K20A3 I/H/E lean on stock tune?

[Lucky16]

03 Base RSX, injen cold air intake, PLM header (no cat), magnaflow 2.25 in exhaust.

I just now put the header in. Before I was running a stock tune 91 octane .kal from the vault, but read somewhere that may run a bit lean after installing all 3 intake/header/exhaust on a stock tune? (And idk how the 91 octane kal is setup to begin with..)

I found 3 kals that looked like they sort of matched my setup in the vault (stock k20a3 rsx except I/H/E) but all 3 ran like complete crap for me - random big bogs, poor acceleration, etc.

I went back to the 91 octane stock tune and it runs much better so I plan on sticking with that. I’m not trying to make the car fast, just safe. And I plan on getting a proper tune after putting in the RBC manifold.

I have next to no tuning knowledge. Humbly requesting someone either tell me what couple of key parameters/values to look at? Or be willing to check out a datalog to make sure I’m not hurting anything?

I could do the lean protection thing but don’t know what values I would put in there? Default is 14.2:1, 3 inches, 250ms. But seems weird to be to have a blanket cutoff regardless of throttle position?

Side question is even though I unchecked secondary O2 I still have a check engine light with P0141 code (secondary O2 sensor malfunction)? I kinda thought unchecking it would inhibit that? If it’s checked the car won’t start which makes sense.

[MondeCruel]

Log your afr and tune your fuel tables it is not hard just takes time if you want it precise. Post your calibration and datalogs

[Lucky16]

Real ignorant question, but for a datalog what should the driving profile look like? Like how long should I record for and should it just be driving around town or doing pulls or what?

[Lucky16]

here's the .kal for starters

91 oct 2 w header.kal

[MondeCruel]

Wait until the engine is warm and drive. It depends of you, what condition of driving you want to cover. Best is long datalog with different driving style to fill all the fuel table cases with samples of your actual afr, you can select different display with shift+f3 or f4, or f5 will show you how much difference your actual afr is from your target afr, like +10% from your target you need to remove fuel or minus -5% you need to add fuel.

[Lucky16]

Thanks for the help. I’ll give it a go and post my data logs as well.

Also figured out the check engine light. I was unchecking sec O2 in Misc instead of in closed loop menu.

While troubleshooting that I found an exhaust leak at one of the joints where I put back together the exhaust so gotta fix that before I do the data logs.

[Lucky16]

Alright, did a quick data log on a 6 min drive. I was able to read it and see the lambda/adjustment values in the tables. It does look a bit lean especially at 30 degree cam (I think?).

jersey mike home.kdl

91 oct 2 w header.kal

questions are:
- is there a way to combine data from multiple logs? or to get more squares filled do you just have to drive longer?

- I think I have a decent number of squares if I include all fuel low/high/cam angles. do those tables overlap at all? like will I ever have different data in a given RPM and column on different tables? As an example, will column 5 and 3k rpm ever show different values on the fuel low speed tables vs fuel high speed (or at different cam angles)?

- I was going to just use "edit all tables" which in my understanding applies the change to all cam angles, but does it also affect both fuel low speed and fuel high speed or do those need to be edited separately?

- lastly and most importantly, after I edit a cell how do I save my work? just resave the .kal?

Thanks

[MondeCruel]

Yes you are right it is lean in those area 16-17:1, i inverted logic in my previous post if you see 10% you need to add 10%.

i dont know if you can add up samples by opening multiple log ??

yes you will have differents results for every cam angles and low cam/high cam, they are seperated, if your vtc goes to 50 deg you have 6 fuel tables x 2 to tune. There is interpolation if you are between 0 and 10 degree for example that is why if you want to do this better you should consider changing your cam angle table to one cam angle at the time, example put all you cam angle low speed squares to 0 drive and tune then run only on 10 degree etc it will get you better result than your cam angle changing all the time. you can read this here also
https://hondata.com/help/kmanager/tuningyourvehicle.htm

all tables is only across cam angles, it doesnt reach the other cam profile

You can park and make your changes then save, then reupload calibration to ecu drive and see if you get better values

[Lucky16]

Tuning is in progress, thanks for all the help!

[Lucky16]

Another question on fuel tuning. Default cam tables for low speed only go to a max cam value of 30. Do I still need to tune low speed tables for 40/50 cam then?


—————————————————————-
—————————————————————-
For anyone else that stumbles upon this post here’s a few other things I’ve found:

Start with a “new calibration” when doing this. I started with one from the vault I thought would be close and found it harder to tune in the tables than just starting from scratch. Also as I explored all the options in parameters there was stuff enabled/disabled/edited that I didn’t realize.

Follow the kpro help and use vtec point to isolate low and high speed tables. Set entire cam angle map to 0/10/20 works way better and will actually zero in faster as mentioned above.

I was originally adjusting fuel tables using the delta lambda on individual cells but got pretty erratic results. In options - settings change sample interval to 1 sec instead of .5 and change sample size to 10 instead of 5. Read up some more and learned it’s more effective to look at trends and apply average increases/decreases to broader sections/full rows/columns to keep it smoother. I’ve had better luck using those techniques.

I made an excel and logged what the lambdas were across 2-4 runs and then averaged them vs just doing one run. If you do this you’ll notice some runs say a bit lean and some runs say a bit rich (outside temp changes?) and that to me means good enough. My goal is to keep everything within about 5% delta. Also note, I would also be skeptical of extreme values that don’t have many samples. Like less than 100 samples is automatically skeptical to me.

Please take all of the above with a big grain of salt because I’m very new to this and just trying to figure it out. Anyone with more experience feel free to update or call bs.

[MondeCruel]

No i dont think you need it, i did copy 40 onto 50 degree low speed that i dont use, it looks more coherent than completly untune. You could tune roughtly your low speed cam / high rpm rows just in case your vtec doesnt engage ( happened to me on a bumpy road : vtec clip fell off ), but i dont know if it would be the same logic on a k20a3.