Jason McWilliams

According to the manufacturer, the operating range of the NGK AFX is an AFR of 9.0 to 16.0 (0 to 5V linear).

I picked up both of Greg Banish's books from Amazon.com.  I am about half way through reading Engine Management: Advanced Tuning and it has been great.  So far, we have covered a lot of the "why".  How engines work.  What all of the relevant terms mean.  The fact that, at its core, the PCM simply tries to guesstimate what mass of air is entering the engine so that it can add the appropriate amount of fuel to hit a commanded AFR.  So much "mystery" vanishes when you look at the core function of a product....inputs and outputs.

I am also spending a lot of time reading:
HPTuners forum
EFI Live forum (I do not own the product, but the theory is identical, so its good reading)

I have also read through Marcin's blog which covers some great topics from a mathematical modelling perspective trying to figure out exactly how something works.

I finally finished up the installation on Monday.  I have everything cleverly hidden...unless you open up the glove box.  I have the AFX control "box" mounted in there so that i can see it when the glove box is open.  I have enough loose wire in there for it that i can move it up onto the dash right in front of my face if I feel the need...although, to be honest, I dont think I ever will.  I may dive back in this weekend and pull that loose wire back so that it isn't filling the glove box up and it has a cleaner appearance.

I managed to run the ODBII cable for the HPTuners box through the dash of the car, which was a real pain in the ass.  I de-coupled the ODBII port on the car and have the cable zip tied such that it and the connector to the car are mounted flush with the bottom of the dash...no more smacking my legs with it.  Plus, you can unplug it and still have access to the connector if/when the car needs to spend some quality time at the shop.  I haven't done anything about semi-permanent mounting for the box yet.  I am concerned about leaving it plugged into the car 24x7 since I think that the ODBII port may always be powered (flat battery).  I'll have to test that out and rig a switch to power the HPTuners box on and off if my assumption is right.

So Tuesday morning I get all excited and try to follow SD VE and MAF Tuning Instructions rev3.pdf.  I got an Open Loop Speed Density (OLSD) tune set up based on what came stock on the car.  I multiplied my VE (Volumetric efficiency...basically a lookup between RPM and manifold pressure to figure out the rough mass of air going through the engine) tables by about 15%, per the instructions since the Vararam intake should increase airflow.  I threw it on the car, fired it up and it ran!  Granted, I have a "Service Engine Soon" message at all times now, but going back to stock should cure that (once I have proper VE and MAF tables to copy into it).

Over lunch and after work I was all excited, logging data after the car warmed up.  I needed to lower the VE table a lot, which I did.  Went back out Wednesday morning and it was raining HARD...like 40mph max on the highway hard.  According to that log, the VE values that I had almost dialed in were off quite a bit, so I ignored the log (due to the weather) and went out over lunch.

I was driving down Shawnee Mission Parkway trying to get a good fill of data for the VE table and I took the RPM up higher than I had before (ideally, I would like to tune VE to redline before tuning MAF, even though I dont need VE to be right over about 4400rpm since MAF should fully take over by that point).  I was about to turn around and I pushed it a little high..then bam.  Reduced Engine Power mode.  it was running SO rich in that mode that the wideband was pegged at 9AFR (which is the minimum value that it will register).  I pulled it over and shut it down.  The log showed no knock or anything, so I fired it back up and it was good as gold again.

The conclusion that I made from this is that I really needed to start over.  My initial assumption to just bump the whole table 15% was WAY off.  Enough so that at higher cylinder pressures (high load/rpm), I wasn't comfortable getting into non-adjusted parts of the table for fear of breaking something.  I compared the stock VE table with the one that I had been adjusting it the amount of change was a little varied, as it should be, but the average change appeared to be 5%, so I took the stock VE table, multiplied everything by 1.05 and put that into my SD tune.  I am MUCH more comfortable with that.

I threw the new tune on the car and drove it home last night.  The log showed some changes, but it wasn't a huge amount.  Mostly 5% or less for AFR error % (which is what I am using to adjust the VE table).  I reduced the max change in any cell to 10%, which only affected a couple.

I put that on the car this morning and took a log on the way to work.  I hit a few more cells again so those new areas still need a relatively major (comparatively) adjustment, but for the region that I updated last night, the new AFR % error is almost entirely under 2%, which is fantastic.  I am going to flash the PCM again and do a little driving over "lunch" to see how it looks.

One follow up note is that after looking at my min, max, and count values, I feel like I need to do some manual data massaging before blindly throwing a change at the VE table.  There are clearly values that need to be omitted.  The filter in HPTuners does a good job of getting most of it, but some slips through at either the leading or trailing end of transitions.

Last night I worked on getting my NGK AFX Wideband O2 kit installed. I have an appointment later today to have a bung welded into the exhaust for the sensor, so I wanted to have everything else taken care of ahead of time that way I can get it installed and not have to get the car up on ramps to take care of the sensor installation at home.

After some more reading, it looks like the two step idea is a bust...something about the computer triggering the rev limiter every time you hit the clutch, which is bad juju for...you know...shifting...

HPTuners arrived on Friday. I had waited until Friday for my paycheck to place my order for the wideband O2 sensor, but I thought that I could get a start on it before that arrives. I had found a particularly good writeup on how to get a good base tune (ve tables and maf). I downloaded the current, stock, tune from my car and dived right in. I got all of the way through configuring it to run SD (speed density...basically it turns off the MAF sensor and does all of its fueling calculations based on the MAP sensor reading) only to realize that I truly had to have that wideband O2 sensor in order to move forward with the writeup.

So I knew that I needed to buy a tuning app. I knew that I wanted to do it right this time and have a wideband 02 sensor connected to it so that I could really dial everything in properly. The two big players for tuning hardware/software for my car are HPTuners and EFI Live. I would have to get the "pro" version of either one in order to use a wideband O2, so naturally the price point jumped. EFI Live is significantly more expensive than HPTuners. I wanted to go with EFI Live, however due to a few features that I liked about it, but I just couldn't justify the cost. Also, it feels like there is a larger user base and more forum support for HPTuners, so I placed my order.