An in-depth talk about FL5 ECU tuning

[julianobl]

My benchmark is Interlagos

i am properly jealous
that’s a top level track

Sponsored

 

[Vito.FL5]

i am properly jealous
that’s a top level track

It is. I love that track as it is in my hometown so lots of love for it. We use to race beetles there it´s a lot of fun. 1000´s of laps in the past 15 years.

Corrida Copa Fusca GT-OIL 27/02/2021 Fusca #880 piloto VICTOR APPOLINARIO - YouTube

watch only if you have nothing else to do lol.



now a little more off topic, I´m freaking broke from buying the car + a shit ton of parts I should have bought over a period of 2 years lol so I decided to be cheap on the downpipe and now I will most likely spend 3 times more than If I bought one online lol.

Either way I have a problem with aftermarket downpipes since most are 3 inch and the midpipe is 60mm.

So I came up with a temporary solution, I cut the flanges for a 3.5" bend up top and 3" on the exit then machined a reduction to make it a little smoother until I have the money to buy the rest of the exhaust (or the time to build one)

Should be welding it on the weekend. based off an OEM one and hopefully it will fit the car when it arives lol.

 

[tezzasaurusrex]

I'll end the post here and get back to work. I have no Idea if this is of anyones interest. But if it is, we can go on and on.

Please do this. Go on and on.
Love this kind of in depth information, we're lucky that you're willing to share!

Interlagos is a track I drive the f*ck out of... in Gran Turismo

 

[Vito.FL5]

Please do this. Go on and on.
Love this kind of in depth information, we're lucky that you're willing to share!

Interlagos is a track I drive the f*ck out of... in Gran Turismo

Since you asked, lets get to some not very useful information lol.

A few days back someone sent me a video of misha beating the crap out the green Dream FL5 and it seemed off that oil temperature maxed out at 150C for such a long period and it didn´t go boom boom.

So I took some tome to check how the car calculates oil temperature and they have a lot to learn from the germans in this subject. The model is very basic and very likely to propagate the error to a point where they should have just put a freaking sensor at the pan.

well, THEY DID IT in some models but believe it or not, I checked every tune I could find and even when the sensor is present, the car uses calculated temps instead of measured.

here is the model function

BUT the hope is not lost, everything needed to get the info from the actual sensor is there above it:

So it seems like the sensor outputs a frequency, that translates to level and temperature.


So in the end it is a matter of figuring out the ECU pinout for the oem level and temperature sensor, and how to change the source of the information somewhere in this function

And bingo this should get us reading measured oil temps instead of bogus calculated values.

Now: if you have the pinout please share, and if your car from some market shows actual values and has a sensor please let me know so we can move forward on this.

It may seem useless but having real info on the screen is a great thing, that beats having a separate gauge to read oil temps and runing wires all over

 

[Vito.FL5]

Now to some info you CAN use:

I was checking how Hondata managed to do their base tune for the R and ok.

I respect them A LOT as a company, and tuned hundreds of cars on S300 and dozens on Flashpros but I´d say their aproach is wrong. And lots of important missing tables. I´m nobody but If at any point HONDATA wants help from some idiot tuning Bosch ecus since carbs lol I would be happy to help. Anyway.

FIrst thing they lack is START OF INJECTION maps, at least for homogen injection, which is what the ECU uses at full load. That could give users the chance to get a little more juice from OEM injectors without experiencing misfires at high speeds or just leaning out.

Second thing is turbocharger model tables. This may look overkill but if you don´t set the turbocharger model properly, pre control will be far off and you will be relying in the PID to get you where you want, and that may never happen lol. Basic shit like a free flowing exhaust system will make boost control a little off.

Third thing and now trust me on this one.

Hondata basemap basically maxed out everything and tuned the car using compaction ratio tables, this ones

while tables like this are maxed out at the 100% TPS region

Its not necessarily bad to do that but why not make it a little more progressive instead?





Now...

FOCUS HERE:

THIS IS WHERE YOU SHOULD LOOK AT!

Why the f*ck is this maxed out when this is the table that calls for the actual torque your car delivers at a specific RPM.

Look at how this is modeled stock:

Recognise this curve??????? THis is your stock torque curve.

