Need some input -> Problem with 3.2 transplant going lean and misfiring after two min
 

Like John Walker a month or so ago, I've got a 3.2 Carrera engine that is giving me fits. I do think that these are probably the most difficult cars to work on because there are no codes or warnings or any real hints as to what the problem is. With most mechanics, you start swapping in and out parts to try to find the problem. Diagnosis with an oscilloscope is time consuming (compare to checking trouble codes on something like a Boxster), and the use of an oscilloscope is an art - if you don't use one all the time, it can be difficult to get it quickly dialed in.

Here's the story, it's a 914 with 3.2 transplant. I did not build this car, and when it arrived at my shop (with the 3.2 installed) it never ran correctly. This was 11 years ago (2009). The car sat for a long time (I was irritated with it and the fact that it didn't run properly). Then it was stolen, and disappeared for about two years. Then it was located, and found. I emptied the tank, changed the fuel filter, and started it up and it ran the same. The problem was a hunting idle. That issue is gone now, but another problem has shown up.

Here is the issue: When you start the car up, it runs fine for about a minute or two. The idle goes up to about 1,100 or so. Smooth running, no misses. Air/fuel mixture is rich (haven't set that yet) and seems within range). Then, after about a minute or two, the idle drops down to about 850 or so (where I set it), and then the car goes lean (to about 18 or so air/fuel mixture) and starts missing. If you rev the engine just a little bit off of idle, the air.fuel mixture goes straight to 14.7 or so, but the car is still flubbing / missing at this point.

Since the car runs perfectly fine for a minute or two and then all of a sudden starts to flub / misfire, it would seem to be a fuel injection problem for sure (not a mechanical or ignition problem). I replaced the plugs, cap, and rotor regardless because the car was stolen and sitting around for a long time, so who knows what happened to it, etc.

It's also a transplant into a 914, so there are some unusual "modifications" that appear to be done to the car (it has two boots with some open ports sealed, etc.). It also has the thermo valve plugged up, and all of the evap stuff removed. It's got a sport muffler on it too, but since it runs fine for the first minute, I don't think the exhaust has anything to do with this issue.

Here's what's been done so far:

- Tested the air flow meter with an oscilloscope (no issues found)

- Smoke tested the car for vacuum leaks when cold (found a very small leak which I fixed, but didn't fix the issue)

- New plugs, cap and rotor

- Jumpered the test plug and set the idle to be 800 rpm.

- Checked each wire with oscilloscope and inductive sensor

- Checked idle control valve -> clean and working fine

- Checked the pulse to the idle control valve (both sides, looks normal)

- Checked TDC sensor -> stumbles and shows odd signal when the car is misfiring (expected?) Appears to show okay when running just on the starter.

- Fuel injectors - sent out to be cleaned and balanced.

- Checked the injector operation (clicking on and off when powered)

- New fuel filter.

- Checked altitude sensor

- Checked idle switch

- Checked voltages - all okay

- Checked the full throttle switch - okay


To Do:

- Check fuel pressure

- Check temp control switch (this is plugged off - may be leaking when warm?)

- Check the vacuum controled fuel pressure regulators (although they shouldn't theoretically change 1-2 min after starting the car).

- Check intake air temp sensor?

- Check cylinder head temp sensor (Temperature Sensor II)

- Smoke test the car for leaks when it's warm not cold


Thoughts:

- It's not the O2 sensor since the problem is shown at idle, where the O2 sensor values are ignored.

- Although I got some wavering results on the TDC sensor, I think that's a byproduct of the misfire, not causing the misfire, since running the car with the starter worked fine, and the car also runs perfectly fine for the first minute. Also, an intermittant TDC sensor shouldn't cause the lean condition at idle.

- Not the DME relay, seems to be working fine.

- Cylinder head temp sensor - if defective, the car wouldn't start.

Again, the car runs great when cold - idles at 1,100 with a slightly rich mixture. Then after a minute or two, it drops down to 880 where the air-fuel ratio goes to about 18 and it starts missing and "flubbing" out the exhaust.

By the way, I bought a really neat breakout box setup on eBay. It's an OTC 3525 breakout box (typically about $100 on eBay) with the Ford adapter 007-00055. Together these use the same connector as the BOSCH Motronic setup for the 3.2 Carrera. A factory "BOSCH" breakout box is near impossible to find, and when you do, they are about $600+. This is a cheap and effective solution.

Let me know what the expert's thoughts are on this...

