Paul

I am about to order a custom dual pass radiator when I read this page
included from Stewart Components.
Everything made sense about newer designs having more flow and higher
turbulence etc. until I got to the bottom of the page where it made the
claim that dual pass radiators like you suggest  reduce the water pump
flow about 33% and double the water pressure. These figures seem a
little high to me and I would like to invite your comments.

For my 300hp engine  I have room to install one 3" thick radiator,  with
dimensions 12x26 inches.  Total 936 cubic inches.  The alternative is a
U-shape dual pass unit with wetted surface 12X24 1.5 thick with 1.5 air
gap in between.  864 cu. inches.

According to your book 'How to Cool a Rotary', the dual pass radiator
would be more efficient in heat transfer both because of: 1. The  higher
delta Ts experience in each heat exchanger and 2. the first 1/4 inch
does 90% of the cooling rule.

You never mentioned the fact about reduction in water flow or increase in
water pressure with a dual pass, which might lower the heat transfer
efficiency if turbulence is lost.

In a single thick 3" radiator if there are (old style) more than one row
of tubes then of course the ones in the rear are  going to receive
hotter air and thus have a lower delta T. This is an obvious
disadvantage.  Below it implies new radiator designs have tubes that are
very thin but wide.  Three inches wide?  As long as there is adequate
turbulence within them water will be forced to contact the outer walls.
That is the theory at least and I don't know what the maximum width a
tube can be.

On the other hand a dual pass radiator will have one half the cross
sectional area of the water passages but double their length.

Which is more efficient?

Doug in Japan

    " Tech Tip #5 - Radiators & External Plumbing

Radiators
Thicker radiators do have slightly more airflow resistance than thinner
radiators but the difference is minimal. A 4" radiator has only
approximately 10% more airflow resistance than a 2" radiator.

In past years, hot rodders and racers would sometimes install a thicker
radiator and actually notice decreased cooling. They erroneously came to
the conclusion that the air could not flow adequately through the thick
radiator, and therefore became fully heat-saturated before exiting the
rear of the radiator core. The actual explanation for the decreased
cooling was not the air flow, but the coolant flow. The older radiators
used the narrow tube design with larger cross section. Coolant must flow
through a radiator tube at a velocity adequate to create turbulence.

The turbulence allows the water in the center of the tube to be forced
against the outside of the tube, which allows for better thermal
transfer between the coolant and the tube surface. The coolant velocity
actually decreases, and subsequently its ability to create the required
turbulence, in direct relation to the increase in thickness. If the
thickness of the core is doubled, the coolant velocity is halved. Modern
radiators, using wide tubes and less cross section area, require less
velocity to achieve optimum thermal transfer. The older radiators
benefited from baffling inside the tanks and forcing the coolant through
a serpentine configuration. This increased velocity and thus the
required turbulence was restored.

Radiators with a higher number of fins will cool better than a comparable
radiator with less fins, assuming it is clean. However, a higher fin
count is very difficult to keep clean. Determining the best compromise
depends on the actual conditions of operation.

Double pass radiators require 16x more pressure to flow the same volume
of coolant through them, as compared to a single pass radiator. Triple
pass radiators require 64x more pressure to maintain the same volume.
Automotive water pumps are a centrifugal design, not positive
displacement, so with a double pass radiator, the pressure is doubled
and flow is reduced by approximately 33%. Modern radiator designs, using
wide/thin cross sections tubes, seldom benefit from multiple pass
configurations. The decrease in flow caused by multiple passes offsets
any benefits of a high-flow water pump.

Cross flow radiators are superior to upright radiators because the
radiator cap is positioned on the low pressure (suction) side of the
system. This prevents the pressure created by a high-flow water pump
from forcing coolant past the radiator cap at high RPM. As mentioned in
the radiator cap section, an upright radiator should be equipped with
radiator cap with the highest pressure rating recommended by the
manufacturer. The system will still force coolant past the cap at
sustained high RPM. "

--------------------------------------------------------------

We have to be clear about what we are talking about here.
Cross counter flow or dual pass.
Cross counter flow is the first figure. Cross counter flow
is more efficient than dual pass or single pass due to the first
half doing 3/4 of the cooling.

