Dano's Boost Tuning 101 + Cool Stuff

[Enki]

Originally Posted by speedms6 why dont u put the load values to unobtainable values, as already stated, and then tune using the wgdc

Because on my car 2.5 was already unobtainable, and I already cut the wgdc, dynamics and everything else BUT load values by a large margin and it changed fuckall in my boost curves. I still hit 20 PSI.

An example from what I was seeing on the AP last night:
Targeted boost: 17 PSI @ 3500 RPM
Actual boost: 20 PSI @ 3500 RPM

Actual WGDC: 70 @ 3348 RPM, falling to 44 @ 3800 RPM and rising again to meet boost targets later in the RPM band.
WGDC target: 28.3 @ 3500 RPM (highest WGDC is 35 @ 5500-6000 with 0 @ 3000 and below)

The only thing I can think of to explain this is that it's trying to bring the WGDC up too quicky as I'm not commanding any until 3500 RPM; I'm modifying my map to test for this and will see if it quells the spikes.

[BlueStreak]

Originally Posted by rfinkle2 Two different means to the same end.

Not quite.

If you use the TRL X Gear tables to limit boost, you're also limiting load/torque/power.

For example, suppose you set your boost targets to 20PSI at 3K RPM. At 70 degrees ambient, this may translate to a calculated load of 2.2.

A few days/weeks/months later, temperatures drop. Calculated load will go up at every RPM point because the air charge is denser for a given amount of boost. If you're bent on hitting your boost targets, chances are you will not in colder weather because the engine hits the load target before hitting the boost target. Essentially, the car will *feel* more consistent power wise with temperature changes but boost targets will vary.

If you neuter the load tables completely, then it's all up to the boost targets to manage engine output (power). So, at 20 PSI in 70 degree weather, the car will feel "X" powerful. At 20 PSI in 30 degree weather, the car will feel "X + Y" powerful.

[Dano]

sounds like your WGDC needs some smoothing the extreme ramp up is causing the spike and hold at 20.

Originally Posted by Enki Because on my car 2.5 was already unobtainable, and I already cut the wgdc, dynamics and everything else BUT load values by a large margin and it changed fuckall in my boost curves. I still hit 20 PSI. An example from what I was seeing on the AP last night: Targeted boost: 17 PSI @ 3500 RPM Actual boost: 20 PSI @ 3500 RPM Actual WGDC: 70 @ 3348 RPM, falling to 44 @ 3800 RPM and rising again to meet boost targets later in the RPM band. WGDC target: 28.3 @ 3500 RPM (highest WGDC is 35 @ 5500-6000 with 0 @ 3000 and below) The only thing I can think of to explain this is that it's trying to bring the WGDC up too quicky as I'm not commanding any until 3500 RPM; I'm modifying my map to test for this and will see if it quells the spikes.

[rfinkle2]

Originally Posted by BlueStreak Not quite. If you use the TRL X Gear tables to limit boost, you're also limiting load/torque/power. For example, suppose you set your boost targets to 20PSI at 3K RPM. At 70 degrees ambient, this may translate to a calculated load of 2.2. A few days/weeks/months later, temperatures drop. Calculated load will go up at every RPM point because the air charge is denser for a given amount of boost. If you're bent on hitting your boost targets, chances are you will not in colder weather because the engine hits the load target before hitting the boost target. Essentially, the car will *feel* more consistent power wise with temperature changes but boost targets will vary. If you neuter the load tables completely, then it's all up to the boost targets to manage engine output (power). So, at 20 PSI in 70 degree weather, the car will feel "X" powerful. At 20 PSI in 30 degree weather, the car will feel "X + Y" powerful.

I should have said in my post, "all else being equal".

I think your post was helpful to those who may not understand what happens when ambient temperatures change though.

[BlueStreak]

Originally Posted by Enki Because on my car 2.5 was already unobtainable, and I already cut the wgdc, dynamics and everything else BUT load values by a large margin and it changed fuckall in my boost curves. I still hit 20 PSI. An example from what I was seeing on the AP last night: Targeted boost: 17 PSI @ 3500 RPM Actual boost: 20 PSI @ 3500 RPM Actual WGDC: 70 @ 3348 RPM, falling to 44 @ 3800 RPM and rising again to meet boost targets later in the RPM band. WGDC target: 28.3 @ 3500 RPM (highest WGDC is 35 @ 5500-6000 with 0 @ 3000 and below) The only thing I can think of to explain this is that it's trying to bring the WGDC up too quicky as I'm not commanding any until 3500 RPM; I'm modifying my map to test for this and will see if it quells the spikes.

Have you zeroed out the additive portion of the load dynamics table?

