The sort of thing that gets my attention

[whittlebeast]

I was just last night test the newest beta version of MegaLogViewer.  I helped talking Phil into adding the ability to do scatter graphs.  I was feeding in data off the TTS data stream.

This is really interesting.  I always knew the MAP was the largest contributor of all of the sensors when running the ECU calcs.  I never had an easy way to plot it.  Check it out.

The differences in the two curves is evidence that the the Y manifold and exhaust is causing issues.  I would think the bigger the difference, the bigger the issue.  I wonder what this looks like on a Harley before and after the smoothing and valve opening stuff?

Way cool stuff.

AW

[cts1950]

That spot at about 2.5 on the injector and about 32 on the map sure stands out I wonder if that is a visualization of cam reversion. It seems more prevalent on the front cylinder.

[whittlebeast]

I have looked at a 100 or so  tunes.  It appears to be that a very tight pattern is a good indicator of a tightly tuned bike.  What is interesting is if this tight pattern should continue on up to the top right corner of the map.  My guess is yes.

The dot that develops on most of the tunes is definitely the bike at idle.  Some bikes develop the dot above the main line and others below.

Some bikes have two distinct lines that appear.  This is really strange and I bet will turn out to be an indication of a serious tuning issue.

All I would need is data logs off a few a few well tuned bikes as they were being tuned on the dyno doing power sweeps.

AW

[whittlebeast]

Here the the same Big Twin before and after tuning.  Notice how much finer the line of dots is.

Before Tuning




After Tuning



Both tunes were done a few years ago with the TPS (Alpha-N Presentation) based VE tables that are far more problematic than the newer MAP (Speed Density) based VE Tables.

Notice that this trace also has the idle dot below the main line of dots.  You can see from to o2 trace on the right that this bike was running closed loop with a high O2 Bias setting making the o2s difficult to "Read".

Have fun tuning

AW

[whittlebeast]

Here is my bike running on the highway in and out of the throttle between 2500 and 3500 RPM.  Notice how tight the pattern is.  In this range, my bike is open loop as the o2 feed back loop is totally useless in this range stock.  In the second trace you can easily see how two distinct lines develop on the lower trace as the o2s add and subtract fuel seeking two specific AFRs.  14.7 plus .5 AFR and 14.7 minus .5 AFR.  The second trace is the same bike totally stock using the same TTS to data log what was going on.  Keep in mind that this program is still doing a rather tacky auto scaling that is common on poorly written software.  I am trying to get Phil to fix it.  You are looking at early beta software that I bet will evolve fairy quick now thit he is messing with it full time.   


My Tune




Stock



AW

[Jeffd]

dang I wish I understood this stuff.  Is there a basic book that would help understand some of this vodoo?

[whittlebeast]

Somewhat unrelated but this is what reasonable throttle response with total control of the DBW, a light flywheel and programmers with too much time on their hands gets you...

It gets good at about 1/2 way thru.

Renault F1 Car playing God Save The Queen

Have fun tuning

AW

[whittlebeast]

Here is a textbook example of an autotuning software that has given the apparently wrong answer.  This tune was generated by following VTune standard procedure to the tee.  We opened up the data to look around and this is what we found.  The rear cylinder is quit normal but the front tune appears to be a mess.  Without looking at the data, the owner would have had no clue something had gone wrong.

Things to look for may be air leaks in the exhaust, bad o2 sensor...



The lesson here is never trust a tuner (electronic or overpaid mechanic) without looking at the data.

AW

[wurk_truk]

Hey Andy

It may be helpful to some (me) mto explain HOW one loads up TTS data to be viewed in the mega log viewer.  Really simple basic step-by-step instructions would be nice.

Start a new thread for this, OK?

[whittlebeast]

Right now the stuff that does these scatter graphs is still in beta form and will most likely keep changing fairly quick.  Once the final (stable) version is released I will be happy to do that.

You can find the standard stuff that can even be used for free by going to

http://www.ideasandsolutions.biz/MegaLogViewer/download/

He only charges $25 to register it.  He only insists (hopes) you pay for it if used for commercial use.

AW

[bbrowncods]

Once you get the initial data run from a bike, if the tune was close (say a good v-tune), you could do a best line fit and back that into the VE table and be spot on.  Just thinking outloud.

[whittlebeast]

Once you get the initial data run from a bike, if the tune was close (say a good v-tune), you could do a best line fit and back that into the VE table and be spot on.  Just thinking outloud.

Are you talking about data smoothing of the VE tables?  The problem with doing that is figuring out if there is a matching dip in MAP that is in sync with a rise in the VE.  The problem may be driven by the code sampling  the MAP signal at the wrong time in the 720 degrees of crank shaft rotation.  I would have to create a special function in MLV to demonstrate that in a scatter graph.

You may be able to get there if you have an old version of TTS that gave the MAP Default in the data feed.  It may still be in VTune data feed.

AW

[whittlebeast]

Here is a typical data log that ca be done with any version of TTS and any version of MLV.

Notice that the rider jumps on the throttle twice with somewhat different results out of the injectors.  Keep in mind that this is .3 sec resolution so some of the detail gets lost like the spike in BPW that should have showed up with AE setting in.



