[themyst]

After observing more AP logs and reading more into load targeting versus directly controlling the wastegates, I am believing that the differences are a matter of execution.

The AP doesn't target boost levels, it targets a load index across the RPM range, which in most cases is around 13psi tapering to 11 at redline. Depending on ambient conditions (elevation, IATs, barometric pressure etc.) the car will trim throttle, reduce or increase boost as necessary in order to achieve the said load targets. On the AP, boost overshoot, throttle closure isn't an issue whatsoever since the load index itself is not affected. The tune trims throttle, dumps boost, whatever it needs to do in order to achieve the load target. If I am not mistaken, the argument from the flash camp is that this method of targeting allows the full utilization of the feature set of the DME.

This is both positive and negative in a way. Positive for those who want consistent performance given the conditions (road racing) versus those who want to reap the benefits of lower DA, colder ambient conditions, etc by targeting boost.

In regards to timing, EVERY single AP log I've seen from Clap or Alan show a spot-on timing curve (2 degrees at low RPM ramping to about 10 degrees at redline) which in my opinion means the curve as designated by Cobb is spot-on for the conditions, if not a little conservative.

Now how does this relate to the Procede?

Since the Procede has direct control over the wastegates and the subsequent boost targeting, it does not target a load index; it targets boost, while offsetting the crank position sensor to reduce, or soon, increase timing over the stock ignition set point. Offsetting the crank position sensor (AKA CPS offsetting), in its current form, allows manipulation of ignition advance, following the stock curve. I am not certain if it is possible for CPS offsetting to target a specific timing curve like the AP does at this time.

The argument from the piggyback camp shows that it is more beneficial to directly control the wastegates, instead of trimming throttle or allowing boost overshoot events in the first place. As we've seen on the various AP logs, the logs are not consistent across the board when it comes to boost. One overshoots, one has major oscillation issues, and one is spot-on, ironically, with full bolt-ons on a map not intended for them. In this regard, Shiv's statement that no two wastegates are built equal appears to be a valid one.

Although Cobb acknowledges the overshoot issues are something that need to be addressed, and the fact throttle trimming is less than ideal, both instances do not appear to affect the smoothness or drivability of the Cobb AP, since the boost is being dumped pre-throttle body to achieve the targeted load index.

The Procede beta maps that are currently out for testing incorporate a new level of control that was only recently discovered, which opens up a new possibility in regards to power gains. Shiv can chime in regarding the specifics if he wants.

In regards to the JB4, the recent addition of CAN access and direct boost targeting is much welcomed. Previously, the JB3's mapping was more a multiplier algorithm on top of stock boost, which is why we saw inconsistent boost levels and curves in a multitude of conditions, all else equal. The JB4 has the most aggressive timing curve out of all three of the solutions presented, as it is attempting to target 10 degrees in the midrange ramping to 14. The challenge we are faced with, especially in the case of a freshly installed JB4, is the car will not be able to achieve the targeted ignition set point, and thus will result in knock, still attempting to ramp to the targeted ignition set point after the knock event. Being that this car has a highly intelligent DME and a STRONG block to withstand repeated knock events, the DME will eventually recognize this, and less aggressively ramp ignition after a knock retard event. In my opinion, a freshly installed JB4 should be placed on a very conservative boost level initially, as knocking at 14psi is far worse than knocking at say, 11-12psi on pump gas. Either way, allowing the DME to correct ignition advance in a reactive fashion is far from optimal. In my honest opinion, running methanol injection at a modest 15-16psi boost level with proper failsafes on this tune would make the timing argument a bit less relevant.

ALL of this information may be obsolete when the Procede tuner software and Cobb ATR come out, allowing those with tuning knowledge to maximize the potential of our N54 motors.

As I stated before, I have ZERO tuning experience on any platform. I am simply trying to put all the information in one post I've gathered from logs and discussing with more knowledgeable folks such as Clap135 and others in this community.

Please feel free to correct any misinformation I may have provided in this post.

[ianbiz]

Good info. This thread was useful for a bit: http://www.e90post.com/forums/showthread.php?t=475789

[hipnotic]

nice information

[themyst]

Forgive me, I have too much time on my hands lately. With my car in the body shop, I feel like a soccer player with a broken leg.

