If anyone remembers, I've recently started experiencing a unique start from my car.

What basically occurs, and you can see it in the AE grid of the data, that the car starts up instantly, she stumbles within the first 5 seconds, the car compensates by increasing the RPMs, she goes up...and smooths out, then the RPMs settle back down to normal. This only happens after sitting for 10-12 hours, mainly in the morning. If I go out to lunch here in 30 minutes, it shows no signs of this strange behavior.

I have my suspicions, but I thought it might help others to diagnose their own, and I'd like to see if the fuel trim experts around here can offer some additional insight...after all, I could be wrong.

File attached is 31MB, avi so you can watch it in Media Player (use full screen or you won't be able to see it) The screen was captured when playing back the data file at a very slow rate. It is approximately 2 minutes long, but the AE data was gathered in about 30 seconds.

Sensors being monitored in this data file include:

• absolute throttle position
• calculated load value
• engine coolant temperature (F)
• engine oil temperature (F)
• engine speed (RPM)
• fuel pressure (psi)
• ignition timing advance for cylinder #1
• intake air temperature
• long term fuel trim - bank 2 (%)
• long term fuel trim - bank 1 (%)
• manifold absolute pressure sensor (psi)
• mass air flow (lb/min)
• mass air flow meter ground (V)
• short term fuel trim - bank 1 (%)
• short term fuel trim - bank 2 (%)
• throttle position sensor

Problem solving, frustrating...but can be very rewarding when successful. Have fun.

If someone wants another sensor added, let me know, I can add it on my next morning startup.

A screen shot for those w/o video ability...but it doesn't tell you what was happening just prior to this point in time.

 

Even though it's at 1400 rpm, MAF @ 22lbs looks a bit high, to me (also prob. the reason for 36% load?) Probably o.k., I'm a little rusty...

STFTs o.k. but LTFT bank 1 @ 14% is off for some reason. Until O2s are warmed and working (as witnessed by STFT activity in your video) LTFT should be close to -0-. Maybe slow refresh due to too many PIDs open causing false abnormal? (kinda thinking out loud here, but heading somewhere)

Scanner aside, your description sounds like a dribbling injector. Dribbles down fuel pressure slowly into cylinder while parked which has to clear out when first started, creating the stumble. I have to defer to experienced Jag techs, I'm not familiar with the air assisted sytem, maybe a leaky o-ring could also be a cause? I'm beginning to think LTFT1 was running high because fat mixture (dribble), cold start misfire created high O2 (no O2 used/burnt one or more cylinders due to misfire) and it's compensating until cylinder ventilates, begins firing and lowers O2.

The more PIDs you select, the slower the scanner refreshes. I'd stick with ECT, LT & ST Trims. I'd also take a look at the O211 & O221 sensors on graphing, it's hard to tell because of the slow updating but you may have a sluggish O2 sensor (unrelated to the stumble issue though.)

 

thanks beav, lets hope it isn't anything as serious as the injector.

First, my 2007 AE doesn't appear to measure air flow correctly, that's nowhere near normal. I installed the new version, and it corrected that error...but lost my enhanced features of course. I may have to do a firmware update or call AE support on that little issue.

I'm going to look specifically at watching for misfires next time, I read the crankshaft rotation speed changes rapidly. I'll try to find that in the software to monitor.

Other than the spike in air, bank 1 has been having a hard time adjusting on the fuel trims. I've noticed this in real time mode also...where bank 2 trim would rise and fall pretty regularly, bank 1 didn't reach the same peak and valley, and it didn't drop as quick either.

This is the freeze frame data in January when it first occurred.



Since the first occurrence was temporarily corrected by a cleaning of the MAFS, i'm throwing a new unit at the car this weekend. After that, I'm going to do the 02 sensors pre-cat...heck, probably all of them. Pretty sure I can handle that job.

 

Don't know that I'd start throwing parts at it, just yet.

Rear O2s exist only to determine cat efficiency, absolutely no effect on running.

Select O211 & O221 (Ox sensor bank 1 upstream & bank 2 upstream, rears are O212 & O222) and watch live graphing. Not sure if this has been covered elsewhere or if you're familiar with how they function, so...

O2 sensors typically measure .1-.9 volts. .1 is lean, .9 is rich with .45v considered the 'threshold' (typically.) On a live graph this voltage is represented as a sine wave as the O2 measures (let's begin with 'rich' as a starting point) .9v signaling 'rich' to the ECU. The ECU responds by reducing the injector on-time (a.k.a. 'dwell'), which reduces the amount of fuel supplied. The O2 then 'sees' this new lean condition and reports .1v to the ECU which in turn increases the injector on-time. And so the cycles go...

