EVOscan 1st time deciphering Evoscan logs need some help from the gurus

[Gwaihir]

I've collected a couple sets of logs using EvoScan. I have an innovate LC-2 wideband. Overall it looks like my LC2 readings are 1 - 3 AFR higher than the ECU AFR map through out the RPM/Load range. At Low Load (Load Column 10) my AFR readings were very high in the log (is that normal?) but show correctly between 14.7 and 15.2 on the gauge.

I have read an exhaust leak around the manifold area could be causing the lean conditions from the LC2 logs. I did the calibration before the install just FYI. What is strange to me is the gauge reads closer to the ECU Map than the logs from the LC2. So my question is, Does the LC2 normally have lean'ish readings in logs or should I be checking for an exhaust leak (which I'm going to do anyway)?

On a separate note I found knock in my logs several instances of 1, 2, and 3 sum. They all seem to be around the 140 load cell. I think this could be due to the fact that I went from a T25 to a 16g that boosts through out the RPM range. So I should back down timing a bit in the 140 column to account for the additional boost higher in the range now correct? Or would the slightly higher AFR (according to the LC2) than target cause knock?

Attached are my logs. If someone has the time to look through them and give me some pointers or suggestions I would greatly appreciate it.

I have a 98 Blackbox ECU with Tactrix ECUFlash and EvoScan, LC2 innovate wideband. My car profile is up to date.

 

[randman2011]

This doesn't help you too much because I still need to investigate for myself, but back when I had the LC-2 but no gauge, I could log the External W/B parameter and it followed very closely with what the serial output of the LC-2 said. When I installed the gauge, the External W/B parameter now stays at zero and instead the input comes through WBO2, which consistently reads 1.5-2 AFR LOWER than seen on the gauge. Every signal has user-configurable calibration functions, so if you trust that the value isn't being mangled in some way before the ECU reads it then you can change the translation function to match what the serial output of the LC-2 controller reads. I need to investigate more on my setup before I take that action.

 

[Gwaihir]

I read something on spooligans eat sleep dsm site about matching up AFRmap and wideband readings:

FR Map is the AFR the ecu is aiming for.
You need to compare AFR Map to WideBand.

• If WideBand is richer, Injector Scaling should be increased.
• If WideBand is leaner, Injector Scaling should be decreased.

So I’m thinking I’ll use the EVO scaling number of 424 for my injectors. I used excel and did a quick calc of afrmap/wb02 and compared to 424/450 and they are very close compared to each other. Maybe my injectors have lost flow over time and need a scaling adjustment to account for that?

Also just some more information on my WB setup. I have the WB wire going to the gauge and I was logging to evoscan over serial.

 

[95REGF150]

The injector scaling is your global fuel multiplier. You put in a smaller number and the ecu will make a longer injector pulse width for the same AFR map value. Larger number and you get a smaller pulse width for the same AFR map value.

So if your wideband is leaner than what the AFR target value shows then you need a smaller injector scaling value. You don't need to use the evo value, just keep lowering the value in the increments allowed by the hexidecimal until you match the target AFR.

Now keep in mind this is done at wide open throttle because that is where you are in open loop and running directly off the AFR map. Changing the injector scaling will affect your closed loop fuel trims as a result because the ecu will now have to compensate the pulse width differently to maintain 14.7 AFR at cruise.

Adding to what @randman2011 was saying, how have you added your LC-2 to the log in evoscan? Have you ran the 0-5v output from the controller into the rear O2 input on the ecu and rescaled that channel to the linear equation provided by innovative? I did that with my LC-1 and always noticed the evoscan log would be about 0.5-1 AFR leaner than my actual LC-1 guage. I ended up altering the equation until in evoscan until I got the log and the guage to match.

 

[Gwaihir]

So if your wideband is leaner than what the AFR target value shows then you need a smaller injector scaling value. You don't need to use the evo value, just keep lowering the value in the increments allowed by the hexidecimal until you match the target AFR.

Roger that, makes sense to adjust incrementally and track the scaling in the logs instead of test an arbitrary number.

how have you added your LC-2 to the log in evoscan?

I selected the innovate serial 2 protocol (same as lc-1) and selected the com port under the wideband menu. I then ticked the box next to external wideband a/f (shows as column wdb in the logs). I made sure my injector scaling in evoscan is set to 450.

