Old 486 Board for ECU??

[Axel Rietschin]

----- Original Message -----
From: "Mike from West Australia" <erazmus at wantree.com.au>
To: <diy_efi at diy-efi.org>
Sent: Monday, May 01, 2000 12:56
Subject: Re: Old 486 Board for ECU??


>
> At 03:34 PM 30/4/2000 -0800, you wrote:
> >"Peter D. Hipson" wrote:
> >>
> >> Par ports are fast--much faster than the typical ECU of today.
Memory...
> >> Let's see, Windows 2000, 128 MB Ram, and a 10 Gig drive. Of course a
flat
> >> screen LCD display, and a minature keyboard would make a reasonable
setup.
> >
> >The I/O speed isn't the problem.  The issue is: can you generate an
> >output signal with one microsecond accuracy relative to an input signal?
> > How about two outputs that might change at once?  Three?
>
> Pardon !?
>
> You don't need to generate an output within 1usec of getting an input.
>
> As long as the output is generated with synchronism and within the
> time it takes for a revolution/4 or 6 then you will be fine.
>
> The CPU's I've seen in ECU's in last 20 years are *slow* 8 bit micros
> that clock at around 4 to 12MHz. In fact in 1982 I did a Z80 based
> EFI using a 4MHz CPU (tried 6 but too new).
>
> The Z80 NMI came from the ignition contact points (through fitlering)
> which fired an injector. In the background the CPU would constantly
> acquire AFM and engine temp readings and generate a number ready for
> the next interrupt...
>
> A 486 which clocks at 66MHz and has the OS/EFI routines in cache will
> *easily* handle synchronous response to external real time events.
> Heck it might be enough to look at the jitter on the crank angle
> sensor. The problem would be MSDOS, not patched appropriately.
>
> The timers on a 486 board are very well documented and the interrupt
> structure is wide enough to set up a simple priority scheme...
>
> The requirements for EFI is *well within* the hardware capacities
> (timers, CPU speed, RAM) of a 486 - even at 20MHz and no cache.
>
> Of course you need an ADC but only one of the inputs could be
> considered as time critical and even that (AFM) can be asynchronous
> with respect to engine cycles... ie Fast sampling helps but
> AFM synchronous measurement is not required...
>
> my 2c worth.
>

You are describing a very simplistic ECU. Why not put one sensor where you
want your injection pulse to begin, and another where you want your spark to
fire? With a little cleverly designed analog circuitry, you probably don't
need a CPU at all ;)






----------------------------------------------------------------------------
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