Tach's & injector timing

Rod. Hiorns R.E.Hiorns at IEE.org
Wed Nov 3 07:22:27 GMT 1999


Hi, all

Just got on the list. Nice to see what's going on. I design electronics, so
apologies if I seem to rely on them too much ...

A simple way of changing 6 or 8 pulse systems into 4 pulse systems is to
use a phase locked loop with a counter ("divider") to set the ratios you
want. This is particularly easy with cheap CMOS 4000 series chips which can
run off 5-18v (good for autos, but maybe not trucks).

Try the 4046A for the PLL (all in one chip) and a dual 4 bit counter to set
the ratios (one in the feedback path and one on the input). I guess an
optional input divider of 2 or 4, and feedback divider of 2 or 3, will
cover 4, 6 & 8 cylinder engines, but it's pretty easy to choose ratios for
more cylinders. The problem you run into is that you have to make a
compromise between how quickly it'll respond, and how steady the final
readout is, and this brings us neatly to designing a high accuracy digital
tach for a dyno shaft.

To get high accuracy on the dyno shaft, you want a lot of pulses per
revolution so you can avoid too much smoothing which forces your readout to
take ages to settle. This just means having many contacts/magnets/marks on
the shaft for your electrical/hall-effect/optical reader. I've had great
success canibalising the wheel and slotted UV opto-switch from inside an
old (dead) computer mouse (although it took me ages to start with to get
rid of 60Hz clocking ... had to build a box round it so the flourescent
lights didn't interfere ... I guess sunlight may cause missed pulses too).
You can chuck the output through a F-V converter, but a neat, flexible
solution is to use a PIC microcontroller (they're amazingly cheap too). Use
the event counter to time periods of the system clock and then take the
reciprocal to get the frequency (algorithm in the PIC handbook) which needs
scaling depending on your pickup system. I like this solution cause I can
drive a digital display using the same chip ... no need to lock up my
voltmeter which I invariably want for something else! The same system can
also be used with a different program for dwell metering, although I
noticed more readout 'wobble' with this. Perhaps if you're really concerned
you could put some filtering into the program to get it down to the last Hz
or degree.

Hope these few thoughts help. While I'm here could I ask how to work out
when injectors should be fired? When the inlet port opens seems too early;
mid 'inrush' would make sense to me for best mixing (but it's rpm
dependant: I'm using long inlet runners to boost lower rpm torque) ... and
is there a recognised best way of getting past the 'stumble' as I
changeover from sequential to batch firing? Any views appreciated.

Rod





More information about the Diy_efi mailing list