Some Questions about Hondata & Interpolation

Hello guys :)

So recently i finished the EFI Fundamentals and also the practical reflash, standalone tuning , understanding AFR and some other courses.

Most of the time im on the Hondata or Neptune side of tuning in our workshop because thats what we love :D

I got a bit of knowledge in all this EFI stuff before because well yeah you cant get around if you want to do it right and im feeling relatively comfy with the usage of Hondata and Neptune Software and Hardware.

But i also feel like im missing some things and i do have some questions about Hondata or Neptune in general:

Target Lambda and adjusting Fuel Cells:

What is the magic behind this ? Haha, because i feel like im doing something wrong. So lets say im driving a datalog with my car on the streets with everything set up correctly and fully warmed up car ( wideband, target lambda etc ), im running in open loop so no corrections are made.When i analyze my datalog im watching the Lambda Overlay to get aware of the AFR in each cell. No problems till here. So i start to correct each cell with the formula measured AFR divided by target AFR, im doing this for every datalogged cell. When i did this im driving the next datalog, and the results are not matching with my target AFR. Like lets say i got measured Lambda 1.10 and target is Lambda 1 so i need about 10 % more Fuel on this cell, when i did this and drive the next Datalog it doesnt match at all, but why ? I also tried to just add or substract the half of the showed difference in percentage ( 5% based on the example ) and it seems like its getting closer to the desired AFR than adding or substracting the complete difference in percentage . I kinda feel like i can't use the methods that i learned here at HP Academy to use it on Hondata or Neptune. Like im driving a crazy amount of datalogs and adjust it over and over again but it never gets really relly close to what i desire and is always slightly off. From what i heared of is that hondata is a little bit different but in which way ? What do i need to do to get close to my target AFR because i feel like i cant get away with adjusting cell by cell ,what is the correct method here ?

I know that in closed loop the ECU is adjusting the fuel related to the O2 Sensor Feedback at light loads till about 700 KpA or over 5000 RPM ( closed loop ) so i do not need to get really really close to the desired AFR , BUT how much is enough ? Like how do i know how much percent difference between target AFR and measured AFR is enough so i do not have to take care about the cells anymore ? My closed Loop correction is set to -20 and +20 percent bevcause the car is running off the Wideband feedback, there is no other o2 Sensor except this one.

Everytime im watching my 3D Fuel Map after adjusting i think yeah this looks not too bad, but when i switch to the 2D Map its looking like the Graph goes crazy, some random deep dips and sometimes random high spikes like, it's not "linear" or a clean Fuel Curve at all.

Maybe some of you guys know the PFI Speed Tuning guys, they have a great YouTube channel and i learned a lot of them before i bought this courses. If i watch the Videos of these guys i can see that they are only working with the 2D Maps but i do not understand where the heck he knows how to drag the 2D Map to get it right. His Fuel Curves are looking absolutely beautiful.

The next Problem is , how the heck is this interpolation feature working. Is it correct to use it to fill out the gaps that werent datalogged ? Also, how do i use it correctly ?

The way i use it : im driving my datalog, adjust every single cell to my desired itarget AFR. Then there a some gaps logically in areas i can't reach. What i always did, was to highlight the last known and adjusted cell, and highlighted the cells all the way to the end of unknown cells so i get a kind of linear progression from "logged cells" to "unknown cells". BUT it seems not right and the fuel curve looks crazy. So my question is now, do i interpolate to the top and to the bottom of the map from the last known cells ( basically per column or lets say interpolate by load) or do i interpolate to the left and right end of the map from the last adjusted cell ( basically per RPM or lets say interpolate by RPM ) or do i also interpolate the known and adjusted cells to get it "smooth" ?

Last thing i wanted to say is sorry for this long and maybe hard readable Topic, im from germany :D

I just feel like there are some tricks and handlings related to the tuning with Hondata or Neptune that i do not know and cant solve by myself, it feels like its Stoneage technic that is not really matching with the stuff i learned.

if it helps i can also upload some sketches or logs/pictures so you guys can better understand my problems.

