Questions on ECU, oxygen sensor signal, injector pulse width for 700 series.

Thread view:

Enable EMail Alerts

Start New Thread

Thread rating:

Duong Nguyen - 30 Jul 2004 02:23 GMT

I have been trying to debug a problem of falling fuel economy on my
760 V6, 1988, LHjet 2.2. The fuel economy for this car is now 5km/l
(11.7mpg) but used to be around 6.5 km/litre (15mpg) city driving.
This despite increasing the tyre pressure from 28 psi to 35 psi. There
is no other symptom, the car still drives very well.

The oxygen sensor signal when the car is fully warmed up and idling is
from .2V to .8V with a lots of .2V noise spikes that seem to come from
the computer (the disconnected oxygen sensor signal is clean.) The
signal makes a complete cycle (i.e .2V to .8V and back to .2V) in
about 4 seconds which I think is too slow. The transition time is
about .4 second. It also responds in the correct direction to a lean
condition (introducing a vacuum leak) and rich condition (removing
vacuum from fuel pressure sensor.) albeit after to couple of seconds
delay (why?). This oxygen sensor is new and produces similar signal to
the old one which I thought was faulty.

The injector pulse width (which I forgot to measure) with this oxygen
sensor signal drifting up and down, is absolutely constant indicating
that the computer is probably still in open loop mode and ignoring the
oxygen sensor. In that case, what is causing the change in the oxygen
sensor signal? Is this new sensor faulty (and the old one also
faulty?) I'm more inclined to think that there is still something that
stops the computer from entering close loop mode instead of both
sensor faulty. On the other hand, the lack of transitions on the
oxygen sensor is the only thing remains that could prevent close loop.
Is there another sensor signal that I haven't thought of? It is
probably not a temperture sensor since in this car there is only one
temperature sensor and the temperature gauge is exactly in the middle.
Also the rpm has dropped down to the correct value of 750 rpm from a
cold value of 1000 rpm indicating the the computer knows that the car
is already warmed up.

My questions are:

1) What are the conditions that must be met before the computer enter
close loop mode?
2) How can I determine if the computer is in close loop mode?
3) Does anybody know how the injector pulse width shoud behave with
respect to the oxygen sensor signal? The range in milliseconds?
4) How many transitions should the oxygen sensor make in a second with
the car warmed up and idling.
5) If I looked at the Air mass signal, oxygen sensor and injector
pulse width what should I see? Pin number on the AMM?
6) And finally, does anybody have a program listing of the LHjet 2.2
EPROM or similar?

As you can see I haven't go a clue what the problem is yet and really
want to make sense of these signals. Any suggestion welcome.

Reply to this Message

AJS - 30 Jul 2004 13:57 GMT

> I have been trying to debug a problem of falling fuel economy on my
> 760 V6, 1988, LHjet 2.2. The fuel economy for this car is now 5km/l
> (11.7mpg) but used to be around 6.5 km/litre (15mpg) city driving.
> This despite increasing the tyre pressure from 28 psi to 35 psi. There
> is no other symptom, the car still drives very well.

I haven't read all your post, just skimmed it (I will reread later). Have
you tried resetting the ECU? Have you then take the car for a run and let it
remap?

AJS

Reply to this Message

Robert Dietz - 30 Jul 2004 14:32 GMT

> I have been trying to debug a problem of falling fuel economy on my
> 760 V6, 1988, LHjet 2.2. The fuel economy for this car is now 5km/l
[quoted text clipped - 45 lines]
> As you can see I haven't go a clue what the problem is yet and really
> want to make sense of these signals. Any suggestion welcome.

This is all based on the ignition plugs, wires, cap and rotor all being
in good condition. Engine settings otherwise in stock or as originally
set condition.

