Got to 183whp - cant get more

[brn12345]

Originally Posted by nelsmar

your tq has dropped on dyno #2 & #3, and if you look you are looking at a fair bit richer on both of these. When testing mods like this you want to make sure LTFT is not affecting your AFR's. Also verify that the oil, coolant, and IAT temps are near the same between pulls. Verify your ignition timing is near the same, if not look into why. Did you play with ignition tuning this time around? When you remove the OEM header and install a new cat-less one the EGT temps tend to drop allowing more aggressive tuning.

From what i am seeing here you just removed parts and made no tuning changes between so the changes will be minimal. With the secondary cat removed you should work on getting your AFR's close to your commanded AFR first. Then once the ecu is doing what you tell it to you can start working on timing and actual afr's.

Very useful info thanks. Moving afr was something I didn't know how to tackle, so thanks for the advice on disabling it. Thanks to @jamesm as well. Will try disabling next time.


I didn't play with the tune at all in this case. Will try fiddling right after I get a resonated mid pipe. I gutted the secondary cat and the car sounds really bad.


In the meantime Ill probably review logs and review maf scaling as suggested.

[8686]

I noticed that in your last run the AFR is going very rich according to the dyno like 10 on 4500 rpm. on the log, it is reading 12.17.
Can any one help in explaining why is that happing.


Sent from my iPhone using Tapatalk

[nelsmar]

Originally Posted by 8686

I noticed that in your last run the AFR is going very rich according to the dyno like 10 on 4500 rpm. on the log, it is reading 12.17.
Can any one help in explaining why is that happing.


Sent from my iPhone using Tapatalk

Looks like his OEM wideband is not calibrated. OEM calibration only reads around that low. You have to re-write the axis on the tune to lower it and make it readable. However below ~12AFR the oem sensor is not terribly accurate and its update frequency is fairly low.

The dyno wideband is a little more reliable. However if you remove a cat on the car the dyno will start to read more accurately. So it might start to look slightly richer at the tail pipe.

[s2d4]

Wideband != narrowband

[nelsmar]

Originally Posted by s2d4

Wideband != narrowband

The OEM front AFR sensor is a Denso limiting-current wideband. Not a narrowband.



The denso sensor just simply doesn't read much below 12 AFR, and below 12 its accuracy is not quite as good. Compared to my LC-2 and the LM2 we were using on the dyno (with the bosch from my LC-2 wired into it) it was fairly close on AFR's. Mind you I have calibrated mine though. Although below 12 AFR the accuracy drops off a bit. And the oem wideband updates much slower than my innovate does. So the detail I get in my logs are more precise on the innovate. However it is not always necessary unless doing very fine tuning to use more advanced wideband especially on pump gas.

[s2d4]

Wonder why it targets around 10 for AFR with the stock ROM for high rpm and load when it is so inaccurate under 12?
Can't remember off the top of my head how far out it was for @jamesm when he did the compensation map for full close loop setup.

[brn12345]

Originally Posted by nelsmar

Looks like his OEM wideband is not calibrated. OEM calibration only reads around that low. You have to re-write the axis on the tune to lower it and make it readable. However below ~12AFR the oem sensor is not terribly accurate and its update frequency is fairly low.

The dyno wideband is a little more reliable. However if you remove a cat on the car the dyno will start to read more accurately. So it might start to look slightly richer at the tail pipe.

@8686 I am aware of the OEM sensor has a floor of 12.17 but I assumed that if I changed it to read lower it would mess up part throttle etc. I am assuming the stock map limits it to 12.17 for a reason. I am going to install a standalone wideband that will then be plugged directly into the OP2 for direct logging. I believe this will have much more consistent and real results.

[Kodename47]

Originally Posted by s2d4

Wonder why it targets around 10 for AFR with the stock ROM for high rpm and load when it is so inaccurate under 12?
Can't remember off the top of my head how far out it was for @jamesm when he did the compensation map for full close loop setup.

Understanding open loop fueling is the key here and is why MAF scaling (or fueling table corrections) are important. When on open loop, the ECU uses the MAF to calculate g/s of air entering the engine and then calculates how much fuel is required to hit the target AFR in the fuel map. The output of the O2 sensor is purely for you to monitor the output and is not used by the ECU in these conditions.

My O2 has been scaled to read as accurate as possible after time spent on the dyno using the wideband to calibrate it as above.

