Custom Bearings for Hostile Environments
Run Cooler Without Lub
Run Faster With Less Wear
Champion Bearings Inc.
Fax 760-778-5184
Wind Turbines
Aero Space

Tests Reveal Significant Advantages
In Hub Performance
By Richard Kay
  Want to finish more than 100 yards ahead (in a 100-mile race) by simply changing the ball bearings in your hubs? Sounds too good to be true, but it’s possible with today’s new bearing technology.

  Available for more than 100 years, the ball bearing has recently benefited from several major technological innovations. As an engineer and manufacturer of ball-bearing systems for industrial use, I was naturally interested in their possibilities for the bicycling market. What follows are the results of a scientifically precise comparative experiment designed to test the utility of these new bearing systems in popular bicycle hubs from Shimano and Sun- Tour. The experiment is repeatable by any trained researcher wishing to verify (or disprove) the results.

  Part of this new bearing technology uses fiberglass-filled PTFE seals for bearings that offer numerous advantages over rubber or metal seals, including a lower coefficient of friction, excellent lifespan and, when applied correctly, significantly reduced bearing-resistive torque and achieved increased seal-ability (Figure 1).

   What has been learned from tribology (the science of friction and wear) can be applied to the bicycle in the form of the ion deposition process. This procedure uses vacuum and an inert gas plasma with high voltage to coat bearing surfaces. It produces an ultra clean bearing surface because of argon ion bombardment just prior to applying a thin lubricant coating. Excellent adhesion is obtained by driving the vapor atoms into the ball bearing metal surface by intermolecular adhesion. This is not a line-of-sight process; it covers components with complex geometry. Of the materials tested, lead was found to provide the lowest torque.

The ball bearing test apparatus. The drive mechanism consists of a DC motor mounted on 1/2-inch- diameter linear bearings. The motor is engaged and disengaged by a mechanical feed-through. The bearings are mounted on four 1-inch-diameter vertical supports, three of which are visible.
Pressure inside the vacuum was 25-inches HG, plus or minus 1/4 inches.


  Inquiries concerning the feasibility of this experiment produced comments such as: “It’s impossible to measure the difference between precision bearings at speed in front and rear axle hubs,” and, “The bearing torque is insignificant compared to everything else.” A challenge was in the offing!

  A preliminary bench test showed that by simply changing the seal material and configuration in just the front and rear hubs, a significant performance increase could be realized, and that something so random as a speck of sand could possibly decide the outcome of a 100-mile bicycle race. An energy analysis equation presented by Chester Kyle, Ph.D, in the August ‘87 BIKE TECH shows how, as follows:

F = W x [(Crr + sin (arctan (S)] + A1V) +1/2PXCXAX(V+Vw2)
F = Net retarding force on bicycle (Ibi)
W = Total weight of bicycle and rider
(186 lbf)
Crr = Tire rolling resistance coefficient
S = Slope of course (assumed 0 for
comparative purposes)
A1 = Variation in rolling resistance due to
velocity (.0000667)
V = Bicycle velocity (feet/second)
P = Air density slugs/Ft3 (.00233 @70
degrees F)
Cd = Drag coefficient (0.9)
A = Projected frontal area of bicycle and
rider (4.5 feet2)
W = Wind velocity (assumed 0, for
comparative purposes)
This equation reduces to the following:
F = WCrr + A,WV + 1/2PCd AV2 +
F’ = WCrr (0 velocity rolling friction)
A1WV = Rolling friction at velocity correction factor
1/2PC AV2 = Pressure drag
Fsg = Retarding force of front and rear axle bearings
= Bearing torque/wheel radius

  The question we were anxious to answer was: What effect do a variety of parameters have on bearing torque; specifically, seal material and design, type of ball retainer, and lubricants?

  All that remained was to accurately measure the bearing torque. Repeatable results were obtained slightly more than a year ago using a rotating disk of known mass and inertia, and a chopper disk attached to the rotating disk passing through a stationary GE H21A2 light-emitting diode (LED).’ Once every revolution, the chopper disk breaks a light beam in the LED. The signal, after passing through a metrabyte P10-12 digital board, provided a graph of torque vs. rpm. The input signal, a first derivative of rpm, was processed with an IBM PC through a series of computations written in the programming language PASCAL.

  Once the test flywheel was taken up to speed (400 rpm) and the external driver disengaged, the computer accurately noted the change in rpm and printed out the rpm vs. torque data for the test apparatus. The torque values were calculated from the T = I a equation, and the rpm taken between subject calculations. Torque values were recorded at 25- rpm intervals from 400 down to 225 rpm.

  The calculations2 suggested an inertia mass of 17.36 in.-lbs. sec 2, with a weight of 105 pounds. The idea of a rotating mass in a partial vacuum offered a fair compromise between variations in lubrication characteristics in a vacuum and test repeatability.3 This was because of daily barometric pressure fluctuations and associated windage and friction considerations. Therefore, as a result of size restrictions (the inside diameter of the available vacuum chamber was 20 inches), the inertia load could be only ‘/4 of the load of a bicycle and rider. But with a radial load of 105 pounds and a rotational speed of 16—32 mph, the most significant performance variable could be approximated.

  The drive mechanism consisted of a DC motor mounted on 1/2-inch-diameter linear bearings while the motor was engaged and disengaged by a mechanical feed-through. The bearings (Figure 2) were mounted on four 1- inch-diameter vertical supports, three of which are visible in the photo. Pressure inside the vacuum was 25-inches HG, plus or minus 1/4 inches.

