Lean transition into PE and lean spot during WOT

[Niemer]

Do you have the ability to log fuel rail pressure? I'd be curious to see if you are having the same pressure drop right before the event that I have.

If you do, set up a custom PID that looks like this so we can compare apples to apples, it'll log only the pressure drop across the injector, taking into account MAP. The AUX will be whatever you have defined for the fuel pressure sensor


On another note, does anyone have a easy way to convert the power supply to the fuel pump to a 0-5v signal? I'm really curious to see if the cause of my pressure drop is related to the pump or not.

[BBA]

Just read this entire post...following.

I know about the 2.1 seconds thing...my TBSS (E67 PCM) does it as well right after stall. I did not notice it the problem the stock transient and PE delay settings, but that doesn't mean it wasn't there: the stock settings are for PE and Transient to come in soft and gently (which makes the thing feel weak on the bottom end before feeling like it 'comes alive). Quickening up the PE and transient makes the bottom end feel like it's supposed to...but also exposes this issue.

Oh, it's not just 2.1 seconds after PE, it's 2.1 seconds after the engine starts accelerating.

[10_SS]

Just read this entire post...following.

I know about the 2.1 seconds thing...my TBSS (E67 PCM) does it as well right after stall. I did not notice it the problem the stock transient and PE delay settings, but that doesn't mean it wasn't there: the stock settings are for PE and Transient to come in soft and gently (which makes the thing feel weak on the bottom end before feeling like it 'comes alive). Quickening up the PE and transient makes the bottom end feel like it's supposed to...but also exposes this issue.

Oh, it's not just 2.1 seconds after PE, it's 2.1 seconds after the engine starts accelerating.

I think someone with the problem (with ID850's) should attach a scope to one or more of the injector power wires and make this happen... According to the screen shots I see the pulsewidth is saying it's not changing.. but there's only one way to know for sure... maybe the pulse width or the voltage, or both are being altered for some reason... scope would tell you.

[brentsten]

Voltage shouldn't vary I wouldn't think. The pulsewidth ranges are semi large in the screenshots, but they are moving from what I see?

I also have a mechanically regulated ZL1 pump. I have a fuel pressure sensor on the rail, I'll log deltaP in a bit and post results.

[Niemer]

I think someone with the problem (with ID850's) should attach a scope to one or more of the injector power wires and make this happen... According to the screen shots I see the pulsewidth is saying it's not changing.. but there's only one way to know for sure... maybe the pulse width or the voltage, or both are being altered for some reason... scope would tell you.

If I had a scope I would do this and take a look at fuel pump voltage, but alas, I do not.

[Niemer]

Voltage shouldn't vary I wouldn't think. The pulsewidth ranges are semi large in the screenshots, but they are moving from what I see?

I also have a mechanically regulated ZL1 pump. I have a fuel pressure sensor on the rail, I'll log deltaP in a bit and post results.

I use a 10µF or 100µF cap to smooth out the signal, without it, it's pretty rough.

[brentsten]

Do you have the ability to log fuel rail pressure? I'd be curious to see if you are having the same pressure drop right before the event that I have.

If you do, set up a custom PID that looks like this so we can compare apples to apples, it'll log only the pressure drop across the injector, taking into account MAP. The AUX will be whatever you have defined for the fuel pressure sensor

Well I'll be damned. Same fuel pressure spike/dip here as well:
Capture.PNG

InjDeltMeas is actual data using your formula, InjDeltaECM is the standard ECM PID.

On another note, does anyone have a easy way to convert the power supply to the fuel pump to a 0-5v signal? I'm really curious to see if the cause of my pressure drop is related to the pump or not.

The easiest way would be to make a voltage divider with a few resistors, but you'd want to take your measurement as close to the pump as possible for the most accurate reading. You're PWM controlled though; you'll probably want to measure duty cycle. Does the VaporWorx kit operate based solely on fuel pressure or does it take MAP into account as well?

I use a 10µF or 100µF cap to smooth out the signal, without it, it's pretty rough.

Thanks for the tip. The analog gauge isn't fast enough to notice, but it was pretty erratic in HPT. 100µF cleaned it up nicely.

[Niemer]

Kinda what I thought, something in the tune that we can't see must be causing this. The standard PID is a calculated value since there isn't a fuel pressure sensor from the factory.

Yeah, I am PWM and I'm new enough at the electronics side of things that I have very little idea what I'm doing. My kit does reference the MAP sensor to maintain a fixed delta across the injector.

