Dano's Boost Tuning 101 + Cool Stuff

[cld12pk2go]

Originally Posted by BigjohnB20 LOL, Had to quote this. What was it, like the day before yours shit the bed that you posted this...

Pretty close...lol

My K04's death was two years in the making though...

[MattJackson86]

Thanks @Dano. While not near 3.0 load, I wasn't how close to that number I would have to be before the ECU tried to do something. Hence why I thought moving it to something really unobtainable would work better.

[Dano]

if you were at altitude I would agree you need to go higher. There is something like a 20% scalar or baro adjustment added to load calculations at altitude which will make a car with a limit of say 3.0 hit that limit at 2.8.

Dustin has seen it a few times on cars he has worked with up in the clouds where he and his unicorn live

[SRTie4k]

Does anyone have a Boost Error table that works well for the K04 and stock BCS? I've been jumping between my own values and Cobb's OTS values and can't find any decent common ground, and I would like some baselines to start over.

I also don't understand Cobb's OTS wastegate values - they all seem extremely high to me. Cobb's OTS values for most of the RPM range are almost double my values.

[Dano]

Cobb OTS DC values are so high because in OTS form the maps still use load tables even with the boost toggle checked. Load error is much faster reacting and has more authority to alter DC. Since it has both those characteristics WGDC doesn't have to be nearly as close as it would with all load removed from the error tables.

in short if you are going to boost target you throw the toggle, zero out LEC and LD then cut WGDC in 1/2 and start tuning.

BEC should be good for the K04 and OE boost controller and they typically don't need modification. With that said, I do like to add .02 to the -3.0 underrun so boost doesn't get stuck at .5 PSI under target. I also like to move my BEC ramp limit to 15% and think the 10% OTS limit is too restrictive It follows my rule that its much better to have WGDC too low than to high and let the ramp limit allow it to grow

[SRTie4k]

Originally Posted by Dano Cobb OTS DC values are so high because in OTS form the maps still use load tables even with the boost toggle checked. Load error is much faster reacting and has more authority to alter DC. Since it has both those characteristics WGDC doesn't have to be nearly as close as it would with all load removed from the error tables. in short if you are going to boost target you throw the toggle, zero out LEC and LD then cut WGDC in 1/2 and start tuning. BEC should be good for the K04 and OE boost controller and they typically don't need modification. With that said, I do like to add .02 to the -3.0 underrun so boost doesn't get stuck at .5 PSI under target. I also like to move my BEC ramp limit to 15% and think the 10% OTS limit is too restrictive It follows my rule that its much better to have WGDC too low than to high and let the ramp limit allow it to grow

You're talking about Cobb OTS values?

[Frequentflyer]

Still getting some knock in the higher gears, which I never really got before. I haven't changed my plugs in a while, so I'll swap them up soon, but I can't see that as a cause.

Still boost tuning. As you can see by my log, right after my shift to 5th, I get 2.00+ KR that doesn't go away so I let off the gas. Only thing I can think of is that boost temps got high enough after 3rd and 4th gears running 19+psi that it's giving me some KR. This was last week, so intake temps weren't that high, although I never logged them. I'll probably lower target boost back to 18 and see what happens.

[MattJackson86]

Originally Posted by Frequentflyer I haven't changed my plugs in a while, so I'll swap them up soon, but I can't see that as a cause.

How often do people seem to change plugs? I am probably every 15,000 miles or so.

[Frequentflyer]

I was doing it once a year, but I drive 24k a year. I think that's still probably overkill, but they're only $25 and this car seems to be hard on plugs. I'm not getting high RPM misfires, so I think the plugs are fine. I've been running these Autolite Iridiums for quite some time now.

[Nataphen]

@Dano, I have a new question for you, buddy. I've been tuning my MS3 for a while and never encountered this problem, but the MS6 has brought on a new set of challenges.

In part throttle CL exit, like slight acceleration on the highway to pass, the car is very jerky. I started with the same tweaked OTS throttle tables, WGDC, CL exit settings, etc. as my MS3. After I discovered what was causing the issue, I interpolated vales in WGDC and DBW to taper up to max more gradually at those RPMs and load, but the problem remains. My CL exit load is 1.25, and as soon as the car hits that load, it feels like it wants to come out from under me. Any thoughts?

