[Diy_efi] are uC requirements really that high?

[William Shurvinton]

It's a combination of things.

For sequential you need to juggle 8 injectors which open 90° apart, but may
be open for 650°. If you are happy with, say 0.1ms resolution, then you can
handle this with system timers and general IO, but if you want more might
need external HW to help you. Depends on the application.

Spark is the trickier one as this needs higher resolution. If you are using
a GM DIS pack then that does all the hard work with the crank sensor and
it's not too bad. If you want Coil per plug or a home brew ignition then you
suddenly run into all the issues of calculating TDC and scheduling dwell and
firing, whilst still dealing with the 8 injectors that need opening and
closing, sometime simultaneously.

If you are happy to count machine cycles and work out nested interrupt
latency then this is all doable, but it's an aquired art. Pay a couple of $
more for the uC and you get loads of timers which you load and forget. For
$20 you start to get TPUs which are dedicated HW coprocessors for handling
exactly these functions.

Or to put it another way, real time stuff is much easier if you start with
something designed for real time operation.

Bill

----- Original Message -----
From: Marcell Gal <cell at x-dsl.hu>
To: <diy_efi at diy-efi.org>
Sent: Friday, August 23, 2002 5:54 PM
Subject: [Diy_efi] are uC requirements really that high?


> Hello Guys,
>
> Many times I've seen on this list that "you will need
> many timing registers to control a - say - 8 cyl seq. port EFI."
>
> It seems to me that 1 interrupt pin, a readable clock-counter
> and 1 timed interrupt at a pre-specified time would be enough:
> (which all uC should have anyway)



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