.1v & .9v are ideal figures, as O2 sensors age and get a little sluggish, their ranges may decrease to, say .3-.8v. They should sweep about 10 times per second (peak-to-peak @2,000 rpm although new sensors should be able to achieve at idle), sluggish could be less. They can have reduced volt ranges &/or response. When watching the graph remember volts are vertical, time is horizontal. OBDII standards require certain thresholds to be maintained, OEMS usually run a bit tighter to ensure compliance. A little sluggishness or reduced voltages are o.k. but if they get too bad the ECU should recognize the issue and toss a code. However, sometimes a flaky sensor can get past the ECU and cause issues. Look for voltage ranges, reduced waves, big wrinkles/flat spots/breaks/etc. in the waves (a little 'itchiness' is normal) and compare waveforms between both banks.

Remember, it takes a bit for them to warm up and report, until that happens the ECU is operating in 'open loop' and from fixed parameters. Plus, most OEMs don't like replacing O2 sensors just because their idle speed waves don't look good. They are only interested in diagnosing them @ 2,000 - 2,500 rpm (its a heat thing - cool sensor outputs are unreliable.) If you watch downstream sensors it can take up to two minutes at 2.500 rpm to warm them up and begin reporting.

That's a lot of cylinders misfiring. Early crank sensors and ECUs would get confused about which cylinders were actually misfiring but that's not usually an issue these days (that's a whole different lesson.) If your scanner will, take a peek at the crank sensor graphing for anything that doesn't look smooth.

I've never seen a MAF only be bad at cold start-up...

 

Will take a look at the upstream 02's, thanks for the suggestion. I do remember seeing flatspots on the bank 1 sensor, and bank 2 would respond much sharper to the peaks and valleys of the sin wave. I may try another screen video to show the curve...stay tuned!

I have noticed, since the MAFS was changed, my long term trims have dropped to below 10 on each bank, where pre-replacement, it was 12+ and 14+. I did not do a hard reset this time after the part exchange.

 

Vid' if you want but a pic of them at idle and at 2,000, especially if you can catch them acting up, would be sufficient.

First post and freeze frame didn't show but one bank over 10%. Good to hear MAF helped.

BTW, did the MAF value drop down vs. the first post?

 

I do think the MAFs did change the initial startup of the car. I just started it after sitting 11 hours, and it started totally different...high RPMs immediately (about 1600 or so), the air flow never spiked liked it did before; then the RPMs fell to a normal idle level of 700 rpms after 8 seconds, roughly. Also, the actual 'value' of the data was corrected when I installed v9.1 of AE, the gm/sec and lb/min were not off. The previous 22 lb/min and idle of 5 lb/min was more than likely a scale factor error in my original v5.4, and probably should have read 2.2 and 0.5. I'm trying to correct that manually or see if an update will fix this permanently, w/o paying for a new software update ($250!)

Here are my graphs and a couple charts. It appears the S1 sensors are doing a pretty good job, the S2 sensors may be indicating my bank 1 cat is not as efficient as bank 2's, I mean look at how flat it is compared to the other bank....slow to respond after throttle is released or applied. Or, it could be the downstream sensors, not sure. but, as long as my long trims stay down where they are +6-7 range, and not in the low teens like before, I think my problem has been resolved.

I've also noticed that I can't even get the car to knock anymore...where all I had to do before was give it a little gas when say traveling at a consistent speed of 25 mph, RPMs maybe in the 1000 range, give it a little gas to steadily accelerate to 40-50 mph, and it would consistently knock for a good 3-4 seconds before it corrected itself (or the RPMs reached high enough)

 

As I said before, downstream sensors have no bearing on engine function. However, here's a head's up on why they shouldn't appear to be the same as the upstream sensors: As the front O2 voltage fluctuates up and down in relation to rich/lean, the downstream sensors should be pretty steady with little 'sine wave' showing. The reason for this is after the cats light up they should be effective enough that there would be no rich/lean activity. Now when you apply more/less throttle you'll see a momentary increase/decrease but nothing resembling a sine wave. In fact, if you see that the downs are mimicking the ups that means either you haven't given the cats enough time to warm up or the cats are dead.

If your downstream chart is warmed up cats at steady rpm they're not looking too good - the cats, that is. You can only check downstreams at 2,000-2,500 rpm after running at that speed for two minutes or more.

Do these cars actually use wideband O2 sensors? Very few cars do. Your upstream chart is set up for O2 sensor current, not volts. You would check the O2 sensor's heater current but to check the sensor itself you need to watch the voltage. Flat lines like those would be more what you would see from the downstreams.

 

Unfortunately, with the latest version, voltage is not an option for the S1, only the S2, so used what I had available...which was the current on wideband or equivalent ration, in percentage. I did run the car at 2500 rpm for several minutes before recording the graph snapshot.

 

When you select the O2 sensor tab do the drop down boxes allow you to select upstreams?

 

upstreams were not an option in the dropdowns, at least not in the generic drivetrain section...v9.1. I am reinstalling my v5.4 to see if there is more in the unlocked, enhanced areas.