[Enicinitas]

Good info Myst. Thanks for taking the time to spell it out for some of us. I'm a tune sampling whore and I have no problems with it. My car has been knocked in an accident as well as in the engine - I'm sure that if it could talk it would moon me and and flip me off.

[Derek B]

Wow, Great reading. This explanation was very helpful in showing me the different aspects of the tuning solutions that we have for the N54.


Thank you

[Imperator]

In this forum: People freaking out over the smallest details as they attempt to identify the best tune capable of extracting the most power on a white, black, blue, silver or grey car on a Sunday morning at 3:00am running on pump gas with the most and least-relevant bolt-on mods available.

Reality: Any one of the tunes will yield the same result. You don't mind driving right off the compressor map, right?

Solution: "Upgraded" stock turbos featuring ported turbine housings, clipped compressor wheels and the same welded manifolds are available for the low, low price of $4199.

[OpenFlash]

Themyst-- Great work collecting all that info. I just added my 2 bits in below...

Originally Posted by themyst

Now how does this relate to the Procede?

Since the Procede has direct control over the wastegates and the subsequent boost targeting, it does not target a load index; it targets boost, while offsetting the crank position sensor to reduce, or soon, increase timing over the stock ignition set point. Offsetting the crank position sensor (AKA CPS offsetting), in its current form, allows manipulation of ignition advance, following the stock curve. I am not certain if it is possible for CPS offsetting to target a specific timing curve like the AP does at this time.

The Procede can target any given timing advance within a 30 degree window (+/- 15 degrees). Which means that is more than capable of offsetting ignition advance in the required range (~+/- 4 degrees). With a conservative low boost tune, the ignition advance setpoint will almost always be lower than the knock threshold. Which means "spot on" timing logs.

Like a flash, the Procede establishes a boost target which its PID boost control system targets, trying to keep error (target boost - actual boost) to a minimum. Whereas the DME (stock, Cobb, GIAC) defines the boost target based upon achieving reasonably constant airflow (ie, more boost in hotter conditions when air is thin, less boost in colder conditions when air is desnse), the Procede defines it's target boost based upon basic tuning fundamentals (ie, the engine can happily run more boost on a cold night but less boost on a hot summer day). Besides intake air temp (IAT), other contributors to Procede boost target calculations:

-Coolant Temp
-Oil Temp
-Historical knock activity
-Barometric pressure

The argument from the piggyback camp shows that it is more beneficial to directly control the wastegates, instead of trimming throttle or allowing boost overshoot events in the first place. As we've seen on the various AP logs, the logs are not consistent across the board when it comes to boost. One overshoots, one has major oscillation issues, and one is spot-on, ironically, with full bolt-ons on a map not intended for them. In this regard, Shiv's statement that no two wastegates are built equal appears to be a valid.

According to customer data (wastegate compensation values), our wonderful turbo/wastegate assemblies result in a 10-20% range of applied solenoid duty cycle. In other words, one particular car may need 40% DC to achieve say 13psi at 4000rpm. While another car would need 45-50% in the same exact conditions. More if conditions are dissimilar. So that means with any 1 given (static) tune, somes cars on one extreme will undershoot and be lazy to build boost. While other cars will overboost and throttle oscillate. While cars "in the middle" will work well and exhibit neither extreme behaviors.

Although Cobb acknowledges the overshoot issues are something that need to be addressed, and the fact throttle trimming is less than ideal, both instances do not appear to affect the smoothness or drivability of the Cobb AP, since the boost is being dumped pre-throttle body to achieve the targeted load index.

Throttle closure is a secondary solution to boost control. It's used only when the primary approach (wastegate control) isn't doing a good enough job. Mild to moderate (down to 50%) short term throttle closures will usually go unnoticed in low boost tunes. More serious throttle closures (ie, blade opening dropping to 20%) will be obvious to the driver. At higher boost (15psi and up), even moderate closures will be noticeable due to intake charge turbulence and engine VE losses.

The Procede beta maps that are currently out for testing incorporate a new level of control that was only recently discovered, which opens up a new possibility in regards to power gains. Shiv can chime in regarding the specifics if he wants.

The new features activated with the v5 Beta maps have mostly to do with AFR targeting and actual ignition advance management. The former makes sure not only the desired AFR target is achieved, but it is achieved with desirable short term corrections, in both steady state and transient conditions. The latter introduces a more dynamic form of ignition control management that better suits a wider range of engine loads. With autotuning active, average required ignition correction and knock retard will be used to adjust global boost pressure accordingly.