Don´t recognise the torque values??? seem a little high?

that is because they are wrong! For some reason whoever translated the data assumed that maxed value was 655 (probaly because it is a 16 bit value but whatever), when it is not. Maxed out value is 100% of 600NM which is the REFERENCE TORQUE VALUE, for all the torque model on the ECU.

EITHER WAY, when tuning this table, do not max it out, start with it STOCK, and then DRAW IT LIKE ONE OF YOUR FRENCH GIRLS, raising it little by little and see the magic happens.

of course you need to raise boost limiters at some point, but trust me, if you start from a stock calibration and raise NOTHING but this values, you will gain a lot of power before you find the pressure ratio (boost) limiters. Keep in mind to raise the torque limit per gear per driving mode because you may hit that limit also

 

[CTRifecta]

Appreciate your dedication and information, agree with others keep on posting, it's great to hear some about more of the technology and less about PPF debates .

Question for you - can you help explain how oil temp is derived? I couldn't really comprehend based on the code posted.

Also - regarding the dyno #s you posted, how do you deal with heat soak especially on the stock intercooler?

 

[tezzasaurusrex]

EITHER WAY, when tuning this table, do not max it out, start with it STOCK, and then DRAW IT LIKE ONE OF YOUR FRENCH GIRLS, raising it little by little and see the magic happens.

of course you need to raise boost limiters at some point, but trust me, if you start from a stock calibration and raise NOTHING but this values, you will gain a lot of power before you find the pressure ratio (boost) limiters. Keep in mind to raise the torque limit per gear per driving mode because you may hit that limit also

Very interesting stuff.

Does it mean that it is possible to tune boost by gear with Hondata?

 

[Vito.FL5]

Very interesting stuff.

Does it mean that it is possible to tune boost by gear with Hondata?

That is correct!

My usual approach (and also how the car is tuned as stock) is to have throttle torque request tables set with the last column a little higher then the main torque request table.

Then once the car is tuned I lower the value on giver gears, for exemple you can limit torque at 350Nm in 1st gear, 400Nm im 2nd gear and call for 500 from 3rd gear on, while requesting 480Nm in the main table.

Now remember that as of TODAY (I´m sure they will correct it at some point,anthough it is wrong on the Fk8 also), the main torque table values are off so 600Nm is interpreted as 655 so it is about 9% off. (values in the requested torque per gear are correct)

So if I wanted to request 480Nm tappering down to 420Nm up top, this is how it should look like:

now for most gears I would tune them a little higher then that so it does not act like a limit:

And if I wanted to limit 1st gear at 350Nm this is what it would look like:

Now when going WOT in 1st gear, R mode, ECU will calculate the necessary boost to deliver 350Nm only. Thus, making it a "torque per gear" that translates to boost by gear in practice.

 

[Vito.FL5]

Appreciate your dedication and information, agree with others keep on posting, it's great to hear some about more of the technology and less about PPF debates .

Question for you - can you help explain how oil temp is derived? I couldn't really comprehend based on the code posted.

Also - regarding the dyno #s you posted, how do you deal with heat soak especially on the stock intercooler?

It basically models how the engine transmits heat to oil at a given load and speed and then how the oil dissipates the heat to the cooling system and to the environment based on factors like ambient temperature, vehicle speed, coolant temperature etc. I can dissecate this sometime later but the goal is to replace this with an actual read from the stock sensor present in some euro vehicles.

To deal with heat I modeled the LOAD limiting tables to limit load relative to intake temperatures, this prevents the system from heating to a point of boom boom. You could use "pressure ratio" tables also but they are a little tricky since table values does not translate directly into boost, as it is a relation between air pressure in the turbo inlet x air pressure in the turbo outlet so it is easier for me to use LOAD instead.

That table comes maxed out from the factory, but what I did is:

I tuned the car using that one line torque table until I reached the expected result. Next I get all datalogs from the dyno pulls to see the LOAD values for a given RPM, and with that I populate the load limit table at reasonable temperatures. At higher temperatures I lower the values so when heat starts to build up, ecu finds a load limit, thus limiting boost, and keeping things cool.

Let me ilustrate.

this is the table in its stock form:

Now from checking datalogs I saw that I run at around 250% load until around 4500 rpm tappering to 210 up top. I then assume that it would be safe to run this much load up until 60 degrees celsius IAT and from there graduelly reducing load so this is what it looks like now:

This way, at lower temperatures this table will not limit the tune, but when it starts to rise, boost will drop to let the system cool again.

 

[Jester04]

Now to some info you CAN use:

I was checking how Hondata managed to do their base tune for the R and ok.

I respect them A LOT as a company, and tuned hundreds of cars on S300 and dozens on Flashpros but I´d say their aproach is wrong. And lots of important missing tables. I´m nobody but If at any point HONDATA wants help from some idiot tuning Bosch ecus since carbs lol I would be happy to help. Anyway.

FIrst thing they lack is START OF INJECTION maps, at least for homogen injection, which is what the ECU uses at full load. That could give users the chance to get a little more juice from OEM injectors without experiencing misfires at high speeds or just leaning out.

Second thing is turbocharger model tables. This may look overkill but if you don´t set the turbocharger model properly, pre control will be far off and you will be relying in the PID to get you where you want, and that may never happen lol. Basic shit like a free flowing exhaust system will make boost control a little off.

Third thing and now trust me on this one.

Hondata basemap basically maxed out everything and tuned the car using compaction ratio tables, this ones

while tables like this are maxed out at the 100% TPS region

Its not necessarily bad to do that but why not make it a little more progressive instead?





Now...

FOCUS HERE:

THIS IS WHERE YOU SHOULD LOOK AT!

Why the f*ck is this maxed out when this is the table that calls for the actual torque your car delivers at a specific RPM.

Look at how this is modeled stock:

Recognise this curve??????? THis is your stock torque curve.

Don´t recognise the torque values??? seem a little high?

that is because they are wrong! For some reason whoever translated the data assumed that maxed value was 655 (probaly because it is a 16 bit value but whatever), when it is not. Maxed out value is 100% of 600NM which is the REFERENCE TORQUE VALUE, for all the torque model on the ECU.

EITHER WAY, when tuning this table, do not max it out, start with it STOCK, and then DRAW IT LIKE ONE OF YOUR FRENCH GIRLS, raising it little by little and see the magic happens.

of course you need to raise boost limiters at some point, but trust me, if you start from a stock calibration and raise NOTHING but this values, you will gain a lot of power before you find the pressure ratio (boost) limiters. Keep in mind to raise the torque limit per gear per driving mode because you may hit that limit also

I don’t have my software yet and will be able to contribute next month when I get it. The way that Hondata did it in my opinion is just max everything out to never get a throttle closure.


The right way to do it is to have the torque limits and airflow/ pressure ratios mimick the natural VE of the engine with the turbo that it has installed. No sense in dialing torque way past what is achieveable just for the sake of it. I prefer personally to tailor my tables and limits closer to the actual VE of the engine so the calculations are actually correct. When I get the software I will play with it and see what I come up with. Hopefully Hondata will do more table definitions eventually but I think they need to dig deep into training with bosch ecu’s they are very different from the older Honda ecu’s which is all they have know for forever.

Here you can see their thought process better.
https://fk8.wiki/Hondata_Tuning_Guide

 

[Vito.FL5]

I don’t have my software yet and will be able to contribute next month when I get it. The way that Hondata did it in my opinion is just max everything out to never get a throttle closure.


The right way to do it is to have the torque limits and airflow/ pressure ratios mimick the natural VE of the engine with the turbo that it has installed. No sense in dialing torque way past what is achieveable just for the sake of it. I prefer personally to tailor my tables and limits closer to the actual VE of the engine so the calculations are actually correct. When I get the software I will play with it and see what I come up with. Hopefully Hondata will do more table definitions eventually but I think they need to dig deep into training with bosch ecu’s they are very different from the older Honda ecu’s which is all they have know for forever.

You are 100% correct in every statement. It is always better to dial everything in toghether. The pedal feel, ride quality, mileage, longevity will all be better doing it that way. OEM´s are not stupid they do it like that for a reason.

I contacted a local rep for Hondata yesterday offering help to get them situated on what is needed to properly tune on a Bosch ECU and he said they know what they are doing and bla bla, and everything is done for a purpose and this and that but hopefuly he takes the message up to Doug and he listens. I love Hondata and I feel many people moved to standalones on the FK8 platform not because it was absolutely necessary, but because of the lack of tables and functions that are there in the OEM ecu just not defined by them.

I also dont know how bad of a struggle it is to deal with CARB bullshit and EPA knocking on tuners doors but maybe there is room for improvement.

this is the torque model base for our car, the percentage values up top are relative to 600NM.

That doesn´t mean the car get get to that much load on the stock turbo, but this is modeled based on compressor efficiency and perhaps a little extrapolated just to make it smooth.

when we request a certain torque value the ecu looks at this table (plus some other efficiency ones) to know how much airflow to look for, then calculate how much boost is necessary based on compressor effiency, turbine flow capabilities etc. Usually we´ll see higher load numbers when requesting a given torque % because it compensates for losses like friction, etc

 

[Jester04]

You are 100% correct in every statement. It is always better to dial everything in toghether. The pedal feel, ride quality, mileage, longevity will all be better doing it that way. OEM´s are not stupid they do it like that for a reason.

I contacted a local rep for Hondata yesterday offering help to get them situated on what is needed to properly tune on a Bosch ECU and he said they know what they are doing and bla bla, and everything is done for a purpose and this and that but hopefuly he takes the message up to Doug and he listens. I love Hondata and I feel many people moved to standalones on the FK8 platform not because it was absolutely necessary, but because of the lack of tables and functions that are there in the OEM ecu just not defined by them.

I also dont know how bad of a struggle it is to deal with CARB bullshit and EPA knocking on tuners doors but maybe there is room for improvement.

this is the torque model base for our car, the percentage values up top are relative to 600NM.

That doesn´t mean the car get get to that much load on the stock turbo, but this is modeled based on compressor efficiency and perhaps a little extrapolated just to make it smooth.

when we request a certain torque value the ecu looks at this table (plus some other efficiency ones) to know how much airflow to look for, then calculate how much boost is necessary based on compressor effiency, turbine flow capabilities etc. Usually we´ll see higher load numbers when requesting a given torque % because it compensates for losses like friction, etc

Yeah and I agree that it could be improved. My reasoning for what they are doing is that they are considering people who tune on the Hondata software. They try to simplify it so the layman can understand it. If they were to go deeper into the bosch ecu logic and try to keep the oem functionality most people here in the states would need actual training on Bosch ecu’s. Since the software is free and its not exclusive to certified tuners in their network. So its basically a workaround so everybody can do it. Which from a business standpoint makes sense. Where I live there is no real elevation changes so I’m a big fan of map based tuning instead of a maf based calibration. I would love to go speed density on the FL5 just for freedom of designing intake options. Or work around could be to move to a blow through maf based sensor placement so the intake is not restricted by the sensor.

 

[Vito.FL5]

I saw you talked about throttle closure lol, I heard you came from a BMW M2, BM3?

the R is not likely to solve things by closing the throttle, it is tuned much smarter then bimmers. Tuning the R is cakework compared to an M2

 

[Vito.FL5]

Yeah and I agree that it could be improved. My reasoning for what they are doing is that they are considering people who tune on the Hondata software. They try to simplify it so the layman can understand it. If they were to go deeper into the bosch ecu logic and try to keep the oem functionality most people here in the states would need actual training on Bosch ecu’s. Since the software is free and its not exclusive to certified tuners in their network. So its basically a workaround so everybody can do it. Which from a business standpoint makes sense. Where I live there is no real elevation changes so I’m a big fan of map based tuning instead of a maf based calibration. I would love to go speed density on the FL5 just for freedom of designing intake options. Or work around could be to move to a blow through maf based sensor placement so the intake is not restricted by the sensor.

yes, a map based tune is the simple way to go. I live in high elevation so I need to worry more but I do tune many VWs this way also.

I´m sure going SD is possible, once my car is delivered I will look into it deeper.

 

[Jester04]

I saw you talked about throttle closure lol, I heard you came from a BMW M2, BM3?

the R is not likely to solve things by closing the throttle, it is tuned much smarter then bimmers. Tuning the R is cakework compared to an M2

yep main way torque was managed was with throttle closures. I have not played the the Type R yet so no clue what logic they are using.

Sponsored