Indication of a bad upstream O2 (wideband) sensor?
 

Ok so I have been fighting this issue for a few weeks now and I am at a loss here as to what is going on.

I have been experencing some random surging at times when accelerating at higher rpms (5-6kish at least). At first I was thinking that it could be my ignition, so I double checked spark plugs (only 1000mi on them), made sure they were installed properly, removed my homebrew D585 ignition and replaced with brand new OEM coils and wires, cleaned maf, reset ECU trims, all to no avail.

I started taking some logs and here is what I am getting:

https://sphotos-a.xx.fbcdn.net/hphot...99530831_o.jpg

and with the commanded AFR:

http://www.normalexception.net/_shared/Untitled.png

Now some have said that this is a somewhat 'normal' looking graph for the OEM tune, but I just can't fathom that the deviation between commanded and actual is within OEM tune spec.

On top of that, it appears that my fuel trims at idle are bananas:

https://www.rx8club.com/attachment.p...1&d=1368408039

although the long term trims are moot in the above graph, watch the short term start tapering off more negative over time.

At this point I just dont know what else it could be.

1. I have checked, rechecked, triple checked, and cannot find any vacuum leaks.
2. MAF seems to be in working order, and clean.
3. Spark plugs are pretty much new
4. Coils are brand new
5. ECU has been reset

The only other thing I can think of is a bad/fouled upstream o2 sensor. But my car has only 24,000mi and I cannot imagine/haven't heard of any S2 wideband sensors fouling out.

Alright here is a new wrench to the equation. Really pushed her hard to work today, and the car gave me a P0101 code.

I think this actually helps shed light on the issue, but I dont think it is a "bad" MAF per se. Looks like from what I have been reading, the ECU is seeing more airflow than expected, which I suppose could be thanks to the AEM intake with the last bend chopped off?

I believe the MAF table compares actual, against the expected, and if it is out of the threshold it will throw a code. I will need to find and monitor the PID that shows the expected MAF reading if there is one. But nonetheless, seems like the ECU is trying to compensate for this by adding more fuel to the mix, which I assume would cause the surging that I am feeling, and would also explain the rich AFR's at the top end. The question is... would this attribute to the STFT at idle ...

Looks like I have two options though... replace OEM intake, or calibrate MAF

MAF sensors tend to fail lean, i.e. they get gunked up and they sense a lower air volume than is actually passing through them. O2 sensors tend to fail rich, sensing more residual oxygen than the exhaust actually contains. The MAF sensor readings are adjusted according to the O2 sensor readings, so even if a dirty MAF sensor reads too little incoming air, the lean exhaust will be detected by the O2 sensor and the mixture will be enriched for future combustion cycles. Since the two sensors fail in opposite directions, their discrepancies add together instead of cancelling out, and eventually the total discrepancy will exceed the ECU's tolerances and an error will be reported.

A critical failure of the MAF sensor will happen when the circuit is broken, the input voltage is short-circuited, or the readings are impossible given the known air-consumption capacity of the engine. Those are easy to diagnose. As far as I know, the only way to confirm a non-critical MAF sensor failure is to replace the O2 sensor and see if that clears the error. This is because the O2 sensor's readings are considered more authoritative than the MAF sensor's readings, because the O2 sensor is measuring the results of actual combustion instead of predicting the results of future combustion.