In regards to the JB4, the recent addition of CAN access and direct boost targeting is much welcomed. Previously, the JB3's mapping was more a multiplier algorithm on top of stock boost, which is why we saw inconsistent boost levels and curves in a multitude of conditions, all else equal. The JB4 has the most aggressive timing curve out of all three of the solutions presented, as it is attempting to target 10 degrees in the midrange ramping to 14. The challenge we are faced with, especially in the case of a freshly installed JB4, is the car will not be able to achieve the targeted ignition set point, and thus will result in knock, still attempting to ramp to the targeted ignition set point after the knock event. Being that this car has a highly intelligent DME and a STRONG block, the DME will eventually recognize this, and less aggressively ramp ignition after a knock retard event. In my opinion, a freshly installed JB4 should be placed on a very conservative boost level initially, as knocking at 14psi is far worse than knocking at say, 11-12psi on pump gas. Either way, allowing the DME to correct ignition advance in a reactive fashion is far from optimal. In my honest opinion, running methanol injection at a modest 15-16psi boost level with proper failsafes on this tune, makes the timing argument a bit less relevant.

We are fortunate that we have strong blocks and a good knock control system. Because those relying on what is essentially a boost controller to double (or at times triple) stock boost pressures would not have successful results in most other engine/platforms. Those clever Germans.

[themyst]

@Shiv - Regarding the V5 beta maps, will the ignition advance parameter be re-scaled to accommodate the ability to advance ignition when conditions allow? e.g. meth/race.

@Cobb - In your Stage 1 OTS maps, are you changing the tables of any features not accessible by any piggyback? e.g. VANOS

One question went unanswered from all parties involved however; is it possible to monitor the knock activity of the hottest cylinder, which seems to be Cylinder 6? From my understanding, both piggies and even the AP are monitoring Cylinder 1 in their respective datalogs. After finding this bit of information out, I think many of us would feel a lot safer monitoring the hottest cylinder for knock activity.

[OpenFlash]

Originally Posted by themyst

@Shiv - Regarding the V5 beta maps, will the ignition advance parameter be re-scaled to accommodate the ability to advance ignition when conditions allow? e.g. meth/race.

Absolutely. Expect the most aggressive meth/race gas maps to target ignition advance values 4-6 degrees greater than that of pump gas maps.

@Cobb - In your Stage 1 OTS maps, are you changing the tables of any features not accessible by any piggyback? e.g. VANOS

I'm not Rob, but we experimented with VANOS mapping (offsetting the cam position sensors) using what were unused digital i/o channels a couple years ago. Quite easy to do with piggbacks and stand-alones. Couldn't find any worthwhile gains. Other tuner's findings may differ, of course.

One question went unanswered from all parties involved however; is it possible to monitor the knock activity of the hottest cylinder, which seems to be Cylinder 6? From my understanding, both piggies and even the AP are monitoring Cylinder 1 in their respective datalogs. After finding this bit of information out, I think many of us would feel a lot safer monitoring the hottest cylinder for knock activity.

Just as a note, I have yet to see evidence that cylinder 6 is the most knock prone. Nor do I recall seeing any failed engine tear downs suggest this to be fact either. Besides that, we just use Ignition Advance on cylinder 1 as a general reference for tune aggression relative to the stock tune. Not as a specific value to apply a discrete offset upon. So not monitoring each individual cylinder for knock isn't necessary unless you have a "bad" cylinder. But I'd say the rattly noises, copious oil consumption, loss of power, smoky exhaust and various DME codes would probably be the first indicator of such a problem.

[themyst]

Shiv - I guess my direct question was how many degrees are we meth guys going to be able to advance beyond the stock ignition set point? 4-6 degrees over pump gas is so vague, as some of us run that much more ignition over pump gas on the stock set point.

[OpenFlash]

Originally Posted by themyst

Shiv - I guess my direct question was how many degrees are we meth guys going to be able to advance beyond the stock ignition set point? 4-6 degrees over pump gas is so vague, as some of us run that much more ignition over pump gas on the stock set point.

About 2-3 degrees beyond stock advance set point. Worth pretty big gains in the 5000-6000rpm when octane permits.

[etiliko]

Originally Posted by themyst

The AP doesn't target boost levels, it targets a load index across the RPM range

This isn't an AP charateristic, it is the OEM ECU logic!

