E38 PE Issue. Lean 2.1 Seconds. Can we solve this?

[sillycon]

An OE application wouldn't be running something like a Deka 220 on anything other than CNG/LPG where it's not going to have the same sorts of issues we encounter on it in a gas/alcohol situation (terrible flow characteristics at low IPW). That's a *HUGE* injector (220-235lb/hr!).

I can't currently think up how emissions would be helped by spraying a 50ms squirt in two 25ms squirts vs a single 50ms squirt on the back of a closed intake valve but I'm not a professional calibration engineer either so... Then again, maybe that's part of the point; the injector squirts part on the back of the intake valve for cooling/transient modeling and part when the valve is open to get a better/more homogeneous mixture.

Seems like an awful lot of aggravation though (I've been battling uphill with low IPW flow issues on a set of Dekas on another car for a while; I can't fathom wanting to have to deal with TWO calculations of valve dwell and bounce -- doing it for one pulse per cycle per cylinder is bad enough!) for a potentially small gain.

[RRRocketMan]

An OE application wouldn't be running something like a Deka 220 on anything other than CNG/LPG where it's not going to have the same sorts of issues we encounter on it in a gas/alcohol situation (terrible flow characteristics at low IPW). That's a *HUGE* injector (220-235lb/hr!).

I can't currently think up how emissions would be helped by spraying a 50ms squirt in two 25ms squirts vs a single 50ms squirt on the back of a closed intake valve but I'm not a professional calibration engineer either so... Then again, maybe that's part of the point; the injector squirts part on the back of the intake valve for cooling/transient modeling and part when the valve is open to get a better/more homogeneous mixture.

Seems like an awful lot of aggravation though (I've been battling uphill with low IPW flow issues on a set of Dekas on another car for a while; I can't fathom wanting to have to deal with TWO calculations of valve dwell and bounce -- doing it for one pulse per cycle per cylinder is bad enough!) for a potentially small gain.

Well one thing's for sure. Cars have been running more than fine without electronic fuel injection for decades so as far as I'm concerned, a single squirt is just fine in the sense that it won't break anything. I'm just wondering if multiple squirts is the next holy grail of unlocking hidden horsepower that I just disabled lol.

[sillycon]

Ok, did a bit more digging on this. From what I've found, it seems like the core application for this is to handle non-linearity of injector flow at high IPW's. By halving the IPW and firing twice, you can keep the injector flowing in a predictable part of it's flow curve. I've not really seen injectors where this was a major issue, however -- usually it's the low IPW's that have the seriously screwed up flow characteristics.

The other application for the behavior looks to be with fuel transients; the "multi-pulse" logic can be triggered to give an extra squirt of fuel on-demand to avoid a lean condition. Seems a bit over-kill to me in this situation though (like a 10lb sledge for driving a brad nail into balsa...)

The papers I found discussing this behavior are from the mid-1980's, however, so who knows that sorts of other interesting uses have been dreamed up since then.

[RRRocketMan]

Both of those mechanisms are totally plausible.

I would argue however that while multiple pulses might be appropriate for transient fueling, it wouldn't offer much benefit for linearity. Imagine you need 96% DC, which is beyond what most injectors can handle well. That's 2 x 48% DC right? The problem is that the time required to do 2 pulses is greater than the time required to do 1. If the off time is 52mS for the sake of the math, then 2 x 52mS is 104mS off time. With 1 pulse the off time would only be 4mS. If the RPM is high enough, there's no time for the extra pulse, or you end up spraying half way down the intake stroke. Add to this the fact that the turn on and turn off events are inherently non-linear so having multiple pulses INCREASES the amount of time that you're in a non-linear regime. I would rather model the non-linearity of the injectors and dial them in once and then just use a lookup table for a single pulse that will be reliable rather than calculate it twice or three times as much.

The transient fueling is legit though. The tau model I've been reading about is a bit overkill in most cases because we're talking about a change to a puddle on the port wall causing a lean or rich condition for a few engine cycles, which pass by faster than most people can perceive it. I'd much rather go back to the good old "acceleration enrichment" model that relies on the throttle plate d/dt to anticipate and squirt extra fuel. Worst case, you err on the rich side and pollute the environment a bit. Sounds like a plan to me.

