Motronic mysteries - a light shines!

["John Fielding" <johnf@domain.elided>]

I have being doing a bit of sleuthing on my Bosch Motronic system to cure a mysterious
intermittent fault (finally traced to a flaky contact in the harness!).  In order to do
this I built an electronic circuit to simulate the two pickups and fool the ECU into
thinking it was driving a real motor! (I needed to isolate the fault to the ECU with
certainty before I went looking for another ECU!)  I used several variable resistors to
simulate the water temperature, air temperature, AFM flap, switches for the idle and WOT
contacts and lamps for the fuel pump and the injectors and coil.  The rpm and crank pulses
were derived from an electronic counter/divider with the correct relationship (ever tried
to divide by 130 - not a simple task with standard dividers!)

The electronics took about two days to perfect and once hooked up to the ECU I could run
the "motor" at any rpm from 300 (were it thinks the motor has stalled!) to 6250 rpm (which
is were the redline is set).  Some of the things I discovered may be of interest to those
anticipating messing with these things.  The ECU is off an Alfetta 2.0L with two magnetic
pickups, one pickup counts the teeth on the starter ring gear and from this it works out
the rpm.  The second pickup is the "reference sensor" and this gives crankshaft angular
position.  The epoch is at approx. 60deg before TDC on #1 cylinder.  (I used an accurately
calibrated tacho to determine the exact rpm settings). An oscilloscope was used to observe
injector pulsewidth and to measure ignition advance.  Several "spare pins" on the ECU
connector, which are not connected to the harness give diagnostic points, such as the
ignition module drive signal and the injector drivers input control pulses, these can be
really useful to connect an aftermarket CDI box!

The fuel pump is switched by the ECU when it sees the starter cranking voltage and the rpm
sensor producing pulses.  (Unlike the other systems were the airflow meter flap has a
switch to control the fuel pump). If either of these signals are missing, the motor can
still be bump started as after approx. 1 second of rpm pulses and crank position pulses
occuring together the fuel pump is switched on despite the cranking voltage being missing.
The fuel pump remains on as long as the rpm is above 300 rpm.   Clever system !

Damien Gaboury in AD V8-#1037 gave some tips on the Bosch Motronic system fitted on his
1986 Spider as regards the VVT operation.  In it he said that the VVT is triggered by the
WOT switch.

Well, on the Alfetta ECU this is totally different!  The VVT is triggered when the motor
goes through 1500 rpm, either accelerating or slowing down.  (This is covered in the
Bosch/Alfa Romeo service manual for the Alfetta system). At the same time as the VVT relay
is operated by the ECU the ignition is cut for a period of approx. 100ms.  This causes the
tacho needle to take a dive until the VVT switchover period has timed out. If you set the
idle to a tad over 1500 rpm the running is very lumpy as half the time the ignition is
missing!  The original tacho must have had pretty severe damping to stop this effect being
observed.  The tacho in the race car is an aftermarket Taiwanese el cheapo (penurious??)
and has poor damping.  The original tacho wasn't working when I bought the basket case car
to convert to a race car so I never saw it do this odd trick.

Another thing which I was trying to work out is the effect the water temperature sensor
has on the mixture.  In a lot of books on EFI the water temperature sensor is said to set
the baseline injector period and by inserting a resistor in series with the sensor you can
fool the ECU into thinking the motor is cooler and so it increases the injector period,
giving a richer mixture.  Very useful way to get a bit of enrichment over and above the
acceleration enrichment built into the ECU software.  (As this is a turbocharged motor I
was planning to use a pressure operated switch which senses the boost level and use this
to increase the injector period).  Well, on this ECU the water temperature does zip under
these conditions.  My guess it is that the ECU simply looks at the water sensor every so
often, maybe every ten seconds, as it doesn't expect it change quickly and so increasing
the water temperature sensor output voltage (which is what inserting an external resistor
does) is not a way to switch in additional enrichment.

However, I did find one way of enrichening the mixture which I haven't seen in any of the
books I read about Bosch ECU's.  When the throttle body switch for WOT is activated the
ECU ignores the water temperature sensor altogether (proved by swinging the variable
resistor I was using to set the output voltage all the way to cold) but the WOT switch
contacts cause the injector period to increase by a fair amount.  At an injector period of
50% when the WOT switch is activated the pulse width increases to 58%.  This I proved by
activating the WOT switch with the ECU on the test bench and it works from just above idle
to maximum set rpm.  The increase seems to be 6 to 8%, at idle the pulsewidth period is
around 20% and the WOT switch increases this to approx 26%.  At 80% injector pulsewidth
the WOT switch increases this to approx. 88%, which is now getting a bit too high for the
injectors.  I measured the pulsewidth with an electronic counter operating in duty cycle
mode.

The water temperature sensor seems to only have control whilst the motor is warming up.
By setting the rpm to 1000 and swinging the water temperature sensor variable resistor
from cold to hot the pulsewidth changes, but once the "motor" has exceeded approx. 70C the
sensor has almost no effect on the pulsewidth,  if you then change the sensor back to
"cold" the ECU ignores it!  Once the ECU sees that the motor has reached above 70C it
seems to stop taking readings and only "looks" at it infrequently.  (The fuel quality
switch is also interesting.  Changing the setting with the motor running has no effect.
It seems the ECU reads the switch setting on boot-up and then changes the maps for fuel
and ignition.  Changing the switch once the boot-up period has expired is ignored until
the ECU boots again).

And this ECU has no 02 sensor input so it has nothing to do with the lambda trying to
exercise its control!  This system operates in "open-loop" at all times.

As I said this ECU is the one fitted to the Alfetta 2.0L motor built here in South Africa
and may be different to the ones fitted to NA vehicles.  The Bosch part number
is:0261-200-044, Alfa part number:195.00.11.042.00.  It would be interesting to try and
cross-reference these numbers to see which vehicles had the same ECU.  Bosch has been very
silent since I asked for further details!

What says JH???

John
Durban
South Africa
Alfetta 1.8L turbo
--
to be removed from alfa, see /bin/digest-subs.cgi
or email "unsubscribe alfa" to majordomo@domain.elided