>-----Original Message-----
>From: David VanHorn [mailto:dvanhorn@CEDAR.NET]
>Sent: 26 May 2004 17:06
>To: PICLIST@MITVMA.MIT.EDU
>Subject: Re: [PIC]:EFI Controller
>
>
>At 12:47 PM 5/26/2004 -0300, Future wrote:
>
>>I can4t believe a 10mips (18f452@40mhz) processor is too slow to do
>>fuel and spark management.
>>
>>Maybe it will not be the best solution, but will always be
>better than
>>any carburetor.
>
>The controller in my car, operates over about a 1 hz loop.
>It's hunting up and down, never stable. (judging by the O2
>sensor voltage)
>

All closed loop mixture control schemes that use a narrow band oxygen sensor
do this.  These sensors do not give a linear output voltage with exhaust
oxygen content, but switch between rich and lean conditions quite quickly.
The only way to have any control is to hunt either side of the rich and lean
states.

Wideband sensors don't suffer from this problem, but cost considerably more
$$$.

Regards

Mike

>Even at 10,000 RPM, that's 166 Rev/sec, so I don't see what
>the computational problem would be. Pics aren't wonderful with
>ints, but still.. That's just not that much activity. Figure
>four sparks per cycle, 664 sparks, or twice that if you use
>four coils and "wasted spark".
>
>I'd figure two ISRs for ignition timing, and the rest done in
>non-critical time.

I'm pretty sure an 18F series PIC could handle a basic fuel and igntion
system with a modicum of optimisation techniques.  Tony Nixon's mapped
igntion design used a humble 16F84 running at 8MHz, all operations obviously
performed with software counters. Jaycar are still selling the kit so it
must have worked pretty well.  With the multiple CCP modules in the larger
18F parts, a considerable overhead would be removed in measuring periods for
ignition and injection calculations.

Regards

Mike

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads