Experimenting with new boost target XDF

[keninger]

Hi

I had some time to play around with this "new" boost target xdf which I have attached. I am running a rescaled N20 tmap sensor tune at 155 load.

First I did a run with values set to 1.0

This is a log with this tune:
Datalog

Then I changed the values at 5000 and 6000 rpm like in the picture:

New datalog:
Datalog

I was surprised to realize that even though I have not touched values bellow 5000, so 4000 should be a solid data point, it has changed the whole boost target, and not only changed, but decreased values accross the whole rev range.

Has anyone been testing things with this xdf?

 

[AzNdevil]

the boost target is affected by other factors such as IAT and BRO
if you change it by a lot it will also affect the actual load target as well

 

[keninger]

I have a constant load target, and I tried to reshape the boost target with these new tables, and I gotta say, they work pretty well, but my findings are, that you can't increase boost target by a lot, because you will get throttle closures even though you are under boost.

at around 5500 I have higher target, and even though I remain under boost, I get throttle closures, like the throttle closure overboost calculations wouldn't be based on actual boos target, but on an other value, which doesn't get modified with these target modifying tables.

Here is a datalog:
Datalog with throttle closure, even at 5500 where boost is under target

 

[carabuser]

I only use these tables with values of 1 or higher now. These tables are used to manipulate PSSOL which is the boost target you see in logs but the DME looks at a different variable when considering throttle closures so unless you increase the leeway in the major throttle closure table you'll get closures.

Before removing the boost ceiling I used them as a method of going above 18.6PSI by turning down the throttle closures.

Now I just set my boost target at about 21PSI and use these tables to taper the target down at higher RPM as I'm still on stock turbos. The benefit of doing it like that versus just changing the load target is that the shifts seem better, I think when the load target changes quickly after a shift it takes a small amount of time for the boost target to catch up as the boost target is filtered slightly and that would cause throttle closures, with these tables the target will jump right back instantly.

 

[keninger]

I only use these tables with values of 1 or higher now. These tables are used to manipulate PSSOL which is the boost target you see in logs but the DME looks at a different variable when considering throttle closures so unless you increase the leeway in the major throttle closure table you'll get closures.

Before removing the boost ceiling I used them as a method of going above 18.6PSI by turning down the throttle closures.

Now I just set my boost target at about 21PSI and use these tables to taper the target down at higher RPM as I'm still on stock turbos. The benefit of doing it like that versus just changing the load target is that the shifts seem better, I think when the load target changes quickly after a shift it takes a small amount of time for the boost target to catch up as the boost target is filtered slightly and that would cause throttle closures, with these tables the target will jump right back instantly.

As I suspected, thanks for confirming. This table is great for tapering boost, but that's pretty much all it should be used for then.

 

[Tiron]

I have a constant load target, and I tried to reshape the boost target with these new tables, and I gotta say, they work pretty well, but my findings are, that you can't increase boost target by a lot, because you will get throttle closures even though you are under boost.

at around 5500 I have higher target, and even though I remain under boost, I get throttle closures, like the throttle closure overboost calculations wouldn't be based on actual boos target, but on an other value, which doesn't get modified with these target modifying tables.

Here is a datalog:
Datalog with throttle closure, even at 5500 where boost is under target

To get around the throttle closures, and still use this table to set your boost curve, couldn't you just raise the load target until it targets the maximum boost you want, and then reduce the other areas back down using this table? This means you're never raising the target with this table, only dropping it, which I'd expect to stop the throttle from stepping in.

 

[keninger]

To get around the throttle closures, and still use this table to set your boost curve, couldn't you just raise the load target until it targets the maximum boost you want, and then reduce the other areas back down using this table? This means you're never raising the target with this table, only dropping it, which I'd expect to stop the throttle from stepping in.

Yes, the only way to use this table is to actually reduce boost.

 

[mj6234]

This table seems useful to me and also very sensitive as well? I am trying to get a flatter boost curve since I have hybrids. The boost target between 3000-3500 is less than I needed so I have been raising it through BRO. Also have about a 0.8psi dip around 5750 that I wanted to smooth. That has other challenges, such as the scalar needs to be raised as raising the target through BRO will result in higher fuel trims in that area.

I rescaled this table and tried 0.99 instead of 1.012 at 3500 and target went up pretty substantially. Maybe 2.5-3psi (from 18.8 to 22.5). That was way more than expected.