All traces can freeze the range. Not just autorange everything like more cheesy software.  Up to 16 traces can be viewed in sync at the same time.  All zooming is dynamic.  You can create any new field based on any of the the available fields.  The new registered stuff does scatter graphs.  With a little coaxing, I bet you guy could get Phil to do auto tuning (at least on the VE side of things) if enough people ask nice.  At least to the point of copy/paste back into TTS.

Have fun tuning

AW

[whittlebeast]

Check out the difference between this stock bagger timing trace vs the timing map the bike above was running.

Stock Timing Map

http://www.ncs-stl.com/SertTune/StockBaggerTimingMap.jpg

Tuner Timing Map  Note the cross in the center is simply the bike running with the cruse control on and getting lots of data at just one RPM and lots of different loads.

http://www.ncs-stl.com/SertTune/BT_TTS%20TimingMap.jpg

AW

[bbrowncods]

In your post 12 the difference in PW between the two full throttle events can be attributed to the Accel Enrichment being more pronounced on the first event verses the second.  It would be interesting to see with Accel Enrich turned off.

[hrdtail78]

There is absolutly no contact number for "Phil"  Do you have one?  I got some questions about this and how to get the scan data from TTS to megalogviewer.  How many computers can I put it on with one registration key?  I'm not sending $25 to no one I can't contact on the phone.

[wurk_truk]

 

I downloaded the beat viewer at   http://www.efianalytics.com/TunerStudio/beta/

But some help on viewing TTS files would go a long ways.

[whittlebeast]

Phil is easiest to get by email.  He allows you to put it on a couple of your own computers.  He is nor really that concerned provided your licens starts coming from all over the country as if you gave your license to all your friends.  He can see that by IP address on updates....

You guys may have grown accustomed to dealing with people that are far more interested in the mighty dollar.  Phil gave that program away for the first three years till it got to be his second full time job.  He was head of the internet security for a HUGE international bank with banks of programmers under him when he retired at about 40 to do this for fun.

AW

[wurk_truk]

To 'unlock' the beta... $49.95

[cts1950]

I have been using the megalog viewer since 2003, back when I was using the DTT ecm. I have found it to be a wonderful tool to look at the data, up to 12 lines at a time.

To look at the TTS data you have to take your scan data file open it in the datamaster program check how many frames of data you have recorded, it will be on the bottom line of the window after the date and time. Example 16590/980 keep the 16590 which is the number of frames collected.

From the file dropdown menu click on export data that opens a new screen called export data. You will see boxes start rec: 0 and count 1. In the count box put the 16590 in the box if you want the entire file.

The next boxes down show the items it will export.

The 4Th column over begins a series of drop down items you can look at. select from the menu what you want to see. They keep going to the right side and you can end up with all options selected.

Now click on the extract button and you will see the hour glass briefly and the program will say finished data extraction on the bottom of the window.

Now from the file dropdown select save as comma separated file (CSV)  a new window pops up and It will ask you to name that file. Example  XYZ  now click on save that window disappears and you are done with the TTS programs you have opened so you can close them.

The file you just created is in my documents, the tts folder, datamaster, export data.

You should now find the XYZ.csv file you just created. You will need to modify the microsoft Excel csv file by opening it in Excel and right click on column B to highlight it and now left click on column B and select delete that will remove the frame time HH:MM:SS  ####### will be directly under HH;MM:SS. The C column now moves over and takes the place of the original B.
Megalog dose not use hours minutes and seconds data, now  save the CSV file and use the same name or add a additional letter or number and have both copy's. So I will say XYZa as the new name and close Excel.

Now from the megalog viewer file drop down box select open and select the location of this saved file in, My documents,TTS, HD, datamaster, export data, XYZa.csv.

I have a registerd copy of Megalog viewer and was able to down load the beta version no extra cost. I just checked mega log viewer is 24.99  I think Wurk_Truk was looking at  Tuner Studio for 49.99 a difrent program entirely, a easy mistake.

[cts1950]

You do not have to have Microsoft Excel to edit the csv file. You can use a free version of Staroffice calc. to do this. I just checked it this morning.

[wurk_truk]

Of course you are right... Thank you very much, BTW.

[whittlebeast]

Keep in mind that in the meagsquirt world ( this includes MegaLogWorld) when playing with the beta stuff, things change real fast and there is very little to no documentation on this stuff.  Once the stuff goes release, big changes will get mentioned on the megasquirt forums.  Keep in mind, in that world, most everyone is either a programmer, a tuner, or an electrical engineer and they tend to thrive on the bleeding edge.  Some are all three.  People with huge egos and full of BS tend to get weeded out fast.

Have fun tuning.

AW

[Steve Cole]

Megalog dose not use hours minutes and seconds data,

If your not using the time stamps how can you tell what is going on with the data? Since the data rate will vary it plays a big role in analyzing the data properly.

[cts1950]

Megalog dose not use hours minutes and seconds data,

If your not using the time stamps how can you tell what is going on with the data? Since the data rate will vary it plays a big role in analyzing the data properly.

MegalogViewer uses the next column over, the time stamp column which is expressed in seconds from the start of the data collection.