[packinDSS]

Shiv', do most car manufacturers ecu target load like ours? Or is this a rarity. Did BMW ecu always do this way? Even on naturally aspirated vehicles? I never heard or read this when I was in the supercharged cobra forums. Im just curious.

[cn555ic]

Thanks alot Soccer Dude for the write up! lol

[boom]

Thanks Myst and Shiv. Good starting point. A couple more points that I would add or suggest:

Wastegates. I don't think there is a big variance in the actual mechanical wastegates. It seems pretty strange to me that "some would be soft and others would be hard, etc." I would suggest that the increased DC required by some vehicles is a result of a minor boost leaks in the IC piping or core. Those with aftermarket FMICs should leak test them - submerge in water and pressurize before installing. I've had two personal experiences where brand new ICs required welding to seal. The stock plastic end tanks also seem like likely leak points to me.

Timing: The Procede is a very advanced piggyback. It does more than any piggyback I've ever heard of by a mile. That said, it is limited to offsetting the stock curve. This is reliable but only if the stock curve is stable. The fact that the DME is still in control is a good thing as it can still pull timing from knock. The AP actually sets the desired ignition advance. COBB apparently has chosen to pull a little timing to back off of the knock threshold. That's a typical thing to do when increasing boost. The DME is very good at monitoring knock and the factory curve adds more timing than it should and relies on the knock sensor to pull it back slightly. This is the best for emissions and mileage but could be potentially damaging with a lot of boost. This was shown when cp-e tested their car on a 50/50 mix of gas/e85. That is plenty of alcohol to prevent knock and yet when adding more than a couple of deg of timing the car would knock. The knock monitoring that the DME is capable of and uses is why the JB3/4 is still in operation today. I don't believe the Procede will ever effectively add more timing that the DME will. My guess is that the car would run 12 deg on race gas/meth with the proceded adding 2, subtracting 2, or nothing at all - the DME will add timing until it knocks slightly. The danger here comes with the AP and ATR. If a user adds timing to the DME curve it will continually try to hit it and may damage the motor.

Load/boost: Nice job here myst. The only thing I would add is that the AP may have trouble at higher boost as the load for a FBO car running 18 psi is probably not defined in the tables. In that case it will require development - actually programming the values for cam timing, ign timing, fueling, injection points, injector duty cycle, vanos operation, throttle operation, wastegate, etc. That will likely take a lot of time and will be aided by simple extrapolation of existing values but there is much danger here in that at those load levels knock can do a lot of damage. Get it right and nothing will be faster. Get it wrong and you could destroy things quick. Giac seems the most advanced here. I'd be damn scared to push things as a privateer with ATR in the 16+ psi range without an engine dyno, 6 egts, 6 widebands, and a hell of a knock amp - all of which I believe COBB has at it's disposal - ref GTR. The procede has an easier time here because it lies to the ecu and offsets some of the DME lower load values to safely run high boost. Because the AP has full command of every value it may make more power and be smoother, better driveability ultimately but this is likely pretty far off. The procede is also more likely to be able to handle upgraded turbos sooner than the AP.

Lastly I really like the ability to change the functions of the dash - boost for fuel, etc. I'm not sure how the AP can compete here. It might be worth Shiv's time to develop a a product that could implement that functionality for the flash guys.

[etiliko]

Originally Posted by packinDSS

Shiv', do most car manufacturers ecu target load like ours? Or is this a rarity. Did BMW ecu always do this way? Even on naturally aspirated vehicles? I never heard or read this when I was in the supercharged cobra forums. Im just curious.

All modern ECU for turbo engine work on target Load (it's called torque model).
From Audi-VW to japanese as Lancer or Impreza, to turbo diesel engines!

[vasillalov]

boom,

Very nice! I agree completely!

[themyst]

Originally Posted by boom

Thanks Myst and Shiv. Good starting point. A couple more points that I would add or suggest:

Wastegates. I don't think there is a big variance in the actual mechanical wastegates. It seems pretty strange to me that "some would be soft and others would be hard, etc." I would suggest that the increased DC required by some vehicles is a result of a minor boost leaks in the IC piping or core. Those with aftermarket FMICs should leak test them - submerge in water and pressurize before installing. I've had two personal experiences where brand new ICs required welding to seal. The stock plastic end tanks also seem like likely leak points to me.

