Motor Runs Rich, Nothing I Do Leans it Out - SOLVED!

[jhh925]

Help, I'm out of ideas and frustrated as hell. Problem is that my motor runs rich and won't pass CA smog. By all accounts, there's NO reason why is shouldn't pass smog.

Here's the motor:
- Basically a bone-stock M30 B35, fully rebuilt about 3000 miles ago, with only very limited mods
- Only big mod is that it has 10:1 pistons instead of stock 9:1
- I'm running the "Ford" Bosch yellow 19# injectors (Bosch P/N = 0-280-155-710)
- I'm running the Bosch standard 4-wire O2 sensor for the late M30 motors (Bosch P/N 13052)
- Exhaust is currently set up so that I'm running the narrow band O2 sensor to the ECU, and I've added a second bung with a wideband O2 sensor running to an AEM AFR meter.

The car won't pass smog. Mark D'Sylva has been helping with a chip, and we've tried multiple times to adjust the tune file on the chip to lean the mixture out. Nothing has worked. I've also cracked open the AFM and adjusted the tension in the flapper spring (clockwise rotation of the coil spring baseplate) in order to try to lean the mixture out. Make no difference.

The only thing that changes as Mark adjusts the tune file and as I adjust the AFM is that the mixture leans out at idle. But as soon as you touch the throttle, the AFR dips back down to 'rich' again.

Here's what I've done:
- Coolant temp sensor is new with the rebuild (Bosch 2-pin blue sensor)
- Correct BMW OEM thermostat valve, 80 degree C, brand new.
- O2 sensor is new with the rebuild (Bosch as noted above)
- Double checked the wiring to the coolant temp sensor - continuity and correct connections to the ECU
- Double checked the throttle position sensor switch - properly adjusted, clicks off with a touch of the throttle, and reads WOT at about 70%-80% throttle, continuity and correct connections checked at the ECU
- Double checked continuity across all pins at the connector between the engine harness and the 02 sensor - everything checks out.
- Double checked AFM - temp sensor reads about 2200 ohms at room temp, then dips to 1000 ohms if you put a hair dryer on it
- Prior tests indicate the flapper voltage varies as expected if I apply 9v and manually swing the flapper
- When I did the engine harness mods to adapt it to this E28, I checked all the wires starting at the ECU, pin by pin. At the time, the connections to the injectors checked out as they should. That is, I'm pretty sure they're not "double firing," and that the ECU is firing them as it's supposed to (1-3-5 then 2-4-6). I don't have an oscilloscope to confirm this.

OK, so I'm starting to suspect a bad O2 sensor. I put a bone-stock stock chip back in, hook up the wideband, and hook up a DVM to pins 1 and 2 of the O2 sensor.

Result: Idles at about 14.7 or so, then dips waaaay down when I hit the gas. O2 sensor dithers from about .1v to .85v with a period of maybe a second or a bit less.

Here's a youtube video of the AEM readout next to the DVM so you can see the readings. Again, this video was shot with stock chip and unadjusted AFM - plain stock, but I got essentially exactly the same results when I put in the leaned out AFM. In the video, you can hear the engine; I start it at idle, then rev it up to ~2500rpm, then drop it back down. The motor had been running for more than 20 minutes, and an infrared temp sensor was telling me the top radiator hose was at about 180 degrees F, so the motor was warm.

Other than this mixture issue, to my ears, eyes and butt-dyno, this motor runs beautifully. Idles rock steady, warms up nicely, pulls strong all the way to redline, pulls like a freight train on the freeway.

WTF is going on? What do I do next? What have I not checked???

[Shawn D.]

What have I not checked???

The word "fuel" doesn't show up in your post, so that suggests you haven't checked the fuel pressure or replaced the Fuel Pressure Regulator (FPR). All of the other stuff will be for naught if the fuel pressure is excessive.

