| Author | Message | ||
     Ferocity02 |
I just got my bike set up with two narrowband O2 sensors and I'm going to do some data logging and hopefully fix some of my bike's rough edges. My questions is, when logging data, should I set the bike up to be in open loop the whole time, or should I just let it operate as normal(open and closed loop based on RPM/TPS)? I've thought a lot about it and can't seem to determine which one would be best. Thanks! | ||
| M1combat |
I'd settle the AFV real nice then turn off all the closed loop learns. I wouldn't reset the AFV before tuning. I wouldn't tune the closed loop area AT ALL. Once you're happy with the tune in the OPEN LOOP areas turn closed loop learn back on. This should take care of the AFV creep issue murdering your map (and maybe other neat stuff). Do a search for Treadmark's "Square Idle" as well. Oh... and don't try to go for a "real" tune with NBO2s. It's just not really going to happen. | ||
| Id073897 |
when logging data, should I set the bike up to be in open loop the whole time, or should I just let it operate as normal As you are referring to logging, I would recommend to put the bike into the same state as you are usually using it, otherwise you are looking at changed conditions which are not the same as those you are interested to look at. Once you identified the problems you could go either way, depending on which one might give you better control over changes applied. I would agree to not set the AFV, for the same reasons as mentioned above. Regards, Gunter | ||
| Ferocity02 |
I understand that I won't get an amazing tune with NB O2's, I'm just looking for something that runs a little better than it does now. I did two data logging runs and used VE Analyzer and it really helped the bike to pick up some power and responsiveness. So I put another O2 in the front cylinder so I can start over and do it again with a little more accuracy. I was planning on locking the AFV to 100 so it would tune to that value. This way the closed loop learn wouldn't affect it. However, there are other correction factors that exist for open and closed loop, so I figured I could just tune in open loop so the correction factors are the same for all of the TPS/AFV zones. Does this sound reasonable, or should I just allow the bike to run in open and closed loop as normal? (Message edited by ferocity02 on August 06, 2008) | ||
| Id073897 |
If you lock AFV and the ECM would like to adjust it but can't do and you release AFV later and the system then changes AFV, it changes fuel also and all your efforts are lost. Regards, Gunter | ||
| Swordsman |
I've been told to set the AFV at 100 and then lock the Min/Max both to 100 as well. If you DON'T lock the AFV and do you're tuning, it's wasted data because the whole system is in flux. Consider this: you've tuned with your AFV at 110%. According to your logging, you've got a rich spot, so you reduce the fuel. Now your tune is running dead on when the AFV is 10% richer than normal. It's like setting the sights on your pistol when you know you're aiming off to the side. The whole purpose of the AFV is to counter lean/rich conditions, and if you're tuning beyond 100% to begin with, there's no chance of getting level. ~SM (Message edited by Swordsman on August 06, 2008) | ||
| Xl_cheese |
I agree with locking the AFV. I did this exact same thing on my bike with narrowband front and rear. If you lock the AFV you can be certian you are tuning each cylinder to the same spot. then when you are back to normal the adjustments the ecm makes from rear cylinder data will hold correct for the front. when I did datalogging before with only a single O2 I got poor results using a couple guessing methods for the front. Now with tunign each one individually I'm getting great results. I'm curious about tuing with a wideband to see how much different it would be from the narrowband. A test of how good the MLV model is with narrowband. Once I was happy with the tuning I unlocked hte AFV and it is pretty stable at 100 on each cylinder. I did notice the AFV dropping to 95 immediately after a couple decels or engine braking from 4-6k rpm. I went back with MLV and worked on the decel regions a lil bit and it leaned them out slightly which is what I would expect. Something else that you can do with 2 o2's is tune the idle. Tune it so that the ego corr while running in CL idle is near 100. | ||
| Id073897 |
Do as you like, but it leads to wrong results. (BTW: the OP didn't ask for tuning, but logging.) What's happening if you lock the AFV to 100, but air pressure or configuration requires another value? You adjust fuel spot on, release AFV and your tune is gone immediately as AFV changes immediately to the required value. If you want to lock AFV, then lock it to the REQUIRED value (which is calculated during calibration) but not to SOME value. Regards, Gunter | ||
| Ferocity02 |
Gunter, I am actually planning on doing some bit of tuning. The purpose of locking the AFV would be so that the ECM does not change it during the logging of data, obviously this would screw up my results. Then tuning the bike with the locked AFV should tune it to that value. Once I'm done I would unlock the AFV and hopefully it will stay near 100. If it doesn't, this this means I did something wrong. I plan to log data in similar atmospheric conditions on every run to get the best data. However, my questions still stands, to whether or not I should tune completely in open loop, or tune in "normal" mode. If the AFV is locked, then the closed loop learn area shouldn't have any affect on my results. Is the AFV the only thing adjusted by the ECM during closed loop learn? | ||
| Swordsman |
"You adjust fuel spot on, release AFV and your tune is gone immediately as AFV changes immediately to the required value. " Er.... isn't that exactly what the AFV is designed for? The AFV itself is calculated off of a 100% baseline. To tune anywhere else doesn't make sense. What you're suggesting is trying to bypass the whole purpose of the AFV... it's GOING to shift, without question. The manual itself says to disable the AFV while dyno tuning. I don't see why this should be any different. Ferocity, I would guess that if you lock the min/max at 100%, it won't matter what loop it's in at the time. ~SM (Message edited by Swordsman on August 06, 2008) (Message edited by Swordsman on August 06, 2008) | ||
| Id073897 |
The purpose of locking the AFV would be so that the ECM does not change it during the logging of data, obviously this would screw up my results I totally agree with you in that. Then tuning the bike with the locked AFV should tune it to that value. ... but not in that. It's still unknown which AFR the CL/LCL region is targeting at. We *asssume* it might be stoich, but what, if not? Then AFV will change and OL regions too. As AFR is constantly dangling between the rich and lean limit (as given in the O2 sensor data), averaging that to one single number is a bit challenging. I wouldn't touch CL at all, mainly because I see no reason to do so. You will never see a straight line of O2 voltage, because the ECM is not made for doing that. It's (assumably) made to constantly change fuel and so adjust to changing environmental or configuration based changes. The control cycle is working only if dynamic, not static. In my oppinion going open loop is great for learning and comparing different setups and if full control over mixture is a must (but only in conjunction with a WB O2). In day to day use, I would stick to closed loop and adjust OL region where necessary only. Regards, Gunter | ||
| Id073897 |
The AFV itself is calculated off of a 100% baseline ... under certain (standard?) conditions. This constraint is almost always ignored. If the AFV is set to 100 under differing conditions every cell in LCL has to be adjusted (to an AFR target we don't really know) because otherwise AFV will change and all OL settings are invalid. OTOH keeping AFV as is doesn't require any cell in LCL touched, which makes things much easier. Regards, Gunter | ||
| Ferocity02 |
That's some good info Gunter, thanks! Here is where I might see the problem, which you may have touched on. I am going to be using VE Analyzer on my fuel tables and it's going to tune them to how VE Analyzer wants the engine to run. But as soon as I'm done and unlock the AFV, then the ECM and closed loop learn will tune the AFV to how the ECM wants the bike to run. The issue I see with this is how do I know that the AFV and VE Analyzer desire the same thing. My guess is that they don't, and I'm not sure what to do about it. |