EDIT: I just changed my APP trans tables to direct in N-5th because I noticed that there was a significant change in translation curve around 20-25% in the OTS tables, and this is about where I'm getting my surge. I'll give it a shot in the morning.

[Meder]

Still trying to fix my unstable afr. Can raising boost change afr? I am seeing anywhere from .2 higher most of the time which is possibly normal to .5 higher a very small percent of the time. I am tuning wgdc at the same time and have went from 12-13 boost to 17 or so in the past few maps.

More boost at X rpm/maf volts = more incoming air that the maf is not tuned for in the range?

[Enki]

The only things that should skew your AFRs are as follows:

1. Dirty/damaged MAF
2. Intake leak
3. Exhaust leak
4. Change in fuel mix (more E = leaner without tune)
5. Tune (specifically, MAF cal)
6. Dirty/damaged o2 sensor
7. Pizza in the intake

[Nataphen]

The change to direct throttle helped keep me at lower load levels when driving around, but as soon as I get to that transition at 1.25, it happens again. Granted, with direct throttle, you're pretty much going to floor it by the time you get to that point. So, it didn't fix my issue, but it helped take it more or less out of reach.

6th gear is unchanged in APP, so the problem is the same there. Direct throttle in 6th makes the cruise control go ape shit, so I'll have to keep trying new ideas until I find a solution. I may scale the throttle position down a little so the translation is smoother up to about 75%, then go direct from there to 100%. I'll report back, hopefully this is an issue that other guys are having and this can be of some help.

[Enki]

You could try remapping your throttle requested load tables (under load section) and see if that helps.

[SRTie4k]

I finally found BEC values that don't make my boost fluctuate like crazy.

0.04	0.03	0.02	0.02	0.00	0.00	0.00	-0.02	-0.02	-0.03	-0.04

These are what I'm currently using and they seem to work perfectly to keep boost within .2psi of target (so long as WGDC is set within reason).

I also got a chance to test the new boost tuning method in extremely cold weather. We got temps around 32* yesterday. I used to get very bad boost creep below 32* with Cobb's previous boost/load hybrid maps, but the little I got to get on it yesterday (it was raining), all I got was a slight spike to 17.5psi at 3000RPM (normal on my map), 17psi held, and the normal ramp from 15psi to 17psi from 2500-3000.

This give me pretty good hope for later this winter, that I won't be badly creeping to 19psi at 4500RPM on those 10* days. Crossing my fingers.

[rfinkle2]

@Dano...

I have a few q's about pre-boost tuning that I was hoping you could answer when you have a free moment.

1) did you find the load dynamics table to have an effect on wgdc?

2) did you guys have access to a wgdc boost error comp table?

[Dano]

Originally Posted by rfinkle2 @Dano... I have a few q's about pre-boost tuning that I was hoping you could answer when you have a free moment. 1) did you find the load dynamics table to have an effect on wgdc? 2) did you guys have access to a wgdc boost error comp table?

Honestly can't remember if LD altered boost. Supposedly now it only effects TPS

We didn't have LEC or BEC nor the ramp tbls.


Tappin

[rfinkle2]

Originally Posted by Dano Honestly can't remember if LD altered boost. Supposedly now it only effects TPS We didn't have LEC or BEC nor the ramp tbls. Tappin

Thanks for your time Dano...

So you must've used the LD to alter the wg...

LOL. Who knows.

As long as in the end you get the job done I guess.

[Dano]

If I had to guess the table names changed. LD must have altered WGDC or its logic back in the day whereas now it only alters TPS.

yes all my load control was done with LD and WGDC. back then WGDC didn't really have to be all that close to get the job done. no where near as close as it does for boost targeting.

