garrett376

On my coupe I get an intermittent engine light, which ONLY occurs at one point after driving it - it's very odd, actually. I will drive to the office, about 8 miles of roads, some up to 50mph. The engine gets warmed up... when I arrive at the office, the car is turned off, no problems on the way there.

Then when I drive home, after about 1-2 minutes of the car running, when sitting idling at the same light (give or take 50 feet), the check engine light will come on when stopped at the light - no driving with the engine at speed. So, I read out the fault with the Hammer but I am having trouble understanding a couple of the readings.

The display is as follows:
1. Closed Loop System rich not present
RPM: 880
Load signal: 2.4ms
Counter 47
DTC: -23-

The DTC (diagnostic test code?) is a #23- which is a rich reading of the O2 sensor - according to the workshop manual, a sticking injector may be at fault here. My question: what is the load signal, counter, and even the main "cryptic" reading supposed to mean by "not present"? Does that mean the sensor shuts off?

By the way, the car runs flawlessly (is extremely fast) although smelling like it runs rich at idle - but with a cat-bypass and a baffle-less primary muffler with no secondary, maybe that's the way it should be.

Any ideas? Thanks guys! And Happy Holidays!

Red rooster

garrett376,

I would hook up your Hammer again and go to Actual values .
Look at the Lamda signal with the motor at operating temperature and idling. Is it switching up and down nice and regular ?
Sounds like it isnt, as the ECU is going open loop rich at idle .

Check through all the values to see what information the ECU is getting - head temp/idle switch /air temp etc etc. Check the values for plausibility. Same old stuff-garbage in , garbage out !

Good luck

Geoff

.

Lorenfb

"My question: what is the load signal, counter, and even the main "cryptic" reading supposed to mean by "not present"?"

As mentioned, the DME unit probably has a problem with the O2 sensor causing it
to not run closed-loop. Remember, the 964 DME unit has an adaptive function to
correct for Lambda values out of range by either richening/leaning the mixture.
This is similar to the OBDII function on the 993, i.e. the TRA idle correction. Thus,
the mixture was richened because the DME doesn't "see" the O2 sensor in its
normal range, i.e. an abnormally low O2 sensor voltage.

The load value in milliseconds is the injector pulse width time.

Adrian

Not that I disagree at all with Loren's analysis here, but just one thing to keep in mind before looking at replacing an O2 sensor.
You have to consider that the mixture is too rich and that the O2 sensor has tried to adjust to tell the DME to adjust the mixture and the DME has been unable to do so.
I suspect also that you are stuck in open loop mode at all times, but it might be too much fuel and not enough air rather than a failed O2 sensor.
Have you noticed a change in fuel consumption?
What other mods have been carried out?
Are you sure the O2 sensor is actually connected and the wiring was not damaged during the installation of the cat bypass pipe? I assume you did this yourself. Believe it or not I know of one case where a mechanic actually twisted the wiring out of an O2 sensor and broke it doing this installation for a customer.
Geoff's suggestion to test some of these things would be a good idea rather than shotgunning and replacing parts not actually broken.
I am also not sure of the impact of your exhaust systems mods to the overall scheme of things either. The exhaust gas flow will be somewhat faster than normal.
Ciao,
Adrian.

Lorenfb

"Not that I disagree at all with Loren's analysis here, but just one thing to keep in mind before looking at replacing an O2 sensor." - Adrian -

Adrian is correct. Although I didn't explicitly state that the O2 sensor itself was bad,
that's implication given by my statement. That's what happens when one doesn't fully
clarify what's posted. My statement should have stated "O2 sensor system".

garrett376

Loren, Geoff and Adrian, thank you for the pointers and discussion, I really appreciate it. I am one who'd rather fix it before replacing it... so no shotgunning here!!
The sensor seems to respond normally according to the Hammer - its varying voltages seem consistent at operating temperature. I have not noticed excessive consumption of fuel, or even black smoke on hard acceleration from the rear end (not a good test, but used to help on my 65 911). The car has been the same for almost 1 year with no changes other than a lightweight flywheel installed several months ago, but the problem did not appear until the last month which is why I am leaning towards an injector or fuel pressure regulator not functioning properly. I even did a track event since the problem first occurred, with no faults occurring during that excercise. I find it interesting that the fault occurs only at idle (closed throttle). Any ideas why?