[Kodename47]

Originally Posted by brn12345

@8686 I am aware of the OEM sensor has a floor of 12.17 but I assumed that if I changed it to read lower it would mess up part throttle etc. I am assuming the stock map limits it to 12.17 for a reason. I am going to install a standalone wideband that will then be plugged directly into the OP2 for direct logging. I believe this will have much more consistent and real results.

As with all sensors, you choose where you want resolution. Actually the stock sensor is good at lambda 1 which is where you spend most of your time in cruise and part throttle. I'll see if I can dig out my O2 scaling as a reference as well as one by a previous tuner, both are different but have worked well.

[s2d4]

Originally Posted by Kodename47

Understanding open loop fueling is the key here and is why MAF scaling (or fueling table corrections) are important. When on open loop, the ECU uses the MAF to calculate g/s of air entering the engine and then calculates how much fuel is required to hit the target AFR in the fuel map. The output of the O2 sensor is purely for you to monitor the output and is not used by the ECU in these conditions.

My O2 has been scaled to read as accurate as possible after time spent on the dyno using the wideband to calibrate it as above.

Yeah, I know about it.
You didn't actually answer the question though.

[Kodename47]

Originally Posted by s2d4

Yeah, I know about it.
You didn't actually answer the question though.

It did kind of explain it though. The map can still target those values as at high RPM and loads it's not using the O2 sensor to validate what it's achieving, it's in open loop so is calculating the output AFR from the MAF sensor reading and fuelling accordingly. That is also why there is LTFT as it works them out in closed loop when the O2 sensor is within good working ranges and then applies them in open loop when it's not.


Remember the O2 sensor is only used for fuelling in closed loop.

[nelsmar]

The ecu doesn't need a wideband to verify its AFR targets. The AFR traget is created based off the maf calibration and other tables combined. The car in open loop can run completely with out a af sensor. The car does not need to know if it is more than say... 1AFR off for adjusting fuel trims, so there is simply no reason to spend the time calibrating it from the factory as it literally does nothing for the ecu. We however are using this sensor to verify our target AFR and as it is only calibrated from the factory for closer to stoich levels... we need to extend this to see further.

The region in which you would be in closed loop with the factory tune is pretty much all 12+AFR. Once you dip into areas in the 11's the car should be in open loop on the factory tune. Open up the fuel table and you will see what i am talking about.

[jamesm]

the fueling table is just a table of lambda offset targets (i.e. internally your '11.2' means lambda + x% fuel). what the computer needs to know is how much fuel it needs to inject to create a stoichiometric afr at any given time. if then applies the offset over this to achieve the target afr you see in the fueling map.

so what this means is that your fueling is only as accurate as your hardware (maf, injectors mainly) calibration. unfortunately there are many, many ways to arrive at the same conclusion, and all but one are wrong and will have side-effects you don't want.

with many cars its common to do 'maf hacks' to overcome load limits or to just not care whether targets were being hit, or just move the targets to compensate. we don't have the same luxury. being that we have two complete, independent, interoperating fueling systems (port and di) their calibration must be as close to real-world correct as possible. whereas you could compensate for slightly wrong injector scaling in the maf on say a wrx, if you were to do that on our car you'd completely screw the direct injection fueling error and wonder why your idle was rough and car tried to stall when you came to a stop.

so, long story short, when you have DI + PI, they both have to be right, which means you shouldn't ever be making 'corrections' to the fueling target tables. if you are, you're doing it wrong, masking a bigger issue and chasing your tail. i know because i eventually caught mine .

[jamesm]

Originally Posted by nelsmar

The ecu doesn't need a wideband to verify its AFR targets. The AFR traget is created based off the maf calibration and other tables combined. The car in open loop can run completely with out a af sensor. The car does not need to know if it is more than say... 1AFR off for adjusting fuel trims, so there is simply no reason to spend the time calibrating it from the factory as it literally does nothing for the ecu. We however are using this sensor to verify our target AFR and as it is only calibrated from the factory for closer to stoich levels... we need to extend this to see further.

The region in which you would be in closed loop with the factory tune is pretty much all 12+AFR. Once you dip into areas in the 11's the car should be in open loop on the factory tune. Open up the fuel table and you will see what i am talking about.

to be accurate, only 14.7:1 can ever be targeted in closed loop, the logic doesn't work any other way. the values in the closed loop section of the map that *arent* 14.7:1 actually arent targets, they're there to generate trims, but that's another topic for another day.

suffice to say that the ecu can only truly target 14.7:1 in closed loop, even if it acts like it isn't sometimes.