Are widebands known to go lazy over time?

Question

I know narrowbands can go lazy with time, but what about widebands?

This is a sequel to the cold-start misfire mystery.

It turns out that we were looking at the wrong PID's for the front O2 sensor voltages, assuming they were narrowbands. Probing the wideband sensor PID's for Bank 1, Sensor 1 and Bank 2, Sensor 1 lambdas yielded a proper signal.

This 2009 Mercedes GLK 280 had an intake leak that was fixed by a mechanic. The LTFT's are now very close to zero (-0.8 %). The STFT's, however, are all over the jungle, cycling between -15 % and +20 % every 15-30 seconds.

What I have noticed is that the STFT's are following the wideband lambdas at idle, which cycle very slowly between 0.8 and 1.2 (11.76:1 - 17.64:1 AFR). Post-cat O2 sensor voltages are around 0.75 V, so rich as expected.

The reason why I'm seeing this car again is because of a CEL that manifests itself intermittently when coming to a stop at a traffic light:

P0171, P0174 - System too lean

My hypothesis

LTFT's are showing the system is airtight
STFT's are cycling wildly because of the front wideband O2 signals cycling
Widebands need to be replaced

Screenshots

Here are the DTC's and freeze frames

STFT's and Wideband lambdas at idle

This is before clearing the DTC's, but the behavior here didn't seem to change much after clearing them

Here is the throttle blip test (if I remember correctly)

Ignore the green line because it isn't Vehicle Speed and smoothing was turned on.

I like this question. I've heard about this over the years but have never really seen any facts to reflect either way. I can see the factual response falling into either bucket and create a made up explanation for it in mind. — DucatiKiller, Apr 14 '16 at 22:42
What do the sensors show at idle in park? What happens when you press the brake pedal in park? Is it the same as when the car is in gear? — Ben, Apr 14 '16 at 23:58
In my experience when Denso AFR sensors go bad it ends up being a range issue and the sensor under or over reports voltage. If it's not completely shorted. I don't believe I've seen voltage cycle (for no reason) on one. — Ben, Apr 15 '16 at 00:04
Not enough data here to come to a conclusion. If you could get a graph the following PIDs: AFR mA, AFR volts, EQ_RAT (actual and commanded), rear sensors volts (BxS2), STFT and LTFT. Actual tailpipe lambda from gas analyzer readings. At both cruise rpm (2500 rpm) and idle. The cruise speed readings are more useful for testing sensors and cats. This gives a more complete picture of the fuel control system and its possible errors. — Fred Wilson, Apr 15 '16 at 05:42
WB sensors are damaged by the same processes that damage narrow bands. They are made out of the same materials. Failures do not show up the same in the data because the controller in the PCM it better at covering the errors. This newer fuel control system will not adhere to the rules seen in the older systems. — Fred Wilson, Apr 15 '16 at 05:43
@Ben I've posted the screenshots I took yesterday. All of these are with the vehicle at idle — Zaid, Apr 15 '16 at 11:59
@FredWilson The downstream O2's look fine (~0.75V) so I believe the cats are fine. I didn't get a chance to take screenshots with the vehicle running at 2500 rpm. I don't have access to a gas emissions analyzer — Zaid, Apr 15 '16 at 12:01
When you did the throttle blip test, did you hammer it? or did you just pop it a bit? — Pᴀᴜʟsᴛᴇʀ2, Apr 15 '16 at 16:28
@Zaid Thanks for the graphs. Whenever I see both banks moving identically like this I first look for a cause upstream of the split in the air flow such as the intake air path or sensors. The same failure of sensors in both banks is extremely rare. Both the AFR sensors look like they are operating normally. We cannot prove their accuracy with out a gas analyzer. — Fred Wilson, Apr 15 '16 at 19:46
@FredWilson I'm thinking that faulty idle air control may have a role to play here. What do you think? — Zaid, Apr 15 '16 at 19:52
@Zaid A good test for the accuracy of AFR or O2 sensors is to monitor each banks sensors and tailpipe mixture all at the same time to see if they all agree on the mixture. For example: Actual EQ_RAT = .98, O2 BxS2 = 0.7 volts and tailpipe lambda 0.980. This way one we get three data points to triangulate if one of the sensors is mis-calibrated. — Fred Wilson, Apr 15 '16 at 19:52
@Zaid I would test for intake leaks, Test the MAF sensor, Test the IAC. My expertise is in Japanese makes so I have no reservoir of practical experience with this particular system to help you. An experienced MB expert could be helpful. — Fred Wilson, Apr 15 '16 at 20:07
Speaking of vacuum leaks what about the brake booster? If the problem only manifests while coming to a stop that would be a good place to start. — Ben, Apr 16 '16 at 00:09
Zaid: I like @Ben's hypothesis better. Test for a leaking brake booster. Watch the mixture sensors and fuel trims while pushing the brake pedal. If they go lean disconnect booster hose and repeat the test. — Fred Wilson, Apr 16 '16 at 05:20
@Ben followed your suggestion on the brake booster - it seems to be an issue here — Zaid, Apr 20 '16 at 13:51
@FredWilson in case you missed it I've provided an update in the answer below — Zaid, Apr 20 '16 at 13:52

Zaid · Accepted Answer · 2016-04-28T20:33:37.303

We found at least two issues

The air filter housing wasn't forming a proper seal with the MAF sensor

This engine boasts a quirky design where the MAF sensor sits at the back of the engine. The air filter housing seal is in the middle of this first picture, MAF sensor in the second picture (stills were taken from this video).

Whoever had installed it didn't install the air filter housing correctly, so unfiltered air was making its way into the MAF sensor and causing all manners of problems.

After seating the housing in the MAF properly, the STFT fluctuations vanished. Curiously, this vacuum leak was actually metered by the MAF sensor. Notice the slightly postitive LTFT's, which weren't registered by the app before.
The brake booster line jiggles a little

We monitored STFT's upon pressing the brake line and it looks to confirm that the brake booster/line is suspect. I think I can make out a hissing sound near the brake booster vacuum port but didn't have carb/brake cleaner handy to confirm it. Here is the trace when my friend tapped on the brakes three times in quick succession (car was stationary):

The CEL still returns upon braking. At this stage we feel it necessary to repair the brake line vacuum leak before proceeding with declaring this problem solved.

I will update this answer with details as an when they become available.

Update : Found the smoking gun

And it wasn't the brake booster line. It was this: a crack in the part-load breather hose:

The way we discovered it was by blocking the air intake and realizing that there was barely any suction. Because this hose was buried deep under the air filter housing, it wasn't visible but some brake cleaner made light work of uncovering it.

It looks like someone tried to repair it with epoxy.

As a temporary fix crack was sealed up with duct tape and the car taken for a spin. Although I didn't get the chance to log the fuel trims, the car felt much crisper in response to throttle changes.

There was no need to replace the wideband sensors after all!

I expect to see a change in mixture when the brake booster is activated, due to the dip in manifold pressure. But not that much motion in the trims. Would need a know good to confirm if this is normal. — Fred Wilson, Apr 21 '16 at 05:53

Are widebands known to go lazy over time?

Screenshots

1 Answers1

We found at least two issues

Update : Found the smoking gun

Linked