Author |
Message |
Vagelis46
| Posted on Monday, January 01, 2007 - 10:01 am: |
|
I have done the following mods on my 2004 XB12R: 1. Open airbox with K&N and removed snorkel 2. Race ECM 3. Micron exhaust system I plugged the bike to the VDST and I see an AFV value of 106.7 when I switch the ignition on. Then I start up the bike and after it warms up a bit the AFV value starts to change as I apply gas. The reading I get for the AFV varies from 102-109 depending the rpm I keep the bike. I thought the AFV=106.7 means a little lean, so I install the snorkel back while keeping the open air box. I take the bike for a ride and put it at learning mode, steady at 3800rpm. I get back and this time the VDST says AFV=96. Which one of the two is the best as far as performance. I need to get the max horsepower possible. If I open the intake even further by removing the plastic around the gas plug, and say I get AFV=115-120, would this mean that I am getting more power by forcing more air inside the engine ? But maybe I get a little overheating ? |
Al_lighton
| Posted on Monday, January 01, 2007 - 04:04 pm: |
|
I'm going to qualify everything written in this post as a best guess. Buell hasn't revealed the detailed workings of their algorithms to me, but Terry and I have developed a working model for how it works, so even if the algorithms we map to aren't perfectly the same as Buells, the logic works. First, you have to understand that the AFV exists to compensate for air density changes. Buell has a map that matches the configuration on the bike, and then the AFV scales the overall map richer or leaner as the air density changes. Pipe and intake changes affect it as well, but pipe and intake changes (especially intake stack changes) aren't linear across RPM and can lead to inconsistent behavior. Buell did not design the AFV compensation mechanism to compensate for pipe and intake tract changes. The AFV is set while running in closed loop AFV learn mode, which is 2-4K steady state light load operation. It is set based on a comparison of the injector duration obtained while running on the O2 sensor with the injector duration calculated from the map at the same operational point. Since the injector duration while running on the O2 sensor oscillates rich and lean (based on the algorithms that maximize the .45V crossings to keep the bike running at 14.7:1 AFR), the instantaneous AFV also oscillates as you saw. But the real AFV is a time averaged value of the instantaneous AFVs. Much of the time the bike is running in closed loop, those instantaneous AFV samples are discarded, they don't affect the real AFV. Only when the bike is operated in a very specific way (certain vehicle speeds, RPMs, temperatures, steady state time, light load, etc) do the samples get time averaged and set the real AFV. The problem when you use a map developed for a different pipe or bike configuration is that the amount of mismatch between the O2 dirived injector duration and the map dirived injector duration is a moving target. For instance, from 2-3K, the Micron and the Race pipe aren't too far apart from each other. An AFV learned in this range is likely to be somewhere near 100. But between 3-4K, the Micron makes good power, but the Buell Race pipe is deep in a midrange hole. Since Buell mapped the race ECM for that hole, it delivers much less fuel than the Micron wants there. When running in closed loop, this results in a instantaneous AFV that climbs substantially above 100. With an AFV of 100, and in open loop operation, and in the 3-4K range where the micron stomps the Buell race pipe, the micron will be very lean. From 4-5K, it will be slightly lean. In the 5-6K range, where the Buell race pipe shines, the Micron will be fueled slightly rich. Above that, it is being fueled slightly lean. If the AFV was learned in a 2-3K learn mode, the AFV would be around 100, and that is how the bike would run. If the bike was learned in the 3-4K range, the AFV would be above 100, the lean spots wouldn't be as bad, but the 5-6K rich spot will be grossly so. There is enough heat there that you probably won't foul a plug because of it, but you'll not be making max power. Higher AFV's merely indicate that during the learn mode in 2-4K RPM, the map was leaner than what was needed to make the bike run at 14.7:1. This means that the engine WAS pumping more air than the configuration that the map was developed for. Whether that higher AFV yields more power or not at higher RPM is dependent on how far the map is off at higher RPM, and that may not match, and likely WILL not match for a Micron operated on a Race ECM map, the same amount off as it was during closed loop learn mode. AL |
Vagelis46
| Posted on Tuesday, January 02, 2007 - 02:12 am: |
|
Al, Good morning. So the real AFV is for compensating for air density. I also know that the more dense the air the better, for performance, right? Consider a bike with the race kit (filter+ECM+pipe)that are matched. This bike, if it runs at high altitudes will have a low AFV value, right? It will also have reduced performance and that is a fact. As this bike will be operated at reduced altitudes its AFV will rise, and so will its performance, since the air will be denser containing more oxygen. So we reach sea level and the bike has AFV 100 (right?) and the bike makes the max power it can possibly can with the given intake, pipe, and ECM. So then I modify the intake (while at sea level)while keeping the stock velocity stack, by cutting the airbox (open airbox) and removing the snorkel. From what I have seen at the VDST the AFV will rise above 100. For me this means that since of the reduced pressure drop across the intake arrangement, is like operating with even more dense air. Because I minimize the pressure drop I feed the engine with more air and since the ECM will match this, the bike will make more power. Correct ? I have read somewhere that the AFV can climb up to 120. Is this correct? So I make the mods and (say) I reach 119. This would be the max power I can get from the engine. Further increases will not be able to be matched by the current ECM's maps and the bike will run lean, having reduced performance. So until I reach the max value of the AFV the bike has extra potential for power. Do the above stand as a theory ? |
Al_lighton
