Wideband with narrowband emulation and > 14.7 AFR

[Kaldek]

Sitting on my bench at the moment is a Tech Edge wideband O2 kit that supports narrowband output. One of the features of this unit is that it allows you to tweak the narrowband signal so it looks like it's tracking to 14.7 AFR when it's actually tracking to something other than that (either richer or leaner - your choice).

For @TME_Steve I'm curious if a tune that was set to, say, 15.2 AFR or something for cruise, and we set the narrowband emulator to also signal the ECU in closed loop mode to hold 15.2, would there be a measurable improvement in cruise economy or is this something that would drive EGTs too high?

 

[lateshow]

I have tried running in openloop years ago with gasoline and then I had better fuel economy. Maybe about 0.5l/100km. It was running much leaner on low loads...like 16 AFR.

With ethanol I think this would a nice idea too.

 

[Kaldek]

I have tried running in openloop years ago with gasoline and then I had better fuel economy. Maybe about 0.5l/100km. It was running much leaner on low loads...like 16 AFR.

The stock ECU uses target AFR as its tuning method. Obviously Mitsubishi have their underlying algorithm based on injector scaling and VE to allow an AFR target to be translated to an injector pulse.

If your car was running at 16 AFR this would mean that it was running much leaner than the target map as I believe 14.7 is as lean as the map goes. Which in turn means that it probably had a nasty air leak or an injector was partially blocked.

 

[lateshow]

I have seen this happen with all of the cars. Fuel trim low is normally maxxed out to positive side in stock cars. Maybe its the way MAF is scaled.

And about target maps I believe that when ecu enters closed loop mode the target is fetched from somewhere else than fuel maps. But as I don't know much about disassembly, I cant tell.

Well technically the tune in the

The stock ECU uses target AFR as its tuning method. Obviously Mitsubishi have their underlying algorithm based on injector scaling and VE to allow an AFR target to be translated to an injector pulse.

If your car was running at 16 AFR this would mean that it was running much leaner than the target map as I believe 14.7 is as lean as the map goes. Which in turn means that it probably had a nasty air leak or an injector was partially blocked.

 

[TME_Steve]

you can turn off closed loop and run lean as lauri did. On 98 in the cruise regions that use closed loop i usually leave it alone and limit the tune to the use of timing for economy and response gains (obviously change fuel in other regions of the map). on e85 i usually tune to lambda 1.10-1.05 on cruise even gone to 1.15 ish for economy gains as the fuel has so little specific energy. ive run plenty of cars on 98 at lambda 1.05-1.08 in cruise as well. saves a bit of fuel so in summary emulating a stoich output while running leaner will save fuel yes.

 

[jungle]

Well technically the tune in the

The stock ECU uses target AFR as its tuning method. Obviously Mitsubishi have their underlying algorithm based on injector scaling and VE to allow an AFR target to be translated to an injector pulse.

If your car was running at 16 AFR this would mean that it was running much leaner than the target map as I believe 14.7 is as lean as the map goes. Which in turn means that it probably had a nasty air leak or an injector was partially blocked.

Nope, Steve would have set the car to run in open loop. I have mine set to run open loop constantly, so it reads from my AFR map. I run 15.5-15.8 in light load areas. So, even my engine with poor compression and leak down still gets just under11km/L

 

[jungle]

Oops I see Steve has already answered it!!

 

[lateshow]

We are thinking of building our own little arduino based computer that would read the wideband signal 0-5V and have a possibility to select different outputs for the 0-1V range (14.7, 15.5, 16) These would be controlled either automatically (interrupt maf of tps signal) or via a switch or possibly both. This would allow to control the ecu in closed loop running leaner. Id say that idle would be 16. Light cruise area 15.5 and when approaching boost closed loop would hit 14.7 marker. Would be nice to see how ecu handles this....

 

[Kaldek]

We are thinking of building our own little arduino based computer that would read the wideband signal 0-5V and have a possibility to select different outputs for the 0-1V range (14.7, 15.5, 16) These would be controlled either automatically (interrupt maf of tps signal) or via a switch or possibly both. This would allow to control the ecu in closed loop running leaner. Id say that idle would be 16. Light cruise area 15.5 and when approaching boost closed loop would hit 14.7 marker. Would be nice to see how ecu handles this....