Hope you guys can help me out here and greetings from germany :)

Sebastian,

Here are a few possibilities for your first question on fuel delivery changes not having the expected impact on measured lambda:

1) If you're working in the non linear region of the injector, without excellent data and compensation table which I don't think you have access to, the change in injector on time will not have a linear effect on fuel delivered.

2) If the engine isn't combusting cleanly, readings won't be super consistent. For example perhaps the engine has head work that causes it to combust less than ideally when you're not flowing lots of air.

3) Perhaps differential fuel pressure isn't totally stable and consistent. Are you logging it?

4) Are you 100% sure the lambda reading is accurate? Meaning sensor in good condition, good placement, controller, zero exhaust leaks of any kind, exhaust is quite long so the sensor is nowhere near a fresh air source...

In terms of trims, expect transient behavior to cause big swings, don't fret. Ensure engine response to throttle transitions feels smooth though. Then in steady state or during pulls try to get it running just right.

Interpolation and smoothing, like most things, should be used in moderation and with care. It's nice to have a smooth looking fuel table, and they often will end up that way naturally. Interpolation can also be helpful while you're building out a tune, as you said to fill in gaps between tuned cells. You can go too far in terms of non linear entries, and over smoothed tables though. There's always a balance. If a table looks a mess, I do suggest trying some smoothing to see how the engine responds, what the wideband data looks like. If it's better, great. When getting to the edges of the fuel table, either at highest load or highest RPM, try to look at the trend and continue it, erring perhaps a little on the rich side until you're able to test and improve your values.

I don't know the particulars of what you're tuning, but in 2d mode generally on a turbo engine if you space load axes evenly, once above atmospheric pressure you'll mostly be taking the same shape per row, duplicating and shifting it up a similar amount. There won't be a great deal of change til exhaust pressure gets pretty high or you run into a significant intake side limitation. The variation below atmospheric pressure is generally greater.

Hopefully that at least touches on some of what you're after!

Hey Mike !

Thanks for your answear :)

about 1 -> Yes , i understand this.

about 2 -> i also undstand this hehe, but its just a stock head. In general the engine has just a few mods, its a stock D14Z1 Engine but with the Intake, Throttle Body, injectors, Airfilter and Exhaust Manifold of a D16Z6. I thought it would be good idea to take a cheap engine for learning how to tune :D

about 3 -> im not logging the fuel pressure as the OEM Fuel pressure regulator is in usage

about 4 -> The Wideband is brand new, its a AEM Uego XSeries and is mounted in the OEM Position at the Exhaust Manifold, leaks may be actually a good call because the manifold had some restrictions inside which we didnt wanted so we took them out, im pretty sure there are no leaks but i will take care about that tomorrow, the reading is fine i tested it. Exhaust is Stock, only Muffler and the D16 Manifold -> our german Polizei is pretty strict soooo yeah haha

Thanks for the other explanations !

I just feel like there are some tweaks and tricks to work with Hondata. Im driving everyday about 2 Log's on my way to work and back. Im logging my AFR and trying to get it right. like , i drive it, analyze it a home, then i correct the Fuel Cells in addition with lambda overlay and the formula measured AFR : Target AFR but it just looks like a big mess, i will attach a picture. I also made sure i got no Fuel Corrections fcoming from my IAT or ECT Sensor, Closed Loop is also disabled.

Sebastian,

Your process sounds good.

Looking at the image you attached, your 100 kpa column seems out of place compared to all other data. Is that part of the table which you aren't able to hit and it's not tuned yet? If so, consider using the 95 kpa column values, copy paste, then apply perhaps 10% increase for now, and interpolate between 100 and 95 kpa columns.

Also, the 2d view nicely shows fuel delivered as load goes up. On that NA engine generally fuel required will be similar to greater as load increases, and any time you see a value drop below a lower load range, you know there's likely an issue in the vehicle or calibration. You could start by making any fuel value that drops below the fuel value for a lower load range match it instead and see how the car behaves.