Re: Speedo error

[Phil and Judy Rose <pjrose@servtech.com>]

Paul,

Thanks for the useful info. I don't know yet if my speedo ('91 200q) is
off, but it doesn't agree very well with the trip-computer speed, which
always reads about 4 mph lower than the speedo itself. Is the computer's
speed reading independent (electrically) from the speedometer's ultimate
reading? If so, then it's possible that changing the resistor you mentioned
might bring the two speeds into better agreement? I suppose though that the
difference could merely be from needle misalignment.

BTW, did you lower the resistance in order to decrease your speedo reading
or to _increase_ it?

Phil


>From: paul.heneghan@bbc.co.uk (Paul Heneghan)
>Subject: Re: Speedo error
>
>From: "Smeins, Larry" <lsmeins@ball.com>
>
>>My 87 5kcst has had a very optimistic speedometer from day one.  The
>>odometer seems to be right on.  I thought I had a mechanical speedo and
[snip]

>I also noticed that the speedometer wasn't very accurate.  30, 60 and 90mph
>should correspond to 55.4, 110.7 and 166.1Hz.  My speedo was close at 30mph,
>about 5mph out at 60mph and more than 10mph out at 120mph (not that I plan
>on doing much of this speed).
>
>The chip that drives the speedometer and the odometer is an ITT UAF2115.  I
>couldn't find any reference to this chip anywhere, so I e-mailed ITT.  They
>sent me a PDF file with the data sheet for a UAF1025 which apparently is
>identical.  This shows that there is a series resistor connected to the
>speedo which is used to adjust the speedo sensitivity.  The resistor on the
>circuit board was 51ohms.  I found that 54ohms gave a much better response
>along with a realignment of the mechanical zero position.  The speedo is now
>within 1mph between 30 mph and 60 mph and within 2mph over the rest of the
>useful range.
>
>Of course, the speedo is only accurate when the tyres have the correct
>diameter, but the anal retentive amongst us could use a variable resistor to
>compensate for tyre wear (just kidding).
>
>Note: These adjustments have no effect on the odometer.  That just divides
>the sensor pulses to give 1mile for every 6644 pulses.
>
>Paul

         **********************************************
         *  Phil & Judy Rose     E-mail:              *
         *                       pjrose@servtech.com  *
         **********************************************