Table 3.1 Average Lift (inches)

Figure 3.2. Downward Wheel Travel with RTI for the Front Right Wheel (inches)

Figure 3.3. Downward Wheel Travel with RTI for the Front Left Wheel (inches)

Figure 3.6. Average Downward Wheel Travel with RTI (inches)

Figure 3.8. Percent Articulation from Front vs. Ramp Travel Index

The 21 Road trail test follows a canyon bottom. The primary obstacles were boulder fields with a few outcroppings of larger rock. There were no steeply inclined, large slabs suitable to test weight transfer/stability effects during steep ascents and descents of flat and flat off-cambered sections. There were also no sharp-angled break-overs. The bulk of the stability observations were due to articulation-induced effects and a short drop-off. Overall, I would rate the trail as a respectable hard.

Jim Allen, John Hinkle, and Richard Hills of the organizational committee and Bill Burke identified 10 obstacles on 7 sections on the day before the tests. These sections were tested with the help of Edward Magoffin and his highly modified D90. Unfortunately, delays in the morning testing required that the number of obstacles to be used during the afternoon trail tests be reduced to 5. The first section (1 obstacle) was a climb up a large rock outcropping, off-centered from the usual line to force the left front tire up a large boulder. The second section (4 obstacles) required maneuvering through a boulder field (at Carnage Corner) along a path which required climbing rocks on the left and right sides and ending with a difficult off-camber section induced by a moderate sized inclined rock slab on the right. All of the vehicles either lifted a wheel (including the two SG vehicles) or required the use of a safety strap to keep from tipping over on this section. The final section (1 obstacle) was a drop off to test stability. Four sections that had been identified the day before were not used due to delays during the morning testing. Overall, the test sections were sufficiently hard that all of the vehicles lost points for various aspects of performance such as lack of clearance, lifting wheels, and lack of stability.

Five judges were used for the trail tests. One of the five judges (Bill Burke) spotted all the vehicles through the sections on narrowly defined lines. The remaining four judges (Jim Allen, Tom Collins, Richard Hills, Steve Watson) observed and scored. At the completion of each section, the five judges met to discuss vehicle performance and scores were assigned. The four scores were averaged for each section.

Points were deducted when lack of clearance forced a change of line (intentionally or unintentionally), when a wheel lifted, when the vehicle appeared to have difficulty transversing a section, and when the vehicle displayed less than perfect dynamic and static stability. Points were also deducted for wheel slip if the locker appeared to be on. In the case of obstacle 4 of section 2, extra points were deducted if the spectators had to stabilize the vehicle through a strap attached to the top of the roll cage. The scores for the three sections are given in Tables 4.1 through 4.3 with each of the 4 judges’ responses appearing on separate lines. A summary of the results is provided in Table 4.4. Note that while the spread in overall scores in Tables 4.4 were not very large, there does appear to be some grouping according to suspension type.

Figure 4.1 shows the final trail score vs. maximum ramp travel index. The correlation coefficient between these two quantities is only 0.16 indicating that there was very little correlation between the judges’ perception of trail performance and RTI. The vehicles with the two lowest RTIs placed 1^st and 3^rd. The vehicle with the highest RTI placed 2^nd on the trail. The trail score as a function of % front articulation is presented in Figures 4.2 and 4.3 for RTI=750 and 1000. Note that there does appear to be a strong correlation between articulation balance (i.e., more toward 50%) and the trail score. The vehicles with more balance clearly scored higher. The correlation between these quantities is 0.91 at RTI=750 and 0.93 at RTI=1000 (note that the OME was not included in this last figure since it did not obtain RTI=1000). These are very high correlations and suggest that, despite the small spread in the scores, there were perceptible differences in vehicle behavior. The statistical level of significance of a 0.93 correlation at RTI=1000 for 5 data points is less that 1%. This indicates that the probability of obtaining such a high correlation if there truly was no correlation is less than 1 in 100. We thus conclude that the correlation between articulation balance and trail score is statistically significant.

Table 4.2. Scoring for Section 2

Table 4.3. Scoring for Section 3

Table 4.4. Summary of Results

Figure 4.1. Final Trail Score as a Function of RTI

Figure 4.2. Trail Score as a Function of Suspension Balance at RTI=750

Figure 4.3. Trail Score as a Function of Suspension Balance at RTI=1000

5.0 Drive and Ride

The Drive and Ride tests were used to evaluate the on-road behavior of the various suspension modifications. Three judges drove the test vehicles (Jim Allen, Bill Burke, and Tom Collins). These three judges have extensive experience with Land Rovers and are very familiar with the handling characteristics of the stock D90.

