Injector driver chips? Where?

[Bohdan.L.Bodnar at att.com]

> > Brad Sheridan wrote:
> > 
> > >  Does anyone have/ know where to get the Motorola MC33293T chip?
> > > 
> Does anyone have any reccomendations for decent
> peak&hold injector driver circuits? In a quad package would be a
> real plus.
> 
>>>Note that the 33293 is NOT peak and hold. It's a quad driver intended 
>>>for high impedence injectors.
>>>
>>>The Harris HIP0081is functionally equivalent, ie. quad current 
>>>limiting driver with serial diagnostics interface. Information on all 
>>>these devices (and heaps more) is on the motorola and harris web 
>>>pages:
>>>
>>>http://motserv.indirect.com/
>>>http://www.semi.harris.com/

Shucks, a *real* do-it-yourselfer will *design* his/her own injector driver
circuit!  I thought about this and said, "hmmm...  this IS interesting!"
How's this for starters?

1). Stock injector takes about 1.2 ms to open
2). For minimum heat dissipation, use bang-bang control (current full on,
    current full off)
3). Strictly digital

Here's what I'd do:

Build a clock circuit which runs at, say, 10 kHz (this is easy) and sync the
injector control signal from the computer to this clock (requires one
edge-triggered flip-flop -- 1/2 of a 7474 dual D FF will do).  We now have a
synchronized control pulse.  Now, get a modulo-12 counter;  this counter will
reset every 1.2 ms (12 * 100 uS).  Now build a simple state machine which will
(1) reset the counter to hex C (decimal 12) when no injector "on" signal is
present, (2) turn the injector fully on for 12 clock cycles when the injector
"on" signal comes and (3) pulse the current to the injector "on" for 100 uS,
"off" for 100 uS, on for 100 uS, etc. as long as the injector control signal
is "on."  What I outlined is essentially what Chrysler did in its TBI engines
in the 1980s.  I'd use a power MOSFET to drive the injector -- this way, I
probably wouldn't even need to heatsink the transistor (unless the injector
shorted).  For more exotic control, I'd stick a small resistor (an ohm,
perhaps) in series with the injector and feed the voltage across it to a
voltage comparator.  If the voltage across the resistor exceeds a threshold,
the driver goes into current limiting.  This can be achieved by reducing the
1.2 ms peak period (i.e., go into "hold" immediately when maximum current is
met).  Since a typical peak current is 5 amps and hold current is 1 amp, a
threshold for current limiting could be 5 volts.

I can think of other variations on this topic (e.g., play around with the duty
cycle during the hold period, modify the peak period for either 1.2 ms or
until threshold current is reached -- this is for the case when there's
excessive resistance in the circuit -- and so forth).

Any comments?

Cordially,

Bohdan Bodnar
Lucent Technologies, Inc.