Tune drift

[rbabos]

Not in the public software. I could not see any reason for it before and I still don't. People will just get confused since they cannot have all the pieces of the puzzle and showing you only 1 part of what it takes to make it work with no explanations, just doesn't buy you anything but confusion.

I agree. Any more I'm not sure even placing a whole lot of faith in integrator readings is all that accurate either. More a general guideline . If O2 sampling in the light load is wrong, none of the figures in adaptive or integrators can be trusted. Might be wrong but that's the impression I'm getting with diy'r tuning.
Ron

[glens]

Steve does Tts have a way to read the adaptive fuel values?

Not in the public software. I could not see any reason for it before and I still don't.

I suppose I could be not lazy and look for it specifically, but the impression I've had is that you'd said there was no way to pull the values...

[Steve Cole]

Steve does Tts have a way to read the adaptive fuel values?

Not in the public software. I could not see any reason for it before and I still don't.

I suppose I could be not lazy and look for it specifically, but the impression I've had is that you'd said there was no way to pull the values...

There is no way to just read what the ECM is doing. I'm not even sure that the values displayed are the right thing to start with, but with that said, anything can be done if you want to spend enough time and effort to do it. The problem is as I stated, at best you are seeing something on a screen shot, you have no idea what it means nor a way for you to test it, use it for testing, so all it does is confuse people. No clear description of what and how it works, leads people to makeup stories and it's off and running on the internet.

There are a lot of smart people out here and they, if given the time, could and would, like to work there way through it.

That said, there are fewer of those people than there is the ones that just want to have something that you believe you just plug it in and it fixes everything for you. So we as a manufacture have to do our best to make sure we provided functions that work, provide the supporting documents and test procedures to use it. If we cannot do that then we are not going to do it. Others just toss pieces of it out there and say have at it, then the internet takes over these days.

Yes, I know that sounds terrible but it is the truth.

Joe is in deep with DJ and here he is asking too, so you tell me why he isn't on here explaining exactly how it all works and how to really use and test with the information in the PV itself? This isn't a shot at Joe, it just shows the way it really is. These are just not cut and dry simple things and much more information is needed than what has been displayed for it to be of any use at all.

You have a bunch of values on a screen and from what I know nothing to tell you what it really means, how to use the information, when the ECM maybe using it and how it is being used. So you all tell me how useful it really is to the end user let alone a tuner?

For the record there is still a fair amount of things in the ECM that I'm not even 100% of why they are doing what they are doing and I've been working with the stuff full time for the past 30 + years. Then, there are the clear mistakes in the code as well but that is just the way it is for all of us.

[Steve Cole]

Not in the public software. I could not see any reason for it before and I still don't. People will just get confused since they cannot have all the pieces of the puzzle and showing you only 1 part of what it takes to make it work with no explanations, just doesn't buy you anything but confusion.

I agree. Any more I'm not sure even placing a whole lot of faith in integrator readings is all that accurate either. More a general guideline . If O2 sampling in the light load anywhere is wrong, none of the figures in adaptive or integrators can be trusted. Might be wrong but that's the impression I'm getting with diy'r tuning.
Ron

Fixed it for you. It all boils down to getting good readings and I've been saying that to you all for years. You also have to know and understand how it all fits together to use any information provided or your just guessing and just because someone tries to say I tuned it from Broad Bands doesn't mean there sampling is any better and for the most part it is worse. Your never going to get it to stay completely still or a perfect rock steady mixture unless you can control mother nature, oil companies and on and on.............

[joe_lyons]

Steve does Tts have a way to read the adaptive fuel values?

Not in the public software. I could not see any reason for it before and I still don't.

I suppose I could be not lazy and look for it specifically, but the impression I've had is that you'd said there was no way to pull the values...

There is no way to just read what the ECM is doing. I'm not even sure that the values displayed are the right thing to start with, but with that said, anything can be done if you want to spend enough time and effort to do it. The problem is as I stated, at best you are seeing something on a screen shot, you have no idea what it means nor a way for you to test it, use it for testing, so all it does is confuse people. No clear description of what and how it works, leads people to makeup stories and it's off and running on the internet.