The road course was developed by Jim Allen and was 16.8 miles long. It included 5 miles of freeway (75 mph speed limit) and an off-camber off-ramp. There was 10.8 miles of mixed city stop and go driving, with turns, stop lights, and a variety of road surfaces. Speeds ranged from 25 mph to 50 mph. There was one mile of rough dirt road that included gravel, major and minor potholes, and corrugations. The potholes were useful for testing bump steer and directional abilities.

At the beginning of the test day, the intent was that each of the three judges would drive all six modified vehicles. To do this, the three judges would caravan three vehicles through the course. The judges would return to the staging area, switch drivers, and go out on the course again. This would be repeated once more for the first three vehicles. Three additional rounds would be required for the second wave of vehicles. The course was designed to require approximately 30 minutes to complete. Approximately 1.5 hours was required to test 3 vehicles and 3 hours was required to test all six vehicles.

Unfortunately, during the first cycle for the second three vehicles, the 3-link suspension of SG2, failed, allowing the axle to toggle under during the pitch stability test. Pitch stability was evaluated by braking the vehicle moderately hard (no wheel lockup) from approximately 35 MPH to zero on pavement. The results of the suspension failure allowed the front axle to roll under, damaging the drive line and shocks, causing wheel lock up, and ultimately rolling the vehicle onto its side. No one was hurt. The delay that ensued in up-righting the vehicle, responding to the emergency personnel who arrived at the scene, and returning to the staging area meant that the tests could not continue as planned. As a result, we had to rely heavily on single judge’s opinions for the second wave of vehicles.

The judges met at the end of the day to reach a consensus of the various drive and ride scores and to evaluate how the failure of the SG 3-link front suspension should affect the scores for the two SG vehicles. It was decided that the SG 3-link front suspension was not sufficiently developed to classify it as a commercially ready design. Because of the impact of this failure on safety, the judges decided to assign a score of zero to the pitch stability for both of the SG vehicles. The performance of the SG vehicles for the remaining 5 categories were judged without consideration of the failure.

The scores for the individual categories are listed in Table 5.1. The "Drive Overall" and "Ride Quality Overall" are simple averages of the three sub categories for those categories. Judges’ comments are listed in Table 5.2. A full set of comments was not generated due to the failure of the 3-link suspension. Note that the DR and RW scored highest on the drive section and OME and DR scored highest on the Ride section. Overall, DR took top honors on the Drive and Ride tests.

Table 5.1. Summary of Drive and Ride Results

Table 5.2 Judges’ Comments on Drive and Ride

6.0 Overall Results

The overall scores are presented in Table 6.1. The road and trail tests each contributed 40% to the final score with ramp results contributing 20%. The top two overall scorers were RW and DR. Note that there is only a 2 point spread between these two vehicles. OME scored third overall and tied for third on the trail. This suggests that very little modification was required to the suspension to make it competitive with the other vehicles for the trail tests performed here. The three vehicles which were in the prototype class scored fourth, fifth, and sixth.

Table 6.1. Overall Results

7.0 Discussion

Introduction

The field of rock crawling is in its early stages and we have a way to go before we know what modifications works best for our vehicles. However, tests like this one provide quantitative information to accelerate our progress. To my knowledge, this test is unique in that we are using fairly rigorous methods to quantify the correlation between observable behavior on the ramp and off-road performance. We were fortunate in that we were able to call on the help of very experienced D90 drivers to assess the changes in the on-road performance of D90s due to these modifications.

The modifications to the 6 fully tested vehicles represented a much broader range of approaches than I anticipated prior to the test. Because of this range, I feel that the data is adequate to support some generalizations on suspension design.

In this section, I will present some engineering observations on the performance of the various approaches. My hope is that these observations will promote discussions within (and outside) the D90 list concerning future developments in suspension designs for rock crawling vehicles. The discussion will be clearly biased toward the off-road performance (at rock crawling speeds) since the D90 list was originally created for the more hard-core off-road D90 driver.

RTI

The quantitative analysis presented in the previous section indicated that there was not a strong correlation between maximum RTI and the trail scores. Since competitors were allowed to use lockers, stability and clearance increased in importance relative to traction. The analysis did show a strong correlation between trail performance and suspension balance (i.e., the front and back axles build articulation equally).

