Tune drift

[glens]

Steve Cole said:

"I do not know what they are doing as I haven't the time to look at there stuff, I only look at what has been provided and so far that was from you and Whittlebeast a year or so agao and BOTH of those logs clearly showed improper data being collected and/or displayed."

Steve,

It would be helpful if you could tell us how you were able to identify "improper data being collected and/or displayed".  I'm not doubting you - just trying to learn and PV data collection errors could certainly contribute (or even be the sole cause of) the OPs issue.

Obviously not Steve but it was common, and if things haven't yet changed with PV, still is common to note in the logs combinations of conditions being presented that are at best unlikely.  It seems (seemed?) as though the PV has (had?) a method of collecting data that intermixes current new values in some areas while repeating older values in other areas.  As if the ECM doesn't put out every requested data channel every time it sends data, so the PV repeats last-known values until such time as they get updated.  Which may or may not be at the same time other data points may or may not have gotten updated themselves over several iterations of data report by the PV.

Such things as one reported data frame to the next showing at the same time increased engine speed, decreased throttle opening, and increased manifold pressure.  Just an example off the top of my head.  I looked for patterns in the logs through the various "channels" and think I recall it was something like 6 frames on average between channel updates.  But in the admittedly limited time I spent on it could not nail it down to anything consistent.  My idea was to use some of my unix text utilities to cull all duplicated data from the log, leaving only new values in it, otherwise just a blank space.  But like I said, I didn't see any specific way to determine exactly how many lines of log frames any given channel was actually regurgitating old values.  Even were it possible to cull them, then the problem arises:  how to correlate the remaining data?

If this is merely a function of how the ECM outputs the data it offers up (quite possible if not entirely likely, since its main purpose is to devote its time to controlling the engine, not telling you everything its doing at any given moment), then Steve has got his own method for correlating the data and presenting it in the TTS log files.  Maybe it's "right" and maybe it's not.  I don't recall him ever discussing it in detail and would be surprised if he would.  This ain't "open source" stuff we're talking about.  But I don't recall ever noting any glaring inconsistencies or representations of impossible situations in the TTS logs.

I just previewed this reply before submitting it and see Steve has chimed in and briefly touched on some of this stuff too.

[GregOn2Wheels]

Steve Cole said:

"Look I came to try and help and it not our product screwing up here. Rabos is going off the deep end due to him ONLY trying to tell people what he thinks they want to here, not the whole truth. That will make it all but impossible to let any of you learn let alone try and fix a bike!"



A person can learn a ton of stuff here as long as they're able to filter out the noise. 

[joe_lyons]

Steve how is the blabber about Ron helping with the issue?  OK so lets focus on the number one thing.  This is a homebrew tune,  which in some cases has done perfectly fine and other cases have issues.  I doubt it has anything to do with the tuner.  If it was to have a direct link or SEPST for a tuner would the issue still be a friggin data log.  It all comes down to bad data in and bad tune out, simple as that.  Look at a log with false data from the O2 sensors and it will still be that way but how will you know as it is still just a log.  He may need to just take it in to have a pro get it done.  If that pro can't do it with PV then he needs a new job.

[GregOn2Wheels]

Good information Steve and gleans and I thank you for it.

[glens]

OK so lets focus on the number one thing.  This is a homebrew tune,  which in some cases has done perfectly fine and other cases have issues.  I doubt it has anything to do with the tuner. ... It all comes down to bad data in and bad tune out, simple as that.  Look at a log with false data from the O2 sensors and it will still be that way but how will you know as it is still just a log.

I need a new job, but that's a completely unrelated story...

I agree that a log with false O2 data could well appear to not have bad O2 data.  But the part about "bad data in and bad tune out" (which I also agree with) might not really be the situation here.  At least not the main one.

Although it might not have anything to do with the tuner (hardware/software) itself, if not then it's at least something to do with the way it's being used in whatever relation to how it "says" it works.

If adaptive control is supposed to be disabled in whichever fashion and the "tune goes to hell" yet is restored if only briefly by "clearing trims", then a log file could at least offer something of interest regarding integrators and "VE new", etc.

[rageglide]

I find that SEPST does not reset the trims even when the box is checked,  so I used my PV to reset them.  I think it would be wise for two gentlemen to bow out of this thread as it is just causing a pissing match.  I have found an issues with using quick tune to mess with a calibration that has been setup for basic auto tune(pvat). Other than that anything with vrod has been fine and dandy.  Granted I don't tune a million vrods, but I've tuned my fair share.

Newer version of PV firmware clears out the Quicktune stuff which used to be retained.  Never got a straight answer as to whether existing QT settings would be applied to any flashed map.   But it sure would screw with the results!

One downside to DJ removing the QT settings once applied, how do you confirm what it was set for?  What if you want to add a bit more VE, timing, etc... You better remember what you changed previously or start fresh... Does a new change get added to existing/previous change or does it replace it?   QT seems like a problem waiting to happen... I don't use it anymore.

<edit - minor rewording cuz my grammar sux>

[joe_lyons]

I haven't used it in quite a while either.


Just remember that vrods cool down much quicker than air cooled motors.  With my bike and all the other bikes I have tuned with PV I have not had an issue with a drifting tune or resetting fuel trims and having it make a difference if the adaptive is shut off.  I myself have not looked at the calibration or logs because of time restrictions.  I will see if I can find anything deeper in the calibration that others can't see.