[Dano]

Originally Posted by Nataphen The change to direct throttle helped keep me at lower load levels when driving around, but as soon as I get to that transition at 1.25, it happens again. Granted, with direct throttle, you're pretty much going to floor it by the time you get to that point. So, it didn't fix my issue, but it helped take it more or less out of reach. 6th gear is unchanged in APP, so the problem is the same there. Direct throttle in 6th makes the cruise control go ape shit, so I'll have to keep trying new ideas until I find a solution. I may scale the throttle position down a little so the translation is smoother up to about 75%, then go direct from there to 100%. I'll report back, hopefully this is an issue that other guys are having and this can be of some help.

@Nataphen

I just ran across this but can't think of anything that would cause it if all your CL exit tables match since the MS6 CL exit is not the "D" table IIRC.

I am sure all your fuel tables match and are interpolated.

does anything show up in any logs around that jerky point? you might need to cull it down to AFR, Load, TPS, ign, and RPM to get higher res.

oh and lastly...you may have better luck getting a response if you PM me...I don't look at my notifications regularly lol

[TexasSpeed3]

subbed

[Nataphen]

The MS6 tuning is on hiatus, my wife wanted the AP back on her MS3, LOL. I guess she appreciates the difference more than I realized. An AP for the MS6 will be my next mod purchase so that I can continue. FYI, a bone stock MS6 apparently won't hold high boost pressures past 4500RPM. It must be because the exhaust is more restrictive than the MS3.

[Meder]

Originally Posted by bmorrisj ok what was the trick to have AFR drop a little faster.. im targeting ~12.1 - 12.2 ... when i hit it (you know 3rd gear around 3k) my AFRs hang around 13.3~13.5 but after a shift it hangs on normally.... but damn i see 2.4 calc load on there!

Originally Posted by Dano Apparently a fairly common occurrence lately. What does your CL max throttle look like? Tappin

Did you guys ever figure this out? My car does this and I've been through the car time and time again dealing with this issue for 6 months now. Good maf, good 02sensor, no leaks. It follows its target after the first pull everytime but will sometimes stray off by .3-.5 on the first pull. It does not do it everytime on the first pull. Does yours bjmorris?

What can I do to tune around this besides run lower afr?

Car is a 08.5 ms3

[Dano]

I haven't seen this happen in awhile now...probably tuned 4 cars of varying models since and it hasn't come up again.

*knocks on wood*

Are you running the latest AP/ATR software?

And are your fuel trims close to zero in CL? Example, if your trims are like +10 right at the transition it might take awhile to get AF down when you go WOT.

other than that the other things mentioned should get the AF down almost immediately.

[Meder]

Yes to all those ^

I cant figure it out. Once I get this downpipe on and new fuelpump in ill see if its still doing it. Hopefully going through both the intakle and exhaust sections again revels some sort of leak.

Its odd though because it only does it 1 in 6 pulls or so.

[Dano]

So I have been having issues of late...or more correctly the issue has been there all along but doesn't happen all the time nor on every car, but since joining PERM tuning I work on more tunes now.

The issue is lazy or non existent boost error compensation on a boost target tune. The recent history can be found starting here:

MAZDASPEED AccessPORT and AccessTUNER Get Big Updates

I think a possible solution may lie in the BEC, Ramp limiter and/or RPM Comp tables.

Post taken from the Cobb thread referenced above

Originally Posted by Dano I am about to get midevil with my BEC table...stay tuned... but yes typically I leave the .15 cell alone because it can cause oscillation. It just depends on how the boost ramps up to target...if its a gentle ramp it can get stuck ~.4 psi under target...if it is a more aggressive ramp it will pass that cell and run target or slightly over. this second example is what I shoot for and I am going to give OEM [not OTS] BEC values a try...or something close to that aggressive. Prab start out at 1/2 the OEM values. I already use -0.14 in the first overrun cell of BEC on problem tunes so I am not that far off. OTS BEC 0.04 0.02 0.02 0.00 0.00 0.00 -0.02 -0.14 -0.22 -0.44 -1.10 OEM BEC 0.70 0.34 0.25 0.22 0.13 0.00 -0.16 -0.30 -0.46 -0.70 -1.40 theory: since boost targeting is less responsive than load targeting why not put in more aggressive values in BEC? if the control loop is always looking at BEC + LEC values in each iteration, and pure boost tuning zeroes load out..we are getting 1/2 the resolution in the control loop...right? we only get correction 1/2 the time since load is ignored. anyway Dustin and I discussed this and think it makes sense. Cobb's OTS BEC settings are designed with load targeting in mind even though its called Boost Targeting. They plainly state to set your TRLs within .05 of logged values. We [by that I mean me] neuter all load from the calculation so I need more power within BEC. should have thought of this before. I will resurrect my Boost tuning thread with my results.