Suspension Balance

To understand the effect of suspension balance, one can compare two hypothetical vehicles. Assume a vehicle with an unbalanced suspension generates all of its articulation from the rear axle and none from the front. Assume a second vehicle with a balanced suspension generates 50% of its articulation from the front and 50% from the rear. Now consider what happens when these two vehicles transverse a section that requires climbing a rock on the right side on otherwise flat ground. Also assume that both vehicles have sufficient articulation that all 4 tires remain on the ground. The unbalanced vehicle with the rigid front end will generate approximately twice the body tilt when its front tire climbs the rock than the vehicle with the balanced suspension. This is because the front axle of the unbalanced vehicle does not articulate relative to the body. When the front wheel drops off and the rear climbs the rock, the unbalanced vehicle experiences no body roll whereas the balanced vehicle experiences the same body roll as when its front tire was on the rock. Generally both vehicles will be able to transverse the rock on flat ground since the rock is within the articulation limits of the vehicles (i.e., all wheels remain on the ground). If this rock is part of an off-camber section that already has a vehicle half way towards its limits of tip over stability, then the additional roll angle caused by a rock can lead to tip over. In such cases, a balanced vehicle will generate less additional roll, allowing it to transverse a significantly taller rock without tipping over since it generates less maximum roll angle during the maneuver. Compounding this is the effect of the angle of the trailing arm of the rear axle.

Trailing Arm Angle

There are several forces that act between each side of the axle and the chassis. These are the spring/shock forces and the forces associated with the trailing arm. All of these forces act along the components (along the spring, along the shock, and along the trailing arm). If the trailing arm is not parallel to the ground, then there will be a vertical force component on the chassis due to this angle. When under power, the arm will be in compression. If the arm is sloped upward (i.e., the wheel is drooping), the force vector will have an upward component. This will cause that side of the vehicle to lift somewhat. With the additional lift, the wheel can drive under the vehicle, increasing the angle more, further increasing the upward forces on the chassis. While all of this is going on, the cg height of the vehicle is increasing. Normally these effects are not significant. However, if a vehicle is climbing an obstacle on a side slope, the vehicle may already be near the limits of its roll over angle. Under these conditions, it does not take much force along the trailing arm to further tilt the vehicle. Eventually, one gets to the point where it feels that all of the drive torque is going into tilting the vehicle and none is going into moving it forward. At this point, the only option is to back down and try a different line (or for the really brave, back down and use some momentum).

How can these upward chassis forces be reduced? One approach is to allow less wheel droop (and hence, less articulation) on the back (OME, SG2) so that the rear wheel cannot climb under the vehicle as much. This can be accomplished either through retained rear springs of sufficient stiffness, or limited axle movement due to shorter shocks. We can also use longer rear arms as this reduces the angle of the trailing arm relative to the ground, reducing the upward component of force on the chassis.

Another approach that sometimes works (this was the case on the last obstacle on Section 2) is to build in more articulation in the front. This will result in less body roll to begin with (SG1 and SG2), reducing the angle of the trailing arm.

How Much Articulation is Too Much?

The answer to this question depends on the trail and the vehicle. When lockers are not used, articulation can have a very large positive effect on traction. However, the use of lockers tends to lessen the traction advantages of articulation. What articulation does provide, when lockers are used, is an improved ability of a wheel to climb a rock. This is because the traction force on the rock, required for the climbing wheel, is less due to the easier upward movement of the wheel, and because the remaining wheels can push the vehicle harder into the rock, generating more traction force on the climbing wheel. Articulation improves off-road performance if there is not so much of it as to negatively effect the stability of the vehicle. Vehicles that spend most of their time in canyon bottoms (such as 21 Road) can probably afford more articulation than those that spend much of their time on off-cambered trails and steep articulated climbs. Vehicles with longer trailing arms, lower cgs, and wider tracks (i.e., buggies) can also possess more articulation. Vehicles with balanced articulation front and back can also afford more total articulation than those that generate most of their articulation from one axle.

Lift

The most direct and important effect of suspension lift is the lift raises the cg height, decreasing the lateral acceleration (or tip angle) at which the vehicle rolls. The high cg (relative to the track), combined with the short wheel base, is the primary reason for rollovers of sport utility vehicles on the road.