Any Bosch Fuel injection system, starting with the 1970 system, reads
the coolant temperature first. In the LH cars this sets open loop and
increments the fuel map values as the coolant temp increases until it
reaches the threshold switching value to go to open loop operation. At
this point priority goes to the MAF. The air mass measurement is
converted to a dwell angle or pulse width signal to the injectors with
consideration given to whether on not the throttle is closed. The
validity of this calculation is compared to the expected O2 value versus
that reported by the O2 sensor. Whatever discrepancy is seen at the
checkpoint time is used as a fudge factor to adjust the injector pulse
width, and so on and so on.

There is a test point located in what appears to be a two pin AMP plug,
IIRC, along the right front fender. There should be SB/W wire that when
grounded closes the idle air control valve. A DMM connected between the
G/W wire and ground should read a fluctuating voltage between ~2v and
~12v. This sweep is obtained by turning the adjustment screw on the
MAF--you may ahve to drill two 1/8" holes in the aluminum plug inorder
to remove it with needle nose pliers. The voltage sweep when reading the
analog scale across the boot, of the meter should cycle evenly about
once per second. A second meter connected across the O2 sensor and
ground will sweep slightly slower but not as slow as .25 cps.

Lazy O2 sensors are mainly caused by tip contamination. Silicone spray
used on intake hoses or seals can be deadly to O2 sensors. Antifreeze
from leaky head gaskets or raw fuel from streaming injectors can also be
a problem.

I'd check the thermostat first, to make sure the engine isn't running
too cool; coolant reservoir for CO, to make sure the engine hasn't at
some point over heated and damaged the head gaskets; coolant temp sensor
for proper value, if the coolant temp signal is low the engine runs
richer. Pull the vacuum hose from the fuel pressure regulator to make
sure the diapragm isn't leaking. If that didn't turn up anything I'd
make sure the injector grounds on the manifold were clean and tight,
retest, then bypass the fuel pump relay to run the pumps with the key
off, then pull the fuel rails and using a plug with pigtails and two
jumper leads (switching the ground side), fire each injector in turn to
check for wet tips, drips and streaming.

Other than a bad control unit, I think that's about it.

Bob

Signature

The goal of driving is to miss the maximum possible number of objects.

Reply to this Message

Duong Nguyen - 02 Aug 2004 04:30 GMT

> There is a test point located in what appears to be a two pin AMP plug,
> IIRC, along the right front fender. There should be SB/W wire that when
[quoted text clipped - 3 lines]
> MAF--you may ahve to drill two 1/8" holes in the aluminum plug inorder
> to remove it with needle nose pliers.

Does this means I can influence the sweep rate by altering the AMM
output? How easy is it to restore the orginal setting? simply remember
which way the screw pointed? If the sweep rate comes up to 1 cps does
this mean the AMM is just about correct?

> I'd check the thermostat first, to make sure the engine isn't running
> too cool;

I have already replaced the thermostat sometimes ago because the
temperature gauge wasn't where it should be (in the middle pointing
straight up.) Don't know how to find out that the engine is definitely
warmed up though.

> coolant reservoir for CO, to make sure the engine hasn't at
> some point over heated and damaged the head gaskets;

Don't quite know how to do this. Shouldn't there be "hard to start",
"oil in coolant" and such symptoms?

> Pull the vacuum hose from the fuel pressure regulator to make
> sure the diapragm isn't leaking.

Already done this, found no leak.

> If that didn't turn up anything I'd
> make sure the injector grounds on the manifold were clean and tight,
> retest, then bypass the fuel pump relay to run the pumps with the key
> off, then pull the fuel rails and using a plug with pigtails and two
> jumper leads (switching the ground side), fire each injector in turn to
> check for wet tips, drips and streaming.

I'll do this ASAP. Perhaps this is the reason why the oxygen sensor
signal looks funny.

> Other than a bad control unit, I think that's about it.
>
> Bob

Thanks Bob for a very detailed response.

Reply to this Message

hk@hk.com - 05 Aug 2004 16:56 GMT

better to measure the exhaust CO reading ,it'll indicate the outcome of
all you mentioned.

Duong Nguyen 寫入：

Reply to this Message