[Rootbeer]

Hi-Thanks for your detailed post. I've got an 88 535 unable to pass smog second time in. It's passed easily the past 4 times by previous owner.
It's a stock M30, close to 300,000 miles, never rebuilt. It starts easily after one or two cranks, even after sitting 6 weeks. Runs strong, and no issues.

To pass second time, I was told to put 40 PSI in rear tires for rolling dyno, add octane boost and make sure car is hot.
I live in Palm Springs and it's 115 now, and smog place a good 30 minute drive so car is plenty hot. My cat is OEM BMW and a good cat, I'm told.
It keeps failing due to running too rich.

Our plan going forward is to adjust the air mass sensor flap so that it doesn't open too much at full power.
My numbers seem to be fine at idle but too rich once throttle added-like your situation it appears.

Per the response to your post, I guess I can check the fuel pressure regulator but hate to buy one if not needed.

I'll keep an eye on your post and situation, and let you know if mine passes in the next couple weeks.
Thanks!

[Mike W.]

For both posters, please post the full numbers or a scan of the smog test. I emphatically don't have all the answers, but I might have a couple, but it all works together, 4 or 5 gasses, saying it's rich doesn't say much.

[jhh925]

Sorry, I'd meant to do that in the first post. Here is my most recent test. I've tried a bunch of times, and the results have been pretty consistent.

[jhh925]

What have I not checked???

The word "fuel" doesn't show up in your post, so that suggests you haven't checked the fuel pressure or replaced the Fuel Pressure Regulator (FPR). All of the other stuff will be for naught if the fuel pressure is excessive.

The idea being that if I put the wrong FPR in there (a distinct possibility) or if the FPR has failed, then one possibility is that the fuel rail is over-pressured and too much fuel is getting in everytime the injectors fire?

I'll test this. Can anyone confirm what the proper pressure readings should be at idle and then at part throttle?

[jhh925]

I'll test this. Can anyone confirm what the proper pressure readings should be at idle and then at part throttle?

OK, never mind. I have the Bentley's for the E34. The testing procedures and specifications are listed there. Fuel pressure spec for M30 is 3.0 bar +/- 0.06 bar, or 43.5psi +/1 .9 psi.

They list three tests after installing the pressure gauge: (i) basic system pressure test: engine not running, jump fuel pump relay, pressurize the system; (ii) residual fuel pressure test: engine not running, turn fuel pump off, watch gauge pressure drop; and (iii) test response to load: run engine at idle, remove vacuum hose, pressure should increase.

None of the tests are really under load or while driving. I'll report back after I have data.

[jhh925]

I went and got a cheap set up at HF. Here's more data on the fuel pressure.

• Engine not running, jump fuel relay = fuel pressure = 45 psi
• Engine running, vacuum hose to FPR detached = 45 psi
• Engine running @ idle, vacuum hose attached = 38 psi (out of spec, too low)
• Engine running @ 1000-1500 rpm, vacuum hose attached = 35 psi

Here's a link to a video of the pressure gauge. Initially, the vacuum hose is detached (0-5 seconds); pressure shows about 45psi. Then I attach the vacuum hose and the pressure drops to about 38psi, and you can see the gauge vary as I rev the engine.

So, I'm puzzled ... it looks like this FPR is out of spec, but the pressure is too low, not too high. How could low fuel rail pressure cause an over-fueling condition at the injectors?

[Shawn D.]

How could low fuel rail pressure cause an over-fueling condition at the injectors?

It wouldn't.

Are you sure you have the right injectors? They might have a higher fuel flow rate.

[jhh925]

How could low fuel rail pressure cause an over-fueling condition at the injectors?

It wouldn't.

Are you sure you have the right injectors? They might have a higher fuel flow rate.

These are supposed to be Bosch "19#" injectors. I listed the P/N above (0-280-155-710). Based on what I'm finding online, they should flow at least 18 pounds and change.

[Edit: According to this site, these Bosch injectors should flow 18.25 pounds at 43.5 psi.]

[Mike W.]