Have you ran the 0-5v output from the controller into the rear O2 input on the ecu and rescaled that channel to the linear equation provided by innovative?

No I’m running serial to my laptop from the LC-2 controller. The wideband 5V wire off the controller is going to the gauge only. The controller does not supply the wideband signal to the ECU.

 

[JPenny]

Is this only not supported by the black box ecu? If I remember correctly I had my lc2 wb controllers yellow 5v wire connected to the front o2 ecu signal wire for logging. Sorry if I'm misunderstanding. Thank you for your time, Jason Penny

 

[Gwaihir]

Is this only not supported by the black box ecu? If I remember correctly I had my lc2 wb controllers yellow 5v wire connected to the front o2 ecu signal wire for logging. Sorry if I'm misunderstanding. Thank you for your time, Jason Penny

You can connect in the wideband signal to the black box Ecu, BUT you would receive 0 - 5V on the ECU. If you plug it into either o2 ECU input you would confuse it because it’s looking for 0 - 1.1 for its narrowband calculations, with black box I do not know of any narrowband simulation you could use to correct this. It would be particularly bad if you plug the wideband signal into the front 02 narrowband input. The LC-2 does have a narrowband output that you can plug into the stock o2 inputs and not confuse the ECU (brown wire).

If you were to remove your rear o2 and use that input you might have more success with getting the wideband signal into the ecu and making accommodations in ecuflash (periphery bits) for the switch.

With Link you can simulate a narrowband signal, but the Link ECU handles that.

 

[95REGF150]

The rear O2 input is simply for P0420 code catalytic converter efficiency check and does nothing for how the car runs. You can repurpose this input for anything you like all you have to do is turn off the P0420 check engine light code in the periphery bits.

Now the front O2 input is a different story. This is what the ecu uses for closed loop so you cannot put your 0-5v wideband signal in here directly. However, many wideband controllers have a narrowband signal output. This is a chip inside the controller that creates the 0-1.1v narroband signal from the wideband O2 (msot likely this is what DSMLink's narrowband simulation is as well its probably just a bit of there code that scales the signal). So if your wideband controller has this option you can put the narroband signal into the front O2 channel and the 0-5v wideband signal into the rear O2 channel for logging.

That's how us black box guys have do it. Unless you want to keep plugging a second USB into your laptop for logging the Wideband as @Gwaihir is doing. I never personally liked that approach though because it brings the computer COM's into the mix as the Openport and Wideband are operating at different baud rates. I prefer that all the sensors are on the ECU that way when the Laptop pulls a log entry all the values are at the same instant and I only have to plug one USB into my computer to log. But that's just personal preference.

 

[randman2011]

For further clarification: on both ECUs, the wideband signals are 0-5v analog inputs. The ECU will see the wideband signal just fine. However, unless you lock it in open loop permanently (which you should not do) then the ECU will see the wideband output and, since it is greater than 1.1v, think you're rich and mess things up. There are three ways around this, though:

1. Hook up your LC-2's narrowband output to the front O2. The controller simulates a narrowband signal from your wideband and the ECU sees this signal and is happy. That leaves the wideband output for your gauge, and/or you can connect the wideband output to the rear O2 and log it through the ECU. You will have to disable the DTCs associated with the rear O2 sensor to not get CELs from this.

2. Use a narrowband O2 for the front and add a bung for your wideband for datalogging. I don't know why anyone would do this, but it's still relatively popular. Like option 1, you can connect the wideband output to the rear O2 input for datalogging.

3. Have ECMLink configured for narrowband simulation and connect wideband output to front O2 input. This is identical in every way to option 1 for those that have ECMLink except that you can also log from jsut the one input. No need to fool with the rear O2 at all. Functionality-wise, ECMLink still converts the wideband input to a narrowband signal and uses that in fuel calculations so you don't gain anything tuning-wise by going this route over #1.

So either you have ECMLink and do 3 or you have a black box and do 1. The results are identical.

edit: quoted the wrong post and ninja'd. Oh well.

 

[stillrunning]

I just leave an OEM front O2 in place as normal for closed loop but have my wideband located in place of the rear O2 and log the rear O2 as a wideband with EVOSCAN using the correct transfer function for AFR reading.
You just scale the rear O2 and AFR map to overlay one another for logging.