Timing: The Procede is a very advanced piggyback. It does more than any piggyback I've ever heard of by a mile. That said, it is limited to offsetting the stock curve. This is reliable but only if the stock curve is stable. The fact that the DME is still in control is a good thing as it can still pull timing from knock. The AP actually sets the desired ignition advance. COBB apparently has chosen to pull a little timing to back off of the knock threshold. That's a typical thing to do when increasing boost. The DME is very good at monitoring knock and the factory curve adds more timing than it should and relies on the knock sensor to pull it back slightly. This is the best for emissions and mileage but could be potentially damaging with a lot of boost. This was shown when cp-e tested their car on a 50/50 mix of gas/e85. That is plenty of alcohol to prevent knock and yet when adding more than a couple of deg of timing the car would knock. The knock monitoring that the DME is capable of and uses is why the JB3/4 is still in operation today. I don't believe the Procede will ever effectively add more timing that the DME will. My guess is that the car would run 12 deg on race gas/meth with the proceded adding 2, subtracting 2, or nothing at all - the DME will add timing until it knocks slightly. The danger here comes with the AP and ATR. If a user adds timing to the DME curve it will continually try to hit it and may damage the motor.

Load/boost: Nice job here myst. The only thing I would add is that the AP may have trouble at higher boost as the load for a FBO car running 18 psi is probably not defined in the tables. In that case it will require development - actually programming the values for cam timing, ign timing, fueling, injection points, injector duty cycle, vanos operation, throttle operation, wastegate, etc. That will likely take a lot of time and will be aided by simple extrapolation of existing values but there is much danger here in that at those load levels knock can do a lot of damage. Get it right and nothing will be faster. Get it wrong and you could destroy things quick. Giac seems the most advanced here. I'd be damn scared to push things as a privateer with ATR in the 16+ psi range without an engine dyno, 6 egts, 6 widebands, and a hell of a knock amp - all of which I believe COBB has at it's disposal - ref GTR. The procede has an easier time here because it lies to the ecu and offsets some of the DME lower load values to safely run high boost. Because the AP has full command of every value it may make more power and be smoother, better driveability ultimately but this is likely pretty far off. The procede is also more likely to be able to handle upgraded turbos sooner than the AP.

Lastly I really like the ability to change the functions of the dash - boost for fuel, etc. I'm not sure how the AP can compete here. It might be worth Shiv's time to develop a a product that could implement that functionality for the flash guys.

I think the Procede already handles upgraded turbos quite well, judging from the members who run the various units without issue.

My personal opinion is, we are close to maximum power levels on the stock turbos today, regardless of delivery format. If Cobb's stage 2/3 maps and ATR prove significant gains over what the Piggybacks bring to the table at the same boost level, then we'll have an exciting 2011. If there's no significant gain, I don't believe the piggy vendors will have anything to worry about. I, for one, don't want to switch tunes for a third time. I love the way progressive meth works on either piggyback, and sure as hell would not want to experiment with a wastegate bypass or monitoring a flow gauge constantly to compensate for a lack of this feature on any flash tune.

There are certain points not mentioned in the original post, as I drew this up with the goal of Clap135 not tearing me to shreds

[Clap135]

Here are some fun facts based solely on data:

Piggyback, simply do not offset timing enough not to knock. They all knock, some more than others.

Cobb: I have yet to see any stage 1 car knock, the timing curve, and the dme ability to raise timing is set below the knock point. Which is alot better imo.
Proof: I have yet to see a procede/jb3 car run 0-2 degree down low at 13psi

Onto the power argument:
Procede and autotuning maps, these maps will make more power sometimes, and knock others

Cobb: These OTS maps will generally make the same power, sometimes less, however won't knock.

Max power potential:
This is where it matters for me. Every piggy back out there is made to work on a wide range of cars. If you are beta testing things, you will quickly realize that some things work on some cars, but not on others. It is the end tuners responsibility to sacrafice some things so his tune works on a wide range of cars.

ATR: The only car I have to worry about satisfying is my own. Simply put that will make it faster, safer, more responsive, more dialed, whatever you want to call it, than any autotuned car that uses logic meant for 1000's of cars.

[BzzzBom]

GIAC>COBB>JB3>Procede IMO

[Clap135]

Originally Posted by BzzzBom

GIAC>COBB>JB3>Procede IMO

lmao, not exactly