I have no problem with either sensor. They both have their place. They both have limitations and both have their benefits.


Narrow band: Can read from 14.2/3 to leaner than stoich. This is a narrower range but it can read this range very accurate. A couple of things that help with the accuracy is that they are much more responsive, and can be read much faster. They can read every spark event. This means that if you are cruising in your cruise range of 3000. You are getting 25 spark events a second and the NB sensor is being polled by the ECM 25 times a second. The ECM can also control when during the exhaust event it is polled.


What we call wide bands: Can read from 10:1 to leaner than stoich. This is a broader range but in order to read this range it gives up accuracy across the entire range it can read in. We also need to add a controller that adds to the latency of the reading. All the sensors and controller out there can only read at 4 times a second. Using the above rpm range. We see we loose about 21 of the 25 events every second. Manufacture of these controllers state that AFR can be 5% off. Which in reality is fine because we are using these sensors to map the richer portion of the calibration, and an error here isn't as critical. Keep in mind that these read when they can and have no input to where the engine is in it's stroke events at the time of reading.


I look at the NB's as a outside micrometer. I can only measure with in an inch but can measure to the 4th place. WB's are like a 6 inch scale. They can measure in a bigger range of 6 inch's but only to the 2rd place. What are you measuring? Bolt length or a piston skirt? Boils down to right tool for the job.


I think it is important to map the VE with NB's in the area that is going to run in closed loop by the NB's. Because basically what we are doing by this is. Centering the VE's in the LT/STFT.


Now we have had a couple of manufactures pushing to replace NB's with WB's and configure the ECM to take the 0-5 volt signal. This is nothing new and has been in the ECM for years or since before it came out. It was proven then that the WB's were not callable of holding the tune in the tolerance that they were looking for. The auto world figured this out a long time ago. (I do understand that auto's do have WB's installed, but they don't only have WB's and look at what those sensors are there to do) There has also been plenty of manufactures that have tried mimicking NB output with WB's. This hasn't been successful because the lack of accuracy/ slow to respond WB sensor.


Where to run open vs closed has many factors that need to be considered. From parts on the bike, sensor location, accuracies of the verified readings..... BUT I do not consider heat in the cruise range. I consider fuel mileage. I don't believe that there is enough actual fuel being delivered between 14.3 to 14 to absorb any more heat in the fuel to make any measurable difference in heat. That might be another topic though.

 

Not exactly.

The PV in basic auto tune mode flashes a tune that sets the entire AFR table to 14.6 or .982 for lambda calibrations.

The PV in pro auto tune mode flashes a tune that sets the entire AFR table to 13.0 or .89 for lambda calibrations.

"Have a read" there are several posts on here you can search up concerning such. I'm pretty sure that is what you told me to do on another post.

 

duly edited with the actual numbers from my PV

 

It is my understanding when using autotune basic after a autotune session, then reflash the ECM, the AFR table returns to the map that I started the session with. Only the VE tables are changed.

I just looked in my Win PV program at my FM map that was loaded in the PV for my bike and compared it with my last auto tune session and only the VE tables are changed. The AFR tables are exactly the same.

 

Different tuners may do things differently but the how Narrow band tuning works is the same. For the PV, I've been flashing the new map to the bike once I finish autouning, because I think that it just returns you to your original untuned starting map. The FP3 for example loads up a map for attuning and has two options during the process - Apply which reprograms the active map with the updates so far, then resets the learned table but keeps autotuning - Finish applies the learning to the map, resets all of the AFR and other temporary changes and exits the process leaving the bike running the newly tuned map.

Correct. Autotune only (typically anyway) tunes the VE tables. The AFR table is us to you and something of a manual process should you want to change whatever it was at stock or supplied by whoever in their canned map. Different people like different AFR table ...a whole topic on its own

 

When you enable the autotune it puts the AFR tables completely in Closed Loop, when you load the MAP back in your bike after you accept the corrected VE's the AFR table reverts back, unless you are running another autotune session, then it keeps it in closed loop

 

This is getting a bit of track from WB vs. NB sensors. What you guys are now describing is how each system address's the narrow band tuning. This really needs to kept to just sensors because no matter what tuning device you are using. The sensors are the same. They are either stock NB or LS4U's.


How the different tuning units set up their cals for NB data collection and their strategy can be a topic in it's self. Some force you into a target. Some let you pick the target. Which is better and why? Why do they pick to do what they do? Is it because all one target is easier for the math and they can't figure out how to do a moving target? Is it because that is how SE does it? What happens when you target stioch and then change AFR table to running target. What happens when you target 13 for mapping and change target to 14.2. Does cubic inch and injector size affect this? Does the EGR/ CDE tables come into play? Two different topics.

 

So why does Harley not run a wide band EFI / sensors as standard then do we think? Are narrow bands common across motorcycle manufacturers?

 

My understanding from the guy that wrote the Delphi code in HD is. They are not accurate or respond fast enough to hold a tight enough tolerance to pass their standards. Yes, NB are the common sensor across motorcycle manufactures. They are the standard for fuel control on auto's as well. Auto are installing WB's but not in the fuel control capacity. Most auto's run 2 sensors on each bank. One before the cat to control fuel and one after the cat to make sure the cat is working properly. Some autos have went to a WB for after cat application.


Look at how the ECM works in closed loop and how the sensor is being used. The ECM is in constant switch from lean mode to rich mode. Watch in a log how fast NB's go up and down and realize that the data we are looking at isn't the whole picture even. We are missing a bunch depending on how fast the logger is collecting data. BUT it is switching faster than every .25 of a second. Easy to see how a WB that can only report once every .25 a second can't keep up.

 

I would never consider tuning one of my race motors with a narrow band. The debate is not even close.

Andy