2008 2.5T Fuel Pressure, Duty Cycle, VIDA, DDX fun! -- Thoughts?

[Got2AV8]

Hey, all! Lots of Volvo forums, but I liked the feel of this one, so hey! I am looking for input/opinions, but hopefully there may also be something interesting and useful for some of you, as well -- especially with VIDA interest. Most of my use of it, thus far, has been fairly basic, but this gave me an opportunity to play a little further. It seems to back up my symptom based suspicions and help narrow down the DDX to.... well, you have to read to the bottom to know that. =) Once solved, I will follow up with that information, as well.

So, 2008 S60 2.5T, 142k miles, (otherwise) excellent condition throughout, developed a fun little problem of conking out after a relatively short time running (say 5-10 mins). Cold starts are beautiful and it doesn't miss a beat... until a little stutter here and there, progressing to a nastiness of misfiring, followed by shut down.

Initial theory was of fuel pressure loss, progressively leaning out mix, etc. But, where is the failure? VIDA statistics support this and seem (to me) to help point toward either the pump unit or the PEM (VIDA refers to as "fuel pump module"). I would argue that it does not point to the pressure sensor (despite their common failure rate), as described later.

A few pertinent (should be) negatives: The PEM is located inside the trunk (by 2008 they had wised up), so would not be affected by corrosion. But, I'm not fully convinced their failure rate is necessarily always because of the elements. So, it remains suspicious.

VIDA fun:
- Pertinent codes: ECM-290D Fuel Pressure - Signal Too Low, ECM-2900 Fuel Pressure - Faulty Signal (this one produced by the prior). After some time running it will produce lean mix codes and misfire codes (as suspected from low pressure).

- I graphed some measurements in two different takes. The first time involved a short drive from a relatively cold start until failure (you will see the progression). The second was parked, mostly (if not all) @ idle. This second was approximately 15 minutes after the first (as in, the car sat off for only 15 minutes before brief recovery).

Screenshots (will attach video, as well, for if you're *that* interested):

Slide 1-1, First take, initial start: Initially Duty Cycle (pulse width modulation -- will call PWM) 42.06%/400 kPa nominal/399.72 kPa pressure (will call "FP"). ECM calls for startup FP (480 nominal) and duty cycle 100.19%, with FP 399 kPa. Then, @ same time pressure spikes to a peak of 593.61 kPa, the ECM begins to gradually lower duty cycle, and well as nominal value (will call "nom"), and pressure follows suit -- leveling off around 42%/400 nom/ 399 kPa FP. This, I'll call "baseline." (Side note, at idle, ideal is 35% (+/- 5%) and ~ 400 kPa.)

Slide 1-2, "baseline" -->freak out: Maintains "baseline" for a time, followed by two notable occurrences. The first (1), all happens nearly instantaneously: FP drops to ~ 360 kPa, ECM increases PWM to ~46%/maintains 400 kPa nom/pressure spikes back to 400 kPa. Between now and the second occurrence (2), the fuel pressure notably varies about (btwn ~370-390 kPa) while nom remains and PWM makes small adjustments -- *all* being relatively close to "baseline." Then the second occurrence (2), which becomes a somewhat familiar sight: FP drops, over the course of ~ 5 secs to ~125 kPa, nom remains 400 kPa, and ECM increases PWM to 85.16% (which seems to be about the running max, until things get nasty later and it's desperately attempting to reach nom).

Slide 1-3: These conditions maintain for ~20-25 seconds. This slide starts at the end of these conditions, showing a gradual increase in FP before a sudden spike to 603 kPa, and ECU immediately decreasing PWM. In short, it re-balances back for baseline for 10-15 secs. But, then comes the familiar sight: FP drops and ECU attempts to compensate, but cannot (though a few notable blips in pressure).

Slide 1-4: Freak out to conk out!: Begins w/ close to "baseline" numbers (42%/400 kPa nom/400 kPa FP). Then ensues the familiar sight of FP drop and PWM increase, but this time without future recovery to "baseline." In the FP decline, note the small blips of increased pressure here and there, followed by drop off. Ultimately, this time, the pressure dropped to around 102 kPa @ 85%. At this, the lowest pressure yet reached, misfires are abundant and restart is unlikely.

If you're super interested (haha), here is a video of it running in real time: https://youtu.be/bLhaExfCeq8

Following are the screenshots of the second take -- 15 minutes after prior "conk out,' all of which is parked and at idle. I won't caption them, but note that, at some point, it pegs its PWM around 60%, no matter what the pressure is. And note the misfires.

