I stated that very early in the discussion: "If the resultant closed-loop AFR is out of range of the O2 sensor (but it otherwise could be within range, i.e. there are no major component/system faults causing the error), it can be reeled back into range and once there fine-tuned, so the O2 sensor operational range (±3%) is not the limiting factor here."

The "adaptive fuel" capabilities allow for bringing the AFR back into the range where the O2 sensor can be used to modulate it. This is not as good a system as wide-band sensors being used in closed-loop everywhere, but it's a good and elegant way to use the cheaper and less controlling-hardware-needy narrow-band units. The range of operation is ±15% against the VE values directly, not against AFR per se, though in the end, I guess, the results are the same.

In the event a closed-loop condition exists where the AFR is out of range of the sensor, by modifying the AFV (adaptive fuel value) from (initially ×1), the "VE NEW" is obtained which will get the AFR back into range of the O2 sensor's capabilities. Once it gets within range, a more perfect AFV is set. This AFV pertains to every consultation of the VE tables to use the resultant "VE NEW".

If you're in closed-loop operation the modulation of the O2 sensor will be centered on stoich (or wherever the CLB setpoint is if different from stoich) because the VE table value for that situation is multiplied by the "perfected" AFV. When you switch to open loop, those VE cells are also multiplied by this "perfected" AFV. This is the way that designed AFR is maintained in open loop operation in the face of external sources of change. It's to prevent squeaked pistons, etc.

Let's use a "real-world" hypothetical example. You have a bone-stock '07 or '08. You're taking a nice day-long ride to the hill climb and back. Along the way on the country roads you encounter farmer after farmer harvesting their beans and corn. There is dust in the air everywhere. Your air filter gradually gets more and more filled with this dust. As it does so, your closed-loop AFR keeps dropping, out of the top end of the O2 sensors' modulating range. Once below about 14:1 the sensor doesn't "know" how much it is below. It could be 13.5:1 or 9.5:1 and the sensor will put out the same voltage either way. The ECU detects the out-of-range condition and starts backing off on the AFV, creating a new derived VE, until the voltage output from the O2 sensor starts modulating again instead of flatlining. The AFV is further modified to center the O2 sensors' modulation at the CLB setpoint. Now, when you open the throttle to pass that grain truck that just pulled onto the highway, your open-loop AFR setpoints are also correct for the dirty filter because those VE cells are multiplied by the AFV to obtain the corrected VE value used to calculate injector duty cycle.

You're getting ready to leave the hill climb and when you go to start your engine, it doesn't light the first time. Instead, for some reason, it coughs back out the intake. This knocks a bunch of the dust out of the filter element. Now the engine isn't being choked as much as it was when you pulled in. When the engine gets into closed-loop, now the O2 sensors are flatlining on the low end of their ranges; the AFR is too high. How much? Who can say, it could be 15:1 or 100:1. So the AFV is walked back in the other direction until corrected VE values are obtained which allow the O2 sensors to once again modulate, and the AFV is f

 

And the winner is ?????????

 

Interesting discussion. May I add this little tidbit that may help?

http://www.nightrider.com/biketech/h...erformance.htm

Click on the link to see the chart. I don't know how to post it correctly here.

Harley-Davidson did publish some interesting power improvement percentages in the Fall 2006 issue of its Enthusiast newsletter. From the information listed below it must be assumed that you can upgrade the exhaust system and intake system on your 96CID engine without any changes to the ECU. Nightrider has historically been a proponent of exhaust system only upgrades with no EFI remap, but now HD has published that HD/SE exhaust upgrades and no EFI upgrades can be added to your bike without warranty implications. The table below summarizes power improvements as published by Harley-Davidson.






Maximum % Power Improvement
No Remap
Maximum % Power Improvement
HD Stage 1 Remap

HD SE
SE Street Power Mufflers
SE Air Cleaner Kit
SE Street Power Mufflers
SE Air Cleaner Kit
SE ECM recalibration

Twin Cam 96CID
3% torque
7% horsepower
7% torque
12% horsepower

Sportster
17% torque
8% horsepower
22% torque
15% horsepower

 

That page was last modified (at least as far as the server is concerned) on this past Sept. 11. I'd thought I looked through that site since then but I guess it's been longer.

