Bosch
Mono-Motronic MA 3.0 type 0 261 204 251.
Used in european
small single point injection catalysed engines such as Peugeot 106
Citroen AX and Saxo, among numerous other.
Analysis and
failure repair
In my case the car is a Citroen AX Spot, 1 liter
engine, delivered september, 1996 and ran 170 000 km on july
2004.
Here is the story of the self repair of an EFI. It took
me slightly more than 3 month, but I prefer to make it myself, to
understand what happens, and anyway, it costs me way less to have a
junk car ready to go and to pay (in France) a minimum insurance, than
the very first step into a garage and saying « please,
repair rapidly my car, I badly need it »
1 System organisation
Inputs
and sensors:
Rotation and phase sensor. Electromagnetic (5V p/p).
Pin 30 & 11, shield pin 19
Odometer (optional) not used in
this Citroen AX (this input send you to the garage after x
km)
Oxygen sensor (aka lambda sensor) output signal pin 10 (lean
0.2V rich 0.9V), ref ground pin 28, shield pin14
Throttle
position dual potentiometer (dual for reliability, 0 to 5V) cursors
pins 7 & 29, ref ground pin 26,
Engine temperature (3000ohm
@25°C) pin 25, ref ground pin 26
Air temperature (2600ohm
@25°C) pin 27, ref ground pin 26
Idle sensor ( 0 or 5V) pin
31, ground pin 19
Outputs:
Ignition
coil 1 pin 1
Ignition coil 2 pin 20
Injector pin 17, tied up
to +12V through 3.5 ohm resistor, located under computer
supportant
Active coal valve pin 5 tied up to +12V
Idle
regulation motor pins 33 & 15
Power
supply:
+12V permanent, pin 18
The DME relay gives the +12V to
computer on pin 37
Ancillary
supply:
One of the double DME relay, directly connected to the
main contact switch, gives the +12V to the computer pin 37 as well
as the +12V to the injector resistance
The other DME relay gives
power to
The two ignition coils
The fuel pump
The
autoregulated (PTC) heating resistor of the oxygen sensor
2
How the computer works:
(deducted from the partial schematics
(appendix)
and from my own measurements)
Powering
on:
Activation of DME1 relay by the main ignition switch
Battery
voltage is applied to the computer pin 37
The computer activates
DME2 relay through pin 3
DME2 relay energizes ignition coils,
fuel pump,and oxygen sensor preheating for half a second (allows
fuel pressure on injection needle,and probably some internal test.
The fuel pressure is NOT monitored)
Engine
cranking:
The engine cranking is only detected by the rotation
sensor (no way to differentiate electrical cranking from car pushing
or from downhill starting) once the computer detects the engine
rotation, it activates DME2 relays and energizes ( permanently ) the
ignition coils, the fuel pump & the oxygen sensor preheating
Idle
running:
On idle mode, detected by contact pin 31, regardless of
the idle regulation motor position, the injection time is 2.5 ms out
of sync of the engine phase (and of the valves). On cold engine the
throttle is corrected by the idle motor, and injection time and
frequency is adapted for best run. The big loop of feedback corrects
the mixture depending on the oxygen ratio in exhaust. This takes
into account the air and the engine temperatures.
On
the road:
As soon as the throttle goes out of idle position
(regardless of the throttle motor compensation), the injection time
and frequency adjusts up to 4 ms. The injection time between two
consecutive injection is controlled by the small loop feedback and
depends from the following and then is ASYNCHRONOUS from the
engine:
- throttle position
- throttle variation speed
-
intake air temperature
- engine temperature
There
is NO air flow sensor (thanks Mr Bosch, when I hear about the poor
reliability of these sensors and their price !! in
the beginning of 2004, due to the increasing amount of customer
complaints, VW reduced in France the price of the frequently
defective airflow meter from the usual 350€ to a customer price
of 75€, some mechanics still sells the hi price),
NOR altitude compensation.
The mixture corrections are made only
by the computer.
I did'nt check yet the functionning of the coal canister valve, which I suppose, make the air venting of the fuel tank going through activated coal, and is purged from fuel residues by an air flow into the intake manifold during the engine run.
