MercenaryX2

mrfred, you got a tester here for ya. I got a couple evo's at my disposal. One is definitely switching to SD. The other one, not sure yet. Gotta talk to the owner to figure out what he wants to do.

RJSP

If you start to matching your fuel table to your wideband AFR via MAF smoothing, it´s important that you resize your MAF Scaling and Smoothing tables so that you can adjust "MAF VE" (in the smoothing table) to all your Hz readings.

I could´t upload my xls sheet so here´s a print out of what a MAF table resized to 2600 Hz looks like, notice in the graph that I´ve extended the "natural" path of original the g/s table to find the new "educated guess" points to the higher Hz values. Now you can adjust fuel very precisely up to what ever Hz is read, on one car I went up to 3300 Hz and resized the MAF tables accordingly. Notice that I´ve taken out Hz points that are caught in between a "line" made by the previous and next Hz point, so the is virtualy no effect to ride quality, just tradring lower Hz adjustability for full range adjustabily.

This makes mapping with the current 3D VE table simpler, as you only use it as to calibrate your SD load to what it sued to be with the MAF, of course once the MAF is out there are further adjustments to be made, but we all know that. The are a couple scalings (the sacling for the Hz in the rom definition) that work well, one is the one I use with round numbers and another that gives slightly off round numbers. There are a couple older threads that explaing the process of changing these scalings so that the MAF tables can be resized.

Cheers,

Ricardo

90zcrex

iTrader: (11)

I'm guessing another reason for this patch is possibly fixing the stutter issue? I noticed your other thread asking about it.

burgers22

Agreed, the MAF scaling is a very nice way of adjusting the fuelling on a more global level in either MAF or SD tune. My MAF scaling looks pretty similar to yours.

03whitegsr

iTrader: (8)

Why would you use MAF comp tables to "tune" fuel? Those MAFComp tables are to deal with the non-linearity of the MAF and nothing more. You've eliminated the MAF and as such, those MAF tables should also disappear from the calculations.

Load is a representation of the amount of air the cylinder has ingested in the form of a temperature compensated pressure. 100% VE, 250kPa, and 298K would mean a load of 250. Increase the temperature or decrease VE and the load should drop because less air is making it into the cylinder.

Everything in the ECU is based on this idea for load.

RJSP

With a true SD ECU, like the Haltech 1000 I use in my STi, Load is straight MAP vacuum/pressure in Kpa or whatever scale available, and the 3D VE table only afects fuel and is used to calibrate the target AFR table. Timing and target AFR values used by the ECU are not affected by the VE table, they follow their own tables and are dependant on RPM and MAP.

What we use in the Evo ECU is an aproximation of this, but as stated by Mrfred, we can not get rid of the calculations made using airflow Hz the ECU uses to determine Load and fuel; thats why the MAF smoothing/compensation table is realy a VE compensation tha affects only fuel for the given Hz reading, which is exactly what we need to adjust fuel without disturbing Load and timing.

We also use a linear relation between MAP and Load, but the Evo ECU doesn´t work with MAP, so the VE table is realy a calibration to get the correct Load that the ECU would expect if it still had the MAF in place; remember Load determines the timing and AFR target in the timing and fuel maps, see how it can get tricky if you only rely on VE changes to affect fuel? You´ll be changing load and timing as well.

Many ways to skin a cat, I just prefer taking advantage of what is available, and the MAF scaling and compensation tables work great if they are scaled to the whole range of Hz readings that are being seen. If your car is seing say 2300 Hz of airflow but your MAF tables are saclet to only 1600 Hz then it will not make sense tarying to adjust fuel via MAF compensation and you´ll end up with funny looking fuel and VE tables. It can work for sure, but it´s not as accurate and flexible as properly using the MAF tables.

My 2 cents,

Ricardo

wreckleford

iTrader: (2)

I agree with you. I actually think it makes tuning SD on the stock ECU easier to do too.

03whitegsr

iTrader: (8)

Why would you use the MAF tables to adjust fuel when you have a target AFR table?

This makes no sense to me. You are adjusting a table that affects an airflow level which means you are linking a load and RPM band together and trying to tune them at the same time. You already have independent control here, why make them linked together like that?

Most ECUs use load to control AFR and timing, along with most other functions. It is the lower end aftermarket ECUs that use MAP vs. RPM look up tables to determine injection and timing. The high end units use an engine model base to how they operate, just like I am suggesting above. OEMs all use it too and they use it because it is a superior way of doing things because fuel and ignition are load dependent more so then strictly MAP dependent.