There are a lot of smart people out here and they, if given the time, could and would, like to work there way through it.

That said, there are fewer of those people than there is the ones that just want to have something that you believe you just plug it in and it fixes everything for you. So we as a manufacture have to do our best to make sure we provided functions that work, provide the supporting documents and test procedures to use it. If we cannot do that then we are not going to do it. Others just toss pieces of it out there and say have at it, then the internet takes over these days.

Yes, I know that sounds terrible but it is the truth.

Joe is in deep with DJ and here he is asking too, so you tell me why he isn't on here explaining exactly how it all works and how to really use and test with the information in the PV itself? This isn't a shot at Joe, it just shows the way it really is. These are just not cut and dry simple things and much more information is needed than what has been displayed for it to be of any use at all.

You have a bunch of values on a screen and from what I know nothing to tell you what it really means, how to use the information, when the ECM maybe using it and how it is being used. So you all tell me how useful it really is to the end user let alone a tuner?

For the record there is still a fair amount of things in the ECM that I'm not even 100% of why they are doing what they are doing and I've been working with the stuff full time for the past 30 + years. Then, there are the clear mistakes in the code as well but that is just the way it is for all of us.

I guess the way to test would be to not reset adaptive and shut off closed loop and go ride the bike and record what the long term/adaptive values are.  Plotting them on a chart through MLV would be an easy way to look at this.  There has to be some stored memory that keeps the adaptive values so they can be used upon next run cycle.  When I get the chance (if). I will try to get some more data on this and see how it pans out.

[whittlebeast]

Wow, it must be really tough to deal with a system and data to you do not trust to be accurate.

[glens]

He's talking about PV, Andy.

[whittlebeast]

Glen

Who is "He"?

Glen, have you ever verified that the reported timing as reported in the logs is what's really happening at the crank?   Is the reported pulse width whats going on at the injector?  What about the warm up logic?  Is that what's going on at the injectors?

After reading some of these posts on this thread and any number of other threads, what can you trust on any of this stuff?

In the Megasquirt world, lots of people question everything and figure out ways to verify the thing in question.  If something does not check out and can be documented and duplicated, the programmers are all over it.

Andy

[Onthefence]

How do you get the PV to display the AFF values.?  I guess ML viewer could show the same info.

[whittlebeast]

As I recall, you simply have to turn them on in the logger screen.  They plot right up in MLV.

It has been a wile since I messed with the PV display.....  My bike is back on narrow bands and I moved the wide band on to a different ride.

Andy

[whittlebeast]

Tune drift is really easy to "see".

Either scatter plot X = RPM, Y = MAP and Z = The Long term fuel correction.  One company calls the AFF and the other calls it AFV as I recall.

or turn on as an optional field, MAPxRPM and DutyCycle1 and then plot

X=MAPxRPM, Y = DutyCycle1 and Z = The Long term fuel correction

Look for changes over time.

Andy

[FLTRI]

Tune drift is really easy to "see"...
Andy

Ok, let's use this senario...
Guy comes to you because he heard you can see tune drift really easily...
He's on his way for a long ride/vacation.
You have a couple hours to tune his bike and determine tune drift, if it exists.
He and his bride have left while you tune the bike.
Can you show us the tune drift and the fix?
You can do whatever you need to to get the bike tuned but you only have a couple hours to accomplish it.
Can you do it? And how?
Thanks,
Bob

[whittlebeast]

Interesting that you mentioned a couple of hours.  I tune with statistics and that simply take data logging time.  If we assume that the guy has data off his bike that includes the fuel trims, the first thing I would do is plot data as I mentioned above.  What I am looking for is anything that is non linear when the motor is in closed loop.  Some motors, it is a simple as turning closed loop where the O2 sensors are chasing bad data.  The long term trims will be all over the place with no apparent reason.  Personally, I have only seen this a couple of times.  It jumps right out at you.

