Have you disconnected the O2 sensor when the car starts to run poorly? At warm idle, the base maps are altered by the Lambda sensor in closed loop operation and barometric sensor. That's it I believe.

Since the car starts and idles well cold (before o2 sensor is warm and working, aka open loop) the pre-programmed enrichment from the Motronic seems to be working. It's getting a cold signal from the temp sensor and an open aux air valve in the intake to keep the idle higher and overcome cold oil/pumping losses etc. The thermo time switch and cold start injector only operates for a brief time whe key is in START position, and a leaking cold start valve will make you run rich not lean.

Once warm, you may be going lean because the o2 sensor is detecting a false 'rich' condition and decreasing the fuel injector opening time/fuel volume.

One trick to see if you are truly running lean is disconnect the vacuum line on the fuel pressure regulator and to give it atmospheric pressure instead of manifold vacuum. This will essentially raise your fuel pressure and add more fuel- see if it improves your running or AFR slightly. This can be done with o2 sensor ON as a way to manually 'force' more fuel into the cylinders.

Most likely fuel, e.g. pressure, AFM, O2, or air leak, and not ignition related. As mentioned, find another DME ECM and rule that out.

Use a short bungee cord, one end hooked to the LM-1 clamp, and the other end to some part of the car like a fender strut, to hold it in. Or some tie wire. The clamp screw is not reliable.

I was wondering if anyone reading this thread knows what the ‘correct mixture’ voltage output would be for the Bosch O2 sensor ?
The info out there is confusing, sometimes I read 450 mV, other times 350 mV for 14.7. I know it is narrowband and has a steep curve around the ideal mixture point but I 100 mV on a max of about 880 mV is a big deal.


Sent from my iPad using Tapatalk

Sunday Update:

- Rechecked the grounds using both the engine and the breakout box -> OKAY.

- Rechecked the idle switch -> OKAY

- Checked the throttle switch - it wasn't working. Since this is a wide-open throttle switch, it shouldn't affect the engine. I had a big issue with this. I checked the sensor using the breakout box and it didn't work. Then I disconnected it and checked the harness. Then I checked the sensor itself (hard to reach). Everything checked out, but when I buttoned it back up together it didn't work. So, I removed the switch (messed it up in the process - it's near impossible to get to on a 914 - probably a 911 too). I ordered a new one. Anyways, the problem was a defective harness with two pins - they were not seated correctly! It actually looks like it may have been defective from the factory. The pins are recessed in this connector. Doesn't fix my problem, but it was another problem that I found (it's good to check *everything*). The idle cutoff switch connector is shown next to it for comparison. See the photos:

http://forums.pelicanparts.com/uploa...1588574046.jpg
http://forums.pelicanparts.com/uploa...1588574046.jpg

If you've ever wondered what this looked like inside, there really is absolutely nothing complex inside. It's criminal that the Porsche list on this thing is over $400!

http://forums.pelicanparts.com/uploa...1588574046.jpg

- Checked the fuel pressure. Showed 2.5 bar with the engine off (fuel pump jumpered), and also kept the pressure after the engine was turned off. Kept pressure during running too. -> OKAY

I didn't have the proper fittings at home, so I used some hose clamps. These worked better than I expected (no leaks!)
http://forums.pelicanparts.com/uploa...1588574046.jpg

Pressure good:
http://forums.pelicanparts.com/uploa...1588575570.jpg

And after the car was off for quite a while, still holding:

http://forums.pelicanparts.com/uploa...1588574046.jpg

- Checked the Top Dead Sensor (TDC) sensor -> OKAY
Photo:
http://forums.pelicanparts.com/uploa...1588574046.jpg

- Checked the RPM / speed sensor -> OKAY
Photo:
http://forums.pelicanparts.com/uploa...1588574046.jpg


- Checked the cylinder head temperature sensor, I think the readings looked okay: -> OKAY
- 747 ohm @ 75 degrees F
- 315 ohm after two min of running
- 155 ohm after ten min of running

