ECU Tuning as per "Down Under".... (LONG Part 2)

[Chris Pronios]

There you have it. In our imaginary ECU, at 5 volts on the AFM, 6800  rpm,
it will pulse the injectors for 16.6ms or around 95% duty cycle.

If you followed all that, then you're now prepared for the good stuff. The
next section will give you all the specifics about the above variables,
where and what the values are in the ROM image, how to interpret them into
meaningful values, and how to use them in tuning your ECU.

--

ECU Injection Calculation - Detailed Analysis

K Required Number: (Nice and simple, some unknowns though)
CA Rom Address: 3F80 - 3F81. 2x 8bit Values. Quanta: 00 B7  seems to equal
0.125ms(!)
SR Rom Address: 000A - 000B. 2x 8bit Values. Quanta: ? ms

As said before, the K required number is the basic injector pulsewidth.  On
a stock CA18DET map it appears to be 0.125ms.  For now that is all you need
to remember.  You'll see how it fits in the "Putting it all together 2"
sections.

How do I know that the K Constant in a CA18DET (00 B7) = 0.125ms?
Well to be honest, I don't.  I received this vital piece of info from a
Japanese CA tuner.  Sceptical (as I am) I deconstructed all the maps on my
CA18DET to calculate Duty Cycle and AFR with the 0.125 ms value and it fits
perfectly.  So, I'll continue with this assertion until I can be proven
wrong/right.

VQ Map: (This could take a while)
CA Rom Address: 3E18 - 3E74. 52 x 16bit Values. Resolution: 0.1Volt.
Quanta: 16bit Precision Quotient
SR Rom Address: 0700 - 077F. 64 x 16bit Values. Resolution: 0.8Volt.
Quanta: 16bit Precision Quotient

If you know nothing about AFMs, or Math, I might confuse you here.  I've
made this as simple as I can, but if I confuse you, or you disagree, email
me.  If you already know lots about AFMs, you could probably just skip to
the "* So how do I read a VQ Map? *" paragraph.

Where do I start.  As I've stated before an AFM doesn't have a linear
response.  If it did, we wouldn't need a VQ Map.  Therefore we need a
mechanism that will convert this non-linear Voltage into a linear Quotient
(or percentage) of the AFM's maximum flow.  Hrmm...

Voltage... Quotient... Map... VQ Map?

What do I mean by non-linear?  Ok, lets say we have an ACME AFM, good for
about 500hp and 0-5v.  At 0v we know its flowing 0hp worth of air.  At 5v
its flowing about 500hp worth of air.  But what about inbetween?  If it was
linear, 1v might equal 100hp worth of air. So:

1v = 100hp
2v = 200hp
3v = 300hp
and so on.
So if you read 2.48v at the AFM you'd know you were flowing 248hp worth of
air.

But that would make this section far too easy to write.  The response of the
AFM is non-linear in that, its voltage gradient decreases as the flow
increases.  Gradient means 'rate of change'.  In other words "it takes more
air flow to increase the voltage, as the air flow increases".  I'll try and
explain a little clearer.  Lets take our ACME AFM, and give it a non-linear
response.
0v = 0hp
1v = 25hp
2v = 75hp
3v = 175hp
4v = 300hp
5v = 500hp
So it takes
25hp worth of air to progress from 0-1v
50hp                               1-2v
100hp                              2-3v
125hp                              3-4v
200hp                              4-5v
thus "it takes more air flow to increase the voltage, as the air flow
increases"

Lets plot this little map out with some bad ascii art to further
demonstrate.
See how the marks are further spread out as you go right along the axis...

Vertical Scale is Voltage
Horizontal is Airflow

V 5                   *
o 4               *
l 3           *
t 2       *
s 1   *
  0   1   2   3   4   5

  Air Flow

So if you read 2.48v at the AFM, how much air are you flowing now?
Somewhere between 75 and 175hp is all we really know!  We could (linearly)
interpolate between 2v and 3v and calculate 123hp (2-3 = 1v, 75-175 =
100hp -> .48v = 48hp, 75hp + 48hp = 123hp)
But thats guess.  Quite a big guess.  And you don't want your ECU making big
guesses.

So what is the solution?  An accurate VQ Map.  See our little maps above
that say 1v = 25hp and so on?  Well that is a crude VQ map.  To make it
useful, we need more precision.  Much more precision.  The CA ECU's VQ Map
has 52 points, mapping the airflow at 0.1v intervals.  The SR's has 64
points, mapping the airflow at 0.8v intervals.

Sounds simple?  Well to make things harder, the VQ map doesn't (directly)
give you a hp reading.  It gives you a quotient.  Think of it as "Percentage
of Maximum Flow".  If you know how to read a VQ Map, you could look at the
point for 2.48v, and say - the AFM is flowing 32% of its maximum at this
point.  If you happened to know that your AFM was good for 500hp, then its
flowing about 160hp at 2.48v.  Simple.

* So how do I read a VQ Map? *

This might be easiest to explain using ROM Editor.  Go onto the VQ Map Tab
and you'll see a list of 52 or 64 numbers depending on what type of map
you're looking at.  If you look at a lot of different cars VQ Maps, you
might notice that all the values range from 0 - 65535.  65535 just happens
to be the largest value possible with 16 bit unsigned integers.  This is
handy to know, as the VQ map uses 16 bit unsigned integers to store it's val
ues.  I still haven't told you how to read a VQ Map have I.  I'm sorry that
you didn't find a tuning website written by someone more to the point, and
you are stuck reading mine, but eh, what're you gonna do?  Ok - Take a value
in the VQ Map - any value.  Divide it by 65535.  What do you get?  Well
that's your quotient.  Multiply that by 100 and you've got your "Percentage
of Maximum Flow".  At what AFM voltage is that % value for?  Well if its on
a CA map, the first value is for 0.1v, the second for 0.2 v and so on.  For
an SR (and most others) the first value is for 0.08v, the second for 0.16v
and so on...

(CONTINUED IN NEXT MESSAGE)