Intercooler Design
This is a documentation of an experimental turbo intercooler system.
It
includes reasoning and math that might help others. This article is being
updated from time-to-time as I obtain more data from system data logs.
Evaluating my car, the best airflow choice was an air-water intercooler. After
searching Internet and asking questions in a few forums, I could not find any
basic articles or calculators for Intercooler size. Since I could not find anything
to calculate size, I used standard heat formulas. My reasoning is that heat is heat, and what works
for heating or cooling a room should work for air in an intake system. The only
unknowns are efficiencies and hardware specifications. For example, the Frozen
Boost system does not give BTU rates. Frozen Boost gives a meaningless
horsepower number.
Air Charge Heat
With a Vortech V2t at 15 psi (running ~800 HP) and a V3si blower at 10 psi (running
~700 HP) data logs showed air charge temperatures in the low 200F range after a
few seconds. Temperature stabilized at around 200F or so no matter how long the
pass. The present turbo system can supply up to 28 psi boost. The measured
temperature rise was about 120F.
Theoretical temperature rise from pressure increase is
(PR0.28-1) x Tabs
Tabs is absolute temperature, which is
the thermometer temperature plus 460. A pressure ratio of 2 (14.7
psi boost on top of 14.7 psi atmosphere) and 80F inlet air would have 2^.28 =
1.214 – 1 = .214 * (80+460) = 116F outlet temperature with a perfect compressor.
If the compressor efficiency is 60%, then the temperature is 116/0.60 = 193
degrees. This calculated temperature is reasonably in line with what I measured with
my superchargers. At 15 psi boost (pressure ratio of 2) and 80-90F air, logged IAT was
just over 200
degrees.
Based on that, the turbo at 28 psi should have 2.9^.28 = 1.347 – 1 = .347 *
(80+460) = 187 with perfect compressor. At 60% air should be 187/.6 = 311.6F
Intercooler Layout
I had some difficulties finding a suitable intercooler system. I
didn’t want to run my car’s SC or turbo air pipe downward or to the side,
then bending forward to the
grill area, across the grill area in a heat exchanger, turn back to the engine
compartment, and eventually back up to a throttle body. Every
bend and every inch of additional pipe length would add flow
restriction. Larger air pipe system volume also acts like an
air storage tank,
there is more air volume built up that has to be built up and bled off. The blow off
valve would have to be larger to release more air volume with long, large
diameter piping, and it could take a little longer to build boost.
After some searching, I chose a type 26 Frozen Boost water cooled intercooler.
Frozen Boost intercoolers are reasonably priced. Unfortunately, other than physical dimensions, Frozen Boost does not
provide very many useful specifications. The only information Frozen Boost provides
for the type 26 is:
100% Lightweight aluminum, No Epoxy Used In CoreFin pitch is 21 fins/inch – 12.5 air side and 8.5 water
side. Recommended for up to 1200 CFM / 1200 HP
Pressure Drop: Less than 0.2 PSI. This is not a typo!
Liquid/Air Intercoolers have incredibly low pressure drop.
Air Inlet/Outlet: 3.5″
Water Inlet/Outlet: 3/4″ NPT
Maximum water/air pressure: 70 PSI
Core Size: 10.5″x6″x5″ (End tanks add to length)
Type 26 link
Absent from Frozen Boost’s information are:
-
BTU heat exchange or heat dissipation capacity at different
air and water flow rates -
Backpressure of air and water over a useful range of
flow rates
The complete lack of meaningful thermal information prevented calculating
water flow requirements and temperature rise in the intercooler. Minimum water flow rate
for a given water system temperature rise, however, is independent of intercooler specification.
The minimum water flow rate for a given water flow rate in gallons per minute
can be calculated.
BTU and Water Flow Rate
I simplified standard proven formulas for horsepower, airflow, heat, water
flow, and BTU to:
BTU per hour = temperature rise * horsepower * 2.1818
after finding BTU per Hour Water flow in GPM
= BTU per hour/(allowed rise* 500)
Example: 500HP, air charge is 110F hotter than inlet air, and 20 degree
allowed water rise we have 500*110*2.1818 = 119999 BTU hr.
119999/(20*500)=119999/10000 = 11.999 GPM
For my car at 1000 HP and 120F IAT rise over inlet air:
1000*120*2.1818 = 261,816 BTU hr.
