Please help diagnosing '80 SC CIS (idle) problem
 

Dear fellow porsche drivers,

I need your help diagnosing the cause for the idle problem of my '80 SC (US version). Since I have bought the car in March (and probably long before that), the car starts instantly when cold but refuses to idle on its own. I have to slightly press the accelerator pedal for several minutes until it warms up. Once warm, the car drives and idles normal (at least I think so, but I never drove another 911 before). During warmup, it first idles very rough (like it wouldn't run on all cylinders). There is quite some grey/black smoke from the exhaust. After a minute or two, it smothes out somewhat but hunts up and down. After another minute or two, I will idle at around 600 rpm which then increases slowly to the normal 950 rmp idle.

One of the first things I learned about the car is that the original engine has been replaced with an '81. However, the oxygen sensor control unit under the passenger seat is still the '80 unit without the cold acceleration control unit.

I have already checked the following:

- checked for Vacuum leaks and found injectors are leaking (waiting for replacement O rings), fixed leaking vacuum advance unit by plugging the retard hose
- checked system pressure (5.1 bar)
- checked WUR (pressure 0,75/1,0/1,4/1,8/2,1/3,0/3,4 bar each sample in 30 seconds intervals, starting at 10 Celsius)
- checked AAR, open when cold, closes from freezer temperatur within 5 minutes when connected to 12 V
- checked all connections to the oxygen sensor control unit (FV, oxygen sensor, 15C switch, throttle switch)

I have made three videos that show the signal to the frequency valve during different states of the warmup.

Still cold (15C switch closed)
15C switch opening
15C switch open

The first video starts when the 15C switch is still closed. It seems odd that the duty cycle starts at around 15% and then jumps to 30%. According to the documentation I read, it should be 85% for the '80 control unit. If I understand the function of the mixture control unit and frequency valve right, this would mean the mixture is very rich which would explain the smoking engine.

The second video shows the duty cycle when the 15C switch opens. THats when the multi meter jumps to 12 V.

The third video is the frequency valve after some warm up. At this stage the engine idled on its own. I don't understand why the duty cycle slowly increases to over 80% and then jumps down to 50%, repeating the cycle again.

Any idea what would cause the FV to act as described? Any additional tests I would do?

Thank you, Manfred

On the duty cycle issue, disconnect the O2 sensor and see if you still get the wild 50%-80% bouncing. Just do this for a test to see if the Lambda unit is actually reading the input from the O2 after the 15deg switch closes.

mANNI,

I had the same issues that have been fixed recently.

I would suggest checking your ignition wires. Are they original? If they are, they are about 30 years old and will need to be changed. This may be all you need.

The second thing I would look at would be if you have vacuum leaks, have your injector seals checked for leaks. If the orings have not been changed, they will need to be. I'm almost sure that they are hard as rock and leak profusely.

Good luck.

Manfred,

I can't see the scope reading in the video to tell much from it. Do you have an analog dwell meter ? Your understanding of the FV is not correct. When the engine is in open loop, before the temperature switch opens, a low duty cycle results in a lean mixture. A cold engine needs a rich mixture during warm up and the high default duty cycle (85%) combined with the lower control pressure of the WUR creates the rich mixture. In open loop the O2 sensor is irrelevent and the ECU doesn't know anything about the mixture. When the switch opens and it goes into closed loop, the system now knows what the actual mixture is and the FV duty cycle that you are reading is the correction factor. A low duty cycle in closed loop mean the actual mixture is rich and the ECU is pulsing the FV slower to lean it out. The duty cycle should dither back and forth around 50% when hot in closed loop. An analog meter allows you to see the rhythm and extent of the swings. If it is slow and erratic, the O2 sensor is usually bad. As the engine wears and compression and vacuum at idle drops, most US 911SC's with lambda will cold run and have better WOT AFR with the hot closed loop mixture set to 25-35 duty cycle. When you adjust the mixture on a CIS lambda 911, you are only really effecting the cold and WOT mixture because the system will bring the average mixture back to stoich below 35% throttle hot. I think you are getting bad data, your mixture is probably set too lean and combined with the vacuum leaks, results in a poor cold idle.