Here is the log after the change. https://datazap.me/u/mj6234/log-1612411009?log=0&data=3-5-13-17-18-19-21-24-25-29-30-31

Log right before. https://datazap.me/u/mj6234/log-1612310191?log=0&data=3-5-21

After looking at another log that was puzzling to me where a tuner greatly increased boost target without changing load (like from 19psi to 25psi), I am almost certain this table was used to raise target. I don't think the only use is to reduce target.

I am also curious how VANOS affects this. You can see in my logs I have a flat spot where boost doesnt move toward target between 3300 and 3700. I am almost certain my VANOS changes there are what is driving that, but will test as soon as roads are dry.

Here is my current normal table.

 

[keninger]

This table seems useful to me and also very sensitive as well? I am trying to get a flatter boost curve since I have hybrids. The boost target between 3000-3500 is less than I needed so I have been raising it through BRO. Also have about a 0.8psi dip around 5750 that I wanted to smooth. That has other challenges, such as the scalar needs to be raised as raising the target through BRO will result in higher fuel trims in that area.

I rescaled this table and tried 0.99 instead of 1.012 at 3500 and target went up pretty substantially. Maybe 2.5-3psi (from 18.8 to 22.5). That was way more than expected.

Here is the log after the change. https://datazap.me/u/mj6234/log-1612411009?log=0&data=3-5-13-17-18-19-21-24-25-29-30-31

Log right before. https://datazap.me/u/mj6234/log-1612310191?log=0&data=3-5-21

After looking at another log that was puzzling to me where a tuner greatly increased boost target without changing load (like from 19psi to 25psi), I am almost certain this table was used to raise target. I don't think the only use is to reduce target.

I am also curious how VANOS affects this. You can see in my logs I have a flat spot where boost doesnt move toward target between 3300 and 3700. I am almost certain my VANOS changes there are what is driving that, but will test as soon as roads are dry.

Here is my current normal table.

View attachment 48910

The reason your boost is not moving closer to the target is due to the limitation of how far PID can go, you need to adjust the base.

Do NOT use values lower than 1.0, because if you meet the target, you will get throttle closures right away, or depending on how far you go below 1.0 you could get throttle closures without even getting close to the boost target.

Use boost ceiling and taper boost with these tables to your advantage.

 

[mj6234]

The reason your boost is not moving closer to the target is due to the limitation of how far PID can go, you need to adjust the base.

Do NOT use values lower than 1.0, because if you meet the target, you will get throttle closures right away, or depending on how far you go below 1.0 you could get throttle closures without even getting close to the boost target.

Use boost ceiling and taper boost with these tables to your advantage.
View attachment 48911

Yea, I know what you mean about WDGC base, I am still experimenting and working on things on my tuneg. The prior log is closer to target than this last one. I will go back to 1.0 or greater, but I still am not sure how else someone would target 25psi @ 160 load without using this table. The same person posted 19psi @ 160 load too.

 

[carabuser]

Yea, I know what you mean about WDGC base, I am still experimenting and working on things on my tuneg. The prior log is closer to target than this last one. I will go back to 1.0 or greater, but I still am not sure how else someone would target 25psi @ 160 load without using this table. The same person posted 19psi @ 160 load too.

Those are not tunes that you want to emulate. People that target high boost at low load are using boost scaling. Basically they make the engine tune act like a JB4, you scale the sensor such that it reports much lower boost than is actually present, so in your example their boost will be reading 15PSI but the actual pressure is 25PSI hence why the car is reporting 160 load. MHD just modifies the boost pressure logging parameter in your CSV files to compensate for this so your logs look pretty.

This method of tuning is trash in my opinion but until recently has been the only way of tuning above the boost ceiling on DCT cars. The currently implementation of the raised boost ceiling is only really suitable for manual cars hence people resorting to fudging sensors.

I altered the DME login in my cars to raise the boost ceiling in a slightly different way to the old COBB Race method so that the boost target can now be raised higher. COBBs method had a couple of errors which made it unsuitable for auto cars hence why shifts went to shit. I passed my version of it on to MHD a while ago so hopefully it'll appear in the next release

 

[TiAgAu]

How are you guys changing RPM breakpoints. Is there an updated xdf?

xdf above is only for main tables

 

[mj6234]

How are you guys changing RPM breakpoints. Is there an updated xdf?

xdf above is only for main tables
View attachment 48968

If you look at the parameters on those tables (right click->edit XDF), you can see the memory addresses of the row/column. Just add another table for each of the rows/columns with the memory address shown on the rows/column tab. Define a 1 row table with 6 columns and check the LSB box and add the conversion for the Y axis (the X doesnt have a conversion). It sounds more difficult than it is.

 

[mj6234]

Actually here you go. Here is the master one that I consolidated random things that came up after the original into one XDF. It has the axes in there.