Twist, Rattle & Shake
by Jim Park
If you think trucks have changed a lot over the years, you should see how they
design and test them. The biggest improvement is in the speed and accuracy of
the design evolution, thanks to the remarkable tools of the trade and the ingenuity
and imagination of the people who use them.
the world headquarters of Mack Trucks in Allentown, Pa. sits the company's Engineering
Development and Test Center. Of the 400 people working there, 75% of them are
in product development.
There are 16 module teams responsible for some element of the
finished product and each team has representation from purchasing, manufacturing,
the lab, finance, etc. They all work side by side - literally - as the design
"In the past, when the design people came up with an idea and the lab
mocked it up, the purchasing department might have looked at it and said no,
it's too expensive. Or the manufacturing people would come along and decide
that it couldn't be built in the plant," says Glenn S. Hinderliter, chief
engineer in charge of durability at Mack's Vehicle Development Laboratory (VDL).
"That whole process used to take weeks. Now, with everyone in the same
wing of the building, we can make it happen in a fraction of the time."
The last complete prototype they built took about three weeks from loose parts
to complete truck. But putting together a one-of-a-kind chassis and standing
it on its own 10 wheels is no guarantee that it's going to work, or stay working.
with a process called finite element analysis (FEA), engineers have only a "good
idea" how it's going to perform in the real world. FEA examines materials,
applications, stress loads, and about a million other variables, and provides
a life expectancy analysis for the component. From there, it's out to the test
cells and/or the proving grounds to see if the computer was right.
Whether it's an entire truck, a particular system such as a suspension, or
an individual component, everything gets tested. That's when the VDL crew takes
over. They're the technicians and engineers who design the tests, calculating
among other things the severity and frequency of the test cycles needed to simulate
in-service life out on the mean streets.
Giving it the Gears
Transmissions take quite a beating in real life. That means the testing
process has to be at least as rough. ZF Meritor director of sales and
marketing Charlie Allen says the object is to design a product that will
significantly outlive the warranty, at 750,000 miles, which demands a
pretty thorough test.
Among the tests ZF Meritor applies to a new design is shock loading,
where an engine is revved right up and the clutch is engaged instantly.
Allen says it's like doing a wheelie with a loaded truck. Or, they'll
hook a transmission up to a big 19L Cummins diesel that produces 2050
lb ft of torque continuously. They'll run transmissions on that machine
20-22 hours a day for days at a stretch.
Allen says he'll let us bring our cameras and notebooks behind the scenes
at ArvinMeritor's engineering and test facility in Troy, Mi., some time
Simulating Real Life
Engineers can't test trucks in real time because the process would take forever.
They must put the truck through a lifetime's worth of normal stress in just
a few days or weeks, and there are a couple of ways of doing that. They can
apply realistic loads to a part or system, but run it through thousands of cycles
over a short period of time, or apply greater than realistic loads over a smaller
number of cycles.
Subjecting a component to a 10,000-lb stress load 2000 times isn't the same
as subjecting the component to a 2000-lb load 10,000 times. The results wouldn't
be the same. So, depending on the accuracy of the engineers' predictions of
the actual in-service environment, and being able to replicate it in the lab,
the testing bears out what the designers were striving for, or it doesn't.
Take a piece of steering linkage for example. You have a solid metal part that
will be under stress for its entire service life. The engineers must be certain
that the part will be strong enough to survive, yet be light enough to keep
the vehicle competitive, and cost-effective to produce.
One of the tests I witnessed was a steering lever being flexed a fraction of
an inch off center - right to left - by the application of many thousands pounds
of force by a pair of hydraulic rams. Somewhere upstream of the test bed, somebody
estimated that over the expected life of the vehicle, that part would be cycled
x times, with x pounds of pressure applied by the power steering pump. The test
reflects that scenario. If life-cycle estimates are off, the part could windup
being over-engineered, and therefore too expensive or too heavy, or under-engineered
and not robust enough to survive in the real world.
Over many thousands of cycles, the part may weaken and develop stress fractures.
Expert analysis can help the designers re-engineer the part, adding material
to the weak areas, or removing redundant material where they can, to reduce
weight and cost. And literally every load-bearing component or system on a truck
undergoes such a test.
In the Lab
They test cooling system effectiveness and hood design by running a truck on
a dynamometer at a predetermined engine load with the equivalent of a 15 mph
wind blowing through the rad. A huge fan mounted directly in front of the truck
provides the wind. Of course, temperature and humidity can be altered as the
need arises. All the parameters are closely monitored to yield repeatable results.