I guess I can't say it's impossible that your aftermarket intake might be causing the problem; the MAF sensor does only sample a small portion of the incoming air, so there's an assumption designed into the system that the total incoming air will be a specific multiple of the air sampled by the MAF sensor. If that assumption is invalidated, then the additional discrepancy in the MAF sensor readings could make a worn O2 sensor look like it has already failed. However, if the diameter of the intake is the same as stock and the positioning of the sensor in the intake plenum is also the same as stock, that assumption should still be correct and the MAF sensor readings should still be accurate -- assuming the sensor is working properly in the first place.

ok so that's starting to make sense. So are we saying that I need to continue looking at my upstream o2 being the possible issue? I just can't make sense of how that could have failed so quickly if it has.

edit: could a faulty upstream o2 also trigger P0101? I wasn't aware it could. Maybe two different issues here?

the issue is mostly your lack of understanding, don't have the time to be more specific

if i didn't lack the understanding .. i would have figured it out by now ;)

Took some time yesterday to dig for a vacuum leak or any other anomaly. No leaks that I could find. Cleaned the MAF real good, let it air dry. Took the very best caution in making sure the AEM intake was installed straight, and all clamps were secured.

Didn't experience any issues today driving into work, but here is a log of my drive right before i got off the highway and drove into work. This is after a 15minute highway drive and a few minutes of going through red lights and such. After I parked I let the car sit idle:

(rpm is on primary axis, the rest on secondary)
https://www.rx8club.com/attachment.p...1&d=1368531294

Here is the attached log of my run from right off the highway. You can see the areas where I hit the red lights when my RPM's drop and the STFT jacks up. Not sure if this set of data is any better than what i posted before.

http://www.normalexception.net/_shar...4_06-40-15.xls

Yeah, looking at the log, this section really looks like a vacuum leak to me:

MAF is too low for idle, but the ECU is adding fuel because it sees more air than the MAF is reporting.

If you aren't finding with normal methods, may be worth it to get the system smoked.

sounds like its the only choice I am left with, i really can't find anything... ugh...

edit:

compare to the STFT that i had in the OP. Maybe I actually created a vac leak in reinstalling the UIM? You figure 5 god damn bolts you can't really eff it up.

You would be surprised.

Think of the fact that most wheels are attached to the car with 5 bolts, and how often that gets done wrong. Now think of those 5 bolts holding your intake manifold on properly, and one of the o-ring seals coming out of place before you bolted it back together.

Just double check your previous work, and all will become clear.
Spray some carb fluid around the different connection points of the intake system starting at the MAF, and working towards the intake manifold, and you should be able to find it.

BC.

yea been dwelling on that all day. I made sure to go 'star pattern' on the mani as well, but who knows. Waiting to get home to start spraying away lol..

Uhhhh.... so i pull into my garage ... keep the car on .... break out my carb cleaner. Sitting there fidgiting to get that stupid straw on, i finally get a chance to start spraying .... spray spray ... spray spray ... nada..... so now im like flipping .... until....

https://www.rx8club.com/attachment.p...1&d=1368561592

I slowly notice the trims going back down ... and down ... and down .... and finally settle at around 2% .... learning?

So i let it keep going. Then i shut off the car. I figure "hey, if im gunna see the trims freak out, its gunna be when you turn on the car (first start, car runs open loop for a bit .. noticed earlier ... rpms dip a little, till it goes closed and the trims fix the rpms) ... so i start the logger ....

the pic above is me turning on the car ... looks like the STFT settled?

So am i in the clear here? Looks like whatever i did yesterday helped fix the issue (before i had weird negative 4-5 STFT) ...i just needed to wait for the trims to relearn?

Full Log Here:
http://www.normalexception.net/_shar...4_15-50-47.xls

Oscillating STFT and AFR
 

I've been trying to iron out my idle issues since I got my engine replaced 2 weeks ago. 2009 MY, 101,000 miles, 1,000 miles on new engine. The first reman the dealer put in was bad from the plant, it misfired, overheated and had low compression. The 2nd reman seems to be doing fine so far, except it threw a code for SSV stuck open while still at the dealer. Dealer claimed I have a lot of carbon buildup in the intake manifold. Code was cleared and hasn't come back though, doesn't seem to be down on power either.