[brentsten]

Something I noticed.. The PID for injector delta pressure never drops below 58 psi, which is what it's programmed to think fuel pressure is at all times. It increases with vacuum, but isn't accounting for boost. Maybe the OS patch will fix that.

Im starting to think it's not tune related at all. It could just be due to the fact that the regulator (or pressure sensor in your case) is so far from the injectors. Since the ecm can't measure actual fuel pressure, a return style system may be the answer.

[Niemer]

It shouldn't matter, the PCM is told what the pressure drop across the injectors is in the tune so it can calculate delivered fuel by that. Boost referenced or not, return or retrunless shouldn't make a difference as long as your data is correct and your system can supply enough fuel.

The fact that you and I are experiencing the exact same thing on different cars with different hardware and the only real thing we share in common is the PCM really makes me think its something in the factory calibration.

[brentsten]

It shouldn't matter, the PCM is told what the pressure drop across the injectors is in the tune so it can calculate delivered fuel by that. Boost referenced or not, return or retrunless shouldn't make a difference as long as your data is correct and your system can supply enough fuel.

I read roughly 65psi delta at idle, but only 58psi delta anywhere from 0psi boost on up.
In stock form, the ECM assumes there is a constant 58psi fuel pressure in the rail and adds/subtracts vacuum to arrive at delta, but won't read higher than baro. The fuel pressure is a fixed number in the calibration and unfortunately, the E38 isn't capable of sensing variable fuel pressure.

To quote Chris@HPT: "I've also been testing a new OS patch to enable the pressure delta to go into boost, instead of maxing out at equivalent barometric pressure. This means the IFR table will actually go into boost areas without the SC Fitted option being set."

I applied the OS patch last night and now delta drops below 58 when going into boost. Didn't do a thing as far as the lean spike though lol.

[BBA]

I think it's something in the PCM. What would cause a 2 second delay before it goes lean? One would think 2 seconds of WOT is enough for the vacuum/boost transients due to throttle changes to have normalized, boosted or not. I think the funny thing is it does this on my relatively stock LS2, about 2 seconds into power.

Let's look at other things not PCM related: Does it take 2 seconds for air pressure waves in the intake tubing between the throttle and the MAF to normalize? Does it take 2 seconds for the resulting pressure wave to be sensed by O2 sensors?

[10_SS]

If I had a scope I would do this and take a look at fuel pump voltage, but alas, I do not.

you can use a sound card mic input, just have to use voltage limiter circuit... or buy a cheap USB scope on Amazon.. links below

http://homediyelectronics.com/projec...illoscope/?p=2

http://www.amazon.com/gp/product/B00...tag=hdetron-20

[griffincox]

Has anyone tried modfying the Prediction Coefficients?

I have mine all zero'd currently.

[brentsten]

Found it.. Sort of.

I've dialed in the non-boost regions of the VE table. I started to work on boost, but the nasty spike still exists no matter how rich I command it.. I found it's coming directly from PE.

Whether I use PE or BE to enrich, the spike still happens. So I disabled all PE/BE, set all OLFA tables to 1.0 and used the airflow gain table to add enrichment above 100kPa. The spike is completely gone and the AFR is rock solid anytime in boost, whether WOT or part throttle.

I re-enabled PE and set all values to 1.0 (still using airflow gain for enrichment), and the spike STILL happened.

Bottom line is, it's something to do with entering PE mode. My fix for now is to use the airflow gain table to reach the desired AFR. The only problem I've had with this approach so far is that the car has to be in open-loop to use it. I haven't found a way to make it go open loop any sooner, so for now I'm just running open loop at all times.

[383_Stroker]

Appreciate the continued digging on this. I've not been able to drive my car enough to contribute, but hopefully that will change soon. Interesting that it follows only when. Switching into pe. What if you force it in pe all the time?

[brentsten]

Appreciate the continued digging on this. I've not been able to drive my car enough to contribute, but hopefully that will change soon. Interesting that it follows only when. Switching into pe. What if you force it in pe all the time?

I didn't try just because PE forces open loop and I'm already running open loop anyway.

The other interesting part is that we're all experiencing the issue based on a factor of time, but changing the enrichment ramp in rate and delays have no effect.

[schpenxel]

Do you have a MAF or VE only?

[Niemer]

I've run it both ways, it seems to be less pronounced at WOT in VE only, but still very much there with part throttle PE.

[BMWMD82]

Any updates on this? I think I might have the "2 second in" lean spike as well. I'm going to try to eliminate a random misfire as the possible cause before I give up, but this seems awfully similar to what mine is doing.