Thx,

Wayne


Here's the car as we're working on it:
http://forums.pelicanparts.com/uploa...1588489309.JPG

Here's the setup in the trunk with the oscilliscope on the left, the breakout box on the right. I also have the Innovate tool with the O2 sensor shown here that allows me to check the mixture of the car.
http://forums.pelicanparts.com/uploa...1588489309.JPG

Here's the breakout box with the BOSCH pinouts labeled. I just saved someone about an hour's worth of work, if someone buys one of these on eBay:
http://forums.pelicanparts.com/uploa...1588489309.jpg

Oscilliscope pattern of the square-tooth wave pattern sent to the ICV (idle control valve):
http://forums.pelicanparts.com/uploa...1588489309.jpg

Another photo of the car:
http://forums.pelicanparts.com/uploa...1588489309.JPG

This shows the inductive sensor (got it with the Innovate tool) that you stick on the spark plug wires to check for consistent spark. Looks pretty good to me. You can also do this check with a simple timing light.
http://forums.pelicanparts.com/uploa...1588489309.JPG

I tested the air flow meter with the oscilliscope. Basically set it up for a slow trace and then pushed the flapper open and closed. No issues seen here:
http://forums.pelicanparts.com/uploa...1588489309.jpg

This was the signal from the rpm / speed sensor. I wasn't too sure about this one, I will take a closer look at what I think it's supposed to look like tomorrow. This and the top-dead-sensor readouts looked okay to me, but I want to double-check them:
http://forums.pelicanparts.com/uploa...1588489506.jpg

This is the TDC output:
http://forums.pelicanparts.com/uploa...1588490064.PNG

Here's a video showing a startup from semi-cold. As you can see, the car runs fairly well in the 40 seconds or so. Then (at around 40 seconds in) the mixture goes super-lean and it starts to run very poorly (this is at idle). Then, I rev it a little (to take it off of the idle switch), and you can see that the mixture immediately comes back to the 14.7 or so range (around the 1:57 mark in the video). It's still not running well at this new rpm (about 1500-2000), but at least the mixture is better. (then at the end of the video the O2 sensor falls out again, and the mixture reading goes to about 20 - which is what ambient air pretty much is).

What is interesting is how well it runs at immediate (cold) startup. This leads me to believe that we have some type of sensor problem (not an engine, exhaust, mechanical, or vacuum leak problem).

I know from reading the BOSCH manuals that there is an immediate start-up / warm-up cycle that lasts about a minute to ninety seconds. It's running fine in this cycle (where rpm is about 1,100). As soon as this is over, and it lowers the rpm to 850 or so, then it starts being lopey, and the mixture leans out quite a bit.

In at least one spot on the video, the O2 sensor fell out of the pipe (I don't want to clamp it down because it would scratch the tailpipe of the brand new muffler - maybe I'll get some soft brass or aluminum foil and mount it there more permanently tomorrow because it doesn't seem like this problem is going to be solved any time real soon.

Thanks,

Wayne

<iframe width="560" height="315" src="https://www.youtube.com/embed/EeqYDWjlbro" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>

Are you sure about the statement on the O2 sensor being ignored. I know it has to warm up and is ignored for about 2 min when starting as it is still warming up, but then I believe it is taken into account.
The O2 sensor is to be replaced every 60K miles. When I did mine at 80K miles I noticed a much better warm idle.

All your signals look Ok by the way. I haven't tested my inductive probe yet. How/where do you use it exactly ?

If you PM me I can send you the official DME test plan document.

1/ You could measure the lean/rich input from the O2 sensor on the scope and compare it to your external o2 meter to check ?
2/ Check the throttle Idle switch - does it clearly connect to GND ?
3/ Does it keep misfiring when you connect B and C terminals on the test port ?

Sent from my iPhone using Tapatalk

I kinda 'made my own' breakout box from a Volvo 940 ABS unit (same connectors). I included some leds to show the switched and see if the idle switch and WOT switch do connect to ground (pull the pin low).

First scope img : speed sensor, amplitude should be >2V
Second scope img : 2 channels of the ICV valve pulsing

http://forums.pelicanparts.com/uploa...1588504516.jpg
http://forums.pelicanparts.com/uploa...1588504516.jpg
http://forums.pelicanparts.com/uploa...1588504516.jpg

Mine won't run if the AFR gets above 16. Never liked the exhaust robe for accuracy. You smoke tested cold but how about during the time when it is going lean. Always sprayed something around the intake looking for the problem in these conditions.
Good luck!

"- New plugs, cap and rotor"

Did you forget ignition cables? At this point I would not assume anything...even simple things.

What's it do with the O2 sensor unplugged? What is the initial CO reading at hot idle with it unplugged? HC reading at hot idle? Tried another airflow meter yet?

Those are key points for this problem! Pin 24 (O2 input) should be checked with the breakout box.
With the O2 disconnected, the measured voltage there should be about .50V.

The O2 function is active at idle, i.e. the O2 voltage is an active input to the DME ECU
once the engine enters the "warm" mode.

