|
Track Testing
When we set up our customer's
racing cars, our main focus is to develop a set up that we can go
to a circuit and test.
This excerpt, says a lot about the
value of steady state testing on a circular skid pad. This is
exactly the situation we are modeling in the Weight Transfer Worksheet.
If you have done your WTW numbers, you are in good shape to analyse what
is happening to your car at the track test.
Handling, Skidpad and Track Tests
Skid pad tests can be the most productive and exciting tests for a
road racing car, but they are highly demanding on driver precision
and consistency. It may take a lot of practise laps before the
driver can maintain a nearly constant throttle and steer angle and
get lap times to repeat within a few hundreths of a second. But a
lot of valuable data can be obtained at low cost and low risk on a
skid pad, so it is well worth the effort.
The test setup will vary depending on the data needed, but driving
technique is simply to keep the car on a precise circular path, at
the highest possible constant speed. To a driver who is used to
clipping corner apexes with the inside front tire, it is probably
easiest to try to keep the inside front tire on the painted circular
stripe. That tire will also be least affected by a change in
friction coefficient due to the painted stripe.
Maximum lateral capability can be measured by one of three methods,
or all three together, if accuracy demands it. The first is to use
an electronic speedometer and record the speed constantly. The
disadvantage is that it is hard to take the average speed, and the
recorded speed is affected by lateral tire slippage. A more direct
method is to use an electronic accelerometer mounted laterally in
the car. But again, it is hard to take an average of all the
fluctuations, and the reading has to be corrected for roll angle.
The easiest, fastest, and probably most accurate method is to take
the lap-time for each complete revolution on the circle. This merely
requires a split-action stopwatch and someone with a very precise
eye and thumb, or an automatic trap switch to start and stop the
watch. A timer's accuracy can be improved by having a distinct mark
on the circle and noting the instant the car's front tire hits it.
It is also a good idea for the timer or someone else to watch the
car and make a note whenever it moves noticeably off the marked
path. As with any other data-taking, any particularly low figure
will have to be backed-up before it is credible. Of course, tires
should always be scrubbed in and
warmed up until lap times reach a consistent and low level, and turn
direction should be reversed frequently.
The relationship between lateral acceleration, radius, speed, and
elapsed time is given by:
(.067) (mph)2 (1.22)(r)
g =
(r)
(t)2
where the radius (r) is to the center of gravity of the race car, or
the inside tire path plus one-half the tread width, plus some slip
angle displacement.
Steady-state stability can be evaluated at the same time as maximum
lateral acceleration. This is the tendency of the race car to
maintain a stable non-oversteering condition at top speed on the
skidpad, as discussed in Chapter 7. As various factors, such as
springs, anti-roll bars, or tires are being changed, the steering
angle necessary to balance the car will change. It should be
possible to see the increase in lap times as the car becomes
unstable and uncontrollable with oversteer. But it is
also possible to measure the oversteer or understeer bv recording
the steer angle. At a very low speed — say 5
miles per hour, or minimum possible,a given steer angle (Ackermann
steering) is necessary to keep the car on the circle. As the speed
is increased,the necessary steer angle will either increase (understeer).
remain constant (neutral steer), or decrease and require quick and
violent corrections (oversteer).
Examples of understeer/oversteer plots are shown in Fig. 45. The
vertical scale is the steer angle required to drive the car on a 100
foot radius circle. The horizontal scale is lateral acceleration in
g's, which increases as the square of increasing speed. (Speed is
shown just for comparison on the lower scale. The speed scale would
change with different radius skidpads, while the g scale remains
relatively constant.)
In the lower sublimit area of the curve — usually up to perhaps 0.3
g — the curve is relatively flat, and can be presented in a single
number of degrees per g. But up closer to the limit of lateral
acceleration is where the interesting things happen. With a normal
understeering car, the curve will continue to climb, usually at an
increasing rate. In other words, it requires more and more steer
angle, until no amount of steering will keep the car on the path. An
oversteering car, however, will reach a point where the curve
flattens out and makes a sudden break downward. This is where "countersteering"
is necessary to stay on the path — a very unstable situation. The
rare neutral-steer car will have a very flat curve from zero and not
break up or down. It is generally accepted that the straightest
curve is best, and with enough rise that it never breaks downward
under any transient conditions.
