Throttle stop mystery

As Mikael says, it's the vibrations wot does it … I'd be very interested to hear any theories as to why the result in this case is screwing in, rather than screwing out in accordance with Sod's Law.

New parts are on order from Burlen; in the meantime some dabs of rubber solution applied to bridge the ends of the screws to the carb body have proved effective.

Chris,

To the best of my knowledge the carbs are as old as the bike so, since it's a '72, nearly half a century !

But there's no sign of the slides sticking, hot or cold (they are Burlen's current hard-anodised type, btw) — the problem, bizarre as it seems to intuition, is definitely the throttle stop screw (it is mainly the r/h one) screwing itself in.

Executive summary: the air pressure differential between the two ends of the screw is the explanation.

It is clear that it is engine vibration which drives the screw’s movement, in the absence of the braking effect provided by the O-ring. The problem is

(A) why does this cyclical motion result in an overall trend in one direction, rather than a random walk around the original setting?

(B) why is the trend in the direction of screwing in, when Sod’s Law predicts screwing out and being lost?

In what follows, we assume for simplicity that the carburettor’s motion is strictly vertical, that the screw is vertical in the carburettor body, and that the effect of gravity can be neglected, and that the throttle is open and the slide not resting on the stop screws.

1. The lower end of the screw is exposed to atmospheric pressure, while the upper end is exposed to the inlet tract which is necessarily at below atmospheric pressure. Thus there is a persistent net upward force on the screw when the engine is running (remember we are neglecting gravity).
2. Now consider what happens when the carburettor is about to change direction at the upper limit of its oscillation. There is necessarily some clearance between the thread on the screw and that on the carburettor body. As the upward motion of the carburettor ceases, the momentum of the screw plus the net air pressure on the screw will tend to force the screw thread against the underside of the carb threads. Wedging of one thread against the other will tend to drive the screw to the right and thus turn it clockwise.
3. At the other extreme of oscillation the screw’s momentum now forces its thread against upper side of the carb thread. However, in this case the atmospheric pressure now counteracts the momentum, so although the wedging action resulting from momentum will tend to turn the screw anti-clockwise the effect is smaller. Thus the net rotation of the screw will be clockwise.

While the net effect per cycle may be very small it can nonetheless quickly produce substantial visible effects. If a gentle test run involves an average engine speed of 2000 rpm, then a five-minute ride means 10,000 cycles. A net rotation of a tenth of a degree would be enough to generate three turns of the screw (in other words 1080º) which was the order of change that I was experiencing.

Experimental verification would seem to involve building a test rig that would oscillate the carb, with a facility to vary inlet tract pressures ± atmospheric. If runs with the inlet tract above atmospheric consistently unscrewed the throttle stop I'd take that as confirmation.

 Lacking the time or facilities, however, I'm just waiting for Burlen to supply some new parts.