If I want water temperature rise to be 10F 261,816
BTU hr. would be 261,816/ (500 *10) = 261,816 / 5000 = 52.36 GPM
Rounding off numbers in BTU calculations will not hurt. A rounded formula is:
| BTU per hour = compressor system temperature rise * horsepower * 2.2 Water flow in GPM = BTU per hour/(allowed rise* 500) |
52 GPM would be required to maintain a 10 degree rise over inlet water in
sustained full power operation, where the full power running time is longer than
the warming time (thermal lag) of the aluminum and water in the intercooler. 52 GPM is a significant water flow rate
requirement, until we think about a pot of water or a big lump of aluminum.
Intuitively, we know it takes
time to heat any dense material (especially water). There is a heat capacity related to material volume and
material type that slows any temperature change with application of external temperature
changes. This thermal lag effect is very much like the mechanical inertia in a flywheel. The thermal
inertia of the water and aluminum in the intercooler slows intercooler thermal rise
during a short burst of power. The heat is applied only for a few seconds,
while
the intercooler and water in the intercooler system start out being reasonably cold. This
thermal lag will reduce required water flow rates in intermittent duty, since the system
heats slowly. Some heat will be stored in the aluminum and water, and that heat can be removed gradually after boost is removed.
There also is some forgiveness if we guess or estimate incorrectly. If the
system runs half the required water flow, temperature rise will only
double. If we try for a 10 degree rise and the actual flow is half what is
needed, peak water temperature will only increase by 10 more degrees than
planned.
I attempted to measure my system flow with a short 3/4 inch ID hose from my
well tank, through the intercooler and then the radiator. Approximating my car
system, filling a 5 gallon
bucket in about 15 seconds required about 3 psi (0.2 bar) at the IC inlet.
The required hose and pump size (52 GPM for 10 degree water rise) is space,
power, and cost prohibitive. Consequentially, I used
the largest pump and hose sizes I could reasonably afford, find, and fit. This wound up being a
Johnson SPX CM90 pump that flowed according to the following chart, with some of
the plumbing 1 inch hose and some 3/4 inch hose. The Johnson pump draws 7 amperes and is fairly
quiet.
I believe the CM90 pump is a logical compromise, since pumps in commercial 650 HP engines with short
hoses and high flow water systems are around 25% of the CM90 size. My system has
fairly short hoses and uses a large radiator.
I used the largest hoses I could fit in each run area.
While far less than ideal, the
CM90
pump has proven to be adequately sized for intermittent short running. Data logs show the peak air intake temperature increasing approximately
45
degrees F over starting idle air during an eight second run. The largest problems
in my system appears to be the lack of forced air across the Volkswagen
radiator at idle and low speeds, and lack of enough water flow rate to keep the
IC water cool. My car’s radiator fan is tied to the engine cooling
radiator’s water outlet temperature, not the intercooler. This causes the VW
radiator to idle around 20-30 F over ambient air temperature. UPDATE: I
corrected this without an auxiliary heat exchanger fan by running my Derale fan
controller at about 30%, with the engine radiator fan controller commanding more air only if
the main radiator cold tank gets warmer than 145F.
On a 90 degree day in sunshine, starting intake air temperature and IC water
temperature is about 112 F. This is prior to running the radiator fan at 30%
speed minimum.
The log graph below is on drag radials, and shows time from brake release.
This low nine second 1/4 mile run is on 275 drag radial tires. I had to
lift throttle at maybe 1000 ft because the car got a little unstable at speeds
over ~150 MPH. As boost built to 17.6 pounds, the car required one
full throttle lift to bring the rear tires back into traction. You can see that
lift where the throttle TPS line drops to zero just before 2 seconds. The logged
data is throttle position, mass air temperature, and mass air pressure. Pressure
is in kilo Pascal’s, with 100 kPa MAP being 0 psi boost. The peak boost in the
log is 240 kPa, which is 20.6 psi boost.
In the log graph below, starting air temperature is 112 F. During the run at
wide open throttle, highest air temperature is 140F. The peak air temperature of
154F occurs after the throttle lift near the end of 1/4 mile run. That peak is
with the throttle closed, the154F is probably just caused by engine heat pumping
back into the intake manifold from the fully closed throttle. This is typical of
the approximately 25F rise I have been logging during a 1/4 mile pass.
UPDATE With my fan controller modified to run 30% minimum, I now start every
pass with the IC at ambient air temperature.
The above graph shows me I still have a starting air temperature issue. If MAT
started at 90 F instead of 112 F, the ending IAT would be 22 degrees
cooler, or about 120F. Still, my ending temperature with 21 psi boost is less
than the typical temperatures logged with the supercharger and no intercooler.
