Re: 88 MB Urq questions

[Orin Eman <orin@wolfenet.com>]

> The problem is - these duty cycle values are calculated, not derived
> from tables (as with ignition and boost values).  The code is complex
> and not yet understood - but it _must_ make assumptions about the
> engine's geometry.  Especially - if our experimental results are valid -
> air intake volume above 4500 rpm is being guessed at.   Temperature
> and boost are still measured, but volume _might_ be unknown and
> assumed.  Change the air inlet duct (K&N?) or the cam, and you _might_
> disturb the whole house of cards.

It seems that it turns into a MAP system once the air flow meter
tops out.  Given MAP (manifold pressure), RPM, intake air temperature
and a table of the engine's Volumetric Efficiency (VE), you can
calculate the fuel requirement of the engine.  There are some
equations somewhere on the the diy_efi site.

Now volumetric efficiency is a measure of how much air gets stuffed
into the cylinders.  It depends on just about everything...  cam,
exhaust, manifolds, intake/exhaust ports, combustion chamber shape
etc..  The effect of a lot of engine tuning is to increase volumetric
efficiency.  Higher volumetric efficiency means more air gets
into the cylinders at a given manifold pressure/RPM.

So, if you increase volumentric efficiency, you've got to find
some way of getting more fuel!  IF the air flow meter has topped
out and the ECU is doing some simple calculations based on its
idea of VE, then you won't be getting enough fuel!  Yes indeed,
the house of cards falls over.

> We think this is what happened to John Robinson's car, which has had
> some major 'tuning' surgery and now stops at 4500 rpm.  It's utterly
> weird - the car accelerates like a rocket to 4500 - and then stops
> accelerating.  Period.

How does it do at less than full throttle?  Better?

As for the US MC engine, it forgets about the O2 sensor pretty much
as soon as you go on boost and uses one of two RPM based duty cycle tables
for the frequency valve... (one table for low boost, one for
not so low boost levels).  There is no carry forword of anything learnt
by the O2 sensor to on-boost behavior.

On my car, I haven't seen any evidence of it leaning out at higher boost
or RPMs using a simple LED bargraph on the O2 sensor, nor using
the voltages from the O2 sensor on my data aquisition board.
However, I have noticed that the MC tends to detect knock
around 4800 RPM...  We have a code change for the MAC11 along with
a simple 1 transistor and LED circuit which will indicate when the ECU
detects knock and retards timing.  There may be a code change soon
that displays this on the check engine light!  The knock LED tends
to come on between 4500 and 5000 RPM on the MC with not much above
stock boost levels.  It takes quite some timing retard to stop this.

(I'll try to find the MAC12 listing and look into doing the knock
LED.  Got to do the MAC14 first tho'.)

Again, if you are getting more mixture in the cylinder than the ECU
expects, its timing tables are going to be wrong too... more likely
to knock... and once it starts retarding the timing, it does
feel like you stopped accelerating!

I believe some US EFI systems do carry forward corrections learnt from
the O2 sensor to areas of operation where the O2 sensor isn't used.
You'd have to search the diy_efi archives to find out more.

Orin.