[hassmaschine]

No, its 268 - 360 was just what it reported at the time of the error. Thre might be hysteresis, but that the trans is rated for 358 is pure coincidence.

Sure, you can do that with testo now. Like i said, i dont know how accurate it is. And it doesnt correct for ambient conditions - its like an uncorrected vs SAE dyno.

[scanspeak]

I assume the limiting factor in the manual transmission is the clutch?

[tlow98]

Originally Posted by hassmaschine

My buddy is looking at the head (BPC is lightly involved - they sent a "junk" head to him to play with, courtesy of Pete) - so who knows, but I agree. People speculate on larger valves, bigger cams, etc - but since we have no baseline at all, there's no way to know if any of that is needed or what would actually be an improvement.

As