68FB

While checking the fuel pressure, I hooked up an Autotap scanner to log the fuel trim changes when I pulled the vacuum line off the regulator. The results surprised me.
The sequence was:
1. Normal idle – vacuum line connected, 20” vacuum, 62 psi (4.3 bar) normal fuel pressure.
STFT’s both bouncing around 0%.
LTFT’s dead steady at 9.4% (B1) and 8.6% (B2).

2. At idle, disconnected vacuum line and plugged it. Fuel pressure increased to 72 psi (5.0 bar) as expected. Plugged vacuum line meant no air leak into intake.
Right after disconnect, STFT (B1) dropped to –5% and STFT(B2) dropped to –8%, i.e. reduced injection pulse width to compensate for higher fuel pressure.
But then STFT’s start correcting and after about 50 secs ended up back at 0%.
No change in LTFT’s.

3. Same as 2. but rpm raised to 3000. No change to STFT’s or LTFT’s.

4. At idle, unplugged disconnected vacuum line, introducing air leak to intake. Fuel pressure remained at 72 psi.
Right after unplugging, STFT (B1) increased to +6% and STFT(B2) increased to +7%, i.e. increased injection pulse width to compensate for air leak.
But then STFT’s start correcting and after about 50 secs ended up back at 0%.
No change in LTFT’s.

5. At idle, reconnected vacuum line to regulator. Fuel pressure returned to normal 62 psi.
Right after reconnecting, both STFT’s go negative, i.e. removing fuel because no more air in-leakage.
But then STFT’s start correcting and after about 30 secs end up back at 0%.
No change in LTFT’s.



My question from all of this is what was correcting the STFT’s each time? The DME wasn’t compensating by changing the measured LTFT’s. They didn’t change. The DME must have been changing LTFT’s but NOT the standard ones monitored by the OBDII scanner. I think there is a hidden LTFT component that we don’t know about.

rao

STFT's are adjusting to conditions a few milliseconds at at time. They are supposed to bounce around. When they are consistent for a specified period of time, the adjustment is moved to LTFT.

Trying to chase them is not possible.

__________________
Rob


Current - 98 540i with Chevy 6.6 LS2/T-56 power - Build Thread
Old - 92 325i with Chevy 6.6 LS2/T-56 Power
Older - 95 M3 with Chevy 6.6 LS2/T-56 Power

Please read the important REVISED DISCLAIMER that applies to this post.

cab

If I remember correctly, there is a additive and multiplicative trim in INPA which equate to LTFTs. Only one of those, "Multiplicative" I believe, maps to what I'll call the "generic" OBD II LTFT...so perhaps it is the "Additive" that is adjusting (note: In INPA the STFTs are represented as the Lambda Integrator).

68FB

I don't think you are understanding what happened. When I introduced a change like increased fuel pressure, the STFT's compensated as expected but, even though the fuel pressure remained high, the STFT's drifted back to 0%. Something else must have taken over the compensation to allow the STFT's to go back to zero.

68FB

I thought the correlation between the BMW (INPA) terminology and the OBDII STFS and LTFT had been pretty well figured out, or at least guessed at in an educated way, although I can't find the thread that discusses it. You were involved in it, cab. Can you find it?

Yes, I suspect the answer to my question is related to the BMW way of doing fuel adaptation. I read the confusing Fuel Adaptation pdf again but it still didn't make sense.

I need to do this test again with INPA to see if any of the fuel trim parameters adapt to allow the STFT's to return to zero.

bimmerule

Think of STFT as acceleration and LTFT as speed, once the injector duty cycle reaches a steady state to compensate for the newly introduced external condition, STFT falls to zero. If you were to continue the experiment much much longer past your test interval of 30 to 50 sec, I'd expect the LTFT to start drifting as well. This is just my hypothesis.

__________________
e60 ssk, zhp knob, brake ducts, ate super gold, ate brake shoes, pagid yellow rs19, pagid blue, titanium backing plates, dinan rsb, mk4 gps, vxx ae, brushed aluminum trim swap, gp2, bell sa2010, sparco tornado, sparco speed+ racing, 3mm front spacers, turner 32mm bolts, x5 thrust arm bushings, OE 9.5" squared 275/35/18 dunlop direzza zii, OE staggered 245/40-275/35/18 direzza z1 star spec, staggered style 66 on 235-245/45/17 blizzak ws70, pentosin mtf2, liqui moly 10w60, mobil1 0w40, ground control camber plates, bilstein b8 monotube dampers, intrax of holland racing springs*tbi
**********************
dynojet 319ft-lb, 341whp, 13.51@104.3 englishtown, lime rock 1'09.9", lightning 1'23.420", thunderbolt 1'44.2", watkins glen *tbd, secret menu maf test, 138+ 2nd gear
**********************
Favorite Quote: "Understeer is when you hit the wall with your front end. Oversteer is when you hit the wall with your back end. Horsepower is how hard you hit the wall. Torque is how far you move the wall when you hit it. "

