Scatter Plots for Beginners

[whittlebeast]

Here is a timing map off a Toyota truck.  Same sort of operating range as a Harley.

http://www.nbs-stl.com/HarleyTuning/Toyota%20Timing%20Map.png

Keep in mind that Mass Air Flow is almost identical to MAPxRPM and you guys have MAP and RPM available to you in the Harley world.  In the Toyota world Engine Load is MAP

Point is..... all this stuff has patterns.  A happy motor give a happy rider.  You guys have the tools now to see all this and the patterns in living color.

Have fun tuning

Andy

[Tsani]

/quote]

Lets. Start a thread on it. There is nothing in the documentation about it.

[FLTRI]

Anybody want to talk about IAC Warmup Analyzer yet?

Been doing the beta testing on that feature!
Bob

[whittlebeast]

Here is how MAPxRPM compares to a factory Toyota MAF

http://www.nbs-stl.com/HarleyTuning/Toyota%20MAPxRPM%20vs%20MAF.png

And to think, that is 2 samples per sec. over CAN.

It is all just math.

Andy

[whittlebeast]

Here is a fun plot.  On the left is the percent of the fuel coming from the front injector based on the reported Pulse Width front and rear

on the right is what the percentages would have been on a pure speed density system based on the VE Tables Front and Rear and the MAP, RPM, Desired AFR and VE

These plots should have been essentially identical.  They are in the same scale.  Color is just for reference for the two plots.

http://www.nbs-stl.com/HarleyTuning/PCT%20Fuel%20Front%20Per%20VE%20Table%20and%20BPW.png

Andy

[Steve Cole]

That should clearly show you and everyone else your missing something in your math or scatter plots! Kind of the same deal as Tsani plotting two recordings and they are find yet when he lets MLV view plot them together the results are different. Clear as a bell that something is going wrong when it combines them!

[whittlebeast]

OK. Does the BPW include the injector dead time?

[Steve Cole]

Since you are leading everyone to believe that you know the system, why would you be asking me that? Look, you need to get out and do the work before you try and teach everyone else how to do it. Understanding the system you are working  with has to come first! If you do not do that then quit trying to teach people things that you do not know.

Tsani's issue has nothing at all to do with BPW or dead time, so what are you doing to resolve his issue where MLV plots each recording fine, yet once he allows it to combine recordings it shows different things that either single previous plotting of the exact same data does. That's a issue in MLV and needs to be looked into before anything else.

[whittlebeast]

I will take that as an "I don't know"  Has anyone scoped the injector signal and compared the dwell time at the scope to the PW in the logs?

Andy

[Just Nick]

This topic needs to die
Andy you still have yet to show a bust in tune and fix of said bust. Mlv seems to make pretty little pictures and that is it.

[Steve Cole]

I will take that as an "I don't know"  Has anyone scoped the injector signal and compared the dwell time at the scope to the PW in the logs?

Andy

Same old Andy! After 4 years of trying to get you to really go look at the HD system you still refuse, maybe others are now seeing it too and that is why YOU refuse to answer any direct questions. To those of you watching/reading and following what he says it will get you into trouble and it's clear that MLV has some issues that need to be solved.

Tsani

I think if you took Andy out of the picture and went to the people at MLV with the data to show the issues they would jump in and try to solve it but Andy is only going to muddy the waters.

[Tsani]

Steve,
After putting hours and hours of data thru MLV, that was the first time that happened.
The data collected has been from a on going tune tweek, small changes.
That was an anomaly. But I am pretty sure its not the data.
There was a recent update to MLV.
Combine that with what showed, I don't trust it.
I think I will just stick with what I know works.
As I said, I can use the graphing in Datamaster to show me the trends.
I was willing to give this a fair look but it keeps going in circles and questions don't get answered
unless you want to consider a unlabeled graph a answer.
And BTW Andy, that Toyota truck is not even close to the operating parameters of a HD big twin.
I am officially done with this mess.

[whittlebeast]

It as been quite some time since I have seen a basic code bust in MLV.  If you really found one, submit the data and a screen shot of the evidence and the guys at MLV normally jump on that sort of thing real quick.

Abnormalities and the root cause are chased down rather quick.

Andy

[Onthefence]

I think I am catching onto this MAP x RPM and finding busted tunes.  I believe this one is really jacked up.

[PoorUB]

I will take that as an "I don't know"  Has anyone scoped the injector signal and compared the dwell time at the scope to the PW in the logs?

Andy

Same old Andy! After 4 years of trying to get you to really go look at the HD system you still refuse, maybe others are now seeing it too and that is why YOU refuse to answer any direct questions. To those of you watching/reading and following what he says it will get you into trouble and it's clear that MLV has some issues that need to be solved.

Tsani

I think if you took Andy out of the picture and went to the people at MLV with the data to show the issues they would jump in and try to solve it but Andy is only going to muddy the waters.