[joe_lyons]

Man, you just can not buy entertainment like this! [emoji14]op:

I'm here all week. [emoji6]

[Steve Cole]

Steve how is the blabber about Ron helping with the issue?  OK so lets focus on the number one thing.  This is a homebrew tune,  which in some cases has done perfectly fine and other cases have issues.  I doubt it has anything to do with the tuner.  If it was to have a direct link or SEPST for a tuner would the issue still be a friggin data log.  It all comes down to bad data in and bad tune out, simple as that.  Look at a log with false data from the O2 sensors and it will still be that way but how will you know as it is still just a log.  He may need to just take it in to have a pro get it done.  If that pro can't do it with PV then he needs a new job.

Joe

I could care less what is gathering the data AS LONG AS IT IS DOING IT CORRECTLY! Since the topic of discussion here is the PV, well guess what, that's what is going to be discussed. IF PV claims to be disabling the Adaptive Fuel then there should not be a Damn thing Rabos or anyone can screw up in the tune to stop it from function as they claim when its turned OFF. So sit back and read what Rabos has said back and forth and there is only one conclusion that any sane person can come too. Either the PV is NOT disabling the Adaptive or Rabos is not telling the whole truth here. You take your pick but no one is going to learn a damn thing until that gets cleared up and off the table. Maybe other logs will be different but so far it 2 for 2 of the PV not working correctly.

[rbabos]

Well the Logs are totally useless then if they don't represent what's happening. I can clearly see the AFF holding at 100 throughout the whole log if switched off even past minimum learn temp. I can see them come on line in the log when minimum learn temp is hit with them switched on. Pretty damn basic I think. Beyond that just about everyone here won't have a clue what the ECM does or could do in the background.
Sorry Joe.
Ron

[Steve Cole]

Well the Logs are totally useless then if they don't represent what's happening.....
Ron

And that there is what has to be determined before any useful discussion and resolution to the problem is going to be found.

[upnorthbiker]

Hi Hrdtail78,
Here goes.
Jeff

I would set closed Loop Minimum and Closed Loop Minimum temp Hyst to 85c.  Your closed loop bias tables are kind of jacked as well.  What were they originaly?

Sorry Hardtail78 just realised you asked a question.
CLB
Front cylinder

RPM                       MAP
               20        35       40       50     70
1200       360      360     420     440   440
2500       360      380     420     440   440 
3000       420      500     460     440   440
4000       440      500     500     440   440
5000       440      480     480     420   440


Rear cylinder

RPM                       MAP
               20        35       40       50     70
1200       360      380     420     440   440
2500       380      420     440     440   440 
3000       380      420     420     420   440
4000       420      440     440     420   440
5000       440      440     380     440   440

Thanks
Jeff

[hrdtail78]

450mv is lambda of 1.  The cal is set up to run the bike leaner than stoich. 

I would suggest something like below.  Yes, it is a direct copy and paste out of TTS 177 cal. 


486   486   486   486   486
525   525   525   525   525
622   622   622   622   622
622   622   622   622   622
622   622   622   622   622

[upnorthbiker]

Thanks Hardtail78, will try your recommend settings in temps and clb over the weekend.
Jeff

[Hilly13]

Who set your CLB tables up upnorthbiker?

[upnorthbiker]

Who set your CLB tables up upnorthbiker?

Hi Hilly13
Factory settings, for both cl temps and clb
My tune posted was standard factory cal that I ran basic autotune on. When it drifted I went open loop using afr of 14.5 and set switches accordingly to turn off closed loop.
That didn't work either so I asked the experts starting this post.
Appreciate the help and advice
Jeff

[rbabos]

Just as another example. This seems to work well for me. Cooler idle, less fan times, addresses the light load areas and decent cruise mpg. 4k cruise could be leaned to 450 without issues.
Ron

[FLTRI]

Just as another example. This seems to work well for me. Cooler idle, less fan times, addresses the light load areas and decent cruise mpg. 4k cruise could be leaned to 450 without issues.
Ron

Hey Ron, I have a quick test for you to try...if you're willing of course.
Try setting your CLB tables to a bit above 700.
Let us know what you discover as far as run quality.
Thanks,
Bob
PS- Please no preconceived notions.

[rbabos]

Just as another example. This seems to work well for me. Cooler idle, less fan times, addresses the light load areas and decent cruise mpg. 4k cruise could be leaned to 450 without issues.
Ron

Hey Ron, I have a quick test for you to try...if you're willing of course.
Try setting your CLB tables to a bit above 700.
Let us know what you discover as far as run quality.
Thanks,
Bob
PS- Please no preconceived notions.

Bob: I think you should, and then you'd discover what I did. I did that a long time ago testing "Potty mouth". This has nothing to do with the tune drift thread.
Ron

[FLTRI]

So what happened to the run quality when you enrichened the closed loop?
Just curious how the change felt to you...
Could have something to do with drift and I do this test frequently.
Bob

[rbabos]

So what happened to the run quality when you enrichened the closed loop?
Just curious how the change felt to you...
Could have something to do with drift and I do this test frequently.
Bob

The same as it did for you.   Cmon Bob, don't matter what I say, it's open to attack and usually is.
So, post your results and I'll see if I agree or not. It's my turn.
Ron