I most commonly see it when people try to run closed loop with injectors running around 95% DC.

Andy

[FLTRI]

Interesting that you mentioned a couple of hours.  I tune with statistics and that simply take data logging time.  If we assume that the guy has data off his bike that includes the fuel trims, the first thing I would do is plot data as I mentioned above.  What I am looking for is anything that is non linear when the motor is in closed loop.  Some motors, it is a simple as turning closed loop where the O2 sensors are chasing bad data.  The long term trims will be all over the place with no apparent reason.  Personally, I have only seen this a couple of times.  It jumps right out at you.

I most commonly see it when people try to run closed loop with injectors running around 95% DC.

Andy

As I asked in the prior thread, can you show us an graphed example of tune drift and the resultant fixed graph.
That way we can actually relate to what you are referring to when you say it's easy to see tune drift. Then please show how you fixed it.
Bob

[whittlebeast]

I don't recall having one off a Harley and this crowd tends to get sensitive if I show data off anything else.  We had a thread earlier that had a motor the was non linear, but that was data off a dyno in open loop.  It would look real similar.

[Onthefence]

As I recall, you simply have to turn them on in the logger screen.  They plot right up in MLV.

It has been a wile since I messed with the PV display.....  My bike is back on narrow bands and I moved the wide band on to a different ride.

Andy

I guess I wasn't very clear on what I was asking.  My phone wouldn't let me quote for some reason.  I was trying to figure out how 2006fxdci was displaying the AFF info on the PV display earlier in this thread.

[whittlebeast]

Do you have a fairly big log off you bike that we can snoop around at?  Let's see what it is doing.

[Steve Cole]

Steve does Tts have a way to read the adaptive fuel values?

Not in the public software. I could not see any reason for it before and I still don't.

I suppose I could be not lazy and look for it specifically, but the impression I've had is that you'd said there was no way to pull the values...

There is no way to just read what the ECM is doing. I'm not even sure that the values displayed are the right thing to start with, but with that said, anything can be done if you want to spend enough time and effort to do it. The problem is as I stated, at best you are seeing something on a screen shot, you have no idea what it means nor a way for you to test it, use it for testing, so all it does is confuse people. No clear description of what and how it works, leads people to makeup stories and it's off and running on the internet.

There are a lot of smart people out here and they, if given the time, could and would, like to work there way through it.

That said, there are fewer of those people than there is the ones that just want to have something that you believe you just plug it in and it fixes everything for you. So we as a manufacture have to do our best to make sure we provided functions that work, provide the supporting documents and test procedures to use it. If we cannot do that then we are not going to do it. Others just toss pieces of it out there and say have at it, then the internet takes over these days.

Yes, I know that sounds terrible but it is the truth.

Joe is in deep with DJ and here he is asking too, so you tell me why he isn't on here explaining exactly how it all works and how to really use and test with the information in the PV itself? This isn't a shot at Joe, it just shows the way it really is. These are just not cut and dry simple things and much more information is needed than what has been displayed for it to be of any use at all.

You have a bunch of values on a screen and from what I know nothing to tell you what it really means, how to use the information, when the ECM maybe using it and how it is being used. So you all tell me how useful it really is to the end user let alone a tuner?

For the record there is still a fair amount of things in the ECM that I'm not even 100% of why they are doing what they are doing and I've been working with the stuff full time for the past 30 + years. Then, there are the clear mistakes in the code as well but that is just the way it is for all of us.

I guess the way to test would be to not reset adaptive and shut off closed loop and go ride the bike and record what the long term/adaptive values are.  Plotting them on a chart through MLV would be an easy way to look at this.  There has to be some stored memory that keeps the adaptive values so they can be used upon next run cycle.  When I get the chance (if). I will try to get some more data on this and see how it pans out.

One needs to separate these various items/terms.