The site in general offers pretty good information, but most of it isn't pertinent anymore with the '07+ HD Delphis. Take for instance, their pushing either the LC-1 or the resistor network to shift the center of the O2 sensor linear output toward the rich side. If this is done, the AFV will result in a somewhat greater VE table multiplication factor. That's okay in closed loop, but when you go to open loop you'll be richer there too. Lean there isn't a problem now in the stock setup, is it? So you'll be less-than-optimum up there.

There's no way I'd change the intake dynamics without flashing the stock ECU or piggybacking it in some way. Changing the exhaust dynamics (that is, altering the balance between the real VE values and the programmed VE table cells) is iffy enough. Doing that to the intake (alone, or especially in conjunction with the exhaust) would involve more sensing devices than just an O2 sensor. Too much chance for me. Now if we had MAF analysis...

 

Here's the table.
One other thing about that page that really makes the hair on the back of my neck stand up is that they attribute the horsepower/torque differences between the 96A and 96B engines to the exhaust system. While I'm confident that's a factor, they seem to totally dismiss the out-of-balance balancing shaft which must certainly require some horsepower to spin. They seem smarter than that otherwise... But it's indicative of the site, so be cautious, just as you should be here.

 

[quote]ORIGINAL: glens

I stated that very early in the discussion: "If the resultant closed-loop AFR is out of range of the O2 sensor (but it otherwise could be within range, i.e. there are no major component/system faults causing the error), it can be reeled back into range and once there fine-tuned, so the O2 sensor operational range (±3%) is not the limiting factor here."

The "adaptive fuel" capabilities allow for bringing the AFR back into the range where the O2 sensor can be used to modulate it. This is not as good a system as wide-band sensors being used in closed-loop everywhere, but it's a good and elegant way to use the cheaper and less controlling-hardware-needy narrow-band units. The range of operation is ±15% against the VE values directly, not against AFR per se, though in the end, I guess, the results are the same.

In the event a closed-loop condition exists where the AFR is out of range of the sensor, by modifying the AFV (adaptive fuel value) from (initially ×1), the "VE NEW" is obtained which will get the AFR back into range of the O2 sensor's capabilities. Once it gets within range, a more perfect AFV is set. This AFV pertains to every consultation of the VE tables to use the resultant "VE NEW".

If you're in closed-loop operation the modulation of the O2 sensor will be centered on stoich (or wherever the CLB setpoint is if different from stoich) because the VE table value for that situation is multiplied by the "perfected" AFV. When you switch to open loop, those VE cells are also multiplied by this "perfected" AFV. This is the way that designed AFR is maintained in open loop operation in the face of external sources of change. It's to prevent squeaked pistons, etc.

Let's use a "real-world" hypothetical example. You have a bone-stock '07 or '08. You're taking a nice day-long ride to the hill climb and back. Along the way on the country roads you encounter farmer after farmer harvesting their beans and corn. There is dust in the air everywhere. Your air filter gradually gets more and more filled with this dust. As it does so, your closed-loop AFR keeps dropping, out of the top end of the O2 sensors' modulating range. Once below about 14:1 the sensor doesn't "know" how much it is below. It could be 13.5:1 or 9.5:1 and the sensor will put out the same voltage either way. The ECU detects the out-of-range condition and starts backing off on the AFV, creating a new derived VE, until the voltage output from the O2 sensor starts modulating again instead of flatlining. The AFV is further modified to center the O2 sensors' modulation at the CLB setpoint. Now, when you open the throttle to pass that grain truck that just pulled onto the highway, your open-loop AFR setpoints are also correct for the dirty filter because those VE cells are multiplied by the AFV to obtain the corrected VE value used to calculate injector duty cycle.

