SoCalRedLine

iTrader: (2)

I know JB has done a lot to get some fairly accurate lb/min calcs for mass air, so I was wondering if anyone could help with a formula for finding afr without a wideband.

Something like:
(mass air flow)/((ipw)(inj flow rate)(2*rpm)) to get lambda, then multiply by stoich to get AFR?

Obviously units need to be converted, but I was wondering the viability of somehing like this to get an idea of afr's without a wideband.

Ceddy

MUT32 AFR_Map

voidhawk

iTrader: (20)

That only reads the AFR value in the ECU map - what the OP is asking (I think) is using the airflow calculations together with IDC values and known injector size to calculate the actual AFR the car is running. It's a clever idea ... AFR=airflow/fuelflow. Not sure how to get the fuel flow from the existing EvoScan parameters though.

warp9

iTrader: (35)

For tuning purposes? Its old school but Ive tuned my Evo and other cars with just the NB.

SoCalRedLine

iTrader: (2)

Yea, that's the idea basically. My iPhone typing doesn't always come out clear

the formula would basically convert the airflow into cc/min and then divide by fuel flow in cc/min.
This gives you stoich ratio (lambda) on the front end, wihout a wideband. This is obviously going to have a certain %error, but it might be feasible.

I was reading somewhere that there was a maf hz to Litre conversion somewhere, but I can't seem to get back to it...lol. Might make for less flops on the calculation and speed up sampling.

TouringBubble

iTrader: (2)

While you may be able to get close calculating from MAF Hz, IPW, RPM, VE and such, there are lots of other factors that affect the actual AFR result, such as efficiency of the turbo, quality of fuel, intake temps, spark plugs, cylinder temps, dirty injectors, fuel atomization, etc.

I honestly don't think you could get something reliable enough to tune by ... at least not on a variety of cars. If you compare to a wideband and tweak it enough, you might be able to get something to work on one specific vehicle.

SoCalRedLine

iTrader: (2)

^thats what I was going at. I have my evoscanGPS in the car, but rarely have my lm-1 installed. It would mostly be for check, not tuning cause I don't tune wih the gps .
I'll try making a formula and checking it on my car.

Is there a way to call up the address for injector size through the MUT anywhere? That would make the equation easier to modify...

On second thought, it would prolly slow down the sample rate quite a bit.

jrohner

Why do you think you can calculate the AFR more accurately than the ECU does? What makes the AFR map value not match up with the real AFRs are the same things that will make this no more accurate. If your airflow sensor is reading right on, and your injector size and latencies are right, the AFR map will be the real AFR, but without a wideband you're only lucky if you get it somewhat close; you'll never get it right everywhere. I've got mine within 2-3% most places (just with injector settings), but my injector values are much different than I expected.

TouringBubble

iTrader: (2)

I cannot agree with this. The AFR in relation to the map changes in different rev ranges and with the injectors scaled correctly for closed loop operation the AFR MAP doesn't usually match the actual AFR.

And, this doesn't even touch on the main fact ... the reason to monitor AFR is to know if something is wrong. AFR MAP doesn't reflect that at all unless load changes due to the issue.

jrohner

That's because the settings are not correct and accurate then. The fuel table is totally ignored in closed-loop. The ECU shoots for 14.7:1, which is all it can even do with a narrow-band. Scale them to be correct in full-time open loop; it's more accurate and, IMO, easier than messing with trims.

I said IF those things are correct, I never said it was realistic or likely that you'd ever get them spot on. I have my AFR map and LC1 within a couple percent of each other at almost every point on the fuel map that I've been able to log (5% is the WORST, in one spot), and now that I can't get them any closer with injector settings, I've started working on MAF settings to get even closer.

The point of all the sensors and all the calculations is to give the engine the right amount of fuel, so you're saying all of it is wasted because it cannot be done, is wrong, or just comes down to being an educated guess (or just a plain old guess). I guess I'd prefer to get things close instead of fudging the crap out of the fuel map to get my AFRs where I want them.

RoadSpike

iTrader: (5)

The fuel maps as it were are just literally raw IPW values before they are corrected by the ecu to take into account for injector size, airflow, temp, blah blah. There is more to it than that in the main fuel loop but thats the jist of it.

Essentially the injector size is a multiplier and what you are doing is skewing the pulse width to compensate for the injector size in order to match what essentially is a fake scaling designed as a help tool for people that don't understand the ECU.

jrohner

If that were true, the values would get considerably smaller as you go across the RPM axis. Some ECU's go by IPW in the fuel map, ours does not.

RoadSpike

iTrader: (5)

Fine don't take my word for it then.

Here's the thread
https://www.evolutionm.net/forums/ec...bles-path.html

jrohner

I would consider the fuel map as just saying how much leaner or richer from stoich you want. You could view the numbers however you prefer, AFR, Lambda, etc.

Ceddy

In the above linked thread Jack_of_Trades, kind of put a AEM spin on the Mitsu fuel maps which isn't entirely accurate. They both do the same thing in the end, but take different approaches in the middle.

The Mitsu fuel table is a compensation. 0x80 = #128 = 100% = 14.7
The ecu kind of always calculates 1 gram fuel for 14.7 gram air and uses the fuel table to compensate for different target AFRs.

Load is actually g/rev (Grams of Air per Crank Revolution).

If you take g/rev x fuel map x constant, you have the IPW that JOT showed.


Taking Load out of the fuel map like Mitsu does has several advantages. First you have easy to read AFR scaling. Second it gets rid of the need for things like Boost Comp, its automatically done, because of how everything is calculated.



To the OP. You are introducing another layer of error trying to calculate WB AFR.
If WB doesn't equal AFR_Map, it means you have some Injector error, but you need to use Injector Size to calculate WB AFR, so the error will be there also. You are calculating something that has already been calculated, so you are going to end up with at least twice as much error as just using AFR_Map.