3 Functional layout
An
U-shaped radiator holds the heat dissipating components
ignition
control transistor 1 to pin 1
ignition control transistor 2 to
pin 2
not used transistor to pin 23
coal valve control
transistor to pin 5
integrated circuit controls the idle motor
(DC motor, two wires) pins 15 & 33
integrated crcuit controls
injector pin 11
same IC controls DME2 relay pin 14
5V
regulator for electronics and « reference »
voltage
One
microcontroller Siemens 68 pins (family 8051 ?) who receives (but
not necessary the only receiver) the oxygen sensor.
The DME
relays wears rapidly out. He drives an important current (maybe 10A)
for the ignition, fuel pump, Oxygen sensor heating. I simply added
another one, in parallel.
The computer connector has a visible
locking but a poor contact insertion (1.5mm, from the 12 available),
associated with a possible electrochemical potential difference (the
pins are tin-covered, the plugs are another grey metal) All this
makes that, under vibration and after a certain time (at least after
the warranty time !!) the pins show a small hole and the connector
is sensitive to the moving of the incoming cables. I re-soldered the
pins. OK, it becomes now insensitive to cable shaking.
4 Failure history
1- Fuel pump failed at 120 000 km
Sudden
failure:
Impossible to restart engine, computer lamp glows on
dashboard
Cause:
Hard point in rotation, not caused by any
object in the pump fins. This may have given overcurrent. Pump
changed
2- Rotation sensor failed at 150 000 km
Progressive
failure:
First difficulty to start, then normal operation, Second
stage, starting only by pushing the car (manual gearbox). Final
stage, starting only by hi reving the engine, but unable to stay
idle unless accelerated idle
I tried to reduce the air gap: no
change
It was internal sensor failure. Sensor
replaced
(intersting finding: the sensors have same part number
by Peugeot and Citroen, labelled ( genuine Citroen part ) and on
other side of the box ( genuine Peugeot part ) but priced at half by
Peugeot (in France)
3-
Oxygen sensor failed at 170 000 km
( usually given for a 150 000
km duration, absolutely no complaint)
Sudden
failure:
incorrect mixture, black smoke, spark plug immediately
covered with carbon
Sensor (ox?) changed, an enquiry over the Net shows that it exists barely two sorts of sensors, the screw fixing are all the same. Next time I will probably save $150 by buying it on US market
4- Computer failed at 170 000 km plus a few days
Sudden
failure:
brutal stop of injection when moving down, even slowly,
the gas pedal, the power comes normally only after one long second.
A second computer bought in a junkyard, died after one hour of nominal work.
The measurements made around the computer seemed correct including fuel pressure. Only the voltage of the common point between throttle potentiometer and the temperature sensors seemed strange, all around 5V, which makes the pots having voltage variation of almost 10 mV in a noisy environment, when higher voltages could be available (seemed strange design choices). In fact this line is a reference ground connected in star mode to the electrical ground. This line was burnt on the computer PCB ( see picture ).
I repaired it with a small 12V 4W lamp (cold resistance around 4 ohm, which gives me a small inaccuracy of the sensors and pots, all of them in the 1000 ohms range, I might remove this lamp after all is finished) instead of a wire. The computer becomes now failsafe for this kind of agression.
Then, IT WORKS smoothly, with a small hickup every half hour or every hour. I suppose that this hickup reflects the primary cause of the computer failure, some sort of intermittent short-circuit.
This short-circuit tends to appear more often
I supposed then that the overcurrent created during the fuel pump failure damaged the wiring harness and that the damage increases with time
I decides to cut all the lines hooked to the reference ground and to rewire it
GOTCHA : total and final repair ! Confirmed by more than 3000 km problemless ride
Conclusion
A failure of the fuel pump was the root cause, with no immediate effect.
With time , a failure appears once (it destroys the computer at the first occurrence), and more and more often, probably due to moisture in burnt isolation, thermal cycling, vibrations
I modified my computer to make it fail-safe (at the only cost of a 12V 4W bulb !! )
I seems to me that this kind of intermittent failure goes through the standard diagnostic tools from the mechanics, but the customer cost could be skyrocketting. A good mechanics could have charged me:
2 computers 1700€
1 TBI 750€
sensors 250€
fuel pump 160€
hourly charge
It took me only some time, some thinking, a minimum amount of money (65€ for a junkyard computer), and another car to drive
|
© Gary Sullivan
Maintainer of this page. |
|