  For comparative purposes, the following configurations were tested. The Sun Tour Cyclone hub was tested with three types of bearing systems:
• Abec 1 precision, 6001-size Conrad ribbon steel retainer bearings with two rubber seals (labeled as Sun Tour No. 2 in the table).


FREE Search Engine Submission - Submit your site to dozens of top search engines for FREE. No strings attached!

Sun Tour No.2 0.1438 278 Baseline
Shimano Dura-Ace 0.1248 289 +96
SunTourNo.1 0.1164 302 +151
CHAMPION TST-L 0.0608 308 +452
Average of three tests from 400 to 225 rpm.

• Above configuration with one rubber seal (Sun Tour No. 1).
• The Champion TST L bearing.

  This is a phenolic retainer, abec 7 precision bearing treated with several-thousand angstroms of lead. It also incorporates one PTFE seal on the outboard side (Figure 3).
The stock Shimano Dura-Ace hub was then tested for comparison purposes, with its two piston-ring-type seals per side, lubricated with Bufishot grease.

  The results of the test are shown in Figure 5. The Champion TST-L bearings consistently offer the least amount of resistant torque for the complete range of revolutions per minute.
Each of the hub, seal, bearing combinations shows a rough linear increase in torque with rpm change, except for the Shimano DuraAce configuration. Above 350 rpm, the DuraAce displays a nonlinear torque increase (this hub uses no retainer to fbc the bearing at constant relative distances). This increase maybe caused by ball-to-ball contact (Figure 4). In general, loose ball bearings are not as efficient as ball bearings in retainers, especially at higher speeds.

A Question of Torque

  What does the torque of just the front and rear hub ball bearings mean to the bicycle and rider?

  And more important, what is the additional distance that a more efficient (i.e., less resistive) bearing, seal, hub combination, would a!low you to travel while exerting no additional energy? This distance can be determined by the following equation:

Fsg= TsgIRw
Tsg the measured bearing friction at the specific measured velocity
Rw = the radius of the wheel
Assume that all energy conserved from the
more efficient bearings goes to overcome pressure drag. Thus:
Vnew= (2(2Fb — Fnew)/(P Cd A))
Vnew = new velocity with improved antifriction bearing system
Fb = baseline resistance forces
Fnew = new antifriction bearing systems resistance force
Thus, the distance advantage obtained from a better antzfrzction bearing can be determined from the following:
AD = race distance x (1 — (Vb/Vnew)) where:
Vb = baseline velocity

  Outgasing of the lubricant in vacuum was not considered to be significant when inspection after testing showed the grease or oil in place.
Keep in mind that, in general, the bearing torque becomes less significant (relative to other frictional considerations such as wind drag) at higher speeds and more so at lower speeds.
Life testing of ion-treated spacecraft bearings has recorded many years of unattended bearing operation.

  More down to earth, bench and field tests have documented more than a 1,000-mile bearing life to date, with no loss in performance (Figure 6).
Contributing factors such as the type of seal, lubricant, and retainer construction can figure significantly in ball bearing performance.

  The results of these tests clearly show a difference in performance with the three types of bearings tested. Precision lead-coated, PTFE-sealed hub bearings could easily mean the difference between winning and losing in a bicycle race.

Technical assistance by Kyle K. Kinoshita, MSME Stanford University; Jeff S. Nelson, computer systems operator; D.L. Blick, photographer

June 30, 1995
MIT entry fights off rain to win 1,150-mile solar-cell car race
DENVER - Ending a sojoum through the Colorado wet, a solar car designed, built and driven by students from the Massachusetts Institute of Technology won the Sunmyce 95 road race yesterday.

Richard Kay (left), CEO and Engineer from Champion Bearing congratulates MIT team captain and driver Goro Tamal on his victory.

  It crossed the finish line of the Sunrayce at midday in Golden. The race is an annual competition among college teams to fabricate the fastest and most efficient solar-cell-powered electric car.

Race started in Indiana

  The 1,150 -mile race, involving 38 cars, began June 20 in Indianapolis and ended at the National Renewable Energy Laboratory in Golden. The MIT car averaged 36.74 mph and drove across Indiana, Illinois, Kansas and eastern Colorado in 33 hours, 37 minutes and 11 seconds. The MIT student drivers were Eric Gravengaard, Goro Tamai, the team captain and Milton Wong.

  All drivers ended Wednesday’s leg at Aurora’s Hinkley High School on Chambers Road. Yesterday, they began leaving Hinkley at 10 am., circling through the south and west suburban areas back to Golden.

  The cars arrived throughout a 4- hour period. Some arrived at relatively high speeds, but a few were pushed across the line - victims of the cloudy weather and the inability to fully charge the vehicles’ solar-powered batteries.

Wheel bearings made a difference.

  Richard Kay, CEO and design engineer for Champion Bearing congratulated Goro Tamai of MIT on his victory.

  The wheel bearings were modified and supplied by Champion over a year ago and have undergone extensive evaluation by MIT and Champion.

   Team captain and driver Tamai stated that “The bearings have been performing fantastically for the past few months” He said that the bearings have made a significant difference in the performance of the vehicle.

All six ball bearings, 2 per wheel, had an exclusive PTFE seal and grease less lubrication by Champion Bearing.

Champion Bearing © . All rights reserved