By calculations using 21 psi boost and 90 degrees F ambient air, with a 60%
compressor efficiency: (21+14.7)/14.7 ^ .28 = 1.28 1.28
– 1 = .282 * (460+90) = 155 F 155 /.6 = 259 F
I have pulled air temperature down from an estimated 260F on a 90 degree day to
less than 150 degrees, even while the starting base temperature was 30 degrees hotter than ambient. The type 26 with
a VW radiator and CM90 pump, without a constant cooling fan on the radiator in
staging, has lowered air temperature entering the engine by somewhere around
100-120 degrees!!
Mounting Position
Trapped air is the enemy of water cooling systems. My Mustang’s most convenient intercooler mounting
position is above the water pump. For height reasons, this location prevented
using a vertical water inlet and exit. With intercooler water
ports out the front and rear sides of the type 26, the area above the ports formed a large air-trapping
pocket. A second problem was installing a fill tank and reservoir. Because of low hood
clearance, a standard water reservoir could
not be located significantly higher than the intercooler. This further complicated trapped air
problems.
The boosted air path is excellent. The air system path is straight in through
the turbo, directly out to the intercooler, and then through two 90 degree four
inch diameter turns into the elbow. It is impossible to have a shorter path with fewer bends
with the same size intercooler. My turbo easily makes over 22 psi boost, and
makes that boost within 1.2 seconds from idle. This system makes 15 psi less
than one second coming off idle. The highest dyno pull peaked 1280 HP at 28 psi,
but until handling problems are worked out, I am keeping boost at 22 psi or
less.
To save space, I added a
Meziere WP311S electric water pump and an electric fan system for the
engine.
fig 1 above home made alternator brackets and
intercooler shelf. If I can’t find it, I make it
The advantage of this layout is I can fit either a centrifugal belt driven
blower or a turbo. The air piping is straight and direct, as shown below:
The outlet side goes through two 90’s total, each are four inch:
fig 2 The air-water-air system allows short, direct piping. The BOV welds into the four
inch aluminum pipe. The BOV only has to bleed down the area from the throttle body to
the turbo outlet.
Heat Exchanger
Nether the radiator nor the intercooler have published BTU or water flow rate
specifications.
I chose the radiator empirically. There are three things working in my favor
to reduce worries about radiator size :
- 1.) The boost will be on less than ten seconds and off for half-hour or
more - 2.) The radiator had to fit existing space and be a closed water system. The
radiator could not have a fill cap - 3.) The intercooler contains a great deal of aluminum, and there is a
good bit of water in the system. The 16-plus pounds of aluminum mass and 15
pounds of water in the cooling system will slow temperature rise noticeably
After much searching and measuring, an early Volkswagen Jetta or Golf radiator looked
workable (and cheap to replace if it didn’t work):
In addition to two large hose fittings on the same end, the VW Golf radiator has one tank vent.
That vent is located at the tank top right next to the upper radiator inlet hose
(left lower corner of pix). The nipple fits a small 3/8 inch hose for purging
air. To ensure
rapid full air purging, the small
hose fitting must remain open to allow air escape while filling. Notice the radiator does
not have a cap. It was designed for an external reservoir and fill cap.
With a draw-through fan on my radiator system, the VW radiator fits perfectly
between my automatic transmission cooler and the engine radiator. The top air
purge vent is at the left. The hose outlets face forward. This is a dry fit
while I made mounting brackets:
Close up of upper vent:
Water Circulation System
I thought about ice, but ruled it out because of headaches. I decided to use
air-water-air. The circulation system looks like this:
Head caused by lift is unimportant in a closed circulation system, as is
acceleration head. This is because liquid flows with gravity as much as it flows
against gravity. Since drop equals lift distance, the two effects completely
cancel. Fluid velocity is also identical at the pump inlet source and the pump
outlet. The pump does not have to accelerate fluid from a stopped position,
fluid velocity toward the pump inlet equals fluid velocity exiting the pump, it
is a closed loop.
Friction losses in hoses and restrictions are the only things causing head in
a closed system. Unless you are dumping the water out to atmosphere at some
greater height than the water source elevation, the system has no head other than
frictional restrictions.
Type 26 Modifications
My vehicle’s under hood height limitations, along with the high intercooler
location above the engine’s water pump, caused a system fill problem. I needed
horizontal inlet and exit locations for water and air, and I did not want to
trap air in the intercooler. This dictated a water reservoir above the
intercooler height, and some way to purge air from the very top of the
intercooler, well above the hose inlet and outlet. Without that, the intercooler
would form an air trap and collect system air.
