Flywheel/Encoder patterns

[Sandy]

Caution... lots of the 2 Cents flowing here ;-)

It seem a lot of programming and additional hardware for the replacement of
a $20 sensor. The only problem with the cam sensor is mounting it! 

I was thinking much of the same thing (and others as well), but I decided
that it would be out of my 'scientific' capabilities, to actually correctly
monitor the spark current and determine, accurately which is number TDC
Cyl1. In the driver board that I designed, I left 'test points' for watching
the drive to the gate on the output IGBT, and a point to watch the output
current, for things like that, but they are currently not used. They would
need some analog circuitry to massage the signal, and feed it back to an A/D
or some comparator that drives an I/O pin.

In any case, it seems that and just using a cam sensor would give me what I
need, without much additional circuitry. If the cam sensor fails, then the
system can just go into batch mode, and still work fine, probably even go
unnoticed by the driver.

For V8's, I would expect most ignition are going to be waste spark type, ie,
dual coil, so for ignition for them it is not that big of a deal to know TDC 1.
For V8, I think some of the numbers the group found was that the dual plug
coils would work great to 8k+ RPM, 4's and 6's the increase time for coil
dwell allow something like an easy 10-12k+ RPM with dual plug coils, not
much to gain by having a coil per cyl. Unless you are at the end of the RPM
spectrum, the dual plugers should be great.

So it becomes less of an issue for ignition to know where TDC is, and only
important for sequential injection to know where TDC is. Their is also the
understanding (consensus?) that sequential injection makes no more power
then batch style, but sequential is cleaner running, and produces less
emissions. 
(Sequential does do better for systems with large dynamic range of fuel, ie,
Turbos, etc)

The only issue becomes absolute timing. And that I think is the direct
capability of the CPU/Software/Crank Sensor... I better stop here, as I am
digging a big trench, but I would say, if you are going to eliminate the cam
sensor, then go for the gusto, and just run the system in batch mode, you
won't have the extra coils, won't have to worry about different starting
modes, etc. Also much of the driver circutry can be simplified. In doing so,
things get more reliable from the get go.

I think I have had too much coffee this morning!

Sandy

At 05:49 AM 12/21/96 -0500, you wrote:
>Based on previous discussions, I think we determined that the most 
>accurate spark system would be driven by the crankshaft pick-up sensor, 
>with a TDC#1 reference from the camshaft sensor.  In the interest of 
>reducing sensor count, I'm interested in determining TDC#1 at top of the 
>compression stroke without monitoring Camshaft position.  The method I 
>have in mind would be based on ionization current in the ignition 
>system.  I think there are several ignition drivers available now that 
>can feedback this info to the processor.  Is this right?  If so, does 
>anyone know any part numbers?
>
>I think by going to a waste spark system and gang fire injector firing 
>during the first few engine rotations, we can compare the ionization 
>currents between the two sparks for each cycle of each cylinder and 
>determine which one represents the top of the compression stroke.  Based 
>on this determination, we can then go to sequential injector firing, and 
>eliminate the waste spark to prolong ignition system life.  Any comments 
>on difficulties I might run into??
>
>Best Regards, Mazda
>
>