so I am going to experiment starting with BEC values and go from there exhausting each table until it makes no progress then move on to the next until my head splits open or I get results whichever comes first.

since this doesn't happen to every car this may take some time....troubleshooting 101 "if you can't reproduce the problem you can't fix it"

I currently have a tune exhibiting this issue but almost have it worked out with traditional means so I don't want to wear this guy out and winter is upon us so it may be spring before any ground in this area is made.

stay tuned....

[silvapain]

I found that changing the A & B values from -20 and 10 to -50 and 20 respectively helped me when I was having issues where the BEC wasn't adjusting my WGDC at all.

I know COBB states that it's just the summation of the two values that matters, but my question is why didn't Mazda just use one value then? And what happens with severe overboosting issues? For those reasons I felt it prudent to increase the positive value some, and still reduce the negative value such that the summation was more negative than before.

My load compensation tables are all zeroed out.


Tapadatass

[BigjohnB20]

Originally Posted by silvapain I found that changing the A & B values from -20 and 10 to -50 and 20 respectively helped me when I was having issues where the BEC wasn't adjusting my WGDC at all. I know COBB states that it's just the summation of the two values that matters, but my question is why didn't Mazda just use one value then? And what happens with severe overboosting issues? For those reasons I felt it prudent to increase the positive value some, and still reduce the negative value such that the summation was more negative than before. My load compensation tables are all zeroed out. Tapadatass

I think you guys are on to something with the rpm comp tables. I was having and issue with strictly boost tuning where rolling into the throttle vs punching it would give me very different results. Basically rolling into it would cause me to not hit boost or WGDC targets, and the ECU didn't really want to do much to address the situation.

It seems to be much better now after they released the new v200 maps. I changed a lot of tables to better match some of the OTS tables, but one of the main ones was the RPM comp A/B.

I think I originally had it set up so the A+b was alwasy super negative. Now I changed it so that A+B equals the max ramp limit. I think I have 16 high ramp limit and A=-20, B=4. In stock on OTS form, High ramp limit is 10 and -20+10 = -10.

[rfinkle2]

I think the entire idea of the 2 comp tables is like a coarse adjustment and a fine adjustment in stock form.

The load error comp tables are really aggressive, and get you in the neighborhood, whereas the boost error comp tables then "catch" the momentum, and fine tune you to a particular boost target.

I think the big differene in the two (load vs comp tables) is the continous operation of the wgdc boost error comp summation.

From what I can see, that is what allows the load error comp tables to apply full %'s to the base wgdc table without overboosting / overloading.

[rfinkle2]

Originally Posted by silvapain I found that changing the A & B values from -20 and 10 to -50 and 20 respectively helped me when I was having issues where the BEC wasn't adjusting my WGDC at all. I know COBB states that it's just the summation of the two values that matters, but my question is why didn't Mazda just use one value then? And what happens with severe overboosting issues? For those reasons I felt it prudent to increase the positive value some, and still reduce the negative value such that the summation was more negative than before. My load compensation tables are all zeroed out. Tapadatass

Cobb says that severe overboost is dealt with by the load error comp tables.

[Dano]

if they are populated

I say make the BEC more "course" at the outer edges to give it the control LEC has but smooth it in the middle...IDK if that will work in practice tho.

[silvapain]

Originally Posted by Dano if they are populated I say make the BEC more "course" at the outer edges to give it the control LEC has but smooth it in the middle...IDK if that will work in practice tho.

Exactly. 100% E85 guys can't use load tuning at all. There are undiscovered tables at play that limit the throttle plate and WG long before we hit load targets. I was getting 0 WGDC and TP closing with fuel cut at 3.2 load when my targets were 4.05 load with the OTS LEC table values.


