What a trip!
Jerry's always liked to make things go fast, and he's still at it!
one of my trips was to the bonneville salt flats to run a land speed record motorcycle i built. a good friend, drew lippolt, rode the silly thing. it was the first time for this bike and we were looking for +200mph and hand some handling problems that limited the speed to the 186mph - fear of crashing was the spped governor. attached are some pics of the bike. i have two videos but am affraid they may be too large to attach.
|attached are pics of my favorate race bike (Armstrong Rotax 250) and drew in 1998.|
|Bonneville Bike Instability Over Rough Surface
Klawitter, 1 October 2004
Rider Drew Lippolt encountered a sever vehicle instability during speed trials at the Bonneville Salt Flats in 2004. Several potential causes have been be identified:
front swing arm not stiff enough in torsion allowing the front wheel to twist out-of-plane with the rear wheel,
bump steer resulting in steering in-puts while the handle bars remain stationary
free-play within the center hub assemble due to use of large diameter ball bearings or the attachment of the central king pin to the wheel axle or something else.
free-play within the draglink steering assembly
other yet to be identified causes
Swing Arm Stiffness.
There is little information establishing a benchmark for stiffness. Drew and I talked with Denis Manning at Bonneville about vehicle stiffness and Denis stated that in his experience, and he has been making motorcycle streamliners for 40 years or so, high vehicle stiffness was critical for good handling. Denis Manning told us it takes 3,000 ft-lbs of torque to twist the frame of his most recent streamliner 1 degree. Mannings liner has a wheelbase of about 11 feet, which results in a unit stiffness of about 33,000 ft-lbs/degree twist/ft. of frame length. Another reference point is the torsional stiffness of a NASCAR chassis. I found an article (http://www.ces.clemson.edu/~lonny/pubs/conference/sae983054.pdf) reporting NASCAR chassis stiffness to be about 6,000 ft-lbs/degree, which results in a unit stiffness of 55,000 ft-lbs/degree twist/ft.
Eugene Podnos used a finite element analysis (ABAQUS) to calculate the stiffness of the FAMVTE vehicle frame and front swing arm. The FAMVTE frame stiffness was determined to be 14,544 ft-lbs/degree twist/ft., see Attachment 1. Based on this analysis the Manning frame is about twice as stiff as our frame and seems reasonable because the perimeter of the Manning liner frame is larger than the FAMVTE frame since rider Rocky Robinson sits inside the Manning liner frame. Furthermore, Manning liner weights over 1,600 lbs with rider, has a turbocharged 3 liter engine that should produce over 600 hp and it is intended to go 400 mph compared to the FAMVTE bike which weights about 650 lbs with rider, has a 150 hp motor and is intended for speed in the 200 mph range.
Eugenes analysis of the front swing arm determined its stiffness to be 960 ft-lbs/degree twist/ft., which is about 15 times less than the frame stiffness, see Attachment 2. As a check on the FEA result, I calculated the torque required to twist the front axle of the FAMVTE bike one degree assuming the front swing arm members to be cantilever beams. The front swing arms are made of a rectangular steel tube having a base of 1, a height of 2 and a wall thick ness of 0.10. The swing arm members are 8 apart on center where the wheel axle passes though and about 11 apart where the tubes joint the frame. The front swing arm is 24 long. To simplify calculations, and that the front swing arm tube members are 8 apart along their entire 24 length. I assumed the front swing arm is fixed at the frame and made a couple drawings, see below, that show the swing arm tubes will both twist and bend.
As the front axle is twisted one degree each swing arm will twist one degree and bend as a cantilever beam with an end displacement of 0.068. The torque required to twist each 24 long swing arm member 1 degree is 143 ft-lbs (twist angle = TL/GK) and the force required to bend each swing arm member 0.068 (D = FL3/3EI) is 126 lbs. Since the center of each swing arm member is located 4 from the center of rotation of the front axle, the bending forces acting on each swing arm member will result in a torque of 42 ft-lbs (.333 x 126 lbs). The total torque required to twist the front axle 1 degree is the sum all torques = (2 x 143ft-lbs + 2 x 42 ft-lbs) =370ft-lbs. Since the swing arm is 2 feet long the unit stiffness is 740 ft-lbs/degree twist/ft740.
twist angle = TL/GK, where:
T = torque
L = length
G = shear modulus steel, 11.6x106psi
K = sectional inertia, 2t2(a-t) 2 (b-t) 2/(at+bt-at2)
deflection = FL3/3EI, where:
F = force
L = length
E = elastic modulus steel, 29x106psi
I = area moment of inertia, bh3/12
My simplified calculation resulted in a unit swing arm stiffness that is 77% that of Eugenes FEA result. The lower stiffness resulting from my calculations seems reasonable since I did not account from the swing cross member that will stiffen the structure. Based on this comparison, Eugenes determination of the front swing arm stiffness will be considered accurate. In earlier discussions, Matt St. Louis suggested boxing the current swing arm to increase material and dimensions and thus stiffness. The front wheel is very narrow and there is room to increase the width of the current swing arm. The K value for the current 1 x 2 section swing arm members is 0.209. Increasing the section to 2 x 2 results in a K value of 0.686, which will increase the torsional component of front axle twist by 3.3 times. Plating the rectangular swing arm portion between the cross member and the swing arm axle will further increase stiffness. Modification will be made to the front swing arm structure to increase the stiffness by ten times or more.