What is the cold start AFR?

Check the fuel quality switch on the DME. Position 3 and 7 remove 4% fuel roughly. Position 0 is stock, which is full counter clockwise.

Two things come to mind with the AFM. Is it missing the silver plug over the CO adjustment port? This changes the mixture at idle. At hot idle, pull the o2 sensor and look at your AFR. It should be very close to to or just under 14.7 . May want to try setting 'baseline' with the Motronic. Hot idle, pull o2 sensor and adjust the CO adjustment screw on the AFM to 14.3-14.7 AFR. Reconnect o2 sensor. Jump the test port plug to disable the ISV, then reset your idle speed. Your AFR should not change much. I think over the years many make adjustments using the CO port to try and get the car to pass. If the silver plug is still in place on your AFM, then this is probably not your issue.

Does the black cover on the AFM look like it was removed? My 3.2 was very lean running when i purchased it. The clock spring in the AFM is factory set, this sets the tension on the 'barn door' of the AFM. I found that in the 160k miles my car covered prior to my ownership, the AFM clock spring was likely adjusted to increase spring tension, thereby leaning out the mixture...no doubt to pass emissions at some point because it was running rich. This should NEVER be adjusted as they are likely factory set with a flow bench.

http://i.imgur.com/FbtnjZ8l.jpg

This thread explains the clock spring well and is what helped me with my issue. For peace of mind, i was able to find a low mileage unmolested AFM i currently use.

http://forums.pelicanparts.com/porsche-911-technical-forum/835870-afm-mixture-am-i-lean.html

Also as mentioned, the o2 sensor is used at warm idle by the DME for a target afr of ~14.7. Its only ignored when the wide open throttle switch is active. Are you sure the O2 sensor is good? If you disconnect the o2 sensor at hot idle, it defaults back to a preset fuel map on the chip (like how the euro 3.2 motronic is setup)... you should see ~14.7 or slightly richer.

Well, the information that I've read on this for the early versions of Motronic indicate that it's ignored at idle and also at wide-open-throttle (WOT). However, I think I may be confusing the early BMWs with these 3.2 Carreras. On this site in particular is where they talk about the Motronic system ignoring the O2 sensor input: Bosch Motronic Basic Motronic 1.1 1.2 1.3 Fuel Injection Technical Article

Still, that is indeed a good suggestion, and I will check the running voltages and map them against my wide-band LM-1 sensor today to see if they match. I will also try unplugging it.

What makes me think that the O2 sensor is okay is the fact that the mixture appears to be fine when the the throttle is raised just off of the idle position. It still has the miss at this point too (or it did the other day). So, if the O2 sensor was bad, *and* it was used at idle, then it would seem to have problems both at idle *and* about 2000 rpm. Instead, the mixture drops right down to 14.7 at 2000 rpm, which led me to (for now) assume that it was functioning property. Still, it's relatively easy to test, and I'll be including it in my test plan for today.

thx,

Wayne

These are good questions, I will print them out and check. #2 - the idle switch does work, but I don't remember if I tested it to itself or to ground. For that matter, I'm going to re-check all of the electrical grounds for the engine today...

-Wayne

As others have said, the O2 sensor is used at idle, it just does not output a usable signal for the first minute when it is stone cold, until either the internal heater of the sensor or the exhaust heats it up, then outputs a signal the Motronic uses. So that said, a faulty O2 sensor can cause strange idle issues, so you can disconnect it and if the issues go away, look to replace the sensor. So how does the car idle with the O2 disconnected? The car does not really need the O2 sensor to run if there is no cat in the exhaust.

A very lean mixture reading on a wideband can be cause by an excessively lean mixture, and excessively rich mixture, or a misfire of some sort. If you can set the base idle mixture, set it with the O2 disconnected to between 14.2 to 14.4 at normal running temp. If you cannot get a clean reading to set it, then as a baseline, turn the idle mixture screw all the way in clockwise, and turn it back out about three full turns. The screw falls out after about 11 turns. You want to take a improperly set idle mixture out of the equation.

Next do you have another 3.2 Motronic unit you can swap with and see if there's any difference? Faulty or failing units can also cause all kinds of intermittent and poor running issues.

Also how does the distributor rotor phasing look compared to TDC on the crank pulley? Turn the crank pulley and as the TDC mark approaches the mark on the case, look at where the rotor points to the TDC mark on the distributor housing. If the rotor is more than 3-4°s off, the rotor will not properly contact the distributor cap, causing various misfire issues that appear to be a lean condition. Improper rotor positioning can occur on engine rebuilds where it is inserted one tooth off, or in some rare instances of very early 84s built in 83, the worm gear on the crank was positioned 180 degrees off which would not allow the rotor to clock correctly in one tooth position or another. The only solution to this was a recall by Porsche decades ago, or take the distributor and remachine the shaft to reposition the worm gear on it to allow proper repositioning.