That seems a little excessive. Be easier to just turn closed loop off and tune cruise for lean.

 

[whats_doing]

That seems a little excessive. Be easier to just turn closed loop off and tune cruise for lean.

Says the guy doing an arduino controller for the fuel tank. :-p

 

[lateshow]

I don't want to reflash the ecu when i decide to run stoich every bloody time. I have been using open loop and it works fine but what if MAF doesn't operate always so smoothly or one decides to develop SD finally. It's always good to have a closed loop if possible.

That seems a little excessive. Be easier to just turn closed loop off and tune cruise for lean.

 

[GLen20]

So I've un-plugged my 02 sensor to run in open-loop because I'm still chasing a detonation issue (between 4200-4400rpm).
My fuel economy had improved quite a bit and the car seems more eager.

(Forgive my ignorance): if closed loop doesn't lean-burn the engine in light loads thus saving fuel, why do they (manufacturers) run it at all? Is it just protection against against incorrect AFR therefore excessive emissions? And am I hurting my car running open-loop long term?

 

[Kenneth]

Things like this were the reason I did 8 possible map combinations in the KSMods 2.0 ROM.
It is fairly easy to setup the maps so that you can run one set for normal operation then use whatever values you want to switch to open loop with lean AFRs.

That said, the more I played with the tuning on mine, the less I actually used map switching.
(Note, I permanently ran 15.2:1 closed loop)
The better I got my tune, the more responsive the car was to small throttle changes, the better I got at throttle control, the better my fuel economy got. A cyclical improvement up to the point where I could easily get 600+ km out of a tank without worrying about whether I was running close to emtpy.

I learn't that the secret to good fuel economy is throttle control. This allows the car to stay in closed loop and minimises throttle and load based enrichment. Having the engine respond very well to light throttle changes has the benefit of making the car fast, as well as (if you can be precise with your throttle) economical.
Running leaner than 15.2:1 drops off torque, so requires more throttle, which negates any benefit in running lean in the first place.

You might be able to do it better than I did by using the TPS / Load delta values to switch in and out of lean mode. You do need to be a bit careful about this though, because very light throttle is where you will notice any poor running and torque drop the most. So you effectively need to work out just exactly when you need torque and not. When you don't need torque, run lean.

You could do this with an external micro, but again its probably easier to do within the ECU using the 2.0 ROM configuration.

Personally I don't think the time and effort would be worth the potential gains, which would be only minimally better than a good 15.2:1 tune. Especially when that gain can be just as easily lost by a couple of mashes on the accelerator pedal.

 

[Kaldek]

That said, the more I played with the tuning on mine, the less I actually used map switching.
(Note, I permanently ran 15.2:1 closed loop)

Assuming you used the Wideband O2 to spoof the signal so it reported 14.7 AFR when it was actually 15.2?

 

[GLen20]

About bikes, but a good read for learning people like me:

http://www.bikeboy.org/open_closed_loop_efi.html

 

[Kenneth]

Assuming you used the Wideband O2 to spoof the signal so it reported 14.7 AFR when it was actually 15.2?

Yes, that is correct. It was a function available on the narrow-band output from the wideband, so easy to do.

 

[Kaldek]

Yes, that is correct. It was a function available on the narrow-band output from the wideband, so easy to do.

Very interesting. I have a Tech Edge WBO2 controller here which @UNARMD owns. Wonder if I should buy it off him, fit that and sell my AEM gauge, then tune it to simulate 15.2 AFR (with accompanying matching tune of course).

 

[Kaldek]

(Forgive my ignorance): if closed loop doesn't lean-burn the engine in light loads thus saving fuel, why do they (manufacturers) run it at all? Is it just protection against against incorrect AFR therefore excessive emissions? And am I hurting my car running open-loop long term?

Catalytic converters work best at Lamba 1.0 (14.7 AFR). It provides the best mix of reductions of the various pollutants. Leaner reduces one pollutant but increases another, and vice versa for running richer.

 

[lateshow]

The number of Nox -pollutants rises when mixture is lean but number of other pollutants becomes smaller.

 

[paulg25]

I have a tech edge unit. I read the serial data output with a PIC microcontroller and display AFR on a small LCD display. I have a potentiometer with which I can adjust the closed loop AFR on the 'fly'. Seems to run fine right up to 16.7:1 which is as lean as my current code will allow it to go.