Subject: Sturmey-Archer 3-Speed Hubs
From: Jobst Brandt
Date: December 23, 2002
Sturmey-Archer hubs have been in service for many years (1887):
Yet they have had a design flaw from the beginning that escaped scrutiny through most of the popular life of the common 3-speed AW hub. This flaw has consistently been "swept under the rug" or laid at the feet of the mechanic so completely that few have questioned why it jumps into free wheeling when ridden forcefully in top gear. I think the symptom and cause should be explained to prevent injuries.
Drive is transmitted from the sprocket through a four slot driver, through which a small movable cross (clutch) protrudes to transmit drive in, low, direct,and high gear positions.
In low gear, the clutch is moved to the right to lift the (high gear) pawls, driving the ring gear directly so that the ring gear drives the planets with output going through the planet cage and its (low gear) pawls at 3/4 the input speed. With one pair of pawls raised and the other pair driving, the typical clicking sound of AW hubs is absent.
In second gear, (direct drive), the clutch drives the hub body directly through the right hand pawls, making the planet carrier (low gear) pawls click as they lag behind.
Top gear engages when the clutch is fully extended to the left against the planet cage, between the four protruding planet (pinion) pins, to drive the planets that turn the ring gear 4/3 the input speed. In this position, the low gear pawls in the planet cage click as they lag behind. It is the inverse of low gear and hence the reciprocal relationship between low 1/3 down and high 1/4 up with respect to direct drive on AW hubs.
Because the pinion pins are free fit in the housing, they are slightly skewed to the hub axis due to canting within the clearance when loaded by the clutch, a "tilt" that has a disengaging bias to the driving clutch. However, because the axle bends slightly from chain tension, depth of engagement between clutch and pinion pins varies during rotation. These two effects can disengage the clutch and pins under high torque, dropping the mechanism into free wheeling forward. The result is that the rider, if standing, dives over the bars, with the bicycle following.
This condition is apparent upon examining the clutch and pins that both show wear, slanting to enhance disengagement. However, replacing these parts does not resolve the condition. SA has always maintained that the shift cable was misadjusted, something that is easily disproven by disengaging the shift chain entirely. With the cable disconnected, the clutch is free to make perfect contact with the face of the planet cage, the best adjustment possible for top gear, and still disengagement occurs.
This problem could have been resolved by putting a slight flare to the ends of the planet pins and a similar matching slant on the clutch faces, giving their engagement a preferential retaining force instead of the opposite. Most motorcycle gear boxes use such features, especially in older non-synchronized sliding gear boxes... the classic clunk of BMW boxes for instance.
Similarly, the spring-less ratchet of the SW (Silent) hub was sensitive to lubricant viscosity and with anything more than 10W oil could freewheel forward, the pawls clinging to the ramps by oil viscosity while not engaging. This hub was discontinued after a short run probably because one could not place blame on user error.
Jobst and I tend to disagree a bit on this topic. While what he says is true, technically, the problem he refers to only afflicts riders who pedal very hard for extended periods in top gear. Most riders of these hubs do not have that sort of riding style, and find the Sturmey-Archer AW to be possibly the most durable and reliable bicycle gear shifting mechanism ever.
More Articles by Jobst Brandt
Next: Shimmy or Speed Wobble
Previous: Electronic Shifting
|Articles by Sheldon Brown and others|
Last Updated: by Harriet Fell