NefMoto

Technical => Data Logging => Topic started by: louie on January 26, 2012, 12:50:35 AM



Title: ECUxPlot Calc data way off
Post by: louie on January 26, 2012, 12:50:35 AM
Attached is the capture log file and the output from ECUxPlot (using latest version 0.9r3.0).  Obviously the Calc HP/TQ numbers are way off.  I loaded the right profile for my car (B5S4) in ECUxPlot, did not change any constants or fueling values.

The boost graph looks reasonable, even the Calc boost numbers for actual PR and desired PR

1. How are the Calculations done for all the other stuff?  HP/TQ numbers are way off, why?
2. Am I loggin the right variables?
3. What is the filter data option? if I click it the graph goes away
4. I see everybody talking about FATS. What the hell is it? The show FATS window under option just shows me an empty window.

Thanks for the help, searching the forums didn't help me much.


Title: Re: ECUxPlot Calc data way off
Post by: Tony@NefMoto on January 26, 2012, 12:10:39 PM
FATS - For Advancement of The S4 - Is a number based on third gear acceleration.

nyet can probably answer your questions better than I.


Title: Re: ECUxPlot Calc data way off
Post by: nyet on January 26, 2012, 01:03:31 PM
The filter option searches for a WOT run of a certain length in 3rd gear. If you don't meet that criterion, you won't see any data if you enable the filter. You can adjust the filter parameters in the filter menu.

If you aren't in 3rd (or aren't logging "gear") or aren't WOT (or logging accelerator pedal position), the filter doesn't work that well.

If your run doesn't start low enough, or isn't long enough, the filter won't let any data through.

FATS is ET from 4200-6500 in 3rd.


Title: Re: ECUxPlot Calc data way off
Post by: louie on January 26, 2012, 01:14:07 PM
Well that explains it.   The data sample I was trying to plot was a 2nd gear pull.  I'll go log 3rd gear pulls and see what the graph looks like.  I am loggin both gear and accelerator pedal position as well as a bunch of other crap.


Title: Re: ECUxPlot Calc data way off
Post by: nyet on January 26, 2012, 01:25:23 PM
btw, you can use 2nd gear data if you adjust rpm_per_mph and change the filter to 2nd instead of 3rd..

however, hp/tq isn't as accurate for very short runs due to inertial effects.


Title: Re: ECUxPlot Calc data way off
Post by: e_pacman on February 09, 2012, 12:45:18 PM
I have a very strange problem with ECUxPlot, could be because I'm missing something really simple. It just won't plot anything. The axis show up as I choose them in the drop down menus, but nothing is plotted on screen. It's the same on two different computers And the CSV file is straight from ME7 Logger.  Any ideas?


Title: Re: ECUxPlot Calc data way off
Post by: nyet on February 09, 2012, 03:05:22 PM
Disable the filter, or adjust the filter so it doesn't filter everything.

http://nefariousmotorsports.com/forum/index.php?topic=1450.msg13765#msg13765


Title: Re: ECUxPlot Calc data way off
Post by: e_pacman on February 10, 2012, 07:47:59 AM
Ah, it was that simple. Thanx a lot!  :)


Title: Re: ECUxPlot Calc data way off
Post by: julex on February 14, 2012, 12:32:24 PM
nyet, maybe the program needs some pop up when you start it for the first time telling people they need to do pulls in 3rd gear for this to work... It seems to be #1 reason why people tell you that your software plots wrong even though it is right since they do pulls in 1st or reverse :)



Title: Re: ECUxPlot Calc data way off
Post by: nyet on February 14, 2012, 12:34:16 PM
for SURE i totally agree.

i was working on a little "helper" dialog that explains why the filter didn't see anything, including advice on how to set the filter, but never finished it.


Title: Re: ECUxPlot Calc data way off
Post by: julex on February 15, 2012, 08:07:01 AM
Can you add "Save Preset" option which would save current graph settings? This is really the only thing I find missing from it currently.