211Ratsbud

iTrader: (1)

I never understood more load more fuel. Is part of the formula for calculation that ipw is based off % load?

Why does the stock ecu even care what load it sees for quantifying fuel?

Closed loop is the only portion that cares what ve is?

I did actual = target in my sd fuel table and my ve table resembled my maf afr target (shape and trends wise)

Granted 20 psi was 300 load vs maf saw 250-260-270.

This new way of mrfreds seems way better though as far as making the process less complex tuning.

Our ecu does not calculate load from map it correlates map to load then corrects, is that how that's interpreted ?

03whitegsr

iTrader: (8)

You directly multiply a bunch of stuff by the MASTERLOAD variable to calculate the base fuel pulse width.

Load is used for more that just look ups, it is part of the fuel flow calculations directly.

This is why I say it should have a physical unit associated with it and why it should not be randomly manipulated to reach load targets that make sense to you. Increase load 10% and yes it puts you in a different section of he fuel maps, but it also adds 10% fuel to the motor.

211Ratsbud

iTrader: (1)

Ah from the advanced fuel thread.

Good info.

How could we meet those stipulations in this stock ecu format?

mrfred

iTrader: (50)

I'm not seeing a difference between using a VE table or a load conversion table for the load determination. Determining VE is no different than determining the correct load directly - either one is a serious challenge to do accurately without some measure of airflow rate. In my mind, a proper SD conversion process would first involve detailed mapping of MAF master load across a wide range of RPM and boost levels. Once that is known, then it seems like a simple manner directly fill in a load conversion table rather than fill in VE values.

wreckleford

iTrader: (2)

This is why RJSP was saying to manipulate the MAF tables to increase or decrease fuel. Adjusting the MAF table only changes the IFPHz value. Doing it this way you do not affect the load value as you would if you adjusted the VE.

03whitegsr

iTrader: (8)

Mrfred, I mostly agree, however having it VE based does a couple things for you.

1. Visually, the map is easier to understand, IMO. For example, lets say your engine has a VE of 95% at 250kPa and 300kPa. This would correlate to loads of 225 and 270, respectively. Mathematically, it is easy to determine that the engine is filling the cylinder with equal efficiency but you would see that easily with a quick glance at the VE map.

2. It's easy to safely map higher load levels before you get there. For the most part, most setups really don't gain VE once they pass about 10-15 psi. You can pretty much take the 15psi load column and copy and paste it to the higher load areas and you'll very likely be safe on fuel while you are tuning it in. This also protects you in an overboost situation as it's pretty easy to add a small amount of VE above your peak boost level and know it will be rich enough that you won't go lean but also lean enough it won't load up heavily. I've seen blown motors from misfires along with detonation and by using a VE map, you know you are within a couple % on fuel when stepping up on boost to where you need to be.

Yes, you could do this all mathematically with load in the map, it's just quicker, easier, and more intuitive to use VE.


I agree though, when it comes to actual tuning, you say tomato I say tomato but it's still all fairly arbitrary numbers that have to get filled in through repeated testing. It is just my opinion that doing it based on VE is better.

Again though, WHY?
You have the fuel table to change AFR separate from the load calculation.

If you change the MAF table, you make a fuel change through a range of RPM and load sites. You increase say the 1600HZ setting by 10% and it might change 5000RPM@250kpa but it also will change 4000RPM@280kPa and 6000RPM@220kPa...as a rough guess. This is a particualrly bad idea when maxing out a turbo as a maxed out turbo will drop load to maintain an airflow (Hz) level. Now a change at that MAF frequency changes then entire fuel curve.

Unless you only tune for one boost level and don't give a crap what is happening when you change boost level or IAT that puts you in a different load site. Sure it may be "easier" for one boost level and to create a tune that is inconsistent...but you are doing it wrong if you want a good tune that is right regardless of IAT, load, RPM, etc.

211Ratsbud

iTrader: (1)

The whole idea from ve table is to set your fuel map @ desired afr and tune ve around that target isn't it? That's how I originally did 3dsd. But my load was 40 or so higher that way.

I'm gonna switch to stand alone soon and I hope that broadens my range of knowledge here.

Mrfreds op idea seems logical to me. Load vs kpa in a 3d table is way less complex than map vs load + 3d table

And quite a bit more obvious too as to what you're tuning for