It is important to mention that many static vehicle parameters can
affect these curves, and so they are sometimes factored out. The
Ackerman steer angle is the steering necessary to stay on the radius
at just above zero speed or g's, and is a function of wheelbase and
steering gear ratio. Therefore it is usually subtracted out,
defining zero steer at just above zero speed. Steering gear ratio
will also directly affect the slope of the curve, doubling the slope
with a doubling of the gear ratio. However, because the change in
steering wheel angle is what the driver actually perceives, gear
ratio is not
normalized out except for engineering analysis of front/rear tire
slip angles.
To get these curves, it is usually necessary to use continuous
(analog) strip chart recording. Vehicle dynamics engineers
ordinarily use recorders which take a minimum of 4 to 8 channels of
data simultaneously. Here, just 2 channels would be adequate: steer
angle and speed. Another less precise alternative is to increase
speed in steps and make a visual note of the steer angle in degrees
at each speed.
Transient tests can only be positively made on a skidpad that allows
entry and exit from the steady-state circle, or on a race track, of
course. However, if there is even a little extra room on the skidpad,
it is possible to get a feel for transient response. There should be
enough understeer so the car can be accelerated relatively rapidly
without losing traction at the rear. In other words, there shouldn't
be any drastic throttle oversteer. Conversely, there shouldn't be a
sudden change in stability due to a complete throttle-off at the
limit. It can also be an educational experience to brake hard from
maximum lateral acceleration while trying to maintain the path
radius.
When it comes to checking fuel and oil pressures, the skidpad is far
safer than a race track, if not quite a perfect simulation.
Naturally the driver should always be aware of oil pressure at high
lateral accelerations, but the skidpad allows him to watch it more
closely over a longer period. If the oil level is ever going to
settle to one side of the pan and starve the pickup, it is better
for it to happen on a skidpad where there is a greater opportunity
to shut the engine down immediately. The is true of fuel pickup
systems, especially in trying to determine how completely they can
drain the tank without picking up air and leaning-out the engine.
The greatest problem with the test is that in many cases the worst
possible situation is compounded by the addition of braking or
acceleration forces (which you get on a track, but not on the skid
pad).
Suspension deflections can be important to know, especially in
development of geometry and ride rates. Deflection measurements on
the skidpad will merely show the maximum roll angle, how close the
components are to bottoming, and how great the jacking effect is.
The test setup is about the same for aerodynamic downforce, except
that the deflection sensors are mounted as close to the wheels as
possible, instead of in the center of the car. Ideally, there would
be a continual recording at each wheel, but if only one channel is
available rather than four, the test must be carefully repeated four
times. The best location to record suspension deflection is probably
at the centerline of the spring, although it may also be useful to
know the wheel deflection. In that case, a swivel anchor can be
mounted to the center of the wheel, and the sensor can be mounted
outside and above me wheel — as on a fender lip or extended bracket.
For real-life conditions such as bumps, dips, and combinations
including acceleration and braking, only a test on an actual race
track will suffice. (If no data logging use shock travel O
rings on the shock shaft).
TIRE TESTS
Tire tests are one of the most valuable uses for a skidpad, once the
chassis has been fairly well developed. There is no other way to
accurately determine the optimum tire compounds, temperatures,
camber angles, or pressures. Results obtained from race track
testing would have to be far more significant to eliminate driver
inconsistencies. The test setup simply requires a race car and
driver that can run all day long at low speeds and high lateral
acceleration, with no fatigue, overheating, or variation in
performance. It is also important to remember that only a pair of
front or rear tires can be
tested at once, since the car will be limited by either front or
rear cornering capability. The best practice is probably to develop
front tire cornering performance first (since the car should be
understeering), and as it becomes better, to keep increasing the
front anti-roll bar rate as necessary to avoid oversteer. When no
more front cornering power is available or the front roll rate is so
great that it lifts the inside front tire off the ground, then it is
time to work on increasing the rear tires' capabilities. The rear
anti-roll bar rate may then be increased to create oversteer, as a
rear tire limiting condition to overcome.
