E46 330i Automatic - M54B30 Turbo

[stefm5]

too low compression for 10psi boost.
also too much overlap with those cams. vanos needs adjustment especially down low.
try minimizing it by setting 126, -105 up to 2000-2500rpm

[NOTORIOUS VR]

I am trying to do a lot!

I locked the cam timing.

Engine is 8.5:1, 272/256 cams, highly ported head,

Fuel is very close to target under load, it's just the transitions and lift-off that needs work.

The timing map I'm currently using is from YOU!!! LOL

Cool, just trying to make things easier from an outside view from past experience... Fully variable VVT systems can get quite tricky to deal with so simplifying things at the beginning especially when trying to sort out issues can easily throw you in a loop. When control systems for PWM VVT get into play in ECU's that are fully configurable like yours (and others) and you're dealing with PID loops and base duty/feed forward systems it's not always easy to deal with that without some steady state/dyno testing which will give you feedback.

Also don't get caught up in lambda readings in transition periods... they're not all that important and remember you're going to be dealing with a long delay between what's happening at the motor and what your sensor is reading. Generally as long as the car is smooth and drives well you can run with it.

Right I remember now... if you're using mine which is conservative I would say you're going to be just fine for now

[dburt86]

I agree with VR.

Rather then plotting 4 different frequencies for knock, id be looking at 1 in the beginning. You can use a frequency calculator to determine a starting point. Getting a base line at no/low load in each RPM area will give you an idea of what % is normal noise.

From there, id imagine the other frequencies you would want to monitor are closely above and below the true knock frequency, although i dont know that for certain.

How is the VVT control setup in those motors? 2 solenoids or 4?

[PEI330Ci]

too low compression for 10psi boost.
also too much overlap with those cams. vanos needs adjustment especially down low.
try minimizing it by setting 126, -105 up to 2000-2500rpm

He he...it's not for 10 PSI, it's for 91 octane pump fuel and as much as I can get a GTX3076R to breath. I've also got nitrous...

Thanks for the tip on the timing.

Cool, just trying to make things easier from an outside view from past experience... Fully variable VVT systems can get quite tricky to deal with so simplifying things at the beginning especially when trying to sort out issues can easily throw you in a loop. When control systems for PWM VVT get into play in ECU's that are fully configurable like yours (and others) and you're dealing with PID loops and base duty/feed forward systems it's not always easy to deal with that without some steady state/dyno testing which will give you feedback.

Also don't get caught up in lambda readings in transition periods... they're not all that important and remember you're going to be dealing with a long delay between what's happening at the motor and what your sensor is reading. Generally as long as the car is smooth and drives well you can run with it.

Right I remember now... if you're using mine which is conservative I would say you're going to be just fine for now

Yup, I totally agree with keeping things simple.

The car runs just fine right now, making more power than my tires can hook up in 1rst, and the midrange of 2nd gear.

On the Lambda, the M1 has phenomenal tools for fuel film, but you have to map it out. I've started to do this, but there are some steps that require a load control dyno to do properly. I'm cheating a HUGE amount by giving closed loop feeling a LOT of freedom. Once the excitement of driving the car subsides a bit, I'll find the patience to start doing open loop fuel mapping. That will help the most.

I agree with VR.

Rather then plotting 4 different frequencies for knock, id be looking at 1 in the beginning. You can use a frequency calculator to determine a starting point. Getting a base line at no/low load in each RPM area will give you an idea of what % is normal noise.

From there, id imagine the other frequencies you would want to monitor are closely above and below the true knock frequency, although i dont know that for certain.

How is the VVT control setup in those motors? 2 solenoids or 4?

2 Solenoids for the M54, 4 for the S54.

I've just done some high load pulls through 4th gear; I'll post the graphs shortly. What I'm looking at it hard to describe, it would be easier for people to look at the graphs to understand.

[PEI330Ci]

This thread is going to turn into a data orgy.

4th gear pull

Meth ramps in 40% DC @ 3k RPM, to 100% DC @ 6000 RPM



- - - Updated - - -

Did a pull from 1st through 4th down my driveway.

Looks like the intercooler can't keep up...




Knock traces from the same pull:



A = 6khz
B = 8khz
C = 10khz
D = 12khz

[jakermac]

That must be one HELL of a driveway!!

[wazzu70]

Why the heck do you use degrees F? Use real units man! Kelvin is std but Celcius is OK.

Boost pressure is also in psi

Just teasing, but put the temps all in the same units at least for easy comparison

[PEI330Ci]

That must be one HELL of a driveway!!

You should see how long it takes to mow the lawn...