[rfinkle2]

Originally Posted by Dano2010 hey finkle...have you tried the new ATR to see if BT is resolved? ya know use OTS TRL xgear and see if it follows Boost targeting. Apparently Cobb believes they have it resolved in 2010-11 ECUs and will soon port it to earlier ECUs. Unless I read the release incorrectly.

Hey Dano, I believe you read the release incorrectly.

[Enki]

Originally Posted by Dano2010 sounds like your WGDC needs some smoothing the extreme ramp up is causing the spike and hold at 20.

Yeah that's what I'm thinking now too; just checked my logs from yesterday and I don't see it holding 20 anymore; don't even actually see 20 PSI in there anywhere so I ramped up my WGDC to near OTS values and interpolated down to 1500 RPM to smooth transitions a bit.

[Dano]

@ finkle

oh well...not the first time. post deleted to remove confusion.

[amoosenamedhank]

I actually had to lower my WGDC in the 2,500 rpm (IIRC) range from the OTS 50% value because I was boost spiking during spool up. I just effed with my load tables so now I have to start all over again but it was doing it's job nicely.

[Dano]

ok just to stir things up.

so from the below post it sounds like Cobb it targeting the fix for Gen2 in June then the rest afterward?


Maps for the 2011 MAZDASPEED3 feature Boost Target logic and include an new tuning strategy that increases low-rpm power delivery. This new 2nd Gen MAZDASPEED3 tuning strategy will be applied to all 2010 MAZDASPEED OTS Stage1 maps and we are targeting to have them available by the beginning of June. In addition, we are working on porting the Boost Target logic and all the Gen2 Stage1 maps to the Gen1 platform

http://blogs.cobbtuning.com/2011/05/06/4106/

[amoosenamedhank]

What about the MS6?!?!?!

[Dano]

Not this again?

we don't care about the 6ers...move on!

hahahah

Maybe when cobb refers to Gen1 they are including the 6es...they were after all the true Gen1 DISIs

[Enki]

Update:

Fixed all but 1 psi of the spike, which I adjusted in boost targeting and took a hair off of WGDC @ 4500 RPM to compensate. It should hit just a bit lower than target boost then adjust up to meet target now.

[R.MS3]

Interesting findings today, when helping somebody with his tune:

He was having problems with overboosting and fluctuating AFR's while i had a tune with boost tuning enabled and everything on the boost tables were to "spec". We couldn't find the reason why it was overboosting like that.

I think the issue happened when i raised most of the load values to like 2.3 to make sure the ECU would follow the boost tables. I raised the following load tables and immediately we were having problems with overboost:

- Abs. load targets (68.75 throttle up)
- All TRL per gear tables
- TRL A/C (68.75 throttle up)

My obvious guess is that by doing this the ECU would just follow the boost tables past 68.75 throttle, which it didn't.

So i reverted all the load changes back to what the OTS tune had and bam, his boost was right on target and AFR's were exactly like the OL/WOT tables.

Funny thing is, i have the same exact tune w/ all load tables jacked up for another guy and he doesn't have the problem. Anyway, i think you guys know half of this already but thought i would share.

[PapaSmurf]

did you modify your boost dynamics at all? They would solve your over boosting issues.

[R.MS3]

Originally Posted by MicaBlueMS3 did you modify your boost dynamics at all? They would solve your over boosting issues.

I don't think that was the problem. The boost dynamics had the right negative compensations for overboost and i haven't touched that table in a long time.

Proof for this was once i reverted just the load tables back to "normal" numbers, his boost was following exactly what i had set on the boost targets table.

[Enki]

I can confirm that changing the load tables affects your boost targeting. I was at 2.5 load requested and was overboosting by 3 psi until I dialed the load back down to 1.75; it behaves now. I changed some other stuff too, but I was overboosting on shift no matter where in the RPM band I shifted at until I changed the load tables.

Looks like the key is as was stated somewhere in this thread; match your load tables to what you actually hit at the pressure you want.

[manelscout4life]

Originally Posted by Meder I seem to be building boost a lot slower than I used to. I am on the oem turbo and I hit 10 boost fairly quick then build up to 16 ish which is my target. What do I need to look into to get spool up down? My wastegate and boost dynamics havent been touched yet, they are the stage 1 ots values. 08.5 ms3, intake only.

Did changing your BD table fix this?

[Dano]

Clearly setting up your ECU to follow boost targeting logic is a crap shoot at this point.

Looking back at all the posts, eliminating some of the earlier mistakes some have made, we still seem to have about 3 different overall categories that the ECUs fall into when switching to BT logic.

A) Everything works as designed [vast minority it would seem]
B) Adjusting TRL xgear tables to close to actual Calc load is required [Cobbs suggestion but why even switch to BT logic if you still have to tune TRL gear tables]
C) Adjusting TRL xgear tables to unachievable values and neutering the additive side of LD table values is required.

