time are
enhanced by the slower flow. Oil  by nature as you know is less
thermal conductive than water.

 This discussion basically boils down to whether it is more prudent to
have a single thick core thick radiator or two thinner ones in a
reverse crossed mode RCM like you have shown.   Actually I have found
that when the air gap between the two radiators in the reverse cross
flow is taken into account, the total installed package is thicker
than a single thicker unit.

 Paul your thoughts that  proper air flow through heat exchanger is
paramount and on that I wholeheartedly agree.  I'm not so sure the old
rule of the first 1/4 of the radiator does 50% of the cooling is
correct anymore however.  That would depend on fin count air
resistance, tube design, number of  rows and how well the unit is
maintained.  A clogged heat exchanger is as good as the dirt it
catches.    Dave Garber showed me a very compact  design unit he had
made that used two rows of  the newer  overside elongated tubes.
These are about 1.5 in" in length but  essentially the same standard
width. The Dave Garber design installed package thickness came out to
about 2.5 inches.  compared to close to 4 inch thick for a similar RCM
design.  That is assuming a one inch air gap between the heat
exchanges  helps isolate infrared heat transmission.

 In Dave's design, air flow had to traverse about three inches to get
through it though.  The tubes were angled as was the heat exchanger
so air flow was actually parallel to the fins and that might be a
deciding factor in the high performance he experienced.   So high that
he had to cover up the first few inches of it as you can see by the
photo. His  heat exchanger wetted area is about 12 inches tall by 28
inches long.  .  Thats my 8cm x 11cm (3.1/8" x  4 3/8") wallet in the
foreground for scale purposes.

 I will try an identical system but lay it on its side under the
engine and make it 16" wide (max 20" with the side tanks included) by
19" or 20"  in depth.  I 'm not sure how to angle or curved the side
tanks to help insure even flow distribution though.  Logic dictates at
center inlet and outlets with the tanks sides curving  out toward the
ends. Understanding how fluids  flow is hard to predict other than it
takes the path of least resistance.

 Doug in Japan

 Well apparently Stewart did not read Kays & London's bible Compact
Heat Exchangers  :)

 ==========================================================================

  "The design of a heat exchanger involves a consideration of both the
heat transfer rates between the fluids and the mechanical pumping
power expended to overcome fluid friction and move the fluids through
the heat exchanger. For a heat exchanger operating with high-density
fluids, the friction-power expenditure is generally small relative to
the heat transfer rate, with the result that the friction-power
expenditure is seldom of controlling influence."

 [What they are saying here is the HP consumed by the water pump is of
minor importance. In fact the Mazda pump only consumes about 3 HP.]

 "However, for low-density fluids, such as gases, it is very easy to
expend as much mechanical energy in overcoming friction power as is
transferred as heat."

 [What is being said here is HP is consumed by cooling drag. Since
cooling drag can be as high as 30% of over all drag the HP could be as
much as 30% of 200 HP or 60 HP. As you can see cooling drag becomes a
very important issue.]

 "And it should be remembered that in most thermal power systems
mechanical energy is worth 4 to 10 times as much as its equivalent in
heat."

 [Here they are referring to the heat balance of most internal
combustion engines that only derive about 25% to 10% of the heat
energy in a gallon of fuel as mechanical HP.]

 ==========================================================================


 Builders have done some really dumb things like use electric water pumps
 with 1/10 the HP and got away with it. The early Mazda pumps, many
people
 still use, have a restricter in them. Mazda woke up with the RX8 pump.

 You need my book or Kays & London's book Doug :)

 Paul Lamar

 Paul,

 I have both and have read both.  Your statements are correct but
pumping hp consumed is not the issue that concerns me. The resistance
of liquid is, whether it's water or air flows.  Dave's design, with
its orientation, may have some merit for the building community.It
places  the effective heat exchanger components parallel to the air
flow without having to place the whole radiator perpendicular to the
air flow


 Doug in Japan

 Dave told me it would not cool. Tilting the coolant tubes at a 45 degree
 angle sounds good but it did not help as the back side of his rads
were blocked.
 I saw no way for the air to get out.

 Besides that he used NACA ducts which are no good for heat exchangers.
 The inventors said so in the original NACA paper which I have.
 Too much pressure build up ahead of the core and in the entrance
 will cause the air to go around and not in. Probably the most
 significant problem.

