In 2002, a series of patents was granted to Ellsworth, describing a recipe for constructing bicycle rear suspensions, called “Instant Center Tracking” or ICT [examples include U.S. Patent 6,378,885 and U.S. Patent 6,471,230].
Before the distribution of Path Analysis and in the interest of fairness, we requested from Ellsworth a complete explanation of their ICT theory. Upon receipt of what Ellsworth claimed were the essentials of the ICT theory, we conducted an analysis and submitted it to Ellsworth and their consulting engineer Mike Kojima, so that Ellsworth could have a chance to express their opinions, and perhaps modify their theories and marketing, before we released the information. Our original public characterization of the Ellsworth claims and experimental work was taken almost verbatim from Ellsworth literature and correspondence.
We have since been able to review some of the Ellsworth patents and have found that, although significant conversations and exchanges of documents took place, regarding ICT, Ellsworth was not remotely forthcoming in explaining the details of their theory (we suspect that they feared fully informed and proper scrutiny of their theory). We have reviewed the correspondence and recalled that we continually urged Ellsworth to provide more detail for key elements of the theory, particularly the matters of “squat” and “anti-squat” (see below), if such existed. These calls were largely unmet, with Ellsworth claiming that they had disclosed the essentials of the theory. But again, we have now determined that forthright disclosure was not remotely the case.
All explanations considered, as well as examination of the bikes Ellsworth has produced over the years, seem to indicate that ICT has evolved over the years. Yet there are core problems that remain in all versions. We here present the problems that are common to all versions we have encountered.
Ellsworth marketing claims, on the basis of ICT, that their dual suspension bikes have “Up to 100% pedal efficiency (in every gear, and throughout the entire suspension travel range)”. The “Up to...” phrasing is very confusing, but through inquiries to Ellsworth we understand that it means that the bikes are almost 100% efficient in all gears.
In advertisements for the “Dare” downhill bike, Ellsworth has gone further, claiming that, “The 2001 Dare, with our patented ICT technology (which offers 100% pedal energy-efficiency by isolating pedal input from suspension activity), will out-accelerate, out-pedal and out-climb any full-travel free-ride bike on the planet.” [See Page 22, Mountain Bike Action, May 2001.] No mention is made of gearing at all.
Ellsworth also claims that chain tension is decoupled from other forces on the suspension, so that these forces won't feed back through the chain to disrupt the rider's pedaling
Ellsworth further claims that their suspension is unaffected by braking forces.
We first present and examine the Ellsworth claims for pedaling, after which, we present and examine the claim for braking.
Ellsworth provides the following primary recipe for a 4-bar suspension, which is supposed to achieve their claims for performance, under pedaling:
Ellsworth first determines what they consider to be an average, extended chain line, through which chain force from the pedals passes.
In order to eliminate the effects on the suspension of force applied through the chain, from the pedals, Ellsworth claims that the linkage IC (with respect to the main triangle) should align with, or “track”, the chain line, through all positions in the suspension travel. In the case of the Ellsworth Truth, the deviation from this ideal is said to be within .5 %, at all times (that is, the deviation is small).
Ellsworth then assumes a particular rider/bike center of mass.
Ellsworth supposes two effects of this mass from bicycle forward acceleration, ultimately due to the force at the tire/ground contact point. These are squat (a suspension compressing effect), and anti-squat or jack (a suspension extending effect). Squat is the inertial resistance of the rider/main triangle to forward movement. Anti-squat is supposed to counter this. Ellsworth asserts a desirable range of IC locations along the chain line that is supposed to balance squat with anti-squat. We could find no place in the patents where Ellsworth states how they determined the desirable IC range of locations (third parties have informed us that the numbers were generated experimentally). The desirable range is delineated by a percentage scale.
The acceleration effects of squat and anti-squat, as well as the percentage scale, are partially explained below in a quote from U.S. patent 6,471,230. Figure 5.3) shows “FIG. 6” of U.S. patent 6,471,230, which is the diagram for the following explanation.
Click to enlarge
The Ellsworth explanation is as follows:
The torque interaction between the pedaling-induced wheel driving force and the ground can also cause rider-energy-wasting suspension compression due to a torque moment transferred to the shock absorbing means via the suspension upper rocker arms and lower yoke. To counteract this moment, the suspension has approximately 10-20 percent anti-shock-absorbing means compression (or "anti-squat") built into the suspension geometry. As illustrated in FIG. 6, this percentage may be calculated by drawing an imaginary line through the center of the rear wheel tire contact patch and the "instant center". Another imaginary line is drawn through the bicycle and rider unit combination's center of gravity, perpendicularly to the ground plane. The point where this line intersects the imaginary line from the rear wheel tire's contact patch to the instant center is called the "anti-squat calculation point". The height distance in units of measure of the "anti-squat" calculation point to the ground is divided by the height distance in units of measure from the ground to the bicycle and rider unit combination's center of gravity. This number gives the percentage of "squat resistance" built into the rear suspension's geometry, where 100 percent equals full cancellation and zero percent is no cancellation.
Ellsworth has further stated a belief that an IC moving far out in front of the bike increases efficiency in a wider range of gears.
While some rationale is given in the patents for not having an IC located too far forward or too far back, the lack of any quantitative explanation for how the desirable range of IC locations was determined prevents a full, mathematical examination of errors in the ICT theory. However, this will not prevent us from demonstrating that the theory is nonsense.
So, Ellsworth has identified two types of pedaling effects: the chain force, and the acceleration effects of squat and anti-squat. The idea is that, if squat and anti-squat are made to balance out, then a lack of contribution from the chain force will create a suspension that is both non-reactive to pedaling and free from pedal stroke disrupting feedback.
(Note: one might wonder if chain tension is somehow a part of the anti-squat effect, since the physical components of this effect are never explained. But, again, Ellsworth is clear that one of the main objectives of ICT is that chain tension be decoupled from other forces on the suspension and Ellsworth even takes some pride in claiming that their suspensions do not use chain tension to counter squat, as some other designs do.)
ICT is supposedly based on sound classical physics, that is, the classical laws of nature. In classical physics, all sound theories based on the laws of nature must hold in their limiting cases, since Nature Varies Smoothly. Physicists routinely look at these limiting cases to see if their theories hold up, since these cases are often more intuitively obvious then the general cases, making them very good tests of the theory. And fortunately for us, we have just such cases.
The parallel, pp-coaxial, and wp-coaxial 4-bars, which were introduced in the “The Natural Mirror Bike” section are all limiting cases for the possible configurations of all 4-bars. Ellsworth should consider the wp-coaxial 4-bar to fall under the ICT prescription, since they use the configuration in one of their technical diagrams.
PA and ICT are in direct conflict. So we will subject these two theories to three tests, using our three limiting cases, as well as the calculations done in the “PCL Problems – Some Further Calculations” section, to see which of the theories holds up. After each test, where feasible, we examine the fundamental problem with ICT causing it not to hold up (we will not be able to fully do this in the last case, due to a lack of quantitative explanation for ICT dynamics).