PDA

View Full Version : Lean spikes commanded vs actual AFR



billyjo
January 13th, 2007, 04:19 AM
First let me start by saying I'm using EFI Live with COS3. I've been working on setting up my VE table but I noticed some cells that are way out of wack with the rest.

I then started looking VERY closely at my logs and noticed that everytime I step on the gas and go into a KPA above 95 I get these lean spikes when my commanded AFR goes to 11.7. Is this lean tip in that people talk about? How can I make it go away or is it just normal and I shouldn't be worried about it?

Thanks,

Bill

Tordne
January 13th, 2007, 06:01 AM
That spike is actually enriching, going richer. AFR's < 14.63:1 are rich, and AFR's > 14.63:1 are lean.

When you step on the gas one of many things (PE, B3647 in COS etc.). It richens the fuel mixture to protect against knock.

billyjo
January 13th, 2007, 07:47 AM
I understand that, the spike I'm talking about is where it rises to 16+ before it drops down to the commanded 11.7.

Thanks,

Bill

Tordne
January 13th, 2007, 08:08 AM
Doh! I was looking at the wrong PID :nixweiss:

joecar
January 13th, 2007, 08:15 AM
Billyjo,

Is your MAF disabled/disconnected (should throw DTC P0103)...?

After doing AutoVE, what are you BEN's...?
After filtering/averaging, they shouldbe 1.00 +/- 0.02 for all cells hit.

If that is the case, then look at the OLFA table and PE vs RPM table and see what AFR's are being commanded (the richer of these 2 tables wins);

In the scan tool position the cursor on the lean spike, and you should see the cell(s) highlighted in the OLFA/PE/VE tables...

Sorry I don't have the ids for those tables, I'm on a different PC right now...

Joe

dc_justin
January 13th, 2007, 08:22 AM
Billyjo,

This is unfortunately a common occurance for us roots-blower guys. These things seem to build boost faster than the PCM can adjust for it.

One thing you can do to help get it to drop your commanded AFR faster (and reduce lean spike duration) is to make sure your PE settings are similar to the the following:

B3608 = 0
B3643 = 1.0
B3644 = 0
B3645 = 0.1 <-- Delays the transition out of PE. Will prevent knock from occuring when you quickly pop off the throttle while in boost.
B3609 = 0 for all values

joecar
January 13th, 2007, 08:28 AM
...Radixed LQ4 with 3.2 pulley...ah... that failed to register in my mind...:doh::hihi:

Whippled 496
January 13th, 2007, 01:19 PM
Billyjo,

This is unfortunately a common occurance for us roots-blower guys. These things seem to build boost faster than the PCM can adjust for it.

One thing you can do to help get it to drop your commanded AFR faster (and reduce lean spike duration) is to make sure your PE settings are similar to the the following:

B3608 = 0
B3643 = 1.0
B3644 = 0
B3645 = 0.1 <-- Delays the transition out of PE. Will prevent knock from occuring when you quickly pop off the throttle while in boost.
B3609 = 0 for all values

Hey Justin, why not have something like 0.1 or 0.01 in cell B3643? You want the transition to be as fast as possible right? Can you explain what you mean, i probably just dont understand it well enough.

dc_justin
January 13th, 2007, 01:28 PM
Hey Justin, why not have something like 0.1 or 0.01 in cell B3643? You want the transition to be as fast as possible right? Can you explain what you mean, i probably just dont understand it well enough.

B3643 is the size of the step in the transition of the commanded fuel.

If your current commanded EQ ratio is 1.0 and your target is 1.25, if you have a step size of 0.01, it will take 25 steps to go from 1.0 to 1.25. According to the description, there are approx 20 steps a second, so it would take more than a full second to reach target.

With a step size of 1.0, it will hit target with the first step.

Whippled 496
January 13th, 2007, 03:58 PM
B3643 is the size of the step in the transition of the commanded fuel.

If your current commanded EQ ratio is 1.0 and your target is 1.25, if you have a step size of 0.01, it will take 25 steps to go from 1.0 to 1.25. According to the description, there are approx 20 steps a second, so it would take more than a full second to reach target.

