"Busted" Tune

[hrdtail78]

Here is a data log off of a problem child XL.  What is wrong with it and what do I do to VE tables to fix it?

https://app.box.com/s/ya6k9cilt45ho7bqpwadb4uzi9c53npn

[UltraNutZ]

got the map and dm3 file?  I'm a visual kinda guy Jason.  lol

[whittlebeast]

First thing is your wideband is falling off line.  The narrow bands look fine.  With the narrow bands jumping up and down, and averaging say 800 MV, the widebands better both be reading close to 14.4 or so.

Andy

[hrdtail78]

Sensors dropping off line is a problem.  When they do this they read their richest setting and on these it's 7.45.  The data collection going into these "do everything" for you tuners don't understand this.  garbage in, garbage out.

Here is the same bike and WB's are not dropping out.

https://app.box.com/s/7x91kprx0921mso4se1dpglz6cx3umfs

Now what is going on?

[hrdtail78]

got the map and dm3 file?  I'm a visual kinda guy Jason.  lol

I wont post the calibration.  It's a 171 MT9, and it isn't my place to post it.  But with all the talk of viewing data.  Is it really needed?  VE and spark is in the CSV and those can all be recreated.

[whittlebeast]

Do you want to autotune this thing with TTS, hand tune it, or use MLV to autotune it?

[hrdtail78]

I just thought it would be a good case for someone to explain how MLV or any log program to diagnos and change a tune for the better.  Rubber is meeting the road.

[whittlebeast]

To generate VE New Front and Rear.....

This all assumes you have the target AFRs set to 13.5 in the tuning software

Create a few new fields in this order

VE Multiplier Front   =   [WB FRT]>19 or [WB FRT]< 10.5?1:[WB FRT]/13.5>1.1?1:[WB FRT]/13.5<.9?.9:[WB FRT]/13.5

VE New Front   =    [VE Front] * [VE Multiplier Front]

VE Multiplier Rear   =   [WB REAR]>19 or [WB REAR]< 10.5?1:[WB REAR]/13.5>1.1?1:[WB REAR]/13.5<.9?.9:[WB FRT]/13.5

VE New Front   =    [VE Front] * [VE Multiplier Front]

We may wind up setting up a few filters but the average VE New Front and Rear can be generated with the table generator.

Have fun

Andy

[hrdtail78]

I have several ways of generating new VE's, and none of the automated stuff is going to identify a sensor dropping off line.

[whittlebeast]

Yep, there is no replacing a reality check.  Your original data looked horrible until I started drilling down to find dead widebands.

[whittlebeast]

Here was where we started once we got the wide band 02s on line.

http://www.nbs-stl.com/HarleyTuning/DynoTuningStartingMapRear.png

http://www.nbs-stl.com/HarleyTuning/DynoTuningStartingMapFront.png

The dyno operator did a great job covering the VE Map

Closed loop was running on this bike and needs to be turned off on this stage of tuning.  Or we need data logging that is not hiding critical tuning information.

Andy

[UltraNutZ]

never fooled with the megalogviewer but this thread reminded me to grab it.. pretty slick.. thanks for the help whittlebeast.

am I reading this correctly to see at record 741 is where the WB is falling off?

[whittlebeast]

never fooled with the megalogviewer but this thread reminded me to grab it.. pretty slick.. thanks for the help whittlebeast.

am I reading this correctly to see at record 741 is where the WB is falling off?

Yes and then later at about 1025 sec both go flatline.

[Coyote]

I graphed it in Excel and it was obvious.

[whittlebeast]

This is fun

I created a new field called PctFuelFront

and defined it as     100*[Duty Cycle1]/([Duty Cycle1]+[Duty Cycle2])

This is essentially a formula to show how much the Y manifold is allowing the cylinders to steal fuel from the other cylinder.

http://www.nbs-stl.com/HarleyTuning/DynoTuningStartingFuelTheft.png

It will be interesting what we end up when this motor is dialed in.

Andy

[whittlebeast]

Here is why you can't run a Sporty in closed loop.  This is an example of a true ECU code "Bust".  Or at least a Harley Hidden table that needs to be exposed.

http://www.nbs-stl.com/HarleyTuning/SportyClosedLoop.png

I was once told from a Harley engineer that this was done on purpose to make the Sporty to feel more "Harley like" after the VRod ran too perfect.  Go figure......

Andy

[glens]

What the hell are you calling "busted" about that?

[whittlebeast]

Glens

Think thru what swings like that in AFR would do to destabilize the motor.  The big twins don't do this to near to this degree.

Andy

[hrdtail78]

Here is why you can't run a Sporty in closed loop.  This is an example of a true ECU code "Bust".  Or at least a Harley Hidden table that needs to be exposed.

http://www.nbs-stl.com/HarleyTuning/SportyClosedLoop.png

I was once told from a Harley engineer that this was done on purpose to make the Sporty to feel more "Harley like" after the VRod ran too perfect.  Go figure......

Andy

There are other things that can make this happen besides a busted code or hidden table.

Finding the sensors dumping off line was the throw ya a bone easy one.

[UltraNutZ]

apparently not for all of us.  LOL

[whittlebeast]

Hrdtail

I would love to see a little experiment the next time you have that sporty on the dyno.  Lock the dyno at say 2000 rpm. Go from near 0 throttle to WOT as fast at you can twist your wrist.

Repeat the test at 2500, 3000 and 3500 RPM  We will get a better idea for the response times of the 02s as it relates to this data rate.

This would be fun to repeat on your TwinScan.

Andy

[whittlebeast]

apparently not for all of us.  LOL

The basic sensors are

RPM (crank sensor, no static)

MAP (100 or so with a dead motor, 50 or so at idle)

TPS (Smooth roll on, no static)

IAT (a cold motor should be close to ambient temp)

Water/Engine Temp (a cold motor should be close to ambient temp)

02 (narrow band, swings from about 150 to 900 MV fairly often)

O2 (Wide band, reads about 14.7 when narrow bands are real active, never flatline if the narrow bands are active, flatline is almost always a problem if the motor is running)

Battery volts (fairly stable and close to 14 volts most of the time)

On the race motors, add water pressure and EGTs

Always do a reality check on those sensors when things don't look correct.  You would be amazed how bad things get screwed up when one of those go stupid.

Andy

[UltraNutZ]

thanks Andy.. I think this app will come in very handy used with my TechnoResearch package I just bought.  I'm sure I'll be posting many questions once I get going.

[hrdtail78]

apparently not for all of us.  LOL

I actually found it easier to see things just using the trace log and going through the data record frame number at a time.   This way helps me follow along with what the bike is doing as a whole.   

The sensors were dropping off line because of a power source problem I was having.   Not enough amperage.   It was enough for normal operation but after a long Decel they would cool and the amp draw would increase getting sensors back to op temp.    Power source fixed and they don't drop out anymore. 

In the second log.  Couple of things I see with WB is the long Decel in the rear reading real lean, and it doesn't like to follow along with its narrow band at idle.

Bike has VH short shots with a lot of reversion with the rear.   ....and no Decel pop.

[glens]

Glens

Think thru what swings like that in AFR would do to destabilize the motor.  The big twins don't do this to near to this degree.

Andy

Closed loop does not hold steady-state AFRs.  It continually swings lean/fat.  While it might appear exaggerated in your trace, it looks good to me.

Part of why the response time of the broadband sensors is slower than the narrowbands is because the broadband's sensor controller is cycling (effectively) rich/lean just like that.  It takes a couple swings to be able to determine just where the center point is...