Why the heck do you use degrees F? Use real units man! Kelvin is std but Celcius is OK.

Boost pressure is also in psi

Just teasing, but put the temps all in the same units at least for easy comparison

Ha ha ha....I changed them for the 'Mericans!!!

I originally had everything in C and Kpa

[wazzu70]

My faith in humanity has been restored

[dburt86]

your knock is weird looking. On my car, the knock % decreases as the rpms climb, prolly because its a dumb push rod motor with a big cam.


Do you get a similar shape of knock graph with no boost? Have you introduced real knock to see what it looks like?

I love data!

[NOTORIOUS VR]

Looking good! Curious as to why you think your IC isn't keeping up? 18 deg inlet temps at the end of 1-4 run seems good to me (without knowing your ambient temp at that time - but will assume it wasn't 0 deg C).

[dburt86]

If you saw my temp rise, you would see you are in good shape lol

[PEI330Ci]

Looking good! Curious as to why you think your IC isn't keeping up? 18 deg inlet temps at the end of 1-4 run seems good to me (without knowing your ambient temp at that time - but will assume it wasn't 0 deg C).

That's with meth.

"Intercooler Temp" is just after the intercooler.....it climbs 25 Deg F from the start to the end of the pull.

And yes......18 Deg C intake temps is pretty nice.

For anyone that's picked it up, the scaling on the Pre-Turbo temp sensor is wrong. (I changed from a slow Bosch air temp sensor to a fast Delphi air temp sensor, and haven't changed the scaling.)

[wazzu70]

your knock is weird looking. On my car, the knock % decreases as the rpms climb, prolly because its a dumb push rod motor with a big cam.


Do you get a similar shape of knock graph with no boost? Have you introduced real knock to see what it looks like?

I love data!

The knock signal would depend more on load than RPM. At higher RPM the signal will be smoother due to the combustion events being closer together, but generally at high RPM on a pull you are also at high load. If you were using a steady state dyno and keeping load constant with RPM you will see the signal smooth out.

Also keep in mind that torque tapers at higher RPM unless peak torque is at redline so you have to keep that in mind. 100% throttle at 4k RPM and 6k RPM are not the same.

Unfortunately most of my experience is with misfire detection based on crank accel/velocity and not frequency based so I don't have all the deets on this type of strategy

[NOTORIOUS VR]

That's with meth.

"Intercooler Temp" is just after the intercooler.....it climbs 25 Deg F from the start to the end of the pull.

ah I see it now, 190 to 90 F... 90F is like 32 deg C... What was the ambient temp at the time?

[dburt86]

The knock signal would depend more on load than RPM. At higher RPM the signal will be smoother due to the combustion events being closer together, but generally at high RPM on a pull you are also at high load. If you were using a steady state dyno and keeping load constant with RPM you will see the signal smooth out.

Also keep in mind that torque tapers at higher RPM unless peak torque is at redline so you have to keep that in mind. 100% throttle at 4k RPM and 6k RPM are not the same.

Unfortunately most of my experience is with misfire detection based on crank accel/velocity and not frequency based so I don't have all the deets on this type of strategy

Im with you 100%, but his graphs are showing the exact opposite. The higher the RPM, the more knock % he is getting. Being that he said the timing is conservative (and i have to assume based off this thread that hes probably right), then i doubt its actually knocking, which is why id like to see what those graphs do with no load at every rpm.

Once he gets a baseline of knock % with no load/knock, he can start making his knock detection graph reflect what he had on his base line, regardless of what the shape should be.

I think at the end of the day, we can theorize all we want, but until you actually log real knock, its a best guess based off decent information.

[PEI330Ci]

ah I see it now, 190 to 90 F... 90F is like 32 deg C... What was the ambient temp at the time?

18 or 19 Deg C





There's going to be a bit of a delay of any more testing/data acquisition. I've started migrating the tune from the V1.03 firmware to the new V1.04 firmware. There about 80 tabs in M1 tune that I have to update....

[Def]

I wouldn't use a 1-4th gear pull as a basis on if the IC is sized right. All that's going to show is how much thermal mass the IC has, not necessarily its cooling. Really beat on the car for minutes at a time, and THEN you can make a determination on if it has enough cooling or if temps keep rising without reaching steady state.

I don't think your IC outlet temps are anything to be concerned with personally.

[PEI330Ci]

I wouldn't use a 1-4th gear pull as a basis on if the IC is sized right. All that's going to show is how much thermal mass the IC has, not necessarily its cooling. Really beat on the car for minutes at a time, and THEN you can make a determination on if it has enough cooling or if temps keep rising without reaching steady state.

