> I need to design a device that interfaces to the speed sensor of heavy
> trucks. These are variable reluctance sensors and will remain connected
> to the vehicle's ECU, my circuit will piggyback. These are two-wire,
> differential signals full of common-mode noise and vary in amplitude
> greatly with frequency and also position as the tailshaft bearing wears
> in the vehicle.

I had a similar but totally different [tm] application whose solution
may be your solution.

I had a variable speed alternator (exercise machine load) whose out
put was a low voltage sinusoid at low speeds but which soon chose to
flat topped trapezoids clamped to the mean load voltage. As this
voltage varied depending on speed and net absolute load desired it
would have been modestly challenging to speed sense from. Now add 10
kHz PWM that takes a fixed resistor and PWM's it across the rectified
alternator output to change the effective resistance seen by the
alternator. Yee ha.

After trying all sorts of things I was offended by the simplicity of
what proved to be a superb solution.
Stops to find a circuit from about 10 years  ago ...
..... Found ...
Hmmm. May 18th 2006. More recent than I would have thought.

2 x BC337 or whatever
Differential long tailed pair.
100k collector resistor per transistor.
Join emitters with 1k to ground,
Input drive to each base via a 100k (one per base :-) ).
Two drive inputs connect across an alternator winding.
Diagram shows base to ground resistors per transistor but notes that
they are O/C.
That's it !!!
Probably Vcc =3D 5V. (Murphy says the value is on the sheet I didn't
load BUT probably 5V).

Alternator is 4 phase custom built. Max output > 200 VAC.

Speed output taken from collector of your choice.

This is "offensive" in that (at least) it does not explicitly deal
with the 10 kHz PWM or the sometimes very low voltage and max voltages
of say 300 V peak  are applied to transistor bases directly with no
attempt to limit voltage or current (apart from the obvious ways in
which the transistors do this. Ibase max=3D 3 mA worst case but usually
much less than that.

The "magic" lies in the differential pair.
It performs as a superb comparator.

Noise immunity  MAY be aided by the driving of base junctions deeply
into saturation. Ib > to >> Ic in most cases.

As I recall, along the way I used 3 or 4 pole Bessel low pass filters
using 1 or 2  emitter followers to  provide a low cost filter. This
was extremely effective at removing PWM but not needed for the long
tailed pair solution.



    Russell McMahon
--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.