- Checked the ambient air temp sensor: -> OKAY
- 755 ohm @ 75 degrees F
- 755 ohm after two min of running
- 700 ohm after ten min of running

- Checked the O2 sensor heater element power (got 12V) -> OKAY

- Checked the O2 sensor by measuring it against the wideband LM-1 tool. The O2 sensor appeared to be jumping all over the place. I've actually never compared these two in real time before, so I'm not sure what the O2 sensor output is supposed to look like. But it certainly didn't seem to be matching the LM-1. So, when I unplugged it, the car improved, but still did not run super well. The lean mixture came down and it ran closer to 14.7, but not spot on. I still need to adjust the mixture / air bypass screw on the air meter.

What is troubling is that when I unplugged the O2 sensor and then ran the car at 2000 rpm, the mixture was off rich by about 10%). When I plugged the O2 sensor back in and ran the car at 2000 rpm again, the mixture was spot on. So, I have the following confusing situations:

- O2 sensor plugged in, runs terrible at idle
- O2 sensor un-plugged, runs better, but not great at idle
- O2 sensor plugged in, mixture is spot-on at 2000 rpm
- O2 sensor un-plugged, mixture is off by about 10% or so at 2000 rpm

So, based upon the results at 2000 rpm, it would seem like the O2 sensor is indeed properly measuring the exhaust. Un-plugging it creates what appears to be an open-loop situation. Plugging it back in makes the mixture spot-on. Yet, there is something terribly wrong with the car at idle with the O2 sensor plugged in. Why would the O2 sensor appear to be failing at idle, but not at 2000 rpm? The answer must be obvious - it's probably not failing, but perhaps would be something else.

- After running and the car was warmed, I smoke-checked the car one more time. No leaks were found anywhere - the intake is very tight (the smoke machine rubber funnel kept popping out because of the 10+ psi in pressure that was building up). I cracked open the air flow meter and also cracked open the throttle just to make sure there was enough flow everywhere.

Here's another odd thing I recently noticed. I pulled the paperwork for the engine from 2003, and it says it was a 1984 engine (I thought it was out of a 1986 car). The engine number is 64E01127 - according to the appendix in my Engine Rebuild Book, that does confirm that it's a 1984 engine. However, the oxygen sensor has the round connector and the integrated three pronged plug, which supposedly didn't show up until 1986. So, something is a bit weird with this - a 1984 car should not have that plug. There's something amiss here, maybe the injection got swapped sometime, or something like that. Or it's a replacement harness. Something's fishy.

Video of the car running rough, while monitoring the O2 sensor:

<iframe width="560" height="315" src="https://www.youtube.com/embed/YK6zr94i5aM" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>

Now the same thing, but then I disconnected the O2 sensor (sorry, the iphone got confused)
<iframe width="560" height="315" src="https://www.youtube.com/embed/9jGPjP127EQ" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>

Short video with O2 connected:
<iframe width="560" height="315" src="https://www.youtube.com/embed/beNoCbfqfTM" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>

Now disconnected (mixture is good, but it's still missing):
<iframe width="560" height="315" src="https://www.youtube.com/embed/LWMJiHiYwwk" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>


http://forums.pelicanparts.com/uploa...1588574649.jpg

It passed the warm smoke test:

http://forums.pelicanparts.com/uploa...1588574649.jpg
http://forums.pelicanparts.com/uploa...1588574649.jpg

This is the ECU that is in it:

http://forums.pelicanparts.com/uploa...1588574649.jpg

Weirdness - this seems to be associated with the following: 1988 - '89 0 261 200 082 - uP PN B57312 4006, Date Code 8808, 28 pin EPROM

That and the newer wire harness indicate that something is odd and this motor is not as unmolested as I had originally thought. Still, I don't have any reason to doubt that there's anything seriously wrong with it - it does indeed run well for two minutes. I have not run a compression check on it (didn't think I needed to). The engine serial number says that it's a 1984 engine, so there's something amiss. It's possible that the person who installed this into the car used a different brain and harness for some reason - I can't recall and I can't find my paperwork right now. Still, it theoretically shouldn't make any difference?

