Karl, do you think your issue with the
accuracy had something to do with the "spring loading" of the
needle?

Yeah, most likely.  I didn't think of unloading the spring way back then
(nor now, but I won't admit to that).

Has anyone tested the accuracy by spinning the speedo on the
work bench with a drill at say 60 rpm and have a value for that in
MPH?

Seems it would be difficult to maintain an even 60rpm, or any other rpm,
without some sort of a readout.  I can't afford a drill like that.

K