But this is one way in which the multiple option could be good. Another user gave me the impression that shifting was smoother (less afr variation?) without the multiple setting. I'm not sure I follow how that works but if true it would be enough of an incentive to turn it off, so long as the computer deals with transients via some other model (like acceleration enrichment).

[sillycon]

I agree completely; I still can't think of a single scenario where the benefits outweigh the hassle. The "stay in the linear region of the injector" thing just doesn't seem to apply to any modern injector I've ever seen. Typically, as long as you're over 2ms they're PLENTY linear enough to use a simple model for. It's below the 2-2.5 mark where all hell breaks loose.

I'm also with you on the "accel enrich" model; to my feel the cars ran just as smoothly and it was a heck of a lot easier to dial in. We may have been giving up .5mpg or been giving off an extra .01% NOx emissions, however, which is being "saved" by using the "newfangled" transient modeling (which, frankly, is pretty cool -- but way overkill for an off-road application IMO).

The *ONLY* way it makes sense in my mind to bother with the multi-squirt tom-foolery at all is if there's some emissions benefit that was needed to meet regulatory constraints.

[Chris@HPTuners]

The makeup pulse mode determines how many makeup pulses can be delivered before the fuel boundary (the latest angle fuel can be injected and still make it into the cylinder). The makeup pulses are used to adjust the fuel pulse width for changing conditions after the first pulse has occurred but before the fuel boundary. None means no makeup pulses can occur and everything must wait for the next injection cycle, single means one makeup pulse can delivered, and multiple means more than one can be delivered. Basically its a way to adjust the injector pulse width within the same injection cycle after the first pulse has occurred.

The desoot mode has a seperate set of injector EOIT etc. tables (not in the editor) that are used when it activates. I'm not a chemist so i cannot comment on exactly why is being changed, injecting later, earlier or even just someone forgot to calibrate the tables (it happens). My guess is it injects earlier to allow more "washing" of the intake valve preventing soot formation, this would also explain (i think) the lean hiccup as the fuel builds up more on the intake valve and less makes its way into the cylinder until things balance out.

[driftnpow]

I'll throw in a couple of logs. And attach my base tune. The logs have varying changes to the tunes they were run off....... but all produce the same result. Just look for the wot runs.... mid run you'll see it. Should be enough info there to give you an idea.

I'm starting to think it's either fuel pump settings...... or a flow rate issue.

Having a brief look at your log files one thing ive seen on the start of "fing_lean_spike" is the drop in MAF readings in relation to the near linear climb of RPM.
I seem to think there is a MAF reversion issue. in a nutshell phantom airflows created by valves closing and sending pulses of air back past the maf, more severe when your place the maf closer to the intake manifold. I notice you've scaled down base maf table but reversion tables may need addressing to iron out these lean spots, or more the point desensitize the maf as to smooth and make commanding fuel more true to whats needed by the engine.
Take a look at these snapshots that ive taken in relation to this instance that I'm talking about and see what you think screenshot.jpgscreenshot2.jpgscreenshot3.jpgscreenshot4.jpg

[Chris@HPTuners]

Having a brief look at your log files one thing ive seen on the start of "fing_lean_spike" is the drop in MAF readings in relation to the near linear climb of RPM.
I seem to think there is a MAF reversion issue. in a nutshell phantom airflows created by valves closing and sending pulses of air back past the maf, more severe when your place the maf closer to the intake manifold. I notice you've scaled down base maf table but reversion tables may need addressing to iron out these lean spots, or more the point desensitize the maf as to smooth and make commanding fuel more true to whats needed by the engine.
Take a look at these snapshots that ive taken in relation to this instance that I'm talking about and see what you think screenshot.jpgscreenshot2.jpgscreenshot3.jpgscreenshot4.jpg

its the desoot mode.

[RRRocketMan]

The makeup pulse mode determines how many makeup pulses can be delivered before the fuel boundary (the latest angle fuel can be injected and still make it into the cylinder). The makeup pulses are used to adjust the fuel pulse width for changing conditions after the first pulse has occurred but before the fuel boundary. None means no makeup pulses can occur and everything must wait for the next injection cycle, single means one makeup pulse can delivered, and multiple means more than one can be delivered. Basically its a way to adjust the injector pulse width within the same injection cycle after the first pulse has occurred.