garrett376

I just checked the running actual values for my O2 sensor - when it's cold (within the first several minutes of running) it registers at 0.02 - 0.09V. After a couple minutes of idling, the value now changes from 0.09V - 0.80V and the check engine light reappears and stays on. According to the manual, the sensor should read from 0.150 - 0.900V. It seems my sensor is reading below the range it should, and based on the manual, the control unit will therefore detect this (value below 0.10V) and run in no-O2 sensor mode. When the accelerator pedal is pressed, thus causing an enrichment of the mixture, the O2 sensor voltage appears to increase. So it seems from those findings that the O2 sensor itself is the problem, right?! I think that if it was the mixture itself causing the sensor to read out of bounds and triggering the fault (due to a very rich mixture from a stuck injector perhaps) I would be reading higher than 1.0V to cause the fault - which the sensor does not. Does that makes sense? Thanks guys for listening and helping. Diagnostics is fun!

Adrian

Idle is basically a fixed mixture. The ISV allows X amount of air into the engine to be mixed with X amount of fuel based on the idle program in the DME. The variable is engine load and that is why inputs from the air conditioning system are required and cylinder head temperature as well.
There is a key figure missing from your test. The temperature of the O2 sensor. The sensor only works correctly at 600F and above. I would suspect that your sensor is just relying on the sensor heater.
To be perfectly honest the data provided above assuming correct operating temperature tells me that the O2 sensor is actually working and the exhaust gas measurements are out of range not the O2 sensor. This indicates an idle mixture problem of too much fuel.
Once you hit the gas and you are in Closed loop mode the DME controls the fuel/air mixture for emissions control.
Enrichment is usually explained as the system adding additional fuel for certain purposes such as hard acceleration or cold starting which in turn removes the emissions control over the mixture.
The other question I forgot to ask is what is your real idle RPM? You have the hammer so you should be able to provide an accurate read out.
By the way if you have been running rich for some time the rear bumper will be yellowed rather than blackened. You have said also that you can smell fuel which indicates open loop mode.
Ciao,
Adrian.

Laurence Gibbs

Not sure if you have made it clear or not but does your 02 sensor cycle or stay at a fixed reading when in closed loop? my understanding is that it should cycle up and down slightly to constantly adjust afm. I am also confused as to why if the 02 sensors full range is 0.1 - 0.9 you have 0.8 . The sensor would in theory be wanting to go to full range registering the rich mixture. If the sensor is not reaching 0.9 it will never back off the mixture enough(I would guess?). This though begs the question hows does the dme decide the mixture is too rich and go open loop. Equally assuming a rich mixture caused by a sticking injector should the 02 sensor not be trying to back off the mixture ?? i.e. be heading towards lean? before it gives up and goes open loop? Just a few questions and theories going around in my head. May be complete twaddle.

dfinnegan

Marc, with regard to the stuck injector theory, can't you test the injectosr by cycling each with the Hammer and listening, or feeling, for their movment?

Cheers,
Dave

springer3

Garrett:

Forgive me if I am stating the obvious, but some may not know the name "Lambda" was given to the O2 sensor because it has an L- shaped response curve. Its output is essentially a step change in voltage when the mixture goes from rich to lean or vice-versa. That is the reason the mixture control is always "hunting". It is almost impossible to control within the narrow mixture range where the signal is between rich and lean. The fact that you are seeing approximately 1 volt changes in the Lambda signal shows the sensor is working.

Things to check:

Fuel pressure - there is a test port on each fuel rail near the rear of the engine. Engine off. Connect pressure gage. Jumper the fuel pump relay. Read pressure while pump is running (~40 psi IRRC).

Fuel pressure leakdown. Remove fuel pump relay jumper (pump off). The system should hold pressure. There is a leak down spec, something like 10 psi after 15 minutes. The one time I ran this test was on BLK964's car, and it held pressure with no measurable loss. This test should identify a stuck/leaky injector.

