DIY EFI

[Bruce]

From: "Huw Scourfield" <Huw.Scourfield at btinternet.com>
> Requirement:
> accurate ignition timing.
> To do this requires at least a representation of where the egine is, ie
cam
> sensor. From this can be calculated engine speed, needed for one dimension
> of the ignition map. Also required is a representation of engine load,
this
> can be from a throttle position sensor or manifold pressure sensor. This
> gives the second dimension of the
> ignition map.
> The values programmed into this map may be derived in all sorts of ways,
but
> they will always have to be corrected for real world/ideal values.

With that said, all of a sudden the need for .00000003d ignition accuracy
just fell to the way side.  You have to leave a slight bit of head room in
whatever timing you want to run to keep the Knock Sensor / Plug readings
right anyway.  So long as the error is a constant then .0000000003d or 2.3d
doesn't matter as far as resolution goes.  If your having to cut things to
1d accuracy, your in some high stakes Pro Racing field or just in it for the
ego satidfaction of being able to claim, .000000000003d accuracy in my book.


> Other
> corrections would be for coolant temperature, air temperature, maybe
> barometric pressure(depending on where you live, halfway up Pikes Peak
> etc.).

All of those are valid.

> For an ECU /OEMs to work out what site in the map it needs to be requires,
> 1, a reading of engine speed,- this merely comes from time difference
> between cam sensor events.(OK it can/should be done from crank events, but
> I'm trying to keep this simple). One important point to bear in mind is
that
> the cam event signal must be sufficiently in advance of the required
> (maximum advance) spark  event. On reading a cam event the ECU must read
the
> load sensor, then read the map value for that load and speed from the
> ignition map.
> Then it must somehow convert the degrees of advance in the map to degrees
of
> delay from the cam event. The ECU obviously does not work in degrees, but
in
> microseconds, therefore for each and every ignition event a conversion has
> to be done from microseconds to degrees, then back again from degrees to
> microseconds. Also bear in mind that the corrections for coolant temp ,
> barometric pressure etc. would have to be added the degrees result before
> conversion.
> It sounds quite complicated but with modern microcontrollers having built
in
> capture(to measure microseconds between cam events), and compare(to output
> the required ign pulse) ports it really is not that difficult to grasp.
>
> For Fuelling, the basic requirement(disregarding sequental injection) is
> merely to squirt the right amount of fuel into each, or all inlet ports.
> Timing is not of great importance, but quantity is. Fuelling is much
simpler
> an issue for the ECU and comes down to reading of sensors and
corresponding
> maps, then adding the necessary corrections. Maybe later.

Maybe there are some issures at low speed where the injector timing is
somewhat important.  Might reference them off of a sensor event.
Bruce


----------------------------------------------------------------------------
To unsubscribe from diy_efi, send "unsubscribe diy_efi" (without the quotes)
in the body of a message (not the subject) to majordomo at lists.diy-efi.org