[Diy_efi] Re: Minimalist micro based EFI ?

[Mike]

mmmm, I wondering now that 10 bits is just not enough,

The MAF signal is log with a DC offset of around 1.5v (which
varies a little +-10% on mine). So from 1.5 to 5v we have
a log output, we could effectively having a doubling of
air flow at the last 10% of scale from say 4.5v to 5v,
hence I'd think 12 bits would be more appropriate with
a DC offset subtraction and amp prior to the ADC input.

In respect of noise, most ADC's are quite quick, so getting
lots of samples and doing a simple binary weighted averaging
to get rid of some noise would be fairly straightforward,
if the ECU MAF signal it used for calculations was 200mS
after the real unfiltered MAF as raw sample then would there
be much of an issue in cruise or even WOT operation, especially
given we'd have a hardware circuit to add pulses to the injectors
on the same basis as the old accelerator pump for carbies...
In any case this 'delay' might just provide the storage type
cat with its lean-rich cycle and help its operation during
throttle changes ;)

rgds

mike



At 08:33 AM 3/11/03 -0500, you wrote:
>> I agree that 10bits should be adequate.  That gives 1024 codes...
>> which means on a 0->5V sensor, that's 4.8mV or similar.  I would imagine
>> that noise on the line would be greater than this, and hence the only real
>> way to remove it would be to oversample, which is not worth it in
>> this kind of environment anyway.
>
>Couldn't you just apply some filtering on the input signal?  The noise in an
>automotive system has got to be at far higher frequencies than any delta
>signal you'd get, wouldn't it?
>
>
>_______________________________________________
>Diy_efi mailing list
>Diy_efi at diy-efi.org
>http://www.diy-efi.org/mailman/listinfo/diy_efi
>
>

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