Well here is the brains of this operation.  An innocent looking little box that controls the whole show.  If it doesn't work right you go nowhere fast.  Luckily with the current level of technology today this hardware is very reliable, especially considering the harsh environment it lives in.  If you have every designed electronic circuits for your ride you would know that the signal to noise ratio is pretty poor.  Trying to keep the harsh automotive environment out while trying to process noisy input signals is no easy feat.  The ECU accomplishes this by some filtering in the hardware and some in the firmware.
       Well this unit is manufactured by Margneti Marelli.  The best I can figure is that this unit was used on European cars (probably Italian - both Margneti Marelli and Weber are Italian base companies).  Some have suggested that this is the same unit as used on 1992-93 Ford Escort RS Cosworth (model IAW48/P8).  They are similar but are not the same.  The Ford unit has an extra daughter board that mounts the EPROM while this unit the EPROM is socketed on the main PCB.  The model number of this pup is IAW4W6.  I would love to find out exactly what it came out of.  Never can have too many spares (I've got none so I feel naked).  There is quite a bit of unused hardware in this unit.  It may be that this unit was a general use ECU used in several different applications.   You can use this to your advantage.  For example if you need to control an electric radiator fan there are spare relay driver outputs that would do the job nicely.  There are also unused inputs that can be used for some user specified purpose.
      Most of the IC's, at least the ones we are concerned with, are manufactured by SGS-Thompson.  Getting information for this company is pathetic.  Using there WEB page is an exercise in frustration, especially for component that are not being manufactured today.  Several key components took quite a bit of work to figure out what they did.  The easiest was the EF68B40PV  timer.  This is functionally the same as Motorola's MC6840, Hitachi's HD6340, or Thompson's EF6840.  The large IC marked MA SOL1.0 turns out to be a PAL.  Never could locate what type/family this part is from.  Most likely all the markings on the device are house numbers of the programed part.  I've figured out the memory mapping function of this part, some of the latch addresses but not all of them.  There are several outputs that go to unused sections of the hardware that I haven't had the need to figure out.  I would bet some could be easily identified as some bits of the latch at $4000.  Two bits I know (one is the fuel pump).  It will take a little more work to identify the address and latch bits for some of the outputs.  $6000 is another PAL latch that mostly reads the status of the L9112D driver.  The last really mystery part is another SGS-Thompson jewel marked L9112D.  At first I though this part was a MUX of some kind but that proved wrong.  It looks to be some esoteric type of driver.  It can be modeled as a inverting op-amp with a 4.7k pull down on the input with a 75 ohm output impedance.  Looking at what it drives it is a mystery why they chose to uses a part like this.  Only two (2) sections of this hex part are used.  One to drive the injector pair (INJ-D) and the other is the ECU serial receiver input.  Most of the other key parts I was able to reason out what their functions where.
        The schematics I have created are functional models.  Most of the values for the capacitors are unknown.  Most have no direct affect on hardware performance as far as we are concerned.  Most serve as by-pass caps and high frequency filtering.  Most all of the resistors I could identify by value.  A few I had to measure.  There are two modules, HY1 and HY2, that filter the sensor inputs that are not entirely modeled. The DC transfer function is modeled but their AC characteristics have not.  I have looked a little at their AC characteristics and what my casual impression is that they are high frequency filtering, maybe even active filtering.  I do believe that I know enough about them and their transfer functions that what is documented in the code is correct.  The values that I've pulled out of the ECU have correlated correctly to the input signals applied.

12/21/1999 -
    While the ECM was out of the car while I was working on the crank trigger stuff I did look a little closer at the MAP and TPS circuits in the HY1 module.  The MAP side is nothing significant but the TPS side is fairly heavily filter.  Schematics have been update.

• MPU, PAL, EPROM and MC6840 Timer
• Original Edelbrock Circuitry - 01/18/2000
• CPU Schematic with 28C256 EEPROM  (in car programing) - 01/18/2000
• Daughter board Schematic with 27C256 EPROM  (in car programing)
• L9112D Interface - 09/04/1999
• HY1 Input Module - 12/21/1999
• HY2 Input Module
• IAC Driver
• Injector Drivers
• Ignition Trigger Input and SA Amp Driver
• Original Edelbrock Ignition Trigger and Ignition Amp Driver
• Original Edelbrock Inductive Trigger Input  (not used in this application)
• New Crank Trigger Input Circuit   (timing jitter)
• New CAM Position Input Circuit  (timing jitter)

• Original Edelbrock Wiring Harness
• Modified Wiring Harness (Distributor Pickup)
• Modified Wiring Harness (Crank Trigger Pickup)
• Modified Wiring Harness (ECM-555) - 04/24/2005
• Ignition Amplifier and Coil Wiring
• Controller DB9 to 6 wire Phone Patch Cord

Note: The connector definitions for the Crosworth ECM were found here: EFI reverse-engineering central