Adaptive Fuel Values (AFV) are stored at there last good value in the ECM when the bike is shut down. As long as the engine is running they can and do change! The ECM determines when to use them, when not to use them, where to use them and when to update them or if they are updated and used at all. You cannot assume that because there is a stored value it is correct or being used for any one condition!

There is nothing else (adaptive wise) stored after the ECM is shut down.

So the integrator is running real time and once you shut the ECM off, it's gone and once you change the operation of the engine at all, when running it, can and does change rapidly.

Since you donot have all those pieces, as they happen, your pissing in the wind and hoping not to get wet.

[whittlebeast]

That is like saying weighing yourself at 6 AM is not exactly what you weigh at 6 PM as exercise or just eating a big meal has an effect.  The conclusion has to be stop weighing yourself, you won't understand it.   

If you are 6 ft and 340 lbs in the morning, you may be 339.8 in the afternoon, but you are still fat.

Andy

[2006FXDCI]

Heres how i pulled the adaptive tables using my PV . Go to settings then code and type in 924107 then save . Then click on read adaptive table then select .

[Tsani]

That is like saying weighing yourself at 6 AM is not exactly what you weigh at 6 PM as exercise or just eating a big meal has an effect.  The conclusion has to be stop weighing yourself, you won't understand it.   

If you are 6 ft and 340 lbs in the morning, you may be 339.8 in the afternoon, but you are still fat.

Andy

Actually, this is quite true. But my weight can vary by tens pounds or more. So I need to know when to drink.
The bike may be running fat. It may be running lean. Depending on what the data is telling you at that moment, you may be able to correct it.
But alas, there is really no point.
Humans never really experience the present.
Every human being is living at least 80 milliseconds in the past. Our consciousness lags behind actual events and that when you think an event occurs, it has already happened before your brain has a chance to create a cohesive picture of the world.
So in reality, give up, you will never get it tuned.

[FLTRI]

...So in reality, give up, you will never get it tuned.

Yep, especially if you try to get the tune absolutely perfect and think there is no need for adaptive fueling to be enabled.
Bob

[Hilly13]

It comes down to sampling, if its crap your gunna chase your tail, if I had $1 for ever time someone has said that on here alone I could pay cash for a new bike easy, sure there are terrible builds that defy logic and make tuning near on impossible but most cases its down to the old garbage in garbage out narrow band wide band TTS SEPST PV or whatever. Tune perfection is a dream, to many variables beyond your control, I set the closed loop area to cruise alone on lumpy cams and no drift, this is my simplistic view anyhow.
Get the sampling right Josh, your task will be easier from there.

[whittlebeast]

And the ability to know when you have one of those situations.  I use closed loop when it is in my best interest and open loop where that is better.   The trick is knowing how to identify the difference.

Andy

[rbabos]

...So in reality, give up, you will never get it tuned.

Yep, especially if you try to get the tune absolutely perfect and think there is no need for adaptive fueling to be enabled.
Bob

Tell me Bob, how does one with DIY tuning ever get it prefect with using basic crap at had to use. By crap, I mean where sensors are located the process is doomed from the start.
Perfect does not exist in tuning and once I replaced that word with acceptable or decent I had a few strands of hair grow back. Turning the adaptive off, hell , almost better then Rogane.
I'm still trying to figure out why a stock bike even though not the best tune, drift is rare. It continues to run the same all the time, day after day. It's when it's reflashed and tuned that all these problems seem to happen. Same sensor location too. Something is really weird with this stuff. Take my v rod cal for example. If I was to cut an paste the clb tables out of the stock cal and place them into the current tune with higher ve tables, I doubt it would run well enough to get to the street it would be so lean. Yet, even with lower ve's and clbs in the 350-400s in the stock cal it seems to run half decent, although leaner , yet no tune drift. Tune the bike, have it run super smooth, it tries to lean itself out in light load with adaptive enabled. Drives me nuts. I gave up and left the adaptive off. No issues since. However, all normal areas in the cal are left closed loop, also with no issues I can hear or feel.
Ron