I'll tell ya what I see: no rollers. Pins rubbing (wearing) both sides of the tooth. Including the back side, which will never exert any forward longitudinal force on a chain unless it's made of magnets. I also see a highly suspect FEA diagram on the right, where teeth not even engaged still show non-zero stresses (lighter blue, not the darkest blue), indicating that the color/stress scale has been changed compared t…
I don't know anything about FEA, but in reading this article and thinking it through, was the first time it hit me how bad a conventional chain and sprocket is. On a brand-new set with perfect tolerances, you might be able to use many of the front side of many of the teeth, but as SOON as the chain begins to stretch (wear) even fractionally, most of the stress goes to the 'first' teeth in any given power stroke, and…
In a worn chain, the rollers engage high up, and then work their way down (under high load) until seated at the base. In the unworn condition, engagement begins at the base of the tooth where the forces are better spread over the roller.
In point of fact, the forces are spread out over more teeth when worn. The first tooth engagement might begin 3mm high, then 2.5mm, then 2mm, etc. over ~6 teeth. Whereas the fresh chain might manage 2 or 3. But the wear and friction of the worn chain is much higher as you have a great deal of motion under high load as the loaded rollers creep down. The unworn chain has perhaps higher loads, but the roller is essentially seated and immobile at the base of the gear for this.