Lotus Seven Sway Bars

By Kiyoshi Hamai

Chapman Report - May 1984

So what new demon tweak did winter bring? Well, next none… Cuz I was so busy… HOW Busy was I? I was so busy that… Anyway, I did finally take a few days to do a little wrenching just before the Feb 18th autocross. The first project was controlling chassis roll in the corners and the second was horsepower!

Dick Rasmussen was adept at finger to snap some quick pixs of me at the first 1984 autocross. INTERESTING! The Seven on racing slicks was LEANING! I do mean LEANING! How much? IT was SO leaning that the roll bar had scratches where a roll bar shouldn't! Well, not quite that much, but how about 5 degrees? How about being nearly fully on the bump stops? That was a problem. Why? First, being out of suspension travel in a corner is a no-no because if the suspension is bottomed out and you hit a bump there is no chassis compliance and the result is the car leaps up (thus a great reduction in adhesion, adhesion is the stuff that keeps a car from flying off the road…) and moves sideways rapidly (that's another way of saying spinning…) Second, in chicanes the car is having to move from 5 degrees in one direction to 5 degrees in the other direction. Fully that is 10 degrees of movement. That takes time, and perhaps more time than it would take to get from one turn to the next. That means the car would be all out of shape and leaning the wrong directions as it entered a corner. The third reason is optimizing the tire patch. In such roll conditions in a Seven there is significant camber change in the front, enough that it lifts a portion of the tire patch off the pavement. The roll in the rear is not of great concern since the rear is a live axle.

The solution to the situation can be remedied in a number of ways. I chose to use a stiffer front sway bar and the addition of a rear bar. It was desirable to make these bars adjustable for roll stiffness as it was difficult to determine exactly the proper balance in stiffness to get the desired handling. On paper a hollow thick wall 5/8 inch front bar (notes: I eventually ended up using a solid 5/8 bar after breaking 2 tubular bars in a short period of time) with 6 to 7 inch arms looked good. In the rear a solid 1/2 inch bar about 33 inches in length in conjuction with 7 inch arms would, with the stiffer front bar induce a bit more oversteer but reduce the roll (theorically) to about 2 degrees.

Below is a diagram of the location and construction of the rear bar. The material used for the bar was 4130 steel and wild steel arms. The front bar is hung below a new 1 inch square tube which spans the front of the chassis. This new cross member is bolted to the chassis using the original mounting points for the original sway bar mounts. On the ends of this new cross member are pick ups for a new adjustable length link. This link goes to the upper track rod and has rod end bearings on each end. A clevis is bolted to the track rod and retains the new link. Caster can be adjusted by changing the length of the link. In this manner the original functions of the original sway bar are divided: Sway bar and Upper A-arm link. Why all this stuff? Simple, the original sway bar was compressed every time the suspension made even the slightest change in ride height. Thus movement of the suspension created bind in the system. The use of a stiffer sway bar in the original design would only resist free movement to a higher degree.

The results were impressive. The car is much more neutral and predictable. The understeer has been reduced unless one attempts to stop and turn at the same time. The sudden transfers to oversteer are gone ( an indication that the car was on the bump stops) and there is a great reduction in rear inside wheel spin. This reduction in rear inside wheel spin is a blessing, previously I had to wait for the wheel to re-establish contact with mother earth before any forward forces could be felt. This meant I was only coasting around the corners in those situations. Now, inside wheel spin is only had when cornering very hard (on slicks, no wheel spin with street tires) and I go over a change in elevation ( crest or bump) which causes the car to go light. The bottom line is carrying more speed out of the corner.

The second big change came with the decision to put the full race cam that has been sitting on the bench since the car was finished. Originally my plans were to put this cam in the engine when I built it. But, at the time I was talking with lots of folks who advised that a full race cam would be lopy at idle, powerless below 3000 rpm and too much power above redline. Based upon these opinions I put the a "hot" street cam (Ford GT) in the engine. The result was a stump puller engine that ran out of breath at about 6000 rpm. The idle was so-so at best and once above 4500 rpm things got slow. Why? That was a good question. The theory is that below 3500-4000 rpm the head having been ported and polished and containing some very large valves, created a situation where the velocity of the gas charge was too slow, thus the lumpy idle. But, then when into the RPMs where the head began to work the cam was giving up. Result, so-so performance across the board.

So, the new full race cam was installed (without engine removal). The result? Well, the engine seems to idle better… and as for pulling at low revs… it's acceptable (it grumbles a bit and certainly not a stump puller, but it is quite strong after 2000 rpm and then once above 3500… Well, shall I say that redline comes awfully fast as does 80 mph from a stand still!

So, the result is much move pleasing and much more potent at the autocrosses. As for my assault on the Corvettes… well, let's just say that in the prior events I was doing my best to get within 2.5-3.0 seconds (2 laps). On the most recent event (after all the changes to suspension and engine) on a course which one would call the typical Corvette course (lots of straights and few turns), the time differential was only 1.5 seconds. So, if given a handling course and some fresh tires…