If you fall into cat B, I would just use load targeting. If you can tune your TRL xgear to within .05 of actual then you are already there with load targeting.

If you fall into cat C, that's fairly easy and the way Forzda and others did BT before Cobb came out with the magic toggle I wonder if the guys that fall into cat B would remove the boost toggle, if their ECU would move into cat C and behave better. It could be that Cobb's "logic" is mucking up the process.

For the cat A guys, consider yourselves lucky

According to Cobb's latest blog post they plan to have the BT logic fixed for the 2010-11 ECUs by June and then sometime afterward the gen1s.

[manelscout4life]

Yay for A

[Boosted Beluga]

Looks like I'm in the C cat. Better then B hahaha

Hey scout4life want to trade ecus?

[rfinkle2]

@Dano2010 , gen2 has a few different tables, more specifically, we have no Absolute Load Target Table.

Do you feel that table is essential to proper load tuning?

[Dano]

negative.

When using gear based load tuning the ABS table is set to unachievable and thus not needed.

on second thought...if the ABS table is indeed part of the Gen2 ECU and Cobb just hasn't exposed it, I could see it causing issues if it is there and set to a lower value than what your TRL gear table are set to. In general the ECU will follow the lowest valued load tables.

so if you are targeting 2.0 load in all your TRL xgear table and the ABS table is set to 1.8 guess what....your target will be 1.8 but you wouldn't be able to change it. I doubt this is the case though.

[Enki]

There any value in me moving from B to C? Should I drop boost targeting and go for the load checkbox?

[Dano]

it would be an interesting experiment. The way you are setup now, as AMB temps rise load will drop and your targeting will be off again.

if you were to remove the boost toggle and set all your load tables to unachievable and zero the additive side of LD and it worked, then you would set for all seasons with little need to change.

Once boost targeting works, I would have two different tunes one in the winter with lower boost and one in the summer with higher boost. My current setup is winter map at 20.5 PSI and summer/current map at 21.75. I would set it to 22 but I am on OE MAP sensor for now.

if you decide to give it a try please post up your results!

[BlueStreak]

Originally Posted by Dano2010 Once boost targeting works, I would have two different tunes one in the winter with lower boost and one in the summer with higher boost. My current setup is winter map at 20.5 PSI and summer/current map at 21.75. I would set it to 22 but I am on OE MAP sensor for now.

And if/when you upgrade the MAP sensor, what will you turn the boost up to?

[Enki]

Alright, I'll post some logs after lunch today

Edit:
Load dynamics is neutered on the undershoot (for now), with targets at 1.75 (should be about 18 psi mid range tapering to 16 top end) with my mods and max load set for 2.0 with 2.25 fuel cut. We shall see how this works.

[Dano]

@BlueStreak hahaha

I was just discussing that with Dustin the other day. IIRC he has worked on 3071s that really start to heat up around 24PSI even with WMI so I would guess 23 to start out and I would want to throw in an EGT before hand to get temps now and afterward.

I originally thought of 24ish PSI with a M10 nozzle and would change over from 100% meth to some mix with more H2O for added cooling.

so to answer you question IDK for sure but 23 for a start

anyone interested in a slightly used 2871 J/k

[BlueStreak]

It's kind've disappointing to know the 3071s start to heat up at 24 PSI. That puts the 2871 and the BNR S3 somewhere under that value (wonder if it's even worth going with an upgraded MAP sensor with 2871/BNR S3?).

So that puts the 3076 at a theoretical 26-27 PSI? Just musing here...

I know it all depends on power goals yadda yadda but 25+PSI just sounds... sexy.

[Dano]

well we didn't get into any great detail on what he called "heating up" but IIRC the ATP turbo's published efficiencies are as follows.

2871 = 20-21
3071 = 22
3076 = 26

I just thought if we push a K04 from 16 PSI to 21 PSI that a similar increase could be done on a BT but apparently they don't scale quite the same...LOL

I think 22-23 is doable on the GT28 and maybe later this summer I'll find out

I think the issue maybe the drop in VE will be exponentially larger than the HP gain more PSI will provide due to all the back pressure in the ATP housing and EM and the engine will be unable to expel exhaust gases efficiently. If you can't get air out you can't put more in, right?

[jwilkins88]

Originally Posted by Dano2010 well we didn't get into any great detail on what he called "heating up" but IIRC the ATP turbo's published efficiencies are as follows. 2871 = 20-21 3071 = 22 3076 = 26

I can testify to the 3076. They're basically dead on with that one. Dustin and I saw power moving to the left after 25 psi really.

My shit also gets really hot, and I'm pouring a shit ton of meth into it

[BlueStreak]

^^^ My dreams are being shattered with every successive post. Haha.