| Posted on Tuesday, January 02, 2007 - 10:16 am: |
|
You're mostly correct. With a bone stock race kit, at sea level, the AFV should be around 100. Might be a few points higher or lower, but some fluctuation is normal. The AFV will go up as you make the modifications you spoke of, not because the air gets denser, but because you'll be able to flow just a little bit more in (more O2 in by mass flow instead of by density). And if you route the crank breathers out, it'll go up a little bit more because you'll be replacing hot oxygen depleted blowby air in your intake with fresh unburnt air. The AFV can go higher than 120, but you're not likely to go that far due to changing altitude alone, as it is hard to go too much below Sea Level. It typically takes changed components on the bike to get it to those values. But earlier DDFI bikes could go a LOT higher than 120 when they ran out of fuel. And even the early XBs with the old fuel pump would drive the AFV higher when the tank was very low. The statement "So until I reach the max value of the AFV the bike has extra potential for power." is kinda true, I guess. If the engine pumps more air, and you give it the fuel it needs with the extra air, it WILL make more power. But it is dangerous thinking. Because by the time you get to a mismatch of that proportion, the likelihood that the fueling curve shapes are grossly mismatched by a different amount at some other RPM is high. With a 120 AFV scalar, it is likely that it will be running grossly rich at some RPM, as most mods that would have that much effect are likely to not be linear. A grossly rich bike will leave power on the table just like a grossly lean bike will. Al |
Vagelis46
| Posted on Tuesday, January 02, 2007 - 01:53 pm: |
|
Al, I will experiment with : 1. removing the snorkel + 2. KN and open airbox + 3. Removing the plastic thing of the airboxcover around the fuel tap I will let you know of the AFV value I can achieve by doing this. As you have state in previous posts intake mods are pretty much linear, so I guess it will not be bad after all. I have a lot of R6 chasing to do, so I need all the edge I can get. Once again it was nice talking to you. |
Id073897
| Posted on Saturday, June 02, 2007 - 04:16 am: |
|
I would like to add a few interesting details we found out at the track. I went to a small track in germany (http://www.motorsportarena.com/?cat=4&content=9) on very bad weather. I arrived there and we prepared the bike (all stock except some remapping of the ecm) for the track - changed wheels and brake pads. AFV after arrival was 115. Air temp. was about 13-15 degC, air pressure about 1005 hPa, rel. humidity 98%. My bike is equipped with a wideband O2 sensor and I used my AFRSpy software to monitor the ECM and AFR on the track. Whithin seconds (on the first turn at the track) the AFV dropped to 104 and then to 98.8 - without any change in weather or climatic conditions. On the following turns AFV was oscillating between 98.8 and 104. On the last turn the track got dry (was always wet before, so the driver acted very carefully until then, as it was not his bike) and AFV went even down to 93.8 - but just for a minute or so. Now, I didn't even expect the AFV to change on the track, as driving there does not usually meet "cruising conditions", which - usually told - has to happen, to make the ECM adopt to a new AFV. If I take a look at the logs, I see rpm almost always above 4000 and throttle was either fully closed or WOT. Especially when AFV was very low, EGO/lambda correction got active maybe 6 or 7 times only, which too indicates open loop. Any idea, how and why AFV does change? I formerly planned to "misuse" AFV to lean/reachen mixture on the track and to not have to change the fuel map, but obviously this will not work. Regards, Gunter PS: if anyone is interested in the logs, just ask for them. |
Al_lighton
| Posted on Sunday, June 03, 2007 - 10:29 am: |
|
Gunter, Not sure about the SW that you are using, but when one looks at the AFV with VDSTS while the bike is running, the AFV that is being read is the "instantaneous AFV" (my term, not Buell's), not the "real" AFV. The instantaneous AFV oscillates all over the place, just as the O2 voltage does as the ECM commands the bike rich/lean/rich/lean to maximize .45V crossings in closed loop. Any time the bike is in closed loop, the instantaneous AFV is being calculated. But it has to be in closed loop learn for the instantaneous AFV samples to be stored and averaged to calcculate the "real" AFV. The bike can be in closed loop, but not in closed loop learn. With VDSTS, the only way to read the REAL AFV is to shut the engine down and read it. When the bike is running, the AFV readout displays the instantaneous AFV. Is this an XB9 (my guess) or an XB12? It sounds very strange to me, I'm not sure how the instantaneous AFV is even calculated, let alone have the "real" AFV learn, if it isn't in closed loop for some period. I'd love to see the logs. Send them to sales at americansportbike dot com, please. Thanks, Al |
Id073897
| Posted on Sunday, June 03, 2007 - 11:33 am: |
|
Al, what we found out is that there is an AFV stored in the ECM's eeprom, which is the last ego correction in the moment, you switch off the bike. I'm not quite sure, whether your "instantaneous" AFV is written to eeprom too, but a newly written/set AFV is definitely used at once in all calculations. I never tried to check whether there is a difference in the AFV the ECM delivers in it's runtime data and the AFV stored in eeprom, as this is a bit difficult to do, although it would be possible. The "instantaneous" AFV (which I use now as a term for the AFV in the runtime data) is used in calculating the pulsewidth triggering the injectors, this can definitely be seen in the logs. It's a bit difficult to explain, I will send you the files and instructions, how to examine them. AFRSpy is small application for a palm pda, derived from the EcmSpy Toolkit. You can learn more about EcmSpy for palm by following this link: http://www.bike-pix.com/forum/viewtopic.php?t=1793 4 and about the pc version http://www.bike-pix.com/forum/viewtopic.php?t=2020 1 Regards, Gunter |
Id073897
| Posted on Monday, June 04, 2007 - 02:52 am: |
|
Al, I've received an error message from the mailserver, so please leave me a short note if you got the logs or not. Thanks, Gunter |
|