Sailor24

It is possible that you need to be monitoring other things other than the trims. One thing I would watch always is Vanos control and valve overlap. The easiest way for the car to adjust to your changes would be to adjust the ratio between air and exhaust gas. You add more fuel DME increases the air in relation to the exhaust gas.
One other thing you might want to watch is the g/s of air going into the car or calc load.
All that said you are not really changing much by pulling that hose, you have very little load on the engine. I would think the engine has more than one approach to it's many goals, power and emission related.

cab

Hmmm, I don't see it right now, but here in an INPA screen shot from my own car last year:



The Lambda integrator rows are the classic STFTs that bounce all around real-time. The Adaptation Value Additive rows (which that PDF says are used for "low speed and idle conditions" similar to what you tested in I believe) do NOT show up in Autotap (which I also have). Finally, the Adaptation Value Multiplicative rows show up as the LTFTs in Autotap.

68FB

Thanks, cab. I'll repeat the test and see if the Adaption Value Additive is the hidden fuel trim that is correcting the STFT's.

Sailor24

Don't know if this is going to help or make it worse, but here goes.
This is from one of the German forums I read often, yes it is a google translate so best of luck.


"Adaptive value is additive and multiplicative adaptation of the value of the fuel supply diagnosis. The additive value compensates for air leakage error, ie it is a leak in the intake system more erkannteb injected when the value is positive. The value multi8plikative like for example the manufacturing tolerances and more signs of aging from the injectors. If your additive value at the upper stop now, it means that your engine will probably pull air leakage. Therefore, no lambda = 1 mixture is used, because the adaptive value of the fuel supply is already at the stop and therefore the oxygen sensor put as much fat requirement, as it wants, it's not just fat ...
So, my advice:
Go to start pilot in the engine compartment, leave the car running at idle and sometimes start pilot spray onto the entire intake path of the air mass meter. If the engine will turn up, it pulls air leakage somewhere.

And precisely this which leads then to the wrong ignition. This is adjusted according to late to prevent the revving of the engine at idle. Subsequent ignition means in this case a worse combustion efficiency and thus holding the idle speed to be achieved despite high filling.

So my advice: false air!

Greeting
Peace"

Many other times I have seen the Integrator referenced like it is short term trim but it has a different range, not totally different. -35 to +25 %.
So best guess is add and multi are both used to make up long term.

Edit A little more reading and searching netted me this post and the thread goes on with out anything memorable to agree with the post.

""lambdaintegrator" appears to actually be "short term fuel trim". i.e. the OP is seeing +28% SHRTFT1 and SHRTFT2 in his pictures (post #37), like I was until I fixed yet another leak.

I have vac leak problems too on my M54 and my short term fuel trims read via a standard OBD2 reader have improved from +28% at idle, which appears to be MAX, and often triggers "system too lean" DTCs, to +10 - 14% at idle. I had the same broken CCV -> dipstick hose as the OP here. I also had broken MAF -> Throttle body rubber.
Those same numbers show up as "lambdaintegrator" in INPA, with 0 shown for all the additive and multiplicative (I had done lots of resetting..).

I think the "explanation of BMW fuel trims" PDF document floating around that says that additive=short term fuel trims, and multiplicative=long term fuel trims (or the other way round, whatever) is bogus. The document (which I think might be linked in this thread.. it's certainly being referenced on a few forums), says that "short term (additive) fuel trims affect mid/upper rev range", "long term (multiplicitive) affect idle", or something along those lines. I think that document is wrong. What it should say is just "additive" and "multiplicitive". It shouldn't use the words "short term" and "long term" at all because these are a different thing to "additive" and "multiplicitive", and in INPA speak, short-term = lambdaintegrator. The language fits too.. ("lambda integrator.. lambda adjuster/tweaker/tie-in"), which is what short term fuel trims are - they are live mixture adjusting to match o2 sensor data.

So, I think, lambdaintegrator=short term fuel trims
additive/multiplicitive are both sub-parts of the long term (learned) fuel trims, which are reset when you reset adaptation values."