Tire test procedure is simply to record average lap time or lateral
acceleration for each configuration change. However, it will be
necessary to monitor tire temperature constantly, since its effect
is great enough to cancel out other test conditions. The first test
with any tire should be a comparison of g's versus temperature, to
determine the optimum and the drop off on either side of the
optimum. Since it is difficult and expensive to record temperatures
continuously, it will be necessary to stop the test at intervals and
check the temperatures as rapidly as possible with a tire pyrometer.
The best
technique is to run in two or three lap increments, with one person
timing, and another taking temperatures as fast as the car can be
stopped from its high lateral condition. It shouldn't be necessary
to take over half a lap to stop, and the tire technician should be
right there at the stopping point. Within 10 or 20 laps the tire
should be at its maximum temperature, or past its peak cornering
capability. It can also be valuable to know just how fast the tire
cools off, to get an idea of what the true temperature is while the
tire is working. (See Chapter 2) This can be estimated by watching
the temperature fell in a given location over a matter of seconds
and projecting the result. Since the test accuracy is poor, it is a
good idea to repeat it after, or as, the tires cool off, and in both
directions around the pad. Of course, the outside front or rear tire
temperature is of greatest importance.
From then on, all tests with that particular tire should be run at
that optimum temperature or at least corrected for any drop off.
This data will also come in handy at the race track, to determine
whether a tire compound is too hard or too soft for the work input
under a given ambient temperature.
When tire compound and temperature can be held constant, then
optimum tire pressure and camber angles can be determined.
Proper camber angle will show up in skid-pad lap times, tire wear
profiles, or temperature differences across the tread, but the last
method is quickest for tire development work. The pyrometer must be
used rapidly to get three readings (inside edge, center, outside
edge) before the natural heat conduction in the tread evens out the
temperatures. It isn't reasonable to expect them to be exactly
equal, however, since the car will be at maximum lateral
acceleration camber angle only for very short periods.
When everything else has been developed to the optimum on a skidpad,
it can be a good place to teach the driver what extreme variations
in handling feel like. If the car is ever going to lose a shock
absorber, or break a front anti-roll bar and oversteer, or have a
tire go soft, or otherwise become unmanageable, the skidpad is the
safest place to learn the feel and the corrections required. Just
knowing the feel of ordinary changes from oversteer to neutral steer
to understeer is an invaluable aid in later analysis of a race car
on a race track.
TRACK TESTING
An actual race course is the last place where any serious or
accurate development work can be done. Only after the car has been
otherwise ideally set up will race track laps be meaningful, and
then primarily with respect to the driver's performance. For vehicle
evaluations, it will still be necessary to break the track down into
braking, cornering, and acceleration segments, as opposed to
over-all laptimes. The timing isn't as difficult (with an electronic
split-action watch) as it is to find a spot where the car can be
seen at many locations around the track. At Riverside, for
example, it is possible to see a car most of the way around the
track from the roof of the timing tower. A sample of segment times
is shown in Fig. 46.
This is also the best way to find out how a competitor's car really
compares — as opposed to average laptimes. If the other driver is
sandbagging, it will probably show up in a particular segment. But
if the other car is quicker, it is helpful to know exactly where, to
know where there are some capabilities to be gained. In the example
shown, car B is apparently better in acceleration, which means that
car A is probably adequate in handling and braking but should have
more power or lower air drag. Some people have also used radar guns
or electronic gunsight tracking devices to record comparative speeds
around an arc. There are other more sophisticated spying devices to
analyze competitor's cars, which are more accurate and more
complete. But, needless to say, they are much more expensive
and complex — and confidential.
Probably the best race driver teaching' device known to man is a
continual recording of speed and horizontal accelerations around a
race course. It won't say much about the car unless there are other
recordings of the same car in another configuration, or other
recordings of other cars, to compare with. But such recordings will
tell a great deal about the driver's ability to take advantage of
the car's capabilities. The best test setup is to have a two-channel
recorder with speed and acceleration inputs. The speed can come from
a front wheel pickup, preferably the outside wheel, to avoid lift or
lockup problems in cornering and braking. The g sensor can be either
a single lateral accelerometer or a combination of two arranged at
right angles. An electronic circuit can be designed to calculate the
net horizontal acceleration in any direction and produce a signal
proportional to the percentage of traction used versus traction
available. A more thorough explanation is given in Chapter 12.
