[Diy_efi] new EFI controller

[Bevan Weiss]

OK, there was talk about this a while ago but things kind of fizzled with
Gars departure.  I've been thinking about a new controller in the same kind
of style as EFI332.  Using the MPC565 from Motorola, or the MPC566 for those
that just want to be different.  It's got the vast majority of required
features already onchip.  And it has the processing power to achieve some of
those dream objectives like ION sense.

Some of the ideas that I've been considering are:
analog->digital sample rate:
The MPC565 has two QADC (Queued Analog Digital Converters) each can run up
to 200KSPS.  I'm thinking that if one is dedicated to ION sense, ie round
robin 4/6/8/12 cylinders, and the remaining channel does a similar round
robin of the remaining sensors ie MAP/MAF, TPS etc...
What kind of rate of change do you think these signals undergo?
Is the round robin 200KSPS enough?

Ignition:
I really don't like the idea of inductive ignition systems, pumping current
through an inductor just to store some energy in the magnetic field seems a
little harsh.  I prefer CDI systems, although I think that perhaps some kind
of multispark arrangement would be preferred.  Have it such that the 565
sends out an ignition pulse, the rising edge causes the CDI system to
discharge the capacitor by a certain level (however much would give ~50mJ
spark), the discharge is stopped (by the CDI circuitry) and the capacitor
charges some more, then reaches a preset voltage and discharges again.  Kind
of like oscillating between two preset voltage levels. This continues until
the 565 lowers the ignition output, then the capacitor is just charged back
up ready for the next ignition event.

Injection:
Nothing fancy that I can think of.  I'd really like to keep this part simple
as, I don't really want anything like a DC-DC converter, I'd prefer to
obtain any high voltage from the actual field collapse, and then use this
energy to jump start the field back up again with the next injection pulse.

ION sense:
this will pretty much just be a current sense resistor with some
programmable gain differential amplifier which then feeds the QADC of the
565.  Certain algorithms could be run on a particular sample (ie one
ignition cycle) to look for such things as peak pressure not occuring at
optimum crank angle, and knock/pre-detonation.

WBO2:
This would use a concept similar to the DDL scheme...

MAP/MAF:
Is there any advantage to having both?
Apart from redundancy (if one fails...)

Other ideas/comments appreciated :)

_______________________________________________
Diy_efi mailing list
Diy_efi at diy-efi.org
http://www.diy-efi.org/mailman/listinfo/diy_efi