Another effect of the suspension lift is it changes the kinematics of the suspension. Raising the front can increase the bump steer because the drag link is no longer parallel to the ground, and decreases dynamic stability because caster is deceased. Most people understand the effect of these changes. A more esoteric change is the possible change in height of the roll centers for each axle. A roll center of an axle is the point that the body appears to be rolling about, in the same vertical plane as the axle. Standard practice is to design the suspension such that the rear axle roll center is higher than it is in the front because this improves straight-line stability of the vehicle. When a D90 is lifted, the rear axle roll center stays the same height (i.e., at the center ball joint), but the front roll center is raised because the center of the Panhard rod also raises. The result is a decrease in directional stability, resulting a bit more darting about on the road in response to road irregularities.

A secondary roll-center related effect is as the chassis is lifted, the average roll center height is raised, but not as much as the cg height. This results in a longer vertical lever arm between the cg and the roll centers. The driver feels this as additional body-roll in corners and off-cambers.

Off road, lifts provide 1) additional body and chassis clearance through boulder fields, 2) the ability to use larger tires to improve differential clearance and increase traction, and 3) increased approach, departure, and break-over angles. Another advantage is one can generate more upward wheel travel before hitting the wheel wells, allowing more articulation without further increasing the cg height of the vehicles due to articulation. There are several disadvantages of large lifts on off-road performance. One is the increased tendency to produce body roll because of the longer vertical lever arm between the cg and the roll centers. One can control dynamic body roll through stiffer shock valving, and dynamic and static body roll through stiffer springs. Unfortunately, stiffer springs and shocks reduce ride comfort, and stiffer springs reduce articulation. Other more serious disadvantages of large lifts are the decreased angles at which the vehicle is stable and the increased weight transfer to the down slope wheels. Vehicles tend to climb better if they can keep as much weight as possible on the front tires. Finally, lifting a vehicle increases the angles in the trailing arms, which leads to the jacking effect discussed earlier.

The 3-Link Front Suspension

Is the 3-link suspension a good thing? A clear advantage of the 3-link front suspension is the design does allow articulation to be increased in the front. This in-turn, allows the articulation to be more balanced front and back which can increase vehicle stability when rock crawling. If designed properly, such a design can also lead to more precision in handling because one can now use much less flexible bushings in the radius/torque arms at the axle.

However, there are downsides to the 3-link design from an engineering point of view:

1. It is difficult to find room for the third link because of the engine. SG’s approach is to put the 3^rd link below the two radius arm links.
2. Under hard braking, a third link located below the two radius arms must withstand approximately 3.3 times as much force along its length as "both" the radius/torque arms on the stock design (depending on the diameter of the tires and the vertical separation in the axle mounting points). This is 6.6 times as much axial force as each stock link sees! Since this is a compressive force, a damaged lower third link can easily be buckled (we see this happen to the rear trailing arms under much lower loads than will typically occur at the front axle during braking).
3. These larger forces must to be carried by the axle mounts and the chassis mounts. The Rover uses well-trussed frame mounting points to support the lesser forces of the stock radius arms.
4. Because the forces are significantly higher, fatigue failure is more likely unless everything is properly designed.
5. The larger forces of the 3-link require stiffer bushings in the 3-links to maintain handling. However, stiffer bushings also transmit more shock loads to the chassis. This will accelerate fatigue effects relative to the stock design and decrease the overall shock loads that the design can withstand.
6. Finally, the roll resistance generated by the bushing in the stock front axle acts as an anti-roll bar, especially at larger body roll angles. This has a stabilizing effect on the road. A 3-link suspension will not have this stabilizing effect and one should either use stiffer springs up front than in the back, or use a detachable anti-roll bar up front to maintain the desired understeer characteristics of the vehicle while on the road.

A well-designed 3-link front suspension has good potential to improve the performance of the D90 off road. There is lots of room available in the front wheel wells of the D90 for upward movement. A design to utilize this room during articulation could significantly improve the performance of the D90 without increasing the cg height of the vehicle during articulation. However, the 3-link suspension must be carefully developed, due to the increased loads, and thoroughly tested under very harsh conditions before such a system can be considered commercially viable.

Acknowledgements and Final Thoughts

I want to thank the D90 group for this opportunity to test suspension modifications, the participants for providing and driving the test vehicles, and the judges for the care they took in helping to set up and score the tests. Hopefully, these tests will provide some food-for-thought as to what can be done to improve the performance of our suspensions. I also hope that the four participating companies (DR, OME, RW, SG) will use their experiences at the Twist-Off to further improve their suspension products while maintaining the on-road characteristics and safety of our D90s.

Vehicle Configurations

Below you will find the vehicle configuration data sheets as provided by the participants.