This aids in the collection of useable data for future use.
Other test cells take a truck frame, equipped with certain components, and
shake the daylights out of it. Again, all carefully monitored and very repeatable,
right down to the frequency of the vibration created by the shaker. The frequency
and severity of the cycle (speed and depth) are modified to suit the intended
application. For example, a dump truck might be subjected to deep slow cycles
to simulate the frame contortions experienced driving across a construction
site, whereas a highway frame might receive light rapid vibration to simulate
high-speed travel over different types of pavement.
There are cells that test suspension articulation, frame twist, cab rigidity,
crossmember strength, and more. There's lots of high-tech stuff going on in
the VDL, and some decidedly less-than-high-tech stuff too. There's the proverbial
slam-the-door-a-million-times test, a roll-up-the-window test, and a rather
amusing one where a fully fitted cab is mounted on a short chunk of frame, lifted
from the rear and dropped. It's lifted only a foot or so, but it falls hard.
The idea is to ensure the cab doesn't fall apart, or that the interior panels
don't rip loose. This one cycles many thousands of times too, and Hinderliter
took considerable delight in pointing out that one of his competitor's trucks
didn't make it past 1/3 the normal test cycle of a Mack. He wouldn't say whose
The Test Track
The proving grounds, as Mack calls the 65-acre site developed on the Engineering
and Development Test Center property in Allentown, is just an over-dramatization
of conditions experienced in the real world. The facility consists of several
areas, each containing a series of obstacles designed to stress a truck in a
manner consistent with a given application.
Testing has to simulate everything from standard highway conditions to severe
vocational situations such as construction sites and logging roads. In developing
the facility, Mack used data collected over thousands of miles of actual road
surface in North America, then created several surface profiles on the proving
grounds that would replicate those conditions, but in an accelerated manner.
There's a 3/4-mile two-lane oval track where highway speeds are possible, combined
with a 1200-ft long, 12-lane wide, gennite-covered skid pad with enough room
for a 500-ft radius curve for testing ABS and vehicle stability. There are several
gradient sections, too, where brake-hold and startability tests are done. Grades
range up to 20%.
But the fun parts are the two durability-testing loops located around the track.
The unpaved area is a 1.3-mile loop indicative of rough stone or dirt roads
found in logging areas, oil fields, etc. Events on that loop include ruts, washboards,
rocky surfaces, articulated turn areas, and more. The off-road area is a 0.5-mile
loop representative of construction sites, landfills, mines, etc. This loop
includes boulder fields, washouts, articulation surfaces, drainage ditches,
etc. To ensure consistency in the testing and evaluation, all obstacles are
concrete, not actual dirt roads.
has invested millions of dollars in developing the testing protocol used on
their proving grounds. In simulating real-world conditions, they have to apply
stress to the truck in a manner consistent with normal wear and tear over time.
Over a four-year period, for example, a dump truck suspension could be expected
to be subject to maximum articulation x times, under x weight. How can a field
of rocks embedded in concrete simulate a lifetime of wear and tear?
Hinderliter says everything from the height of the embedded rocks, to the depth
and the distance between the ruts on their washboard road is carefully calculated
to approximate real world conditions, but slightly exaggerated to simulate wear
over time. Through a calculated number of cycles over the different obstacles,
he says his team can simulate, say, four years of normal wear and tear in just
a few weeks.
He and his people devise test cycles with different exposure ratios to certain
obstacles that simulate conditions in different chassis categories. For example,
a highway truck might spend 90% of its test-track time on-highway and 10% off-road,
while a severe-service chassis might do 25% highway, 50% off-road, and 25% on
the unpaved loop.
If Hinderliter's team concludes that 4000 test track miles (not the actual
number) is equivalent to the expected life of the truck, they'll run the truck
around the course enough times to clock 4000 miles.
But it's not always complete trucks that need proving on the test track. Often,
components such as torque rods are fitted with dozens of stress sensors and
mounted on a truck for a trip or 50 around the course. The sensors measure the
stress values and record them for collection by the engineers who use the results
in their computer modeling. This helps reduce design and testing time by giving
engineers real-world data to work with.
There's a lot hanging on how well Hinderliter's people do their jobs. Build
too light and test too tough and you'll have a product that appears to fail
prematurely. Build tough and test light and you'll have a truck that appears
bullet proof. It's a delicate balancing act. Warranty and service depend on
accurate measurements of life expectancy, and when you warranty a product, you
had better be sure that it will live up to expectations. Otherwise, it'll cost
you a fortune a few years down the road.