Tuning Targets for Fuel and Air (Lambda)

hulkss

Administrator
Sep 29, 2013
622
2
18
Our bikes use an Engine Control Module (ECM) that controls Air/Fuel ratio (AFR) with feedback from Oxygen (O2) sensors in the exhaust. The ECM fuels the bike as commanded by a table of Lambda values. Lambda = AFR/AFRstoich. The stoichiometric air/fuel ratio is the chemically correct ratio of air to fuel for complete combustion to take place (about 14.7 for ordinary gasoline). The ECM uses lambda control so that it will still fuel properly with different blends of gasoline and those mixed with alcohol.

A value of one in the lambda table is stoichiometric. A value of .95 would be 5% rich, a value of .90 would be 10% rich. The stock narrow band sensors only function at about .975 or greater. The rest of the time the engine must operate open loop (no O2 sensor feedback is used).

The proper Lambda value is selected based on engine load and rpm. Richer settings are used at higher rpm and high engine load. Leaner at low power for emissions and fuel economy. Engine load is approximately proportional to Manifold Absolute Pressure (MAP). It is measured by a sensor in the intake manifold.

Here is the H-D Lambda table plotted in three dimensions (3D) for my 120ST engine:

Lambda%20120ST.png


You can see the engine is kept lean most of the time and then switches rapidly to open loop at the high speed and load points. This can be made a little better with a performance tuner. Here is the 120R (race) engine Lambda table:

Lambda%20120R.png


A little better but not much. This is a limitation of narrow band O2 engine control. High-tech lean burn water cooled engines have no problem with this, but it is not the best fueling for a hot running, high power, air cooled, high compression, good old fashioned Harley Twin Cam (used off road of course).

How to do better? Switch to wideband O2 control with "Target-Tune" from Dynojet Powervision. This system uses the factory Delphi ECM with all of its advanced features and mates it with wideband O2 sensors for Lambda feedback over the entire engine operating envelope. Now a fuel map can be designed that provides a smooth transition across all engine operating points along with the speed-density control provided by the Delphi ECM.

Here is the fuel table I developed and am running now in my 120ST engine. This was easy to build in WinPV software that works with a Power Vision tuner. The engine now runs as smooth as this table looks. I am super happy with it:

Lambda%20Custom.png


Converting to AFR for gasoline the top of the chart is 14.3 and the bottom is 12.8 AFR.
 
very interesting post... wish to learn more about it so I subscribe!

Inviato dal mio HUAWEI NXT-L29 utilizzando Tapatalk
 
Yes very interesting post and it's nice to understand and see what your after.

Thanks Hulkss
 
Nice Work Hulkss!

Did you build the table in WinPV and view in MegaViewer? Just curious on how you were able to determine the AFR's... realtime and watch the table or calculate... input and then view?
 
Did you build the table in WinPV and view in MegaViewer? Just curious on how you were able to determine the AFR's... realtime and watch the table or calculate... input and then view?

These tables are the Target AFR. That is what YOU decide you want the bike to run at. Rich, lean, or whatever. When you tune your bike, the volumetric efficiency tables are adjusted to realize the target table commands in your engine. I'll post about that later.

So to make this table, open the AFR (Lambda) table in WinPV.
Set these cells to your desired cruise AFR:

Cruise.PNG


Set these cells to your desired high speed and load AFR:

WOT.PNG


Now interpolate vertical (under edit command):

Vert.PNG


Interpolate horizontal:

Hor.PNG


Continued next post.....
 
.......Continued from previous post.

Interpolate (this looks both directions):

Inter.PNG


Tada!! done. The 3D plotter in WinPV is 'effed up due to the uneven row and column intervals. My plots in the earlier post were made in Excel with extra rows and columns inserted to create evenly spaced data. This makes properly proportioned surface plots.

Now I wanted .871 Lambda at 6000 rpm (my set rev limit) so I adjusted the bottom row and re-interpolated as above until the average of of the 5500 and 6500 rows was .871. Working in Lambda is a little more accurate. There is a switch on the right of the window in WinPV.

6000.PNG
 
Thanks! I got some playing around ahead of me!

Sent from Note4 via TapaTalk
 
Bringing up an older thread....

HULKSS

Dealing with spark knock events... I decided to richen my AFR up in the affected areas starting at 2750 rpms up to 4000rpms and 50-80Kpa range... took it 5% ( avg is around .904 now in the middle of this range vs .952)

I did smooth it out and interpolated both vertically and horizontally....so... we shall see how she runs now.

Seems that altitude messes things up for me as well... I can ride from 5500ft elev. to 8000 elev. with no events.... shut it off for a stop... get back on and get a multitude of events on front cylinder only when under a load... mainly between 3000-4000rpms and 50-75Kpa...

Gonna run this a bit this weekend and see if I still have any events or at least a reduction in them. Datalogging is your best friend here. IF things go well.... I will run more AT sessions and reset trim values..... of course comparing tables as I progress

BTW... still running TPS base and not MAP base at this point.
 
Here's my current spark table (same both cylinders) and fuel table. These should work for you. Flash them in and go for an auto-tune ride. Flash the new auto-tune calibration and all should be well.

Lambda.pngSpark.png
 
Here's my current spark (same both cylinders) and fuel tables. These should work for you. Flash them in and go for an auto-tune ride. Flash the new auto-tune calibration and all should be well.

Interesting.... Here is my front spark and Lambda.... I am richer already... timing may be an issue for sure....

AFR.jpgspark.jpg
 


Latest posts

Forum statistics

Threads
7,978
Messages
69,245
Members
15,071
Latest member
Mikes469