Finally found the post on Camaro5:

Cheapest way to upgrade fuel system

--------------------------------------------------------------------------------

Cheapest way to upgrade your fuel is via tune, if your FPCM will allow. Up the fuel pressure setting to 65psi. That's it mainly. Else +1 BAP wires coz it's a PITA to get pump basket in and out.
Then turn off extra pulses, correct delta and adjust injector flow and timing in the tune.
Swapping in LSA (supercharged spray pattern) is gonna stuff transitions esp on ethanol and waste fuel.
--------------------------------------------------------------------------------

This post from a guy named Pirate, indicates "turn off extra pulses" is a good idea and when I asked, he followed it up with this:
--------------------------------------------------------------------------------
Quote:
Originally Posted by rrrocketman View Post
One at a time because this is a very dense and cryptic post.

1) Would be really nice to know how you are able to tune your FPCM to get 65 PSI. I don't think anyone has done this before but by all means if you have, please share with the class.

In EFIlive for e38 PID is F0586-F0589. In other words the settings are there available in the tune, my car is a factory flex fuel vehicle so maybe my FPCM allows for it (although it wasn't set that way from factory).



2) "Turn off extra pulses" - What does this mean?

Basically by default the e38 is set to deliver Multiple extra injection pulses after the main injection pulse, the options are Multiple, Single or No Extra injection pulses. Allows the accurate timing of injection versus puddling at back of valve etc. Some related info as to why we have a stratified pulse option here.

3) "Correct delta" - What's this mean x2?

Raising fuel rail pressure affects the tune, injector flow correction with e38 is handled by correcting pressure delta across the injectors, mainly.

4) +1 on LSA injectors. I went this route and Bosch told me not to use LSA injectors on an N/A engine because the spray pattern is not symmetric. It's 10 degree offset.

Yes it still works to a large degree if you simply think of FI injectors as an upgrade to flow capacity / hp but the fuel consumption is terrible as are handling transitions. It's because the atomisation and spray pattern of NA vs FI are set with different objectives as target / throw, spray angle, shape, width.
--------------------------------------------------------------------------------

So all I get from this is that having multiple pulses somehow impairs pulse timing.

What do you say to this assertion? Any truth to it?

[Chris@HPTuners]

stratified pulse? this is PFI not DI.

As i said above, the makeup pulses allow the ECM to add more fuel within the current injection cycle. eg. the fuel code decides it should have injected more fuel in the main pulse, the makeup pulse allows it too, instead of waiting for the next intake cycle.

[sillycon]

Chris, what sort of logic is running in the "main thread" which would cause the ECM to want to revise the fuel squirt? Seems completely dodgy to me to not just "do it right the first time" as a second pulse is just going to exacerbate valve dwell and bounce issues inherent in the injector. Is this perhaps leveraged as a way to help mitigate burst knock or such?

[Chris@HPTuners]

Chris, what sort of logic is running in the "main thread" which would cause the ECM to want to revise the fuel squirt? Seems completely dodgy to me to not just "do it right the first time" as a second pulse is just going to exacerbate valve dwell and bounce issues inherent in the injector. Is this perhaps leveraged as a way to help mitigate burst knock or such?

the logic for the makeup pulse calculations are contained within the TPU microcode and not visible. At low RPMs the ECM code is running much, much faster than the engine is turning eg. at 2000rpm each intake stroke is 60ms so the ECM can update the injector pulse width multiple times in this time. The code has a minimum injector off time that must be observed to prevent non-linearities in fuel delivery. The TPU microcode handles the rest. If you think makeup pulses are an issue, turn them off.

The point of this thread was to solve the PE mode leanout that some people were seeing, the desoot mode is the answer to this.

[GHuggins]

Chris, sorry to keep beating a dead horse here, but would it then be a safe assumption that multiple pulse modes only work during small pulsewidth or lower rpm scenarios? Or is it duty cycle biased for wot scenarios and why sometimes you can have injectors operating correctly above 100% dc?

[Chris@HPTuners]

Chris, sorry to keep beating a dead horse here, but would it then be a safe assumption that multiple pulse modes only work during small pulsewidth or lower rpm scenarios? Or is it duty cycle biased for wot scenarios and why sometimes you can have injectors operating correctly above 100% dc?

i don't know precisely because the code isn't visible to me, only GM could answer absolutely. If i was to guess then yes i would agree that it is most likely for low RPM transient corrections.