[Dano]

don't know how DJ had your 3076 boost curve setup Joel but I wonder if targeting higher boost in the meat of the curve then tapering off to the efficiency range in the upper RPMs would generate more power with minimal heating? well nevermind, on a 3076 the "meat" is up to redline. LOL

I would probably try something like 21PSI at 3K then 23PSI by 3.5K and hold to 4.5 or 5K then taper back down to 21 by 6.5K. Ya know as the turbo starts to heat up ask less of it so you basically flatline the heating at an acceptable level. This "theory" may only work on the smaller turbos if at all.

IDK that's my theory for now anyway and as I said the VE drop due to the ATP housing may negate any advantages.

Originally Posted by jwilkins88 My shit also gets really hot, and I'm pouring a shit ton of meth into it

Is it safe to assume you increased the H2O in the mix when pushing the PSI? Water does provide more evaporative cooling than meth.

[jwilkins88]

Originally Posted by BlueStreak ^^^ My dreams are being shattered with every successive post. Haha.

If you're wanting to run higher boost on a 3076, I'd just get a t3 turbo with a larger hotside. The thing that destroys the ATP turbo is the tiny, tiny, tiny, tiny hotside.

I think that a .82 a/r 3076 would be sittin' pretty @ 27psi, and probably have a bit more room to grow.

Originally Posted by Dano2010 don't know how DJ had your 3076 boost curve setup Joel but I wonder if targeting higher boost in the meat of the curve then tapering off to the efficiency range in the upper RPMs would generate more power with minimal heating? well nevermind, on a 3076 the "meat" is up to redline. LOL

My boost "curve" is very flat. My current EBC acts a lot like an MBC, so when I hit 26 psi, I stay at 26 psi. My new boost controller (it's on the way... I'm so giddy ) is a lot smarter, and it has more intelligence built into it. With it, I could do some boost tapering.

Originally Posted by Dano2010 I would probably try something like 21PSI at 3K then 23PSI by 3.5K and hold to 4.5 or 5K then taper back down to 21 by 6.5K. Ya know as the turbo starts to heat up ask less of it so you basically flatline the heating at an acceptable level. This "theory" may only work on the smaller turbos if at all. IDK that's my theory for now anyway and as I said the VE drop due to the ATP housing may negate any advantages.

I think that theory is sound with any turbo. This isn't the same thing exactly, but when I was on the dyno, my max PSI was 27. I saw a little bit more power (not much... like 7-10 whp iirc), but the power dropped off a lot sooner. It's not so much about the size of the turbo as where the efficiency drops off, imo. Tapering boost to redline would probably be a good thing for me because my power wouldn't taper off as soon in the RPM range. I'd rather be able to hold 450 whp to 7k than 460 to 6.5k.

Originally Posted by Dano2010 Is it safe to assume you increased the H2O in the mix when pushing the PSI? Water does provide more evaporative cooling than meth.

Yes sir. I'm running a 50/50 mix right now.

I just got crazy idea syndrome. Dynamically mixing your meth/water ratio. Have two independent WMI systems. Each would have its own tank. You'd need a fancy controller, but you could set the mix to be like 80/20 below 4K, 60/40 from 4k-5k, 40/60 from 5k-6k, and 20/80 from 6k on.

I don't know if there would be any benefit to that, but that sounds like a damn cool idea

[cld12pk2go]

My approach on Meth is that if you need more cooling then nozzle up. Water is simply going to cost you power compared to meth...

[jwilkins88]

Originally Posted by cld12pk2go My approach on Meth is that if you need more cooling then nozzle up. Water is simply going to cost you power compared to meth...

Well granted... But you have to admit that my idea was pretty fuckin cool

[Dano]

Originally Posted by jwilkins88 Well granted... But you have to admit that my idea was pretty off the wall

fixed

[cld12pk2go]

Originally Posted by jwilkins88 Well granted... But you have to admit that my idea was pretty fuckin cool

Yep.

Originally Posted by Dano2010 fixed

LOL.

[Dano]

Originally Posted by cld12pk2go My approach on Meth is that if you need more cooling then nozzle up. Water is simply going to cost you power compared to meth...

IDK the trouble with the nozzle up approach is too much meth down low. Even on a progressive controller you want SOME spray around 8-10 PSI and still want full spray before max PSI. If you move that entire window up the PSI curve you lose spray in the lower ranges.

I know you run off of MAF but it would be similar right?

I guess with enough controller tuning it could be done.

Maybe a better approach is E85 mix in your tank for the power aspect and some sort of H2O/Meth mix in the WMI tank.

or just turbo up LOL

[jwilkins88]

Originally Posted by Dano2010 or just turbo up LOL

+1