Just an FYI...
All of the motors you tested (so for as I can discern) were induction motors. Voltage fluctuation has very little effect on induction motor RPMs, rather this type motor is most sensitive to frequency (60Hz, which is nearly always with +/- 0.5%) and slippage under load (with respect to the rotating magnetic field). Your 3500+ RPM motors are really 2 pole single-phase induction motors with an internal magnetic field rotating at 3600RPM (60Hz multiplied by 60secs). You typically see posted RPMs that range anywhere from the theoretically perfect 3600 RPM (an impossible zero slippage) to as low as 3500 RPM which would represent considerable slippage under heavy load (likely at or near rated HP).
An ammeter (Amp meter), either the clip on type or an inline plug-in style, will provide far more useful diagnostic information. While severely bad bearings will eventually slow an induction motor, which would register on your tachometer, that same resistance/friction from the bad bearings will show more readily and much earlier on with the ammeter. As the motor's load increases so will its amperage, but the rotor will still want to maintain an RPM as close as possible to that of the internal rotating magnetic field.
Universal motors, on the other hand, are much more influenced by voltage and load than they are frequency (which would show with your tachometer). They also exhibit far greater variability in their RPMs under load. However, universal motors also incur an increase in measured amperage as their load increases, just as with universal motors, so again an ammeter can be used to diagnose such a motor, but atleast a tachometer would also register an observable change with this type motor so long as the variables of voltage and load are properly accounted for first.
Just my $0.02
Enjoy your new toy -- there are plenty of other applications where a tachometer is genuinely useful.