Ignition timing reference points

[Kalle Pihlajasaari]

Hi Mark, Stephen, Martin & All,

Long message on adding mapped ignition to older vehicles.

(Mark, putting our thread back onto the list, seems like there is 
general interest :-)

==========
From: Mark Pitts <saxon at zymurgy.org>
 
> More or less exactly what I'm trying to do... My dizzy arm 'pops' on and =
> off, why not get a 2nd arm to modify... if you are going to have to take =
> the cap off to remove the 'point saver' clip... change the rotor arm as =
> well.

Yes, easy fix but if I want to keep running compatibility then I can't
use the dizzy for the timing input so the centrifugal advance will
be close enough to the ignition map for the distributor rotor and I 
won't need to stretch the rotor lobe as you suggested.

> My system is going to be timed off the cranks front pully with some =
> small drilled holes in it, and MAP from a T piece in the vac line, all =
> I'll have to do is switch between points and electronics on the line =
> into the coil. (I might just let the points get beaten to hell anyway!

I am reluctant to fit a sensor that is not enclosed as someone/thing
will bash it.  If I have the point saver clip off one could indeed switch
between the two timing systems if the original is left totally as is.
On the 4 barrel Alfa I am not real comfotable with manifold pressure
sensing as there is no common gallery, I was thinking of using just
throttle position for the other map input (on top of RPM).

> Next stage after this I to go to wasted sparks, using a pair of Citroen =
> 2cv coils, never had one fail in years of motoring in those cars.

Yes, I would just stick in 4 standard bosh coils and in this event
I would rip out the whole dizzy and replace it with one that has
no cap or rotor or centrifugal advance, just the optical encoder.
With 4 (or 2) coils one can get 90 deg dwell by turning the next coil
on as you fire the current coil.
----------
From:  Kalle Pihlajasaari[SMTP:kalle at device.data.co.za]

> > Just file up a bit of brass plate to extend the edges of the rotor arm,
> > and screw or solder it in place! Dont make it any more than the gap
> > between the electrodes! Stupid comment but I eneded up with sparks to
> > the WRONG cylinder!
> 
> Sounds good.
> 
> I am a little bit of a paranoid delusional maniac at times and am =
> thinking
> that I would like to leave the original igition untouched in the car
> and have a totally separate circuit for my new electronic ignition.
> 
> I figure I could set the timing on the points and dizzy correctly
> and then pull the points open with a clip of sorts to avoid wear.
> 
> If I don't put the encoder in the dizzy I would still have standard
> centrifugal advance on the rotor and would not need to modify it.
> 
> Then add somewhere on the crank or better yet the camshaft a sealed
> optical encoder with a non-glass wheel that is tightly attached
> with a 'something' to the cam shaft (DOHC) to give me a new reference.
> 
> A friend suggested the electronic ignition from a V8 Mercedes as =
> something
> that would always give high energy sparks and then I could use all the
> other stuff as standard and if my map fails I can just plug back into
> the points to get the car home :-)
==========
From: Stephen Dubovsky <dubovsky at vt.edu>

> >I was thinking of adding a 256 or 360 slot optical interrupter disk into
> >my distributor to give double the that number of edges every rev 
> 
>   The only problem I see w/ this is dirt.  The guys in the motor industry
> very rarely use optical encoders because you get a few specs of dust on the
> encoder wheel and things go crazy.  They almost always prefer some sort of
> magnetic/hall sensor type pickup or use a resolver (and then a resolver to
> digital converter) if higher accuracy is required.  If your interested and
> dont know what a resolver is, I can type up some sort of "short"
> explanation;)  Both of these methods are relatively insensitive to dirt and
> grime (unless its a ferrous material...)

Resolver I know but not realy the guy for the job as it will give you
good indication in a alalog fashion and I am a digital person.

Another alternative that I discussed with Mark was to couple a
sealed insdustrial optical encoder to the overhead cam shaft.
(I do not want to take the engine out, real lazy, and beyond me)
==========
From: Martin Hill <EAXMJHI at ean2.mecheng.nottingham.ac.uk>

> > From:           Kalle Pihlajasaari <kalle at device.data.co.za>
> > 
> > I was thinking of adding a 256 or 360 slot optical interrupter disk into
> > my distributor to give double the that number of edges every rev 
> > of the dizzy (or the same resolution on the crank).  Is one degree
> > on the crank accurate enough for a totally digital one or two of
> > design for a mapped ignition.  Interrupter disks like this are
> > available alone and have usually got an index mark at another
> > radius.  I could then just replace the cam section with a nice
> > shaft without the centrifugal advance (will the rotor still point 
> > close enough to the correct cap electrode if the rotor is not
> > advanced at all and the timing is ?
> > 
> > Just a few thoughts.
> 
> This is far more than enough for a digital system.  Most systems 
> these days use a 36 tooth wheel with one tooth missing to signify 
> TDC.  With the computer it is possible to get accurate enough timing 
> form this.  360 teeth would provide so much information you would 
> need a faster processor to keep up.

As discussed on one of the other threads, I would make use of 
extra circuitry to grab the high speed timing, could be as simple 
as a flip-flop to gate on the timer input from the index mark
and then do 4 compares of the timer value in the Micro and then 
gate of the signal and reset the timer and wait for the next
index to start the input pulses again.

There are only a few things I am not sure of one is that would
a discreet resolution of 1 or 2 degrees crank angle be enough
for timing and it seems probable.

The second is that is there going to be too much jitter and 
backlash on the cam shaft, the distributor shaft or the crank
shaft to accurately count one degree pulses, in other words
would any hypothetical occilation exist that is greater than
one degree that would cause multiple counts on the sensor ?

Would I have to use of direction information on a quadrature 
encoder to subtract angle if there is a back fire on start
or do I just assume every count edge is one more degree in the
correct direction ?

I want to do this the brute force way and will probably only
ever make two units so am perpared to invest a bit in good
encoders unless they actually make the system unworkable.
I would however like to keep the system fully digital with
timing changing stepwise at the encoder resolution.

I would make provision in the circuitry to support the 
variable inlet valve timing system that could be canibalised
from one of the 159i engines.  Does anyone know off hand what
else would be affected on a carburetted motor if the cam
timing were to be changed.  It is a step change of around
5 .. 15 degree if I recall.

Yeah and what does one do if there are 4 butterflys and no
common gallery for manifold vacuum if I were to consider
a fuel injection system.  The 159i had a flap flow sensor
and one butterfly so avoided this problem.

Cheers
-- 
Kalle Pihlajasaari     kalle at data.co.za
Interface Products     Box 15775, Doornfontein, 2028, South Africa
+27 (11) 402-7750      Fax: +27 (11) 402-7751