I just went through this. My cold idle problem was an airbox leak at the seam and a few old vac lines that were rotten and leaking. With new ox sensor and all the leaks fixed that I could find, it starts excellent cold and runs at 1400 rpm for a few minutes and settles down to 950. Duty cycle is as suggested both cold and warm and it runs excellent so far! Other leaks to check for are vac leaks at the hose that runs from the air filter rubber boot to the oil tank, injector sleeves and orings, all vac lines,the intake runners and most likely the airbox at the seam.

Scott,

I've disconnected the O2 sensor as per your proposal:

- as long as the 15C switch is closed (engine still cold), the control unit sends the odd 15% duty cycle to the frequency valve
- as soon as the 15C switch opens (engine warms up a bit), the frequency valve signal changes to a stable ~50% duty cycle. In other words, no bouncing of the signal anymore.

What is the conclusion from this?

Thanks, Manfred

Paul

thank you for the explanations. Sure you are correct, the 15% duty cycle will cause a lean mixture.

What I do not understand is why the 15% duty cycle when the engine is still stone cold (the 15C switch is still closed). In this mode, no other sensor should influence the control unit and per my understanding, the frequency valve signal should be fixed at a duty cycle of 65% (85% for '81).

Would it be a valid conclusion that the control unit is faulty or is there another explanation?

One more question: what exactly do you mean by "bad data"?

Thanks, Manfred

Well, I have the same exact problem on my 80' So maybe we can solve it together. The O2 sensor is sending an incorrect signal to the lambda computer. I see two possible reasons for this:

1. The air/fuel mixture is outside of the range of the O2 sensor and therefore the information being sent to the O2 controller is wrong.

2. The signal from the O2 sensor is "dirty" somehow and again it's not the expected output when it arrives at the o2 controller.

So in my case the mixture is constant for a US 80 at 0.8%, I do not have any vacuum leaks and all of my CIS components are functioning normally. Once my car is warmed up the signal is constant coming from the O2 sensor and the duty cycle is spot on.

So I can fix the issue in a few ways, one would be to override the 15 deg switch until the car is completely warm. But in a bypass switch I suppose. I was told that a company made an adjustable 15 deg switch for this very purpose but I can't find it.

The second would be to override the O2 sensor until the car is entirely warm.

I'm still looking but it appears like the 80' at least has a "lazy" O2 sensor or could possibly benefit from a heated O2 sensor.

The thing for me is this is like a 5 min long event during warmup and after that my drive ability and power is excellent, so is my start up.

Tomorrow I will use the storage oscilloscope to record the signal from the O2 sensor. That is, after I have repaired the =2 sensor connector that is broken (when disconnecting the sensor today I ripped off what was left of the connector.

- Manfred

Would it be a valid conclusion that the control unit is faulty or is there another explanation?

One more question: what exactly do you mean by "bad data"?

Manfred,

15% duty cycle is " bad data". Either your test equipment, connection or ECU is bad. I suggested you use an analog dwell meter to confirm the reading. I would also check the duty cycle reading at pin 17 of the ECU. If you get 15 % duty cycle on both instruments, the ECU is bad. I would avoid assumptions. If you are getting 50 % duty cycle hot, that is "good data", in range, and the erratic results when the O2 sensor is hooked up would suggest that the O2 sensor is bad. The wire to the ECU is shielded and the conductors must be seperated. After you repair the wire, I recommend you do the voltage and rich stop (ground O2 sensor wire=95% duty cycle) and lean stop test (connect O2 sensor to a 2 volt battery= 10% duty cycle).

I'm leaning towards a bad ECU for my particular issue, just trying to find a way to rule it out without having another on hand. I've been in mine to check for bad solder joints and etc, but so far nothing.