Heck, even I don't know most of what you guys are talking about most of the time I even I have figured out Andy likes to post pretty pictures and talk tech terminology B.S. with no real information!

I just like reading the threads to see if he will post anything we can really use!

Andy reminds me of a guy years ago from a ATV forum that could recite terminology about building engines, turbo charging, tuning, the whole deal. Then he actually raced jetski in the World Finals on Lake Havasu and bragged about coming 15th place, racing against riders with much more racing experience. Then the whole thing fell in on him. Someone started piecing information together on him and finally tracked him down to his beater truck and trashed out trailer house. Then someone else happened to be at the World Finals and yes, he finished 15th place, dead last, 2-3 laps down. The only other rider that placed farther down was a DNF.

I gotta wonder if this is the same person??

[ToBeFrank]

Then he actually raced jetski in the World Finals on Lake Havasu and bragged about coming 15th place, racing against riders with much more racing experience. Then the whole thing fell in on him. Someone started piecing information together on him and finally tracked him down to his beater truck and trashed out trailer house. Then someone else happened to be at the World Finals and yes, he finished 15th place, dead last, 2-3 laps down. The only other rider that placed farther down was a DNF.

OT, but I would think just getting to the world finals would be a huge achievement, or do they let anyone into the world finals?

[PoorUB]

OT, but I would think just getting to the world finals would be a huge achievement, or do they let anyone into the world finals?

I think they would let anyone that showed up race.

[Jamie Long]

We like to use as many tools and resources available when tuning and scatter plots are far more than just pretty pictures. Using a scatter plot histogram can work really well if you know what you are looking at, it allows you to take a look at the big picture before playing back the log, searching for an issue blindly or having to look frame by frame. A cool feature in MLV if you see an area or cell in the scatter histogram but would like to view it in the chart you can click on a cell in scatter and you can switch screens and go directly to this frame on strip chart and work from there. You can also search/jump to min/max values or use search query's such as greater than/less than or a specific value.


Here is one example where a scatter plot histogram allowed you to very quickly diagnose an issue for a customer remotely. Note front & rear closed loop integrator was chasing its tail badly, swinging sharply from 20% lean to 20% rich. This is a sign the O2 sensors mixed up front to rear. This could then be verified by referring to the strip chart, the CL multipliers were applying corrections inversely.




Here is a histogram after the sensors were switched, see any difference?

[GregOn2Wheels]

I think I am catching onto this MAP x RPM and finding busted tunes.  I believe this one is really jacked up.

That's hilarious!  So that dude there is Jack?

[azlou66]

Jamie
Thank you Thank you and Thank you.

[FLTRI]

We like to use as many tools and resources available when tuning and scatter plots are far more than just pretty pictures. Using a scatter plot histogram can work really well if you know what you are looking at, it allows you to take a look at the big picture before playing back the log, searching for an issue blindly or having to look frame by frame. A cool feature in MLV if you see an area or cell in the scatter histogram but would like to view it in the chart you can click on a cell in scatter and you can switch screens and go directly to this frame on strip chart and work from there. You can also search/jump to min/max values or use search query's such as greater than/less than or a specific value.


Here is one example where a scatter plot histogram allowed you to very quickly diagnose an issue for a customer remotely. Note front & rear closed loop integrator was chasing its tail badly, swinging sharply from 20% lean to 20% rich. This is a sign the O2 sensors mixed up front to rear. This could then be verified by referring to the strip chart, the CL multipliers were applying corrections inversely.




Here is a histogram after the sensors were switched, see any difference?

Thanks Jamie,
You did in one post what Andy couldn't do in hundreds of posts! Lol
Bob

[Jamie Long]

Here is the same data from the above scatter plot histogram displayed as a strip chart and zoomed in to a problem area.

In these examples the bike is under mild accelleration in both frames. Look at the difference between these charts, first chart is with the O2's mixed up the second with them corrected. The first thing that stands out is CLI, you can clearly see in first chart that CLI is driven inversley proportional. You can also note this same effect with the front & rear injector PW as well. Note how stable these values are in the second chart.

O2 sensors mixed up



O2 sensors correct

[azlou66]

This stuff makes A LOT more sense with a bit of explaining.
Thanks Andy oh wait thanks Jamie.

[Just Nick]

Thanks Jamie


Andy Look Jamie a guy that does not even sell MLV just jumped in and Explained for us what we have been asking you for 4 years to show us. Where did you go Andy ? Why couldn't you just get down to the meat and taters like we ALL kept asking and asking and asking.

[Sunny Jim]

Well , as a learner tuner DIY person, I read all my PV data logs on MLVHD , a bit like the second lot of graphs Jamie posted.
It's great. Reasonably Easy to read and adjustable to your requirements. ( color, subject matter, time frames )
And INEXPENSIVE.