My90Iroc

iTrader: (26)

How exactly does the computer determine the actual incoming air (g/sec) between the maf and the dynamic airflow (VE right?)?

I exported my log data to excel so I can break it down easier. At a given maf frequency recorded at different rpms is showing different g/sec readings. Here's some of my data:
rpm____ maf hz__ maf g/sec__ dyn. air g/sec____ map____ a/f error
5246___ 9088____ 250.59____ 244.02_________ 88____ -2.69 (rich)
4660___ 9097____ 251.47____ 236.67_________ 95____ -6.25 (rich)
4525___ 9043____ 228.66____ 228.77_________ 97____ 0.048 (lean)

2652___ 5950____ 58.73_____ 59.53__________ 65____ 7.99 (lean)
2568___ 5950____ 60.03_____ 59.57__________ 68____ 2.12 (lean)
2497___ 5952____ 58.84_____ 58.74__________ 70____ 0.82 (lean)
2431___ 5952____ 60.55_____ 60.45__________ 74____ 3.49 (lean)
1844___ 5948____ 57.09_____ 40.05__________ 67____ -4.37 (rich)

My maf table in the tune shows 223.93 at 9000 hz and 234.74 at 9125 hz and 57.93 at 5875 hz and 61.50 at 6000 hz. You can see the differences in a/f error and maf g/sec reading based on rpm.

In order to fine tune where should I be making changes to get the a/f where I want it to be across the board?

12secSS

iTrader: (30)

Your MAF table should mimic the Dynamic Airflow. As you posted, look at the reported g/sec of the MAF vs Dyn Air, and it correlates to the AFR Error. If the MAF is reporting higher then Dyn Air, you will be rich.

Also, this is for steady state or above 4000 rpms. Where the MAF is used solely. If you are in MAF mode, you need to calibrate the table just a bit more.

RedHardSupra

iTrader: (1)

ideally they should be the same, however that will never happen.

in SD, dynamic airflow is calculated from IPW, IFR, AFRcommanded, MAP, RPM, Cylinder Volume, VE, IAT, ECT, and the temperature blending bias table.

MAF itself isn't the most precise sensor, just look at some charts.
DYNAIR also has issues, stemming from the fact that it depends on a lot of variables, and unless they are all perfect, the resulting airflow is going to be off.

in a regular MAF/SD hybrid mode that we run, you need both to at least somewhat agree, that's why tuning just MAF makes no sense (unless you completely nuke the MAF/SD blending, just ask sspdemon). if you want to permanently run SD, then you have a lot of tables to adjust, which is difficult to get right.

the way i tune MAF is in SD kinda wacky, but here's how it's done: first i tune VE completely and as well as i can get it. then i go for multiple rides still in SD, but also logging MAFfrequency. it usually takes few rides to really get a lot of data throughout the full range of MAFfreq. i export all the logs to Excel, and chart MAFfreq vs DYNAIR. then i make a 3rd order polynomial trendline of that chart, turning on the option for to display the equation. unfortunatelly the default resolution is quite wrong for these numbers, so you gotta right click on the box with the equation, and ask it to display full 30 digits of precision.

then just use that equation to calculate the airflow for every MAFfreq calibration point, and that's your new calibration.

MAKE SURE YOU USE THE SAME UNITS FOR MAF AND DYNAIR

there's also more involved methods if you want to clean up your data from obvious noise, but that's a lot more involved, email me if you want the gory details.

My90Iroc

iTrader: (26)

Thanks for the info.

This is the only part I couldn't figure out how to do.

I originally tuned my ve in SD as well as I could with changing weather conditions. I've completely hogged out my 85mm maf so I really didn't have a starting point. I used to a histogram in HP Tuners plotting MAF airflow vs. Output frequency with Dynamic Airflow as the plot value. It wasn't exact but got me damn close and I could make changes based on a/f. Now I'm just looking to fine tune some areas and use the car to get the best fuel mileage that it can.

Above 4000 rpm my a/f error is very small, naturally since the only variable is the maf tune.

Below 4000 a/f errors are larger and now that it's getting hot and muggy here overall a/f is richer. Below 4000 how much is the car looking at VE and how much at maf?

In my original post look at the 2652 and 1844 rpm logs. They're within 2 hz of each other and 2 kpa. They're also within 2 g/sec according to the maf but off by almost 20 g/sec in dynamic airflow. Now look at the a/f errors and see 2652 rpms is 8% lean while 1844 rpms is over 4% rich. I didn't post it earlier but these points are also within 2% of each other in tps so it's not some weird transition point either.

So how should I make changes in the tune to get the a/f where I want while keeping my g/sec numbers logical?

I'm also considering moving my maf point (high rpm disable) down to like 1800 to see what those logs look like. Truthfully I'd like the maf to function at that point to adjust for weather changes and help fuel mileage. But I'll wait to hear the responses to my above questions first.

RedHardSupra

iTrader: (1)

that depends on tps and map changes if i recall correctly

look at other data, i bet you there were different temps involved

eh, i dont like the MAF/IFR tweaks sspdemon been playing with lately, it's simplifying a system that's not simple by design, so you're bound to miss things.

post the full log and tune, i can probably show you some explanations on what happens, this needs to be explained anyway.

My90Iroc

iTrader: (26)

Just slightly at rpm 2652 ect was 176 and iat was 79 while at 1844 ect was 174 and iat was also 79. Keep in mind this is an 85mm maf with the iat in the housing.

I understand your feeling about this. I'm just looking for the best way to account for changing weather, especially humidity, I'm surrounded by water, lol.