Compression. God forbid this is the problem, but cylinder leakage affects the idle long before the engine loses power at higher RPM. It is better to get this bad news before spending a wad on things that won't help.

I have a fuel injection diagnostic test kit if you want to borrow it (I would love to have a credit on a Hammer loan next time I need it). There is a fuel flow rate test in addition to the pressure and leakdown. Run that one while you are at it. It is impressive how much fuel a 964 fuel pump can move.

Lorenfb

"Idle is basically a fixed mixture." - Adrian -

The idle AFR (mixture) is controlled at idle just as it's controlled under load with the
additional feature (as mentioned before) of the adaptive correction of the O2 range,
i.e. if the DME determines that the O2 system is too lean it will richen the AFR by
increasing the injector pulse width, 2.2ms to 2.4ms. This only occurs at idle and
not under load as does the 993 OBDII system.

Be aware that the system does not enter closed-loop (reading the O2 sensor)
until the engine reaches a certain temp and as such ignores the sensor.

The basics need to be checked:

1. the cold/hot head temp sensor value
2. the cold/hot fuel pressure
3. the cold/hot open loop O2 sensor value via voltmeter (determines actual AFR)
4. cold intake air leaks causing a lean condition
5. the cold/hot AFM air temp sensor

Note: Read page 2-02 of the Carrera 4 '89 manual WKD 495 121
for additional info on the idle AFR adaptation.

garrett376

This morning I took a drive - one eye on the road, one eye on the Hammer display!!! As usual, no warning light on the way to to the office. It was fun to watch the characteristics of the O2 sensor reading. At idle, the sensor did the usual: varying from 0.06-0.80V. On hard acceleration, the number usually climbs, or stays up higher in the range (0.40 - 0.80V). The car accelerates so fast, that I can't watch it for too long !! When letting go of the accelerator pedal (so the car is coasting but at RPMs of 3000-2000) the sensor reading sits at 0.06-0.09V - always below 0.10V. It was neat to watch.
When I arrived at the office, the engine was clearly warmed up after the 10-12 minute drive, and I checked all the sensor values.
The injector pulse width varied at idle from 2.25 - 2.50ms
The cyliner head temp was 124 C or 255 F
The AF Sensor was reading at idle .21 - .22V
The Intake Air Temp was 23 C or 74 F
The idle RPM was jumping from 800-840-880-920
The O2 sensor values at idle were in clusters - a low value for a bit, then a few higher values:
.06 - .08 - .15 - .75 - .79 - .80V
after a few more seconds: 0.15 - .14 - .09 - .58 - .72 - .79V

Something interesting I did while driving this morning with regards to the check engine light... Last night the check engine light came on. I never cleared the fault. When I was driving for the first half of my drive (about 7 minutes) when I came to the stoplight, I cleared the fault code. The O2 sensor readings before and after clearing the code remained the same. Does this mean that when the engine is restarted after having a fault, the O2 sensor is being used even though previously it was ignored?

Sounds like I need to find myself a pressure gauge to check fuel pressure. I still find it odd that I get no fault code when driving in the morning, only at the end of the day! We'll see - I'll repeat tonight on the way home (glad the Hammer has an illuminated display!).

Lorenfb

Key Point:

Because of the adaptive idle AFR correction, the O2 sensor values are the CORRECTED
values and really can't be used for analysis, i.e. only the open-loop values provide
meaningful data.

Example:

The DME determines the O2 sensor range is (.05 - .50 too lean) so it now increases
the injector pulse width and the range (.20 - .80) is now acceptable for O2 control.
So, possibly because of an intake air leak (lean AFR) the DME made the correction
but also sets a fault code and CEL.

Lorenfb

"Forgive me if I am stating the obvious, but some may not know the name "Lambda" was given to the O2 sensor because it has an L- shaped response curve." - springer3 -

Not true. The term Lambda was used because of the shape of the Greek letter Lambda
which has a top in the form of a kinda horizontal "s" which is very similar to the narrow
band O2 sensor voltage response curve used on the early cars.

Check here ( www.systemsc.com/graphs.htm ) for the narrow band O2 sensor voltage function.