OK, I'll buy that it's running rich. It really wants to run off the O2 sensor which is why adjusting the AFM isn't having the desired effect. Other than maybe dithering a little slow at times, and the high readings coming down to idle around the 42 sec mark, the O2 readings look pretty good. 0% O2 doesn't mean much, I've had cars pass with it before, but it would be consistent with being rich, just not necessarily. Nox is low, very low for an E28, but that too is consistent with being rich. The report notes the cat has been modified or tampered with, what's that about? I mean that's an automatic fail even if it's an air purifier.

I do have a couple of long shot ideas though. Is there an exhaust leak anywhere? Even though you think it's leaking out, it's a pulsating in/out thing where it could pull in oxygen causing the O2 sensor to think lean and richen it up. The injectors. If one of them is plugged, partially, that cylinder is going to be lean, again the O2 sensor would see that excess oxygen and richen it up.

Going back to the cat, what kind is it, how old is it, what's the history? Aftermarket cats are noted for a short life. The requirement they be OBDII rated, in spite of the fact that the cars were long before OBDII is supposed to minimize that, but I think they just went to about a 4 year life instead of a 1-2 year. Factory cats do well, mine still passed at 330K when I sold the car, but they can be damaged and no longer work or work well. You might think about trying a can or two of that Guaranteed to Pass stuff you see at some auto parts stores. I've known a couple of friends who had smog licenses and they both said it can often help. They even kind of shake their heads at it, saying they don't know why, but it can often make the difference in passing and failing, and your numbers aren't terrible, just not passing. I mean I've seen cars with over 2,000 HC and you're not even in the gross polluter range, which isn't all that high anyway.

[jhh925]

Mike:

Thanks for the thoughts.

• It's an aftermarket DEC cat. Unmodified. The guy at the last place said: "it's the wrong part number." I bought this CA-compliant DEC downpipe+cat assembly brand new with this rebuild, and it's been inspected by multiple shops in the same circumstances and passed every time. I'm nearly certain that the last guy was reading the CA government publications incorrectly - he kept getting confused about what BMWs this cat applied to. Anyhow, the cat is nearly new. That said, both O2 sensors are upstream of the cat, so I'd assume that the cat performance has no effect on the readings at the wideband. Correct?
• I don't know for sure if there's an upstream exhaust leak, though I'm not hearing anything odd. And the shop that installed the muffler and did the custom downpipes used one of those stethoscope-things to listen for leaks. As I write this out, an exhaust leak isn't impossible. Is there any way for me to check this out at home?
• Injectors were new with the rebuild. Clogging isn't impossible. Suggestions on how to test? I thought I saw another post from years ago where the owner of a 528e pulled all the rail with the injectors still installed and then ran the fuel pump to confirm all the injectors were flowing. Not very scientific ...
• Guaranteed to pass stuff? Do you mean stuff like this?

[Mike W.]

Mike:

Thanks for the thoughts.

• It's an aftermarket DEC cat. Unmodified. The guy at the last place said: "it's the wrong part number." I bought this CA-compliant DEC downpipe+cat assembly brand new with this rebuild, and it's been inspected by multiple shops in the same circumstances and passed every time. I'm nearly certain that the last guy was reading the CA government publications incorrectly - he kept getting confused about what BMWs this cat applied to. Anyhow, the cat is nearly new. That said, both O2 sensors are upstream of the cat, so I'd assume that the cat performance has no effect on the readings at the wideband. Correct?
• I don't know for sure if there's an upstream exhaust leak, though I'm not hearing anything odd. And the shop that installed the muffler and did the custom downpipes used one of those stethoscope-things to listen for leaks. As I write this out, an exhaust leak isn't impossible. Is there any way for me to check this out at home?

It sounds like it's probably tight then. Cat performance has no effect on upstream sensors, but they or the conventional one of course affect and control things. I've never run into that problem, I've just read about it. I have been known to check for exhaust leaks, quickly, on a cold engine by starting it and running my hand around the suspect area. You don't have much time before it warms up, but a little. I did that on a Bav when I was getting exhaust fumes inside and couldn't figure out where. I wasn't hearing anything, but I could feel it.