-Wayne

Wayne...you have to make the videos public in options no they are private we cannot see it

Ivan

I've got a new wide-open throttle switch that I will get tomorrow along with a new O2 sensor. But in the meantime, I'm open to further suggestions. I'm not very hopeful that a magical new sensor will fix the problem?

-Wayne

Thanks, had to set the "permissions for kids" setting, etc.

-Wayne

now there are no videos to see....i`m sure you did this already just a thought... reference sensor .08mm distance from the flywheel is set ,right?...

Ivan

Hit refresh? I just checked it from an anonymous browser and it worked?

Thx

i did and no videos anywhere to see???

Ivan

Videos work for me


Sent from my iPhone using Tapatalk

all ok now thanx.....

the last video ...how come suddenly it runs ok with the oxy disconnected...do you have to have it in the system ..or can you just bypass it at the ecu and not use it at all?
Ivan

i assume the ref sensor gap is ok ,right?

Back to basics old school here. Grab a can of carb spray and spray it into the intake with the engine running. Verify that the O2 sensor reading on your breakout box goes up (Using a DVOM set to DC volts). All it takes is a little bit, we just want to verify that 1) the O2 sensor responds by reading an output closer to 1.0V, and 2) see if the engine smooths out. If both things occur, look for a vacuum leak somewhere, or possibly the AFM is off. In any case, if it smooths out, we KNOW that it's lean and fuel related.

O2 sensor should read about 450mV unplugged (Key on engine off), close to 1.0V full rich, close to 200mV really lean (engine running). The number of times it crosses the 450mV threshold while in closed loop is what we used to call "cross counts." The more the better. A "lazy" O2 sensor will hardly ever cross the threshold and new (or good) O2 sensor will cross the threshold constantly. The 450mV is actually an output from the ECU, and the computer just cares if the voltage goes above or below that point in order to adjust the A/F mixture.

It is typical on engine transplants that components are cobbled together from different years, so yeah you have a early harness, and a later brain, but it doesn't really matter. The brain you have can be either an 87 with a 24 pin chip, or 87-89 with 28 pins, for which the base idle speed is to set and should idle around 880 rpm.

Your O2 sensor readings being fixed around 0.49v are symptoms of a worn or dead O2 sensor, so replace it. Normally it should switch rapidly several times per second from 0.2 to 0.7v around stoichiometric, or stick mostly under 0.2v is the car is running lean, or above 0.7v is idle is running rich.

You didn't mention what the exhaust was, but I'm guessing some kind of aftermarket header system configured for a 914? If so it's important that there's no exhaust leaks especially around the flanges, which can suck in air and fool the O2 sensor into sending out incorrect readings. Aftermarket headers often have poorly planed or warped flanges, though the stock manifolds can too if the gaskets were reused.

Since it runs smoothly when cold and the mixture is rich, but runs poorly when it warms up,
you may have a poor injector spray from some injectors. You need to remove each injector connector
separately and determine it's effect. You can also pull each spark plug connector to find a missing/weak cylinder.

The head temp sensor readings with a cold engine should be between 2.2 to 3.3k ohms. Low readings here will cause a leaner mixture at startup. The hot readings after 10 minutes running are normal.

If by ambient air temp sensor you mean the air temp sensor in the air flow meter, those are also off. Low readings here will cause a richer mixture. The NTC sensor in should read as follows:

32F between 5.0-6.2k ohm
68F between 2.2-2.8k ohm
86F between 1.5-1.9k ohm

If you don't have another air flow meter at the moment to substitute, you can try substitute a 1.7k ohm resistor here temporarily here to assist in debugging if the sensor really is found to be faulty.

Use an IR temp tool on each exhaust tube. Did you get a CO reading out the tailpipe? HC? The mixture has to burn before you get CO. High HC shows unburned fuel. But you Knew that.