You're getting ready to leave the hill climb and when you go to start your engine, it doesn't light the first time. Instead, for some reason, it coughs back out the intake. This knocks a bunch of the dust out of the filter element. Now the engine isn't being choked as much as it was when you pulled in. When the engine gets into closed-loop, now the O2 sensors are flatlining on the low end of their ranges; the AFR is too high. How much? Who can say, it could be 15:1 or 100:1. So the AFV is walked back in the other direction until corrected VE values are obtained which allow the O2 sensors to once again modulate, and the AFV is furth

 

This could go on for ever!

 

[sm=bicker.gif] [sm=popcorn.gif][sm=alcoholic.gif]

 

The site in general offers pretty good information, but most of it isn't pertinent anymore with the '07+ HD Delphis

Their harleyinformation came from the fall '06 Enthusiest Magazine, which would, I would think, then pertain to the '07 Models which were introduced in the summer of '07. Note that it refers to 96" engines.

However, in Joe Minton's series of articles he started earlier this year in American Rider, he showed that changing both the air cleaner and the mufflers did virtually nothing to gain power in the useable range. (April issue). In fact, the best gain came when installing the screaming eagle mufflers, but leaving the air cleaner stock. He attributed that to a carefully designed aircleaner that sets up a resonance, and recommended just replacing the element with a K&N, not the whole air cleaner.

It was only when he changed the AF ratio (using the Cobra unit) that they got appreciable gain in the midrange, where we ride most of the time. (August issue).

 

The SERT users manual states "switching sensors can only control within a narrow range (+/- 0.5 AFR)". +0.5 AFR (from stoich, the center of the range) is 15.2 AFR or -3% fuel [14.7 ÷ 15.2 = 0.97] and -0.5 AFR is 14.2 or +3% fuel [14.7 ÷ 14.2 = 1.03]. That's the limit, which I'm not ignoring. It's also not the function I'm describing. But remember that, the usable reporting range of the O2 sensor is only 3% in either direction from the midpoint.

The SERT manual mentions the VE-New how? It says it's calculated when running closed loop. Does your copy explain how it's calculated? My copy doesn't. Does that mean it's not really happening or does it just mean that it's impertinent to the SERT because it's yet another of the things the SERT cannot touch?

I'll attach an image I picked up from some post here a while back. Note the "VE Front" is 86% while the "VE New Front" is 81.5% (94.8% [=AFV for that cylinder] of the set value) while the "O2 Integrator Front" shows no deviation from setpoint which appears to be 0.56V. That's -5.2% deviation from set VE, almost the total range of "control" available from the O2 sensor! Yet the usable range of control from the sensor in establishing the "VE New" in any direction is only half the total O2 sensor range available, which makes the deviation represented nearly twice the range of the O2 sensor for that direction. How does the VE New get that far away if it's being "controlled" only by the O2 sensor, especially when the sensor is at its CLB value (evidently)? I contend that the sensor is at its CLB value because the VE New, and not the VE, was used to calculate the injector duty cycle. The VE New controls the O2 sensor output, yet the sensor output helps establish the VE New (via the currently-adjusted AFV). None of how that operation takes place is described in the SERT Manual yet it obviously happens.

Look at that again. The VE New is almost twice the distance in percentage of what the O2 sensor can measure in that direction from its midpoint (which it's not even set at; I believe where it's set and the direction of the VE New would result in greater than twice the available range of the O2 sensor in that direction; but even if the "headroom" in the sensor output toward the other direction from its setpoint was used it's still less than the resultant AFV).

You're the one of we two who has a SERT-equipped bike. You could data-log and see for yourself. So far the only instance of that I've been able to scrutinize is included here and discussed above. That's hard data/proof. Though not proof that my specified range limits are correct. But you could easily verify that yourself. I'll leave it to your imagination as to how to go about it. (something like zip-tying some non-flammable fabric around the end of the muffler and data-logging a trip around the block ought to suffice and be harm-free enough)

As I've previously stated, I don't have access to the HD Delphi engineering documentation so I can't provide you with proof that way. I do know, however, from other creditable sources it probably wouldn't do much good to drag into this, that the range