I have not done the ignition cables yet. I did replace the plugs, the cap and the rotor because I have been burned so many times in the past on these. I had a jet ski that I spent hours trying to track down a problem - it would cut in and out almost randomly from 4 to 3 cylinders. I thought it was the fuel supply - had the pump cleaned and rebuilt. No dice - turned out a new set of plugs did the trick. That's happened at least 3 times in recent history to make me try them regardless (besides plugs are cheap). I didn't think that would work in this case, and it didn't.

The reason why I don't think the plug wires are the culprit (or anything part of the ignition) is that the car runs fine for the first two min or so, and then starts to miss. It also goes lean at that point. Ignition problems would show up at the first moment, and they wouldn't necessarily cause a lean mixture to occur immediately following a normal warm-up period. Still, the wires could be suspect, but my gut tells me (at this moment) that the problem lies elsewhere?

-Wayne

Thanks for your thoughts and photo of the AFM. I did test it on the scope and I didn't see any issues there. I forgot to mention that I bought this engine from Robert at Dutch Treat (here in Torrance, CA) in 2003. It came off a smashed car that had roughly 65,000 miles on it. Engine and trans were $6,800, and what a great find that was. I was able to listen to, hear, and inspect the engine while it was still in the car. Finds like that are difficult to locate - usually the engine is out of the car on a pallet, and you have to take someone's word for it that it's any good.

Interesting story, this engine was a replacement for the 2.7 that "blew up" in the car previously around 2001. That engine then became the "rebuild engine" for the 911 Engine Rebuild book I wrote. That 2.7 engine was disassembled and sold off in pieces.

The 3.2 engine sat until 2009, until I had someone go thru it and put it in the car for me. When it came back in 2009, it never quite ran well, and it sat until about 2016 when it was stolen. Then I recovered it (tracked the guy down - he's back in jail) back in 2018 and we had it painted (that took a year). Now it's back and almost done, and I'm now tackling the engine running issues that I knew I had in 2009 (they didn't magically fix themselves unfortunately). I'm not sure what happened to the car while it was out of my control (it was left outside for about four months at the auction company where I had to buy it back). So, this is almost the last thing that needs to be done with it - I'm very close.

-Wayne

Steve, you're one of the uber experts on this, so I appreciate you correcting me on that. Indeed, I think on the earlier version of the BMW injection, the O2 sensor input is ignored, but I can't find any other additional info on that (doesn't really matter anyways). I've got my test plan for today which will spotlight the O2 sensor front and center...

Thanks!

-Wayne

Test plan for today:

COLD:

- Re-check the grounds to the engine and the pins on the DME using the breakout box.

- Re-check the idle switch to ground at the DME (I'll also recheck the WOT switch). Install new throttle-return springs (just got them in, mine were a bit slow to return the throttle switch to closed).

- Check fuel pressure (I hope I have the correct adapters here at home)

- Check the vacuum controled fuel pressure regulators by pulling a vacuum on them and then looking at the fuel pressure gauge (although they shouldn't theoretically change 1-2 min after starting the car).

- Mount the LM-1 O2 sensor into the tailpipe too - I've just been sticking it in there for the time being, but I think this problem is not going to be sovled any time soon.

- Check the proper operation of the TDC sensor using the starter (the procedure detailed in the factory manuals)

- Check the proper operation of the rpm / speed sensor using the starter (the procedure detailed in the factory manuals)

- Jumper pins B&C on the test port and test / check mixture and misfire.


START / WARM-UP:

- Check the cylinder head temperature switch (use the breakout box to map this on the oscilloscope). I'm going to hook up my laptop to the scope to record the data in real time to review it later on (this scope is pretty awesome for $200 - Chinese made, but it has a *ton* of features).

- Check the proper operation of the O2 sensor (both the heater, and map the voltage valves against the readings on the LM-1 meter). I'll probably need to take a video of this, as I don't have the LM-1 setup to dump data to the computer right now (that would take a bunch of time to do).

- Unplug the O2 sensor to see if forcing the car into open-loop mode will improve things


POST RUNNING:

- Check temp control switch / Thermo valve (this is plugged off - may be leaking when warm?). What I'll do is take the car and perform my measurements until it's hot, and then I'll bring it back into the garage and put the smoke machine on it and see if we have any leaks at temperature.

- Analyze results and post back on the forums!

-Wayne

Correct based on this;

This problem typically occurs at higher RPM, i.e. the result of the phasing of the electronic advance & the centrifugal advance.