Fw: Programming 101 Part 2

[Bruce Plecan]

-----Original Message-----
From: Bruce Plecan <nacelp at bright.net>
To: diy_efi at efi332.eng.ohio-state.edu <diy_efi at efi332.eng.ohio-state.edu>
Date: Monday, March 30, 1998 9:57 AM
Subject: Programming 101 Part 2


Read no further until you read part 1

If your writting me use  nacelp at bright.net,  don't include this
whole thing, in a reply........

Some of this will be old news to some, and new news to
others.
If you don't have a chip burning program, you will need one
to write, to a chip.  You'll also need an eraser (ie Ultra-Violet
light).  You'll need chips, and their holders.
You will want a file editor function on your burning program.
When looking at a file with an editor it should look something
like this:
               0   1   2   3   4   5   6   7   8   9   A   B   C   D   E   F
000000 aa ab ac ad ae af a1 a4 a6 a3a4 a6  a7 aa  ab ae
000010 ba bb bc bd be bf 06 90 ff  fd  dd  a0 a9
000020 9a b7 a9 c9 c5
000030,
down to,
000FFF


This is just a small sample to let ya know what ya should see.
 So an address of 000001 is ab
 An address of 000001a is dd
Address 00000f is ae
the vertical column also goes in Hex like the top row
The top row is complete, the vertical column reads to
FFF.
On a 1227747

000000 is part of the check sum
000001 is also part of the check sum
000002 is chip ID
000003 is chip ID
000004 is enable checksum
  On this chip finding these was easy.  Just starting at the beginning
change one character at a time.  You could make a change with the
first few entries, and not get a prom error, code.  And changing the
next 2 changed the chip id on the scanner.

  If you've read the stuff about tuning then you've seen some
approximations for what tables, look like.  Having discussed the
importance of timing, and knowing what the curve "looks" like
then ya now know, that they if graphed out would look sorta
like saw teeth.  You would know that they are about 14
characters, in lenght.  You get that by counting from 30-100
by fives.  That is the increment that the vacuum is measured
in, since remembering our chart that it was K/Pa.  You would
as know that it is 12ish rows tall.  One of the stickier parts is that
these characters are hex numbers, so be prepared for
spending some time looking for them.  So by looking at the bin
you'd see that there is a area 15x14 where they characters
are saw toothed shaped, at
  000035-0000106, and that would apprear to be a timing table.

What would be another major area that we can find easily.  Well
how about a fuel table.  Now, the fuel is going to be a rough
number in closed loop so the computer can use the feedback
from the O2 sensor to compute a pulse width that will produce
the proper AFR.  So it doesn't need to be say any finer than
about half the resolution of the timing table, and again by
looking back at old posts ya see that it goes lean to rich.
And then repeats this pattern.  Again being a computer is
going to be written in hex.  So now we are looking for a table
that is 7-9 characters, by 7-9 characters.  If you continue
reading down the table then at about 0380, and running to 03D6
you'll again see a repeating patern.  Good candidate for being
a fuel table.

This is going to be all for a while now, since I got to come up with
some pulse generators for benching the ecm.  It would be quicker
thou if some EE spent a few to come up with some simple
555, and maybe a flip flop for a 50% PWM, for a VSS,
something for a tack signal,  and a sloopy O2 would probably
work.

Cheers  Bruce    Maybe it's the Cone Shaped Hats that suck
                             the smoke outta them little chip thingies