I don't think your IC outlet temps are anything to be concerned with personally.

Agreed. It was just a statement that the IC was not staying constant, despite airflow increasing.

Regardless, the IATs are so low that they could be mistaken for an N/A car.

On another note, I'm having issues with 2 of the EGT sensors reading off. I might order 2 new ones just to prove the issue is the sensors. (They were brand new from Motec) I'm also looking at ordering an ultra-high speed air temp sensor to replace the OEM one in the intake manifold. (It's the same one that T1 uses on their GTR)

[NOTORIOUS VR]

Instead of buying new ones, just swap them with another cylinder and input..

[PEI330Ci]

Instead of buying new ones, just swap them with another cylinder and input..

Done both.

There's a difference at ambient.

Coolant Temp = 62
Oil Temp = 62

Cylinder 1 = 59
Cylinder 2 = 59
Cylinder 3 = 65
Cylinder 4 = 59
Cylinder 5 = 60
Cylinder 6 = 60

The issue with Cylinder 3 follows the sensor....

Going take apart the connectors....

[NOTORIOUS VR]

Done both.

There's a difference at ambient.

Coolant Temp = 62
Oil Temp = 62

Cylinder 1 = 59
Cylinder 2 = 59
Cylinder 3 = 65
Cylinder 4 = 59
Cylinder 5 = 60
Cylinder 6 = 60

The issue with Cylinder 3 follows the sensor....

Going take apart the connectors....

Yeah resistance issues with the thermocouples can make or break it.. if the issue is at ambient at least you know it's either sensor/wiring/ecu input related.

[PEI330Ci]

Confirmed, I'm getting 2 more thermocouples sent to replace the 2 "Duds". Also having an ultra fast air temp sensor sent to replace the OEM one in the intake manifold that I don't trust. (I had to build my own CAL file)

On another note, I think all the noise on the knock traces is because of the solid engine and transmission mounts. What is showing on the knock trace, is what the car sounds like when you drive it. There is so much noise form the engine vibrating, that you can't hear the exhaust at all. I'm waiting for the CES engine mount to be released, but in the meantime I'm thinking of putting back in the 2 Poly mounts for the transmission, and the driver's side Poly engine mount to try to reduce the vibration (and clean up the knock signal) Because of where the turbo sits with the SS TS, I can't fit the passenger's side Poly mount. (But I'm hoping the CES mount works?) I had to hammer the heck out of the turbo inlet pipe to clear the relatively small solid engine mounts.....

Back in 2009 I put solid engine mounts on the 330Ci, and it was pulling timing like crazy on the dyno. (With a stock ECU) By changing to Poly engine and trans mounts, the ECU stopped pulling timing and the the engine would make a clean pull.

[NOTORIOUS VR]

Confirmed, I'm getting 2 more thermocouples sent to replace the 2 "Duds". Also having an ultra fast air temp sensor sent to replace the OEM one in the intake manifold that I don't trust. (I had to build my own CAL file)

On another note, I think all the noise on the knock traces is because of the solid engine and transmission mounts. What is showing on the knock trace, is what the car sounds like when you drive it. There is so much noise form the engine vibrating, that you can't hear the exhaust at all. I'm waiting for the CES engine mount to be released, but in the meantime I'm thinking of putting back in the 2 Poly mounts for the transmission, and the driver's side Poly engine mount to try to reduce the vibration (and clean up the knock signal) Because of where the turbo sits with the SS TS, I can't fit the passenger's side Poly mount. (But I'm hoping the CES mount works?) I had to hammer the heck out of the turbo inlet pipe to clear the relatively small solid engine mounts.....

Back in 2009 I put solid engine mounts on the 330Ci, and it was pulling timing like crazy on the dyno. (With a stock ECU) By changing to Poly engine and trans mounts, the ECU stopped pulling timing and the the engine would make a clean pull.

Yeah NVH is a bitch :P

[PEI330Ci]

Yeah resistance issues with the thermocouples can make or break it.. if the issue is at ambient at least you know it's either sensor/wiring/ecu input related.

I think most people use 1 sensor of "a" type, so they don't see deviation.

Every single sensor I have reads a different value at ambient temp/pressure.

In an effort to standardize the data output, I used a couple of reference sensors, and then skewed the calibration table of every other sensor. So while this moves the curve at a starting point, it doesn't address the issue of sensors having a different shape to their curve. All the AEM pressure stuff for example is supposed to perfectly linear, but I'm not finding that in reality.

Now I'm starting to see the value in $500 pressure and temp sensors......