[Diy_efi] are uC requirements really that high?

Marcell Gal cell at x-dsl.hu
Fri Aug 23 18:01:25 GMT 2002


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) 

The interrupt pin would handle the crank position and
possibly other Digital input signal changes
(any number if they are XORed together).
In the interrupt routine first we read the time,
then check what changed,
at the very end of the interrupt check if there's perhaps
another change to note or an output change is due. 

Even if the interrupt is rising edge only, we can XOR an
output pin to generate interrupt at falling edge when we want.. 

The timed interrupt would serve for any output change
(start or stop injecting on any cyl, start or stop spark impulse,
even reset or set RS232 TX! ).
Note: At any time, there is always a first thing on the TODO list.
(I think this is called event queue) 

Timing:
rpm = 6000
teeth=60
octave:3> us= 10**6 * 1/(rpm/60) / teeth
us = 166.67 

On a 20 Mhz uC this is plenty of (>3k) clock cycles to do a simple
"read time and note in a queue" event and get back to work.
The RS232 has similar number of events to serve (not at 115200Bps)
and the number of inject/spark events are an order smaller (800/sec). 

The interrupt rutin could be made short enough not to
delay output with an unwantedly large amount.
(Although I do not really understand why >10 teeth crankweels
are good for, when we have quartz oscillators and
divide capable uCs  anyway to measure
time and time events with great precision) 

In normal (non-interrupt) operation it could do its daily
lengthy calculation duties, AD data gatherings etc... 

What is the problem with this calculation? 

thanx:
 Marcell 

An AD, an input driver,
an output driver, and perhaps a many-input XOR
would be needed besides an ATMEL or PIC or 68k uC. 

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



More information about the Diy_efi mailing list