EGO Control vs Incorporate AFR Target

[luftvagon]

If you are set to incorporate AFR target, should you also be running EGO Control?

These two sound like competing algorithms.

[MarioVelotta]

They are separate.

Ego control is just meant to keep you on track while driving.

IncorperateAFR is meant to change your AFR in one are or the entire map without changing your VE table once the VE table is tuned to the AFR's on the AFR table, if that makes sense

[luftvagon]

so OPEN LOOP with automatic VE table calculation should you make AFR TABLE change VERSUS CLOSED LOOP TO ENSURE YOU ARE ACTUALLY HITTING THE DEFINED TARGET in AFR TABLE.
Got it

[Piledriver]

The CL or open loop O2 control is not fast enough to be of much use except at steady state.
It is almost useless if your VE table is horked, same for accell settings.
It does work for it's intended purpose... fuel variance etc.

If you have "incorporate AFR" enabled (default) and have a well massaged VE table, lambda delay and injector dead time etc...
...the resulting fuel calculation will be dead nuts on to whatever you have the AFR target set to.

Ideally, you will run around autotuning the VE table with ~13.1 or so AFR everywhere (or stoich if you only have a NB sensor), then once the VE table is solid, change the AFR target to whatever works best for that cell//area of the map.

You should NOT have to continuously tweak the VE table (once dialed in) with VEAL.
It is tempting (and i am guilty) but there is always noise in the measurements and calculations.
The flash memory in the CPU has a limited number of write cycles, do not waste them needlessly.

The latest MS3 FW has long term trim, they were looking at storing that on the SDcard.

[luftvagon]

PD, have you done any calculations on injector dead time and v compensation based on logs. I wonder if that could be automated.

[subwoofer]

It is so easy to measure flow rate and dead time using the test mode and a spreadsheet, several versions including mine exists over on msextra.com. Should take you less than an hour to do.

[Piledriver]

PD, have you done any calculations on injector dead time and v compensation based on logs. I wonder if that could be automated.

Just realized the subject--- it's really false, there is no "vs.", you run both.
(I use the "simple" mode with a fast control time, it does what I expect. Never really put any time in getting PID closed loop working)

What subwoofer said, but you can get some feedback if dead time is high/low by tuning the ve table well, then changing your targets rich and then lean.
If your dead time is set low or high, the resulting AFR will run lean or rich...
You can also change # of squirts to get similar effect.
(Which way is which I don't remember, I have slept since then)

When I set a 19:1 target, I get 19.0:1, same at 12:1, so I must be reasonably close.

There are apparently tools to help determine lambda delay in the 1.30 firmware I asked for but didn't notice to play with.

[thedrew]

(I use the "simple" mode with a fast control time, it does what I expect. Never really put any time in getting PID closed loop working)

Pile -- whats a fast control time? I see the default is like 50ms.

Thanks!

[Piledriver]

I was thinking MS2: MS3 settings are called out a bit differently
The sensor response time (in MS3) is actually a physical characteristic of the sensor/controller.
(in meatspace it actually varies, AFR, EGT, sensor temperature, exhaust backpressure have significant effect... The high end 14point7.com units can correct for sensor temp and backpressure IIRC, some uber-$$$$$ ECUS can probably do it too, given the sensors)

MS3: I have my simple ego set at 1% step size, and use it with the in-flash EGO delay table, 20% authority.
(I created that table from data I collected to generate the tiny 9 cell delay table in TS that allows VEAL to work ~perfectly)

MS2E: 10 ignition events per step (That's "response time" in the amusement park of my mind)
0.1 step size, 20% authority.
20% authority due to running E85, which varies in % wildly tank to tank, but he calculations with the flex fuel sensor seem to be dead nuts on...

Last night I ran the tank as low as I sanely dared before filling up with gasoline, and am now at 13% ethanol, from 64% last tank. (temps have dropped and heads overcool running a high% ethanol blend below ~70F OAT)
AFR still on target with minimal EGO correction, even as it changed over about 2 minutes, real time.

It's kinda fun to watch the ethanol % slowly change after a fillup and watch the ECU compensate in real time automagically after startup, at least the first time or three.

Recent GM ECUs calculate ethanol%/corrections based on EGO feedback after a fillup, and lack the fuel sensor, saving a few $ per vehicle and somewhat simplifying things.