Adjustable electronic ignition

John T Stein JSTEIN at dpc2.hdos.hac.com
Wed Jan 10 16:55:54 GMT 1996


> Date:          Wed, 10 Jan 1996 09:34:54 +0800
> From:          Adam Tate <at at multiline.com.au>
> Subject:       Adjustable electronic ignition
> To:            DIY_EFI at coulomb.eng.ohio-state.edu
> Reply-to:      diy_efi at coulomb.eng.ohio-state.edu

> Just a quick question,
>                                 My Holden Commodore runs a Buick 3.8 litre
> V6 and uses a crankshaft sensor for the ignition timing. What I want to do
> is vary the timing, preferably from inside the car. One way to change it,
> although being non-variable, would be to offset the sensor slightly to
> obtain a fixed advance or retard in the timing. What I would prefer however,
> is to intercept the signal before the ECM, and somehow modify it
> electronically to make it appear that it is advanced slightly. 
> 
> Can I do this with a simple capacitance type delay, or would I need to use a
> specialised circuit? Also, any ideas what the delay would have to be -
> milliseconds, nanoseconds ? Ideally I would like to have some sort of dash
> mounted box with an adjustable dial or something along those lines.
> 
> All input is most appreciated.
> 
> Adam Tate.
> [at at multiline.com.au]
> http://www.multiline.com.au/~atate/index.html
> 
Adam,

While "playing" with the ignition timing by introducing delays into 
the crank encoder signal will work, a better place to introduce the delay
 is in the electronic spark timing (EST) command line which is generated by
 the ECM and received by the the Direct Ignition (DIS) module. 

The desired ignition time in represented by the shift in crank angle between 
the pulse waveform on the EST command line and that on the "reference" 
line which the DIS creates from the crank encoder and sends to the ECM.

The problem that results from simply intrtoducing delay into the EST signal 
results from the fact that the mapping between delay (time) and 
firing angle (crank degrees) is a function of engine speed -- and
what you really need is to "delay" by a constant percentage of
reference-waveform PERIOD, not a constant TIME INTERVAL
Delaying by a constant number of cycles of the oscillator of a phase-locked loop
that tracks a multiple of the reference waveform frequency is 
one  way to attack this in hardware.

John



More information about the Diy_efi mailing list