british-cars
[Top] [All Lists]

Re: Speedos

To: british-cars@encore.com
Subject: Re: Speedos
From: garnett@theory.TN.CORNELL.EDU (Roger Garnett)
Date: Wed, 13 Feb 1991 09:35:53 EST
Dick Nyquist (and others) wrote:
} Subject: Re: Speedos
} |  
} | > > ????????????? how ???????????
} | > > one turn of a sticky cable is still one turn.
} | > > A sticky cable can cause a speedo to "bounce"
} | > > but the cable will not turn faster.
} |  
} | What may happen is that the binding cable will cause cycles of fast and
} | slow revs, as the cable twists enough to overcome whatever is binding it.
} | While this may cause the needle to bounce, if the frequency is correct it
} | can result it a higher reading, due to the speedos tendency to "average"
} | out the bounces.  If you think about it, that is how a speedometer works,
} | the spinning cable causes pulses on the needle that are counter-acted by
} | a spring pulling the needle back.
} 
} "if you think about it" the average of the high speed and low speed should
} come out to a just about right speed, certainly not 25% high anyway.

Ah- this might be true, if the speedo needle is as likely to go down as up.
But, lets throw in  some damping on the return stroke. This will allow the 
needle to quickly rise, but fall slowly. (this allows quick response, 
without jerky action. If you have smooth, constant rotation of the cable,
you can calibrate the system- the rise is proprotional to the speed
of the cable, and fall times are slow, as real speed changes.  If your
cable gets sticky, you begin to get a series of fast pulses, which can
make the needle go up rapidly, but not down, due to damping. This would result
in a high reading. If the pulsing is excessive, or the damping fails, the
needle will jump more. 
Seeing as how speedos aren't direct-drive mechanisms, there is damping in
the system, so this failure mode is quite possible. The *rate* of damping
in the speedo is designed for slow, gradual transients. When we get
large, rapid transitions (Pulsed cable rotation), we exceed the systems
capibilities.   

______________________________________________________________
Roger Garnett  (garnett@THEORY.TN.CORNELL.EDU) (607) 255-2522
Cornell University Agricultural Economics       Ithaca, N.Y.


<Prev in Thread] Current Thread [Next in Thread>