Lets assume we modify the front swing arm so that the torque required to twist the front axle 1 degree is 10 times greater; 4,800 ft-lbs (unit stiffness = 9,600 ft-lbs/degree twist/ft.). One degree of front axle twist will result in 0.19 side displacement of the front wheel where the rubber meets the salt. The front wheel is 22 in diameter and thus will require a side load of 5,236 lbs (where the rubber meets the salt) to produce a torque of 4,800 ft-lbs. Questions to be considered are: 1) How much out of plane displacement of the front wheel can be tolerated and, 2) how much force can be transmitted through the rubber tire to the metal wheel?
The FAMVTE bike uses a drag link steering system. The steering assembly consists of a pitman arm that moves front to back as the handlebars are turned right and left, an off-set arm attached to the front wheel hub and a drag link with rod end bearings at each end that attach to and connect the pitman arm to the front wheel hub. If the handle bars are turned to the right, the pitman arm moves forward, The rod end bearing connecting the drag link with the pitman arm is located about three inches above the front swing arm axle and the rod end connecting the drag link to the wheel hub is located about two inches behind and above the front wheel axle. As the front swing arm moves up and down over bumps, the front wheel hub moves in an arc having a radius of 24 inches and a center located at the swing arm axle centerline. During swing arm motion, the rod end bearing attached to the wheel hub moves through a arc that is not concentric with the wheel hub arc because its center of rotation of the rod end bearing is displaced from the swing arm center of rotation and the length of the drag link is shorter than the swing arm. The result is that as the front swing arm goes up and down over bumps, the divergent arcs of motion of the wheel hub and the wheel hub rod end bearing cause the wheel to turn the co-called bump steer. In the case for the FAMVTE bike, the wheel will turn to the left as the swing are goes down and right as the swing arm goes up. The idea is to minimize the amount of bump steer so that it does not cause instability. The fact that the bike ran well over a flat surface and became instable over a bumpy surface points to bump steer as a potential cause.
I do not know how much bump steer there is with the current drag link steering geometry and plan to measure it. I intend to do two things: 1) measure the bump steer in degrees as a function of vertical swing arm motion while the handlebars are fixed and, 2) determine how much handlebar motion results as a function of vertical swing arm motion while the front wheel is fixed.
To determine the bump steer in degrees as a function of vertical swing arm motion while the handlebars are fixed, I plan to lock the handle bars and place a dial indicator on the swing arm to measure the side motion (displacement) of the wheel as the swing arm is moved up and down. The bump steer in degrees can be calculated from the side motion displacements and plotted against wheel axle vertical motion. Maximum vertical wheel motion can be limited to say _ 2 and this will define the limits of bump steer possible. The center hub is designed so that the king pin inclination can be varied to change the castor angle of the wheel. Changing the king pin inclination will change to location of the rod end bearing attached to the hub with respect to the wheel axle. Thus bump steer, defined, as degrees of steering per unit of swing arm motion, will be a function of the castor angle of the wheel. Measurements of bump steer will be made for various wheel castor angles.
The question is how much bump steer is too much? One way to answer this question may be to use the rider as the measuring stick. The idea is to fix the front wheel so that the handlebar will move back and forth as the swing arm is raised and lowered and ask the rider if the resulting handle bar motion represents a significant steering input, one that would cause a remarkable change in direction.
Center hub Free Play
The center hub of the front wheel provides the means for the front wheel to steer left/right and rotate about the front axle. Two different bears types are used to establish the axis for steering and front wheel rotation. Steering is accomplished by mounting the center hub on a centrally located king pin that is rigidly attached the front wheel axle. The center hub rotates about the king pin on bronze bushings. Wheel rotation is achieved by mounting tow large diameter (4 I.D.) single row ball bearing around the outside diameter of the center hub. The front wheel is mounted on the two ball bearings allowing it to rotate about the rigid front axle. Ideally, both of these bearing (bushed king pin and ball bearing) are sufficiently tight to not result in free play that could cause wheel wobble. A dial indicator will be placed on the outer most circumference of the wheel and used to measure free play as lateral force is applied to the wheel. Engineering judgment will be used to determine if there is too much free play.
Drag Link Free Play
Front wheel steering is achieved by connecting the center hub to the handlebars by means of a drag link and rod end bearings. Ideally, the connection of the handlebars to the front hub is tight enough so that no steering motion occurs when the handlebars are held firm. Free play in the multiple connections of the drag link steering assembles could result in a steering motion of the wheel even though the handlebars are not moved. The handlebars will be rigidly fixed and a dial indicator used measure any wheel steering motion as the wheel is forced right and left.
A careful examination of the entire front frame and wheel assemblies will be conducted to look for other causes of directional instability.
"building bodies and bikes for a better tomorrow"