My son Johannes is building Suzuki SJ413 off road vehicle. One part of the project is EFI engine transplant. Donor vehicle is 1989 Suzuki Swift with 1.3l TBI efi engine. SJ413 project can be found from here. I pulled the ECU from the Swift and started to investigate if it can be modified. Denso processor looks very much like Motorola 6801U4 but Xtal is 12MHz and E-clock is Xtal/6. Processor can not be replaced with 6803. Quick verdict is that this ECU is not suitable for modifications. Anyway I spent few nights and draw schematics out of the PCB. I figured out some interesting details.

D3 ECU Label | ECU box

D3 schematics I/O sheet | CPU sheet

High resolution pictures about the PCB Component side | Back side

Injector driver is huge Darlington transistor 2SD1678. It seems to be saturated driver with RC snubber circuit. I have not managed to find data sheet for the 2SD1678. ISC (idle air valve solenoid) driver is 2SD1590. Other power drivers are 2SD1491 NPN transistors. Powersupply is build around Toshiba TA7900S chip. See schematics for details. Connector is 30-pin unknown AMP type. Connector is divided into two sections. Roughly one is for power connections and another for signals.

Ignition and speed pulse conditioning. Toshiba T2333 9-pin sip circuit is used for pulse conditioning. I have not found any information about this chip. It seems to be dual comparator with some additional functions.

MF291 chip. MF291 seems to contain injector watchdog and some other unknown functions. I am puzzled with the LL labeled circuit. It affects directly to the injector pulse width.

Analog to digital converter. Chip is labeled D 151811-0050. It seems to be six channel A/D converter either with internal reference or VCC is used as reference. Later is most probably true because in the PSU there is two trim resistors (RJ401 and RJ402) for +5V fine tuning. Interesting detail about A/D converter interface. It is connected with SCI (ordinary serial bus) with the CPU. CPU transmits channel numbers to the converter and converter transmits results. The chip may be another microcontroller with A/D inputs.

Partial Suzuki Swift wiring diagram 

ECM connector signal description

Throttle body 13400-63B00 9J8

Injector series resistor. 33950-60B11 079950-0581 Resistor is 2 ohm and about 50W

ISC or IAC valve Aisin 81640-5060 / Suzuki 18117-60B10 DC resistance 30 ohm and inductance 300mH

Odd solution. In addition to ECM controlled IAC valve there is a water thermostat controlled air valve in the throttle body. Detail picture.

Yet another odd solution. TPS sensor 179950-0260. This is not a potentiometer but two switches. Idle switch 1.5degree and WOT switch 60 degree.

MAP sensor 18590-60B00 079800-1540 5V

Water temperature sensor (WTS) Green 13650-84101 9J8

Air temperature sensor (ATS) Blue 13650-60B00 9J14

O2 sensor. Single wire Zirconium sensor, no markings left (rusted)

Speedometer Reed switch. Swift meter calibration seems to be 637rev/km and reed switch gives four pulses/rev = 2548 pulses/km.

Suzuki Swift ignition is a separate system build inside the distributor. Picture of the reluctor and electronics module. Again something odd. Vacuum advancer is two phase unit. Exact behaviour yet unknown. Engine speed information is taken from the ignition coil primary.

Another D3 ECU
Another D3 ECU Label | PCB at glance PCB look exactly same as the hacked one 

Another same generation ECU
I managed to get another maybe 1.0l engine controller. PCB seems almost the same but there is very odd 42-pin ! processor chip. Chip is marked D MG0640. Excluding pins 1 and 42 rest of the pinout seems to match with Motorola 6801. Xtal is 8MHz and E-clock 2MHz.

Picture of the A3 3-MT-FED 33920-61B6 3 112000-2040 board.

D3 and A3 boards together.

91 Vitara JLX1.6 ECU (Mitsubishi brand). Some interesting construction details.

Repair work of Vitara ECU

Broken ISC valve, centering construction broken one piece missing. Looks like also plunger spring is missing. I think valve should look like this.

TA7900S

RN100x

RN200x

Home