And the video of this is: https://youtu.be/uU5PA3Nf2Jk

SO, my thoughts and wonders:
- Symptoms indicate a fuel delivery issue. Statistics back this up.

- I relatively rule out FPS, as I understand if an error is detected in it, the ECM has a back up -- I've read it's 100%, but I suspect this is the 60% that it seemed to lock into, eventually. Either way, at PWM of 100% or 60%, at idle (and most driving), it should be symptom free, but is not -- lean mix codes, misfires, stalls (note: I have tried it with FPS disconnected -- no change). Further, by looking at the shape of the variations, it seems that it's recording as would make sense -- potentially fast spike, with a sloped drop.

- All of the above leads me to suspect either the fuel pump or the PEM (or "fuel pump module"). Discerning which seems nearly impossible, as they're so closely related. Further, my understanding is that PWM readouts in VIDA are coming from the ECM, not the PEM, in which case this is far from proving that the commands are reaching the pump. *This said,* there are plenty of occurrences which appear to show the car varying/compensating with the PWM and, indeed, a pressure response. This would indicate that (at least sometimes!) the PWM is, indeed, being delivered to the pump.

- Ergo, I'm leaning toward a failing fuel pump (or possibly, though unlikely a wiring issue between the PEM and pump). It appears that I can source a PEM much more inexpensively than a pump, so I'd rather it be that. But, so far, I'm not feelin' that.

- Having said this, I am interested in opinions, especially if anyone has had a very similar experience.

Slainte!

[jbeebo]

Nice set of data. Do you have access to an oscilloscope? I'm thinking you could further diagnose the PEM vs. pump issue by tapping the lines to the pump motor and monitoring PWM. If the PEM is bad I'd imagine a strange, erratic or missing signal going to the pump module. If the pump is bad and PEM is good you'd see consistent PWM signal from the PEM.

Have you confirmed the VIDA data with a good old fashioned analog measurement of fuel rail pressure?

[vtl]

Where's your PEM is located? If it's still in rear right wheel area, it may have oxidized connector. Volvo issued TSB some time later to relocate PEM into trunk.

[Got2AV8]

Thanks for the responses. Unfortunately, the only diagnostic tool I have on hand is VIDA. The oscilloscope is a great idea, though, because from the data I'm getting it really (I think) is impossible to tell which of the two are at fault, since the data is coming from the ECM. Hmmmmm... However, I have two multimeters, the better one of which seems to have hidden itself in a move. I think that one may measure frequency -- am compelled to find it now.

I've not run any more particular diagnostics (nor taken apart yet, as it decided to become Winter suddenly), but I have run it a couple of times (no VIDA - just to get it home) and find the behavior wholly erratic. Previously, I thought that it was allowing some window of time -- I think this still holds true to some degree, as after sitting it does run again -- but the trip home consisted of two legs, where the first was much shorter till shut down and the second (about a half hour later), was perhaps 2.5 times as long as the first. It quit right as I turned into my driveway, but immediately restarted (with a little stumble, not surprisingly) to get up the driveway. I then let it idle for a few minutes out of curiosity and it idles smoothly for about that long until I sensed a miss here and there, so turned it off. This doesn't strike me as the old pump overheats and binds sort of idea. It also doesn't distinguish *where* the issue is, though. So, now it's snowing. Next chance, I'll pull out the rear seat, inspect the connections, and if I can find the multimeter, attempt to tap in and see if I get a clue. Thank you for the idea!!

To the latter comment: The PEM is in the trunk, under the flooring and next to the spare tire. It's a pristine shiny new looking PEM, but I don't believe it! haha. I am, though, aware of the TSB about them, but weren't they relocated from the outside of the tank to under the rear seat? My (possibly poor) understanding is that they had wised up by 2008 and the location of this one is from the factory.

[Got2AV8]

One interesting thing that I can't entirely confirm, but seemed this way, is that it seems to be happier when going -- such as cruise/on the throttle. This may be entirely not the case, but it seems like when it has begun to act up has always been around the time of off throttle, idling, or trying to accelerate after off-throttle (such as slowing down for traffic, then accelerating again). Can't confirm it and may not even mean anything -- but it's slightly interesting.

