I am trying to write an EFI general tuning guide....

[whittlebeast]

That is why EFI tuning is a big game of statistics and probability under the covers.  You are basically watching all of the sensors and saying the last 3000 times the motor was in this combination of sensor readings, the motor needed X amount of fuel to hit the target AFR so X is the best guess the next time this combination comes up.

Andy

[Coyote]

Ron is pointing out that in CL, if the bike ain't happy, no one is happy. And you can't just change the VEs to get what you want as the ECM will just undo it with corrections. You either have to fix the sampling error or adjust the target afr.

[rbabos]

Ron is pointing out that in CL, if the bike ain't happy, no one is happy. And you can't just change the VEs to get what you want as the ECM will just undo it with corrections. You either have to fix the sampling error or adjust the target afr.

Yes, that's what I was saying. Try as I might, I've never been able to outsmart that damn ECM for long.
Ron

[rigidthumper]

Ron is pointing out that in CL, if the bike ain't happy, no one is happy. And you can't just change the VEs to get what you want as the ECM will just undo it with corrections. You either have to fix the sampling error or adjust the target afr.

Yes, that's what I was saying. Try as I might, I've never been able to outsmart that damn ECM for long.
Ron

[whittlebeast]

Yep, here is what things look like when closed loop tuning goes really bad.

http://www.nbs-stl.com/pc5wb/MAPxRPMIssue.png

This should be a relatively straight line.

Here is fueling by Motec but running open loop and logged on the same PC5 as the first plot.

http://www.nbs-stl.com/motec/MAPxRPM%20vs%20DC%20-%20AFR%20Bosch.png

And here is the same motor running a Motec with the Motec high speed data logging.

http://www.nbs-stl.com/pc5wb/MAPxRPMIssue.png

Andy

[Ohio HD]

Ron is pointing out that in CL, if the bike ain't happy, no one is happy. And you can't just change the VEs to get what you want as the ECM will just undo it with corrections. You either have to fix the sampling error or adjust the target afr.

   

[whittlebeast]

Does anyone here have a log of a bike doing this?  Evidence of tune drift caused by bad sampling?

[rbabos]

Does anyone here have a log of a bike doing this?  Evidence of tune drift caused by bad sampling?

Hmm. I'd think bad sampling would give you a bad ve to start with. This can also be from reversion which sets the table wrong, not just O2 location.  Either can start the problem with ill running or the so called tune drift.  Then comes along the magic finger and software to fatten certain areas up in that ve area of operation that never seemed to be high enough in the first place. Bike runs fine until some short term or long term learn comes into play and it again goes lean. Reset trims and it's back to normal for a brief period because it is now functioning on the richer ve edits. What you will see at first is the integrators in the low- mid 90s with a good running bike. Adaptive will be in the 90s as well. Over time it works it's way up to 100 and the bike is starting to run like a  Ok, so what's that tell you? Sampling for whatever reason created the ve too low to run smooth and you are stuck with it, thanks to the ECM and the sensor feedback it receives. It's a no win situation.  Time to move in another direction at that point.  Corrective action is to open loop that area and hit the afr table. It is now stable or can be made so with the correct afr.
I do not think real tune drift can happen if the ve table is correct to start with. Be carefull how you blend and extend cells after the tune. It can cause issues as well.
Ron

[Karl H.]

Yep, here is what things look like when closed loop tuning goes really bad.

http://www.nbs-stl.com/pc5wb/MAPxRPMIssue.png

This should be a relatively straight line.

Here is fueling by Motec but running open loop and logged on the same PC5 as the first plot.

http://www.nbs-stl.com/motec/MAPxRPM%20vs%20DC%20-%20AFR%20Bosch.png

And here is the same motor running a Motec with the Motec high speed data logging.

http://www.nbs-stl.com/pc5wb/MAPxRPMIssue.png

Andy

Andy

The first and the last link show the same chart!?

Karl

[whittlebeast]

Yep, here is what things look like when closed loop tuning goes really bad.

http://www.nbs-stl.com/pc5wb/MAPxRPMIssue.png

This should be a relatively straight line.

Here is fueling by Motec but running open loop and logged on the same PC5 as the first plot.

http://www.nbs-stl.com/motec/MAPxRPM%20vs%20DC%20-%20AFR%20Bosch.png

And here is the same motor running a Motec with the Motec high speed data logging.

http://www.nbs-stl.com/motec/MAPxRPM%20vs%20DC%20-%20TPS%20Motec.png

Andy

Fixed

[glens]

I followed the link to the document when you asked a few posts back about understanding the image at the top of page four.  Here's a couple of annotated screenshots from when I was scrolling down to page four.

I didn't get any further than that...

[whittlebeast]

What are you questioning in the big circle?  edit:Never-mind, I just caught it.....

[dynablack07]

Very helpful & detailed, I like this site for this very reason.

[whittlebeast]

A question for you guys that tune with TTS and a dyno, how do you guys verify that the VEs are spot on? Are you using O2 eliminators and then use widebands to feed data back to some sort of data logger/autotune to generate VE New? When VE New is fairly close to VE, you return to narrow bands?

Just asking questions

[lonewolf]

With the Green TTS you use both at the same time. Watch them both with datamaster on the monitor. Also watch the integrators.

[glens]

It's been a while, but I don't recall TTS having/using "VE New" for several years.

[whittlebeast]

With the Green TTS you use both at the same time. Watch them both with datamaster on the monitor. Also watch the integrators.

[/b][/i]

Can those be seen on the screen but not logged?  The long term and the short term integrators?

[Coyote]

It's been a while, but I don't recall TTS having/using "VE New" for several years.

But it is in their logs I believe. Use, IDK

[masstch]

Trying to catch up, here. I just downloaded MLVHD (trial period for now) but have already hit a snag. I can't get it ti run the PV .csv log files, reporting "entire file not loaded"
tiny log or large one, no joy yet.
can open them in excel and even .txt and they *look* normal.

I'm missing a step...

[whittlebeast]

The unregistered version is limited to like 500 records.  It is designed for people to see if it will read their logs before they purchase the SW.

[masstch]

I'll have to look at the entry counts but I have some tiny log files that it won't open.
I would've assumed that if it was a permissions issue that the software would tell me "purchase to do this" but it doesn't.
I'm not sure that's what's happening.
So the answer would be "no, it won't run my files."?



Sent from my iPhone using Tapatalk

[whittlebeast]

Post the log that you are trying to load.  It will not read a PV autotune file but it will read a normal PV log file.  That may be what you are hitting.

[Durwood]

I'll have to look at the entry counts but I have some tiny log files that it won't open.
I would've assumed that if it was a permissions issue that the software would tell me "purchase to do this" but it doesn't.
I'm not sure that's what's happening.
So the answer would be "no, it won't run my files."?

masstch, more than likely it's a setting issue. Power up PV, go to Datalog>Datalog settings and be sure it is set to "Simple" not Pro XY.

[masstch]

I'll have to look at the entry counts but I have some tiny log files that it won't open.
I would've assumed that if it was a permissions issue that the software would tell me "purchase to do this" but it doesn't.
I'm not sure that's what's happening.
So the answer would be "no, it won't run my files."?

masstch, more than likely it's a setting issue. Power up PV, go to Datalog>Datalog settings and be sure it is set to "Simple" not Pro XY.

I'm betting that's it. Its on a different machine, not in front of me, but I recall from looking at text from the .csv that the phrase "pro-x-y" appeared in the log file.

Yes, it's the logfilexxx.csv that I'm trying to read.

If I get a chance I will get the laptop in here for a look.

[Durwood]

Let us know what you find out.