In some installations, the
water inlet and outlet could have been placed vertically. The mounting position
I used placed the intercooler top about 2
inches below the Boss 2″ rise hood on my Mustang. Because of close hood clearances and a lack of mounting
tabs on the water face, the intercooler could not be mounted water flow vertically,
placing water inlet and water outlet ports
mid-way up the sides. Since water
is much heavier than air, all the air migrates to the highest point of the
intercooler. This causes the entire upper area to become air locked. The high
location also causes a system filling problem. A normal water reservoir and fill
tank could not be located above the intercooler without creating aggravating
fill problems.
To solve air pocket, fill, and water tank problems, I
decided to modify the type 26.
First I
drilled a 3/8th inch bleed hole on the higher pressure inlet tank. This is the side where
the cold water pushes in under pressure (it is also the best air outlet side):
I drilled several holes along the upper edge of the outlet tank. The
outlet is the low tank pressure side, where the pump is pulling water out into
the cooling radiator:
I cut and bent a water tank out of 0.120 inch thick aluminum:
I carefully fit all edges to the intercooler surface:
I welded in a radiator cap fitting:
I welded all of the water tank seams, and welded the tank to the cooler:
The square bung in the lower corner is for the overflow/vent fitting.
Final System
The final system was
insulated with an adhesive backed thermal wrap. The intercooler is a tight
squeeze, with only 1/2-inch clearance to the electric fan for the upper two
inches of fan area. The small hose vents to the radiator water overflow
reservoir:
Intercooler inlet side. Air is straight in, no bends :
Intercooler water pump mounts on a home made bracket below the right front
corner below the inner fender liner, but above the lower radiator support and
bumper:
Final System
This system easily runs over 150 MPH in the 1/4 mile. Because of good
plumbing layout, the turbo spools from fast idle to 15 psi in one second, and
off fast idle to 20 psi in 1.1 seconds. Due to traction and handling, I’m
limiting boost to 20 psi. I’ve had boost as a high as 28 psi.
kW, BTU, Joules, and Temperature
I wanted an idea of how much water flow is
required, but I don’t know much about thermodynamics. A list of relationships or
conversions between units follows. I gathered this information from various
textbooks:
For Water:
One BTU is the energy
required to heat or cool one pound of water
1° FSpecific weight of water is
62lbs/ft^3Water weights 8.35 pounds
per gallon62 BTU raises temperature of
1 cu ft of water 1 degree
For BTU, joules, or
watts:One kilowatt-hour is about 3412
BTUOne joule or watt-second is 0.000948 BTU
One BTU is approximately
1055 joules or watt-seconds
For air:
0.018 BTUs per cubic foot changes temperature 1° F1 BTU changes 55 cu ft. of dry air
1° F0.24BTU raises one lb. of air by
~1deg F
One cubic foot of dry sea level air weighs .07 lbs.
For Engine Airflow:
One HP uses approximately 2 CFM of air
Specific heat of water is 4.1813
times specific heat capacity of air (this means a pound of water
absorbs about four times the heat of a pound of air, water is one of the
best heat absorbers in nature)
Water flow was calculated above
based on steady operation and 10F rise, with 262,000 BTU per hour. Drag racing
takes around 5 seconds in the 1/8th mile, and 8 seconds in the quarter mile. A
ten second full load time would be conservative. There are 360 ten second
periods in an hour. This means the total heat dissipation would be 262,000/360 =
728 BTU.
Water temperature increases 1
degree F per pound per BTU. My system holds approximately 2 gallons of water, or
16.7 pounds. 728 BTU would increase temperature of 16.7 pounds by 43.6
degrees. Without any cooling from the VW radiator, if water circulated all 2
gallons in ten seconds (a 12 GPM rate) the water would reach 44 degrees over
ambient at the end of the pass. 4.4 times that rate is 52.8 GPM, which tracks
with the 10 degree rise calculation of 52.36 GPM. Using this entirely different
way of estimating GPM, we find close agreement with earlier formulas. This
assures us the numbers are correct.
My worse case water rise should
be somewhere around 20 degrees F. I wound up with about 30 degrees air
temperature rise over a 1/4 mile pass at 21 psi boost, so I expect the water
rise is pretty close to estimates. Water temperature rise has to be some amount
less than air temperature rise. The type 26 seems to work adequately at
over 1000 HP.