Tapadatass

[Dano]

sounds like if you really want to load tune you'll need to implement load cap fixes

negative MAF scaling
altered AF targets
altered load targets
timing table changes
blah blah blah


it can be done though...been there done that...although not sure if there is even enough resolution to make it feasible with those load values...you might just fall right under the new load cap.. IIRC I just went 15% lower on my MAF and stayed under the old 2.0 cap.

[silvapain]

Originally Posted by Dano sounds like if you really want to load tune you'll need to implement load cap fixes negative MAF scaling altered AF targets altered load targets timing table changes blah blah blah it can be done though...been there done that...although not sure if there is even enough resolution to make it feasible with those load values...you might just fall right under the new load cap.. IIRC I just went 15% lower on my MAF and stayed under the old 2.0 cap.

Not worth it at all. I'd rather do the extra datalogging work to dial in my WGDC and stay straight BT.

P.S. Those are all the things I already have to do for E85.

P.P.S. I did make sure my max calculated load tables were over 4 as well.

[Dano]

roger that.

Got logs back tonight from my current tune and my BEC change did nadda.

his error starts out around .2 psi underrun and increases from there to 1.5 PSI underrun with no appreciable DC change

so I added .02 to the .15 underrun cell of BEC and absolutely no change.

time to get midevil...sooner rather than later.

I only want to change one table at a time to see what actually works....I'll continue to ramp up BEC until I reach OEM values and if still no change I'll move on to RPM Comp.

[Lex]

Ugh that BEC table does do very little unfortunately. Hope you nail this one Dan.

[cld12pk2go]

Originally Posted by Dano time to get midevil...sooner rather than later.

Speaking of going midevil...

I am fighting the AP Load Cap V2™ where load falls off up top down to ~2.1 at 6500 RPMs regardless of any boost or load table values (using either of the two load targeting strategies).

The map I am loading this morning is going to test to I can get around this by limiting the "WG Duty Boost Error Ramp Limiter - Low Limit" and "WG Duty Load Error Ramp Limit - Low Limit" from -100% to -20% such that my top end shouldn't have authority to limit power up top as much. This is obviously playing with fire, but my WGDC tables are configured such that I should only be ~3% higher than my logged WGDC from yesterday if this works.

TBD if this is a workable solution. If not, then I guess I am back to the old school MAF curve lightening bolt + rich fueling tables to try to work around it.

[Dano]

lol @ load cap v2tm


GL with your experiment!

[Lex]

The interesting and worrisome issue here is that the compensation tables imply we only have access to a proportional aspect of the PID controller. On top of that we have two competing/parallel strategies - boost and load ... and if that was not enough, we also have strange limiting tables on what the controller can do.

In spite of what it might appear like - the ECU is actually set up logically although our explanations and ways of "tricking" it fall a little outside of the logic that should be there.

First of all, I don't think the BEC and LEC tables necessarily imply what we see - or they are not the only compensation tables (ie, there is also a table for the integral and derivative part of the control algorithm).

If you've got a K04 car running mostly gas and the stock WG solenoid, the stock logic can stay largely unchanged.

If you don't the next step is boost tuning .... but that is limited as well in terms of solenoid/boost response. So an MBC may be in order and then you just have to worry about the throttle closing.

The best solution is to discover the rest of the tables and logic allowing us complete freedom with load tuning. Otherwise, feels a little like boxing in a closet .

[Dano]

oh we are def boxing in a closet...with the light out sometimes

but even now the high HP guys can't load tune do to some other cap yet to be uncovered.

woe is us.

David did label each part of the PID controller but that definition may change at some point in the future. It was in his logic flow chart but I suspect there is something missing there and why BEC seems to "sleep" sometimes.

I would "guess" and I stress guess..that at some point and only some times a limit or process is reached where the BEC portion of the loop is skipped and only the load portions are calculated into the loop. This is why I get no correction on maps with LEC/LD zeroed

Wastegate Controller is a Proportional Integral Derivative controller (PID).
P = Boost/Load Error Compensation
I = Boost Error Limiter
D = RPM compensation switch