Actually, I have measured the duty cycle on pin 17 on the ECU. Tomorrow I will try the tests you propose. One thing that puzzles me is that the 15% duty cycle when cold is actually the inverse of 85% which would be correct for cold enrichment. One of the previous owners of the car has hacked the wiring in the engine comparment when installing an MSD ignition such that I can't even find the test connector for the CIS. I will have to find out what they did with that pin17 wire to make sure I'm measuring the right thing (signal).

That is what I also would like to verify or rule out. Re-soldering all components in the ECU is one of the things I still have to try.

This happened today on my way home from work, so when I got home it was still doing it and I could try a few things. I could get the cycling to stop by changing the mixture from 0.8 to 1.5. I wonder if it's not this crummy ethanol gas causing this for me?

You just bought the car. I would suggest waiting a bit before I started doing anything.
By your description of how it starts your issue is minor if at all.
It is a little rich on startup .What's the ambiant temp on startup? You need to check your wur pressures first of all. That will tell you lots. If it runs fine once warm ,then just wait a while and drive it ,you'll find that it starts differently everyday depending on what kind of day it is.
If anything a minor adjustment of the mixture screw at the fuel distr is all you need.

Thanks to all who contributed with their help and knowledge to this thread!

Today I checked some additional things. I found the test port wire in the trunk. A PO has cut off the connector, used the 12V and ground for the MSD ignition and isolated the blúe/white wire that goes to pi 14 of the control unit. There was nothing connected to that wire so still no explanation for the 15% duty cycle when cold. Further, 12V and ground at the control unit checked out OK. Repaired the oxygen sensor connector and have a 0.5 V signal at the control unit. As a last resort, I re-soldered the complete control unit with no luck.

The only conclusion for me now is that the control unit itself is faulty and I need to have it repaired.

Can anyone of you recommend a repair service?

Thank you, Manfred

Lonewolf,
as per Pauls reply further above in this thread, the 15% duty cycle when cold causes a lean condition. Anyway, the effect is that the car stalls and doesn't idle. When approaching a stop due to traffic conditions, this makes driving unpleasant when the car is still cold.
When driving the warm engine with a steady speed, I've found a slight hesitation that repeats with a period of 10 seconds. I didn't report this before because I only noticed it today during a test drive. This hesitation/surge corresponds to the strange cycling of the FV duty cycle I've measured and described above.
To make sure I'm not hunting after some kind of meausrement problem I've traced the signal for the FV directly at the control unit output in addition to the test signal on pin 17. It is exactly the inverse of the test signal which is due to the fact that the control unit switches the ground pin of the FV to 0V. In other words, I'm totaly sure the FV gets a wrong signal during warmup.

No, I haven't adjusted the fuel distributor. I have no CO meter available and I can't use the duty cycle due to its wild bouncing.

I don't feel offended at all but appreciate any help I can get.

- Manfred

"15% duty cycle is " bad data". Either your test equipment, connection or ECU is bad. I suggested you use an analog dwell meter to confirm the reading. I would also check the duty cycle reading at pin 17 of the ECU."

Too much confusion on the Lambda ECU, FV, and duty cycle on this thread,
e.g. the FV pin is 15 & not pin 17. Read here for more info; http://forums.pelicanparts.com/porsche-911-technical-forum/673097-need-recommendation-oxygen-sensor-control-unit-repair-service.html#post6701976

Sorry .I disagree.
Your making assumptions. when at speed and you get surging.this can happen with O2 hooked up. it's common. just as an example.It can be annoying and why a lot of people disconnect them.

Your in Bavaria.what's the altitude?
Did you buy the car in the same area?
Your rich so either too much fuel or not enough air. If the car was tuned at sealevel and your at 3000FT that could be it right there just as an example of how simple it could be.
You don't need an exhaust analyzer to adjust the FD

I know that pin 15 drives the FV, pin 17 is the test signal going to the three pin test connector in the engine bay. I have conencted both signals to an oscilloscope. When cold, pin17 is at 12V for 15% of the cycle time. Pin 15 is 15% at ground, it is the inverse of pin 17 because the ECU activates the FV by switching the ground side of the coil. This means that the FV is activated during 15% of the cycle and not during 85% as it should be.