With a step size of 1.0, it will hit target with the first step.

Ahhh, got it. That description for table B3643 takes a little reading into to completely understand it. Thanks man, by the way....you better to get to work on redirecting that domain name, still says its for sale...LOL

Whippled 496
January 13th, 2007, 04:13 PM
Also, I use AFR not EQ. So basically the higher the number, the quicker the return right? If my commanded was 12.0 and i had say 2.7 in my cell, that would put my back to 14.70 very fast. Is an extremely quick transition like that the best method or should be a little longer?

dc_justin
January 13th, 2007, 04:18 PM
Also, I use AFR not EQ. So basically the higher the number, the quicker the return right? If my commanded was 12.0 and i had say 2.7 in my cell, that would put my back to 14.70 very fast. Is an extremely quick transition like that the best method or should be a little longer?

AFR is derived from EQ ratios in the software. A step size of 1.0 is plenty sufficient and will allow you to richen up to as much as 7.35:1 in 1/20th of a second.

billyjo
January 14th, 2007, 05:38 AM
I made the changes that Justin reccomended and it looks like it made a big difference when I slowly transition to boost but when I do a quick stab of the pedal I'm still seeing the spike.

I'm also wondering if the fact that I have a manual transmission vs an automatic makes any difference or is compounding my problem. My thinking on this is when I'm down in the lower RPM ranges I just create more load with the manual where as an auto would shift to a lower gear.

Is the answer to modify my VE table in these areas so that there is a huge spike? Or would commanding a richer AFR ealier help to alleviate some of the problem?

Thanks,

Bill

hquick
January 23rd, 2008, 11:50 PM
I know this is like a year old guys but I just dug it up and have a couple of questions. Firstly...I just changed my PE fig's to the same as Justin has posted in this thread. Looking forward to trying them tomorrow.
I was/am having the EXACT same issue as Jaybo.
OK...the question:


A step size of 1.0 is plenty sufficient and will allow you to richen up to as much as 7.35:1 in 1/20th of a second.
Where does the 7.35:1 come from? I sometimes get AFR's of 10 +/- and I don't command it. Where does it come from?

Thanks
__________________

dc_justin
January 24th, 2008, 03:56 AM
I know this is like a year old guys but I just dug it up and have a couple of questions. Firstly...I just changed my PE fig's to the same as Justin has posted in this thread. Looking forward to trying them tomorrow.
I was/am having the EXACT same issue as Jaybo.
OK...the question:


Where does the 7.35:1 come from? I sometimes get AFR's of 10 +/- and I don't command it. Where does it come from?

Thanks
__________________

The 7.35 was just a reference of how quickly PE mode could richen up to that point. In stock trim, it can take a few moments for it to transition from stoich to your PE AFR.

The AFRs in the 10s that you're seeing is occurring when you let off of the throttle when you're in boost, right?

hquick
January 24th, 2008, 06:35 AM
Thanks Justin,
it actually seems to occurr when I get on the gas.
I'll pay closer attention this morning in a few moments on the drive to work.
I'll post my log in half an hour or so.

hquick
January 24th, 2008, 08:57 AM
Here's my current tune and log if you guys don't mind having a look?
Thanks

dc_justin
January 24th, 2008, 02:38 PM
Haven't looked at the tune, but looking at the log, everything looks typical. Best I can tell, it's due to some sort of fuel evaporation/boiling off of the intake runners.

The part that scares me more is the 18:1 AFR spikes as you (and all other 2 bar positive displacement guys) jump into boost.

On a side note, I may be in your neck of the woods for a day in March. Wife has to spend 5 or 6 days in Canberra and I'm going to try and tag along. :cool:

hquick
January 24th, 2008, 02:43 PM
Thanks Justin.
Hey...if you make it over here and can get down to Melbourne...we could hook up for a beer.

hquick
January 24th, 2008, 04:58 PM
Here's a log I just took on the way home.
Still getting the lean spikes...and you're right Justin...the AFR's of 10+/- are mostly as I let off the pedal.
1's mark some slight bucking.