Dual pass is the second and 3rd pictures. I recommend the dual pass
only because it makes the plumbing simpler for a typical
aircraft installation. Both inlet and outlet are on
the back end of the rad. It does however increase the water
velocity at the expense of consuming more HP from the pump.

Yes to be sure the cross counter flow is also a dual pass.

Much of what they say on the Stewart Warner web site is correct.
However I don't buy the 16X increase in water pressure required. If that
were true dual pass rads would not work at all. VW would never
have used one in the Scirocco. I also doubt the 10% more
airflow resistance than a rad half as thick. Easy to test
however. Stick both out the side door of a van mounted on a lever
and measure the forces.

As far as I know there are no rads available with 3 inch wide tubes.
Usually the tubes are about one inch wide and they come in rows. One row
two rows or three rows. Therefore a cross counter flow rad is
more efficient with one row in front and one row in the rear
than a non cross counter flow with the same over all thickness.
Most automotive rads are now one row with large frontal area.

As Kays and London pointed out however increasing the HP
on the pump is outweighed by increasing the water flow velocity
increasing heat transfer and reducing the cooling drag.
Kays and London is the bible so I would rather believe them.

Quote:

"One disadvantage in the use of highly compact surface geometries
is that the resulting core shapes are characterized by large flow
frontal area and short flow lengths for the gas-flow path."

What they are implying here in the use of the term "compact surface
geometries" Is the most cooling for the smallest core volume.
In other words the most efficient rad in terms of heat exchanger volume
is a large frontal area with a thin core.

In most cars the rad is placed upright between the grill and the engine.
If you want more cooling your only choice is to use a thicker
rad. The designers of the RX8 fully understood what Kays
and London were saying so they mounted the rad at a 45 degree
angle allowing a thinner rad with a larger frontal area.
Stewart Warner is mostly concerned with dealing with most cars
and trucks.

Kays and London again.

"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. [both air and water] For a heat
exchanger operating with high-density fluids, [water] the friction-power
expenditure is generally small relative to the heat transfer
rate, with the result that the [water] friction-power expenditure is
seldom of controlling influence."

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

For an example a 10 pound increase in drag at 300 fps (205 MPH) is
5.45 net HP out of the prop. With a 80% eff. prop that
is 6.8 HP more from the engine. While the total HP on the water
pump is a max of about 3 HP.

Total drag HP with 2 square foot equivalent
flat plate area is dynamic pressure 109 pounds per square foot
at 300 FPS times 2 or 64,400 pound feet per second or 225 HP net
out of the prop. 281 engine HP. Reducing the cooling drag pays
off big time compared to 1 or even a 2 HP increase in water pump
power.


I highly recommend getting a copy of the Kays and London book and
reading the whole thing. It is very comprehensive. Hard to understand
for me but worth it. I had to read it and then re read it several times :)
It is a complicated subject.

Paul Lamar


Paul,

Yes it is a tricky subject.   I have a few chapters of the Kays and
London book and I might have to get the whole book.  Still I need to
test.  and look at the water flows. Air/water heat exchange rates aside,
water velocity is a parameter affected by pressure and apertures
restrictions. Pump pressure has to increases and consumes more
horsepower to  reach the upper design limits. As pump pressure increases
flow drops off.
Beyond a point low flow and high velocity has no merit.

Pumps that are design to operate with a large thin automobile radiator
will not flow the same when the liquids have to go through  more
restrictive setups like a dual pass heat exchanger whether the
configuration is side by side or U shape.  If pushed beyond their rated
flows centrifugal pumps just start to cavitate. They can also cavitate
if the inlet flow is not adequate.

I also wonder how they got the 16X pressure numbers. Even localized
within the radiator that would exceed the burst strength of the aluminum
tubes I would think.  Perhaps the decimal place is wrong. 1.6 times.

Cheers

Doug in Japan