[jbeebo]

Was thinking about this some more. Oscope would be definitive, but I think you could do with a multimeter; in frequency measurement mode if it has one, else I think in AC voltage or AC current mode might work too. Although the RMS calc would be inaccurate with a square wave PWM of unknown frequency (probably 1kHz), it should give some indication of "average" voltage/current, and thus be a good enough diagnostic tool for this purpose. Just another idea/thought. Good luck!

[abscate]

Make a resistance measurement on the pump to see if it's aging

Ditto request for a second independent fuel rail measurement. If you have time pressure, Maybe try a junkyard FPS PEM as that might be faster than trying to diagnose

[jbeebo]

Unfortunately static resistance measurements on DC brush motors are not very meaningful. There's too much variation in the brush/commutator interface and you can't get accurate results, especially with a multimeter which is pumping milli/microamps thru the circuit. The correct way to measure DC resistance is quasi-static, driving the motor close to stall, measuring voltage and current and calculating DC resistance. Has to be close to stall to minimize motor back EMF.

This is not possible in situ, so the best you can do is measure voltage and current and see if it deviates from known good operating characteristics. As the motor will run the pump up the pump curve to a steady state operating point, the current draw will be proportional to motor speed, and should be pretty consistent if the system is operating well.

Unfortunately we don't know typical operating conditions, but in this case it's not critical as we're looking at: is it operating or not, and what is the PEM doing. So voltage and/or current measurements between the PEM and pump should suffice.

I like abscate's idea of getting junk yard parts and just swapping them out. Indeed might be faster than proper diagnosis such as confirming rail fuel pressure, and monitoring PEM - pump interaction.

[Got2AV8]

Okay -- finally got to it some more! Long story short, it has a new fuel pump. But, quickly, here's the story:

I was able to ensure that the duty cycle signal was getting from the ECM to the PEM. I opened up the PEM and using the freq setting on the multimeter, was able to match it with the same numbers VIDA had been reporting (as above). Of course I was also able to get a visual of the circuitry, which looked like it was made yesterday. I wasn't able to assess the PWM to the pump -- I'm really not sure why, but whatever.

So, I pulled the back seat, opened up the cover and started poking at the pump terminals. Of course, at the time, the car didn't want to run, but I lucked out at some point and it came to life. Again, I wasn't able to assess the PWM, but I observed this: Knowing some expected trend (thanks VIDA), the longer it ran, the more it would increase duty cycle to compensate for falling pressure. Sometimes the simplest assessment is the needed clue -- the pump motor accelerated along with the trend. The pressure dropped, it got cranky -- that was enough to point at the pump, as well as determine the PEM was, indeed, getting the signal to it.

Then, I learned that the pump had been replaced before (before my ownership) with an aftermarket brand -- something like Swedish Performance Parts ('ish). I'm not opposed to aftermarket parts, but I am not so sure this was a high quality one. Firstly, I suspect that part of the failure had to do with the install -- it *seemed* that the hoses across the tank had been forced in and kinked. This may have been overworking things. But, upon removal, the nipple on the bottom of the plastic container the pump is within broke off. The pump, clearly, is not very old -- but, sometimes that means little.

So, I replaced it with an OEM Bosch, and it's a whole new world.

I had failed to mention, prior, regarding an "analog" pressure measurement. I didn't bother with this for a few reasons. Partially, I don't have a gauge and last time I rented one, it didn't have a fitting that worked on a Volvo (though I fabbed one and was able to use it). But, the main reason is this: It wasn't necessary/would have been unnecessarily redundant because of the combination of data VIDA was receiving along with the symptomatic presentation. That is, 1) I was receiving apparently stable data from the FPS (not bouncing about, not out of range, etc.). 2) The pressure versus duty cycle, at best, was near to ideal (what I had labeled baseline). 3) When strangeness came about, the readings (looking at the graph) indicated what I would expect from a functioning sensor (the peaks and falloffs). 4) Lastly, and definitely most importantly, the data combination alongside the physical symptoms all said the FPS was working -- that is, the car ran well when it indicated the pressure was appropriate (or even during some of the fall off), but poorly when it indicated the pressure was not, despite the increased duty cycle commands. So, there was plenty of data regarding pressure to say that the FPS was working well (enough).

Slainte!

[jbeebo]

Great work Got2! Super glad you figured it out.

And good to know VIDA reports real physical parameters. Too often the data reported by scan tools is scaled or otherwise obscured, which makes it tricky to interpret.