[*]Injectors were new with the rebuild. Clogging isn't impossible. Suggestions on how to test? I thought I saw another post from years ago where the owner of a 528e pulled all the rail with the injectors still installed and then ran the fuel pump to confirm all the injectors were flowing. Not very scientific ...

I've done the same to test for a leaky injector, pull the whole fuel rail. On an L-jet engine I was able to cycle the distributor to pulse the injectors, but I don't know a similar trick on motronic. And you're right, trying to measure in a calibrated manner would be difficult to say the least.

[*]Guaranteed to pass stuff? Do you mean stuff like this? [/list]

Yep, that's what they were talking about. Never used it myself, but they said it often helped. But they didn't like it because they didn't know why! I make no claims it's a magic bullet and I've never used it myself, but two different people in the trade have said there is a good chance of it working.

[LeiseyJr]

You check fuel pressure, and vaccum line to the regulator?

[demetk]

Test O2 sensor reading with it both disconnected and connected, both at idle and part throttle. If the O2 sensor is pegging close to 1 volts then obviously you're running rich and the motronic is unable to compensate.

Are you running the 179 ECU?
What AFM number are you running? Got a pic of the board?
Do you have a pic of the injectors?

[demetk]

You check fuel pressure, and vaccum line to the regulator?

His pressures look good to me.

[jhh925]

Test O2 sensor reading with it both disconnected and connected, both at idle and part throttle. If the O2 sensor is pegging close to 1 volts then obviously you're running rich and the motronic is unable to compensate.

Are you running the 179 ECU?
What AFM number are you running? Got a pic of the board?
Do you have a pic of the injectors?

Thanks for taking a look at my post. Here are my responses:

• Test O2 sensor reading with it both disconnected and connected: Just to confirm, you're talking about taking the voltage reading from the narrow-band O2 sensor, yes? Did you see the link to the youtube video where I did show the narrowband O2 sensor output at both idle & P/T? Here's the link again. It fluctuates between about .15v to .85v on a pretty regular cycle. I have not tried it when disconnected. I will do that and post another video.
• Are you running the 179 ECU? Yes, the 179 ECU. This is a full Motronic 1.3 set-up with the wiring harness taken directly from a '91 E34. I have a small stack of ECUs and I'm getting pretty similar results with all of the boxes.
• What AFM number are you running? Got a pic of the board? I will get you a picture of both the part numbers off the bottom + a picture of the board.
• Do you have a pic of the injectors?[/quote] Yes, see below ... it's a bit challenging to read all the numbers off the body, so here's a link to the full resolution.

Injector:


Also, ahab had some offline comments, and he had me hook up noid lights to the injectors. Here's a link to a video in slo-mo showing noid lights hooked up to all six of the injectors. My son turned the key as I filmed, and he pressed halfway down on the throttle, so this was the injector behavior with the "idle signal" from the TPS turned off (that is, the idle position switch was open). Ahab and I are both interpreting the audio (one flash of the lights for each "boom" on the audio) to mean that the injectors are all firing once each rev. I'm going to re-run that test with an inductive timing light added to the video to confirm that.

[jhh925]

You check fuel pressure, and vaccum line to the regulator?

Yes:

• Engine idling, vacuum line disconnected = 45 psi
• Engine idling, vacuum line connected = 38psi
• Engine revving to 1000-1500 rpm, vacuum line connected = 35 psi

[demetk]

Those O2 sensor ditherings in the vid look good to me. The wideband looks to be slightly rich. The NO measurements are on the low side indicating a rich condition. NO needs to go up before the CO goes down. There's a fine line between the two. Maybe 1-2 clicks lean on the AFM and then check out wideband again?

OK I see that you did that. I would baseline the AFM. Loosen the spring all the way until the flapper is loose then adjust the tension 5 clicks and go from there.