Other tests that can be performed on a race track were previously
explained under straightaway or skidpad testing. However, in general
they tend to be tests of the track configuration rather than of the
vehicle. Unusual suspension deflections are mostly dependent on
surface condition or grade changes or bankings. Different tire or
brake temperatures are a function of track coefficient and
distribution of time spent on cornering, acceleration, and braking.
Vehicle transient response characteristics will change with respect
to the types of corners on a given race track. If there is enough
time, a race car
can be set up to the optimum for each particular track's predominant
characteristics, but it is likely to take at least a few days of
track rental and exclusive running. It helps a great deal to have a
lot of experience at a particular track and at a wide variety of
tracks. But for those with no experience or for a new track, a
careful analysis of speed and g recordings can work almost as well.
Of course, in the absence of any recording instrumentation at all,
stopwatch times, in various track segments will probably indicate
the worth of any vehicle changes. Handling or tire improvements
should show up in the low-speed cornering segment times, aerodynamic
downforce should show up in high-speed cornering segments, engine
power should show up in the straightaways and so on. The overall
improvement will probably show up in total lap times, but less
significantly. This leads to the question of the relative
value of various vehicle improvements. Racers tend to concentrate on
making
improvements in the areas they understand or enjoy the most rather
than those areas where the potential gains may be greatest. It isn't
that hard to determine just what these relative values are, however.
A computer can be — and has been — used to put a numerical value on
various race car improvements, but any racer can find the numbers
for his own car on any particular track by the handicap method. If
it isn't easy to improve a car's performance, it is all too easy to
diminish it.
All that's necessary is to know the amount by which the effect is
reduced and the increase in lap time which results. The effect
will be linear enough over a reasonable range to project from a
decrease in performance to an increase in performance due to a
positive change.
The most obvious example is in determining the effect of reduced
weight on lap-times. All that is necessary is to plot lap times
versus fuel consumption in pounds. Say that a 2000-pound race car
consumes 200 pounds of fuel during a race, while its average lap
time decreases from 90.0 seconds to 88.2 seconds. Then, assuming
that all other factors remained equal during the race, a person
could project that a 10 percent reduction in vehicle weight would
produce a 2 percent reduction in lap times — at that track. (This is
true for race cars with centrally mounted fuel tanks. For extremely
rearward tank locations, weight distribution changes are an
uncontrollable variable.)
Other factors may be as easy to degrade as weight: power reduction
with throttle-stops, air drag increase with a flat plate, tire
braking capability reduction with lower temperatures. However, it
may be somewhat difficult to quantify the exact value of the change
in the factor. But once the relative positive effect of various
changes can be estimated, the proper concentration of efforts can be
allocated. Of course, even the most scientific approach must be
adapted to fit the capriciousness of racing regulations and the
availability of time and dollars.
Most racers feel that durability tests are also beyond their limits
in available resources, so they tend to use experience, intuition,
and luck instead. The only way to really know whether a car can be
raced hard for 24 hours or 500 miles is to race it hard for that
time or distance. If it is done in a test session and nothing
breaks, then the car can be totally rebuilt as new for the real race
— and still fail due to some random faulty new component. Even if
experience is the best judge of durability, that sort of experience
can be bought. A brand new design may be faster but it definitely
doesn't have a history
of reliability. The best insurance is in knowing that a particular
design or component has been around for a long time without unusual
failures. At the very least, it is a good idea to know a car or
component's history, and its average life expectancy before
inspection, rebuilding, or replacing is necessary. On the other
hand, if durability testing is feasible, the biggest mistake is to
try and make the car survive under those conditions. Instead, the
idea is to try and break the car under reasonably severe simulated
usage, rather than pussyfooting it around. However, it is wise to
test on a track where a failure doesn't have consequences as
critical as they could be in a race.
|
|
