How To Fine Tune Acceleration Enrichment Values?

[TGP-37]

1996 Grand Prix SE
PCM: 97 Lumina for HPT support
Turbo: TO4E .50/.63 @ 9 psig
Spool Up: 2800-3800 rpm's
Throttle Body: 52mm (ported to 58mm equivilant)
*IAC Steps remain at 35
Injectors: 36lb SC 3800v6
Gas: Pump 93 with up to 10% Eth (true stoich is @14.25:1)
Ported Heads, Pocket Smoothing, Gasket Matched, Valves Deshrouded, Exhaust Smoothed and Polished.

How does one go about fine tuning the AE parameters?

What are the basics, as in...?

What table do I generally make the AE richer or leaner?


What do you guys do to get the AE in tune?





Should I Increase the AE vrs. TPS Max Delta table by multiplying 1.01 and decrease by 0.99 as often as needed to adjust?

Is there any trick like, getting a dead spot when snapping the throttle, increasing some setting until the dead zone disappears?

Could I command a steady AFR in PE mode at idle (stoich) and get a better reading of AFR snapping rich or lean?

I understand a proper AE can make a big difference in take off. Seems like a very important setting for 1/4 mile times. Any delay at the snap of the throttle is just a big waste of time and we know how that can impact 1/4 mile times.

So if anyone wants to get technical, please do. I can handle some serious math or logging. I considered logging only the NBO2 to gain greater resolution but I haven't read anyone doing that so I wonder if it a feasible application of the scanner or is it pointless?

Also, I am pushing 9-10 psi of boost. Boost starts at 2800 rpm and peaks at 9psi at 3800 rpm. So far I haven't had any lack of airflow at 5500 rpm's so the turbo can handle the airflow just right. T04E .50/.63 trims works great for the 3100v6 small port.

[99ssei]

I would like to know this as well. If we want to increase AE to accommodate for burst knock, which tables do we adjust?

[TGP-37]

Anyone? I need educated in AE operation. If I knew how to do the rough math myself I could go from there. I get the concept of coefficients and how it lags.

I don't want a fish, I want to be taught how to fish...lol...get it?

Oh well, in time I assume someone will mosey along and post up.

Maybe I should refine my questions...

How long is 1 refcount? Is 1 refcount 12.5ms?

AE is confusing.

[99ssei]

AE is confusing.

Agreed. I too want help on this.

[leti]

I am sure that this subject on how to go about fine tuning acceleration enrichments parameters will be something that a lot of people will want to know about and we already have three people (including me) who want to know about it, so hope we’ll get some info on the subject!!

[TGP-37]

I developed a system to tune my AE and DE values. I now have my AE and DE spot on target. I will share my method but it does take time and many reflashes.

You will need Microsoft Excel. You will need to log Commanded AFR, Wideband AFR, TPS%, ECT and Vehicle Speed.

TPS = Throttle Position in %
F1cmb = AFR commanded before tip in
F2cmd = AFR commanded after tip in
F1wb = Wideband AFR readings before tip in
F2wb = Wideband AFR readings after tip in (usually the spike)
AE = AE vs. TPS MAX Delta (tps specific)
Ratio = Correction Muliplier to AE vs. TPS MAX Delta

( (1/F1cmd) - (1/F2cmd) ) / ( (1/F1wb) - (1/F2wb) ) = Ratio

Ratio * AE = New AE Value

Get a good log of you engine before changing values. While you are driving depress the throttle to the target TPS% quickly and hold it there. I.E. go from 0%TPS to 20%TPS without delay and hold 20%TPS for at least 2-3 seconds. Long enough to log the entire AE Enrich process and to stabilize the AFR's to log the DE part. Now release the throttle from 20%TPS to 0%TPS with out delay. Allow the car to coast at 0%TPS for a few seconds or until the AFR's stabilize at 0%TPS. Do this for all throttle ranges, 10%, 25%, 75%, 95%, 100%, ect, ect.

The TPS log should appear box like with no gradual throttle changes.

Go through the log and hand pick each instance the TPS goes from 0%TPS to X%TPS and figure the amount of fuel change from both Commanded AFR and Wideband AFR. Lets say the Commanded and WB were at 14.7 AFR with 0%TPS. When the throttle is depressed to 50% the commanded AFR goes to 12.5 but the Wideband goes to 10.5 (spikes to 10.5 briefly and returns to 12.5) We have too much AE in this example.