[demetk]

Also, ahab had some offline comments, and he had me hook up noid lights to the injectors. Here's a link to a video in slo-mo showing noid lights hooked up to all six of the injectors. My son turned the key as I filmed, and he pressed halfway down on the throttle, so this was the injector behavior with the "idle signal" from the TPS turned off (that is, the idle position switch was open). Ahab and I are both interpreting the audio (one flash of the lights for each "boom" on the audio) to mean that the injectors are all firing once each rev. I'm going to re-run that test with an inductive timing light added to the video to confirm that.

That vid is with the engine cranking or actually running? Is the #6 plug pickup connected?

[tschultz]

To help illustrate the theory:







http://bigcoupe.com/phpBB2/viewtopic.php?t=19661

[jhh925]

Also, ahab had some offline comments, and he had me hook up noid lights to the injectors. Here's a link to a video in slo-mo showing noid lights hooked up to all six of the injectors. My son turned the key as I filmed, and he pressed halfway down on the throttle, so this was the injector behavior with the "idle signal" from the TPS turned off (that is, the idle position switch was open). Ahab and I are both interpreting the audio (one flash of the lights for each "boom" on the audio) to mean that the injectors are all firing once each rev. I'm going to re-run that test with an inductive timing light added to the video to confirm that.

That vid is with the engine cranking or actually running? Is the #6 plug pickup connected?

Video is taken with the engine only cranking, not running.

Yes, #6 cylinder pickup connected. Regarding the pickup, the resistance measured across pins 1 and 2 is under 100 ohms.

[demetk]

Video is taken with the engine only cranking, not running.

Yes, #6 cylinder pickup connected. Regarding the pickup, the resistance measured across pins 1 and 2 is under 100 ohms.

So the motronic batch fires at crank and it should do semi batch of 3 cylinders at run? When the pickup is not connected I know it reverts to batch fire mode. I run the engine without the pickup and it has passed emissions.



Interesting that CT has more stringent emission limits on CO and HC than CA.

[jhh925]

To help illustrate the theory:

Thank you! I think I'd seen these graphs a couple of times before, and this time, reading the post you link to is also helpful, and gives me more to work on:

Causes of rich running can include:
- Bad plugs [I'll pull plugs and post pics]
- Mis set AFM
- Excessive fuel pressure [not an issue here]
- Bad coolant sensor (the ECU senses the temp and richens the mix when the car is cold), or bad wiring between the coolant sensor and the ECU [Maaaaybe ... the wiring is definitely good (checked from sensor to ECU multiple times, part is new, and part is correct based on research here & online - but maybe it's bad.]
- Bad air temp sensor in the AFM (the ECU richens the mix when the air is cold, because cold air is denser, so it takes more gas to maintain the 14.72:1 air fuel ratio

One interesting thing ... my Nox is low at the tailpipe, indicating the Cat is working. But HC and CO are high as shown in your "pre-cat" charts ... so maybe it's just that the fucking cat is bad.

Edit: One thing I wanted to highlight is that when Mark & I played with the AFM (increasing tension on the flapper door) and the chip mapping, we had no problem leaning out the motor at idle, but then as soon as the throttle is pressed (and presumably the TPS switch idle switch opens up), the motor goes rich again. I'm nearly certain that I have the correct TPS in there, and the switching behavior at idle, part-throttle and WOT all correctly match what I'm expecting based on the wiring diagram ... so I feel like I'm looking for something else that's causing the ECU to look at the TPS, getting the correct info from the TPS but then "drawing the wrong conclusion" and going rich. And the fact that we can lean the motor out at idle also indicates to me that the coolant temp sensor is working correctly?

[Mike W.]

One interesting thing ... my Nox is low at the tailpipe, indicating the Cat is working.

No, not necessarily. 3 way cats have 2 different sections, the front does Nox and the rear does CO and HC.

Something else that occurred to me after my last post is does the ECU have a fuel quality switch inside like 059s do? It used to get quite a bit of talk but I haven't seen it referenced in years now. That 3.1% leaner could by your ticket.