When the car is warm, and idling, STFT does this slow sine wave, between about 0 and 4%, with a period of about 5 seconds. At the same time lambda from the wideband O2 sensor is oscillating about 7%, so for example between 1.0 and 1.07. Maximum lambda (max AFR = 15.7) corresponds to minimum STFT, at which point the idle is a little rough (increased vibration). Then ECU corrects by increasing STFT and the cycle repeats. With the AC off, RPM does not really change much during all this, but it used to swing +/- 250 rpm before I cleaned the MAF and ESS.

LTFT at idle is +7%, and at cruise it is 1.6%. MAF reads 3-4 g/s at idle. Before the engine replacement, there were no idling issues.

1. Could this be a sign of a fouled front O2 sensor? My reasoning is that even with a MAF that reads low, or a vacuum leak, the ECU should be able to correct for that with LTFT, without oscillations. But if the wideband O2 sensor is fouled and is taking too long to respond, then the ECU is always chasing the AFR. I suppose if the 1st reman was dumping fuel, and who knows, maybe oil and coolant out of the exhaust, it could have fouled the sensor...

2. Am I unduly obsessing over this? I know 15.7 AFR at idle is not the end of the world, but I want my replacement engine running as close to perfect as possible. Is this even normal ECU behavior? I guess when you have access to OBD-II engine data, you start finding all sorts of problems you never knew your car had! :eyetwitch

3-4 g/sec at idle is a vacuum leak. That is why it's running lean and adding fuel through STFT. It should learn over some time and move the STFT to LTFT.

Get a can of carb spray and start spraying. The idle will jump up noticably if you spray where it's leaking ..

In my experience with the OE map there is essentially no meaningful relationship between commanded AFR and actual AFR

I would disagree with that. The calculated MAF is never affected by the health or presence, or lack of presence, of the front O2. It doesn't matter what you do with the front O2, it will never change what the MAF is reading. However, the calculated AFR / lambda is DEFINITELY affected by what the MAF is reading.

If your MAF is well below what it should be, solve for that first, because any other solution could end up with you just chasing your tail. If you ever have tried to tune a MAF based car with a vacuum leak, you will know the pain. Even if you aren't touching the ECU, the ECU is trying to tune it. It won't ever really get it right.

I suspect the oscillations are electrical load changing somehow, changing the drag on the e-shaft, changing the load on the engine. I suspect this because of the comments you have made around the AC. A large RPM swing with the AC on is the compressor cycling on and off. It should be minimal, but if you have a vacuum leak and the fuel trims are wacky, a wild swing is entirely reasonable to expect.

Solve for your vacuum leak.

No, it's not normal, you have a problem :)

Alright guys, I went at it with a can of carb cleaner but couldn't find a vac leak in any of the obvious spots. The only place I couldn't get to was the back side of the intake manifold. I gave the MAF another cleaning, which didn't affect things that much. I also disconnected the VFAD and sealed off the vacuum line, so I know there's no leak around the VFAD actuator. (Why I had to do this is a different story.) :doh:

Anyways, even assuming a vac leak, the numbers don't add up. When fully warmed up, I just saw MAF rate of 3.4 g/s, with 10.2% LTFT and lambda of 1.05 (still with sine wave oscillation). So according to the ECU the true air flow rate is 3.4g/s x 1.102 x 1.05 = 3.9 g/s, still far below what it should be. Should a Renny even be able to idle at that flow rate? :confused:

Also, idle lambda oscillation gets slightly worse when engine is hot after a drive (but the fans are not yet cycling on)

EDIT: I guess I'll check is the spark plugs are torqued properly next time I'm under the car. After that, I'm out of ideas.
EDIT #2: Checked vacuum with a vacuum gauge. It is pulling 16-17 in Hg at idle at about 2,500 ft elevation. As I understand it though, this does not discount the possibility of a vacuum leak. Also have a rattle from what sounds like the intake manifold when getting on the throttle at any rpm. Thought it was the SSV (because it had thrown a P2070, which was cleared and hasn't come back), but when I tested the SSV actuator, it moves quite smoothly and within the specified vacuum range.