 I suggested two P51 style scoops and he rejected that idea.
 Probably on the basis of cosmetics but maybe because of
 the frontal area increase.

 Paul Lamar


Paul,

He also told me it initially wouldn't cool until he modified it.  I'll
have to admit I didn't realize he had the oil heat exchangers on the
inside of the water heat exchangers as well.  That would be to much
restriction to flow.  Other points on lack of planned air flow, and
heat exchanger orientation I totally agree.

Doug in Japan

Angled tubes would be nice but very expensive.
One thing I notice about cooling system design is
any change in one thing has unintended consequences
in other things both good and bad. That is why
you see so many variations that all seem to work.
On the other hand it is also why it is very difficult
to optimize.

For an example if you force the air to make a right
angle turn to go through the exchanger in induces
a positive effect in that turbulence is increased
between the tubes. That enhances the heat transfer
from the metal to the air.

Another example. Intuitively you might want
to use turning vanes to get the air going in the
right direction. What happens is a boundary layer
is built upon the turning vanes which actually reducing
the amount of air getting to the exchanger. Page 48.

I have a paper on a comprehensive  study that Honda did on
automotive cooling systems. The percentage of cooling
drag  in a car is far lower than it is in an airplane
so that changes things as well.

Cooling systems are very complicated.

This is why it took Kay's and London 30 years
to write the book :)

Another good book with some cooling stuff in
it is Hoerner Fluid Dynamic Drag. Klaus used
to date his daughter and he signed my copy
for me back in June 1966.

Paul Lamar

Hey Doug in Japan,

The Stewart Components web site is speaking of adding baffling inside an
existing radiator design to create the double or triple flow.  For
example, if you have an 18 tube radiator and you add a baffle in the
center of the inlet end that divides the flow into a double flow then
the fluid will have to flow twice as fast and twice as far through nine
tubes.  If a second baffle is installed in the return end 2/3 of the way
across the tank and the baffle in the supply end is moved to 1/3 of the
way across the tank such that the flow is forced to flow through six
tubes three times the distance you will have a triple flow radiator.

Increasing the velocity requires increasing the pressure head.  By
reducing the area and increasing the velocity the force required
increases by multiple factors.  Half the area, double the velocity.
Double the velocity, more than double the flow resistance.  Sort of like
2x2 = 4 times the force.  1/3 the area, three times the velocity.  Sort
of like 3x3 = 9 times the force.

The RCM design that you are talking about does none of this.  The flow
distance is doubled.  The flow area is unchanged because you are using
two radiators in series.  Because the flow velocity does not increase
the flow resistance only increases slightly due to the doubled flow
distance.  The increase in resistance is much less than a double flow
radiator.  Sort of like 1x2 = 2.

The increase in efficiency by taking advantage of counter flow will far
exceed the increase in cooling efficiency by doubling the thickness of
the core.  This is also true of double and triple flow radiators because
they do not take advantage of counterflow.

The cooling efficiency of any heat exchanger device depends on the
temperature differential between the two fluids.  That difference in
temperature is called delta T.  With an increased core depth the air is
heated as it flows through the radiator.  The delta T at the back of the
radiator core is less than the delta T at the front causing the front
tubes to be cooled much more than the back tubes.  The same is true with
a double or triple flow radiator because there is no way to cause the
air to flow counter to the flow direction of the fluid inside the
radiator.  The first pass of the muti-flow radiator is much more
efficient than the subsequent passes because the delta T is decreased
with each additional pass.  The water is cooler with each pass but the
air temperature supplied to the different sections of the radiator is
the same.

With counter flow the delta T is maximised across the entire radiator
assembly.  The coolest part of the air flow is exposed to the coolest
part of the water flow.  The same air that just cooled the cooler part
of the water flow will now cool the hottest part of the water flow.
There will be a slight decrease in delta T between the air that cooled
the cooler water in the front of the radiator and the hotter water in
the back of the radiator in the counter flow radiator but that is much
less of a loss than that experienced by the tubes in the back of a
thicker core radiator that are now filled with hot water and are being
cooled by air that was just used to cool water just as hot in front of them.

Mark LaPierre

Good explanation Mark.

Paul Lamar