So now we must translate the numbers to get our data we need. divide 1 by the afr to get a fuel value. 14.7 = 0.068, 12.5 = 0.08 and 10.5 = 0.095.

The AE over shot the target AFR of 12.5 from throttle tip in. 0.08 - 0.068 = 0.012 and 0.095 - 0.068 = 0.027. The AE value for the specific TPS% we are measuring is 0.561. Since we need what was commanded (0.08-0.068=0.012) but AE value of 0.561 commanded (0.095-0.068=0.027).

0.012/0.027=0.444 The 0.444 is the ratio offset. So multiply 0.561*0.444= New AE Value.

Once you get the hang of it, it is easy to do. In a few hours time you can have your AE and DE lock in spot on AFR target.

One thing I did figure out was this. The AE or DE value seemed fixed as in it will command the same AFR regardless of the fueling being off slightly during normal operation. Like if you tipped the throttle in and activate AE during DE. It isn't a % injection on top of current IPW. As far as I can tell, I may be wrong.

Just repeat the process for each TPS% position.

1) Log the car
2) hand pick each TPS rating, 0, 5, 10, 15, 20, 25, ......, 100 and get the ratio offset.
3) multiply the AE or DE Multiplier vs TPS% by the ratio acquired
4) Repeat as often as necessary.

Another method is to adjust the TPS% value by 10% at a time (multiply by 0.9) where needed. Again, hand pick each TPS% value and determine if it is too rich or lean and adjust accordingly. Relog the new settings and again adjust by 10% either way needed until there is no rich or lean spots.

HINT: Always make changes from data gathered at operating temp only. Warm up temps will required different values and thus is bad data at first.

HINT: Once the operating temp region is tuned, log the car warming up while driving and adjust the ECT vs GEAR Tables. Simply adjusy either way by 10% until there are no rich/lean spikes.

HINT: I tuned my AE to hit the AFR 12.2 across the TPS range as 12.2 produces great power. That makes my throttle snap very well and the rpms accel much faster.

My DE is just right too. I can tell you it is hard to notice the effects from changing the data. But when you get the AE and DE values tuned just right you can Definately Feel the difference.

Having the AE and DE tuned just right resolved many minor annoyances. Like a mild idle RPM hunt after boost. Nailing the idle AFR with DE stopped that. Now mt RPM's drop to idle smooth and sticks at idle well. My idle RPM varies +/- 20-35 rpms max, real smooth.

HINT: Excel is your best friend.

[Phantom]

Great work I need to get a WB O2 so I can start doing some real tuning.

[Phantom]

How long is 1 refcount? Is 1 refcount 12.5ms?

I found the answer here

Based on what I figured out, not quite.. Reference pulse is called reference as it is dependent on what the PCM is referencing. In the case of our cars, it is referencing the crank position sensor (different in 60* engines though). Which pulses 3 times per engine revolution.. Care to figure out the math to show how many seconds 50 reference pulses are at an engine speed of 1600 rpm?


1st, if your engine is running 1600 rpm, that is (1600/60) = 26.66666666666667 revolutions per second..

2nd, 50 reference pulses will occur in (50/3) = 16.66666666666667 engine revolutions

3rd, that will be (16.66666666666667 revs / 26.66666666666667 revs per second) = 0.6249999999999999 seconds

or... 624.9999999999999 milliseconds

Basically, each reference pulse will occur every (624.9999999999999/50) 12.5 ms.

The example above, as it turns out, indicates a reference pulse occurs every 12.5 ms (funny that, I picked that as an example without even trying ). However, you will find that as RPM increases, the duration of time between reference pulses decreases.. At 2000 RPM, a reference pulse will happen every 10 milliseconds..

BTW 5000 RPM, a reference pulse is occurring every 4 milliseconds.

So, 50 references pulese at 3 RPM values:
1600 RPM = 625 milliseconds
2000 RPM = 500 milliseconds
5000 RPM = 200 milliseconds

http://www.clubgp.com/newforum/tm.as...mode=1&smode=1

[TGP-37]

I know its been a while since I posted. Got caught up in a manual trans swap and lost any reasons to be tuning for the time. Took a whole freaking 8 months to get it done!