Title: Re: ECUxPlot Calc data way off
Post by: nyet on February 15, 2012, 11:46:32 AM
that and "export all" :)


Title: Re: ECUxPlot Calc data way off
Post by: julex on February 20, 2012, 02:04:46 PM
I will test it today. Thanks Nyet!


Title: Re: ECUxPlot Calc data way off
Post by: nyet on February 20, 2012, 02:58:57 PM
Also, Julex, if there are specific things you want it to be able to detect as "known" columns but they are named differently from ECUx or ME7L (e.g. from VCDS logs) let me know.


Title: Re: ECUxPlot Calc data way off
Post by: julex on February 21, 2012, 11:03:38 AM
So far so good, it's been a big help already.

Some suggestions/requests:

* add ability to scroll (with mouse and left/right/up/down) zoomed in graph.
* shortcut to zoom in/out, ctrl + '+', ctrl + '-' and mouse scroll wheel would be awesome.
* ability to add actual values on top of the curves/points. This one is tricky as you don't want them all as they wouldn't be readable. Maybe have a choice of every n-th somewhere
* make graph remember settings for colors, fonts etc. Maybe make it a part of preset but I think it should be a separate item. Just having one that automatically saved and restores without any explicit menu for it is fine.


Title: Re: ECUxPlot Calc data way off
Post by: nyet on February 21, 2012, 01:13:00 PM
Unfortunately, all of those might be tricky, because they're part of the graphic package (jfreechart) which I didn't write.

I will try though.

Question: which is more important: those features, or a "batch" mode which opens up a set of presets?

I'm currently working on the "batch" thing, but if the other stuff is more important i can work on those instead.

BTW THANK YOU so much for testing and giving me feedback :)


Title: Re: ECUxPlot Calc data way off
Post by: julex on February 21, 2012, 01:51:04 PM
Batch is more important since one can limp around manually re-zooming the chart as needed.

Having ability to see multiple presets in their own tabs would be just awesome and would nullify any need for excel stuff.


Anyway, food for thought for later:

Mouse scroll wheel  zoom (just set up the property and done, no further code needed it would seem):

    org.jfree.chart.ChartPanel (or org.jfree.chart.panel? they list both structures):

    void    setMouseWheelEnabled(boolean flag)
    Enables or disables mouse wheel support for the panel.
    void    setMouseZoomable(boolean flag)
    A convenience method that switches on mouse-based zooming.
    void    setMouseZoomable(boolean flag, boolean fillRectangle)
    A convenience method that switches on mouse-based zooming.

Sample of panning, zooming and scrolling which I believe is included with JFreeChart, but that's a nice piece of code which might or might not be a quick port...:

http://www.koders.com/java/fidD9EBAC7EA9414FA80C9B16F620BA607556FC99DB.aspx





Title: Re: ECUxPlot Calc data way off
Post by: julex on February 23, 2012, 07:02:32 AM
small bug:

If you load the log file first and then switch presets, items in y-axis/y-axis2 don't get automatically checked off as per preset so you really have no idea what's enabled in the plot unless you carefully study the legend.


Title: Re: ECUxPlot Calc data way off
Post by: nyet on February 23, 2012, 10:28:22 AM
Thanks for the bug report, i'll check it out.


Title: Re: ECUxPlot Calc data way off
Post by: julex on February 26, 2012, 07:09:20 AM
Cool, will give it  areal exercise tomorrow once I get new logs to revise.


Title: Re: ECUxPlot Calc data way off
Post by: julex on February 27, 2012, 10:05:52 AM
Works well!

Something I noticed though, which is not a big deal really but... It appears that you are detecting which log style (ecux vs me7l) you're loading based on headers and make some assumptions in how "TIME" column should be parsed. There are obviously differences in format between two loggers so I routinely edit me7l logs to use with excel stuff (still!) which requires no colons between S and ms part of time stamp. Now, I believe that previous versions of ecux logger  were capable of reading in such modified me7l log and still parse the time stamps properly while the new one is not? If I remove all the headers and only leave description one in log, it works since now it thinks it is ecux log. Maybe I am wrong about previous versions' capabilities though but I swear I loaded such logs few times in he past and stuff was visible once passed through filter.

I know my situation is special but just wanted to bring it to your attention :)


Title: Re: ECUxPlot Calc data way off
Post by: nyet on February 27, 2012, 10:36:16 AM
Julex: if in doubt email me the logs, and describe what you want :)

In general, though, I'd rather ONLY support unedited logs...


Title: Re: ECUxPlot Calc data way off
Post by: julex on March 01, 2012, 07:34:42 AM
That's cool man. So far so good, I like the changes you've made. It has already made your software much more user friendly even though you wiped my custom presets with newest version  ;)


Title: Re: ECUxPlot Calc data way off
Post by: nyet on March 01, 2012, 10:12:11 AM
Julex: thanks for testing.

I'll be adding the "load all" presets next, then a release.


Title: Re: ECUxPlot Calc data way off
Post by: nyet on March 01, 2012, 05:46:27 PM
ECUxPlot 0.9r4.0 released

http://nyet.org/cars/files/ECUxPlot-0.9r4.0-setup.exe
http://nyet.org/cars/ECUxPlot

Added support for "load all presets"

Thank you everybody for your support and feedback.


Title: Re: ECUxPlot Calc data way off
Post by: Timtheguru on April 11, 2012, 12:36:04 PM
I'm just starting to get used to ECUxPlot, I can certainly help find bugs if I knew what I was doing. Is there a manual for it? I tried finding something, but just end up back at the download page. I've been using MarkP's spreadsheet method since Excel is like second nature to me, but this thing is awesome!

Is there a place I can download all of the standard presets that I see in the MarkP stuff? Sorry if this was covered before.



Title: Re: ECUxPlot Calc data way off
Post by: nyet on April 11, 2012, 01:28:00 PM
no, no manual :(

i have been meaning to make the rest of the presets in the markp template but never got around to it.


Title: Re: ECUxPlot Calc data way off
Post by: setzi62 on April 11, 2012, 02:14:41 PM
no, no manual :(

i have been meaning to make the rest of the presets in the markp template but never got around to it.

That would be really great to have more presets.
Probably some people have made already own presets which they would share.
Also profiles for more cars or some more information on the data I need to put in to get a profile
matching other cars than S4 would be very nice. Some data is clear, but for some things
like e.g. drivetrain loss I have no idea how to find out or if I can use just a ballpark number for it.


Title: Re: ECUxPlot Calc data way off
Post by: nyet on April 11, 2012, 03:34:58 PM
Some data is clear, but for some things
like e.g. drivetrain loss I have no idea how to find out or if I can use just a ballpark number for it.

DTL is pretty controversial; you have to make the same guess as you would if you were using a dyno to estimate CHP...

I don't think you'll find any canonical numbers for it...


Title: Re: ECUxPlot Calc data way off
Post by: azaiats on May 19, 2012, 06:54:46 PM
Nyet: you software is really great. It reminds me my kart Mycron4, which I can log and plot a lot of things in kart racing... Keep the good work.
One question: based on my gear ratios and my wheel size, how can I calc my RPM_per_mph? In mycron software we only need to input gear ratio and inflated tires diammeter. I have no idea how to calc it to input on your software :) 


Title: Re: ECUxPlot Calc data way off
Post by: nyet on May 20, 2012, 01:29:25 AM
Nyet: you software is really great. It reminds me my kart Mycron4, which I can log and plot a lot of things in kart racing... Keep the good work.
One question: based on my gear ratios and my wheel size, how can I calc my RPM_per_mph? In mycron software we only need to input gear ratio and inflated tires diammeter. I have no idea how to calc it to input on your software :) 

the best way is to use gps in 3rd at a fixed rpm... hold 3k rpm or something and divide 3000 by mph.

PLEASE feel free to email me (or post) bug reports and/or feature requests.