But it was worth it. I was able to ditch the 4T60e Auto and put in a reman Getrag-282 with a race class Engineered Performance Limited Slip Differential!

WOoo Hoo!! Add in a 3100v6 w/ 8-12 psi boost max and I got a mean machine. It could probably handle 15-20 psi but I keep it down for longetivity and reliability.

Anyways, Phantom that is great info you posted! I am now reconfiguring my PCM for the changes I made. During the swap I rebuilt the turbo exhaust cross-over manifold. The new design is far more streamlined and uses Venturi forces to smooth the wastegate flow. So I got some tuning to get done all over again.

Armed with that info is great news for me. I lost all my excel documents in a PC Wipe. When I rebuild them as needed I'll post up the important one(s).

I found a way to partially automate and calculate the Accel Enrichment. It takes two logs and the input of changes made to the table and tells you how much more or less of the last change you need to make to reduce fueling spikes, rich or lean. No more fueling spikes means AE is almost finished. Kinda like dividing a number in half. it goes on forever but after a few times it ends up being less then 1. Close enough where we can't tell the difference.

I'm glad I posted this saving the formulas above. Time for me to get to work!

[TGP-37]

I found the answer here

My 60v6 has 2 CPS's. A 3x and a 24x

I'm inclinded to believe it uses the 3x but what about the 3x dividing the 24x and THAT being the reference pulse? 8x per rev isn't to unreasonable either.


To the thread linked by Phantom. A lot of people responded as to why even use AE in the first place. Well, I say this.......

Don't knock it until you driven a car with a well tuned AE. The crispness of the throttle will make you smile. The revs are MUCH faster and the PCM/ECM is happier. I know an engine will appear to run 'fine' w/o it. But it will not run at it's best.

When the engine goes from idle (vacuum) to any throttle above 10% (vacuum loss) and the air/fuel mix currently in the manifold will literally rain gas onto the manifold walls. As vacuum decreases the mixtures ability to hold fuel vapor drops. AE is the millisecond compensation for the fuel lost in vacuum transition. The fuel will again re-evaporate from heat at the intake runner/pocket walls. Adjusting the decay profile will be able to account for that re-evaporation.

This is why the deacy profile spikes and then has a low spot half way through. the idea is to replace the lost fuel from precipitating out of the mix. But only enough to replace and hit the desired AFR for throttle up AND limit that replacement of fuel when the precipitated fuel re-evaporates back into the mix.

Complicated at first but it makes sense once you relax, close your eyes and visualize it.



Typically the AE is used at lower RPMs, just above idle RPM. The AE length can be anywhere from 2.13 seconds at 600 RPMS to 0.53 seconds at 2400 RPMs. When logging at full PID capacity, that is roughly 3-11 frames in length.

Ever hit the pedal and the engine bogs for a brief second? Bad AE

Same thing with DECEL. Since a vacuum is rapidly introduced, and often greater then idle, a LOT of fuel gets sucked up from the manifold walls. Flash evap and briefly runs rich.

I'm not too sure but the popping sounds at throttle back disappeared when my DE was tuned just right.

I would go as far as saying, AE/DE are far more important then what people give credit for.

[TurboGP33]

The 96 3100v6 has 2 crank sensors, but if you unplug the 24x sensor the car still runs. So I believe the referenced pulse is the primary 7x sensor. However, that is more like 6 pulses because one of the 6 has a second pulse only a few degrees out as cyl 1 reference.

So now I wonder, is 1 revolution worth 7 pulses...or 6 pulses (minus the extra trigger notch) or 1 pulse if it uses the double notch as the reference to the PCM.

I bet analyzing a log will reveal 1 of the 3 options if the decel is increased to become apparent in the wideband log. Nailing the exact length of the AE/DE operation will make a huge difference in my ability to fine tune them.



Until then, my best advise is to hand craft the modifiers until they stop spiking lean/rich.

[RobertISaar]

reference pulse = the 3X signal created by the ICM that it creates based off of the 7X crank sensor's signal. you'll get 3 pulses per revolution. at 6000RPM, it takes 10mS for the entire engine to revolve once, in that period 3 pulses will be created by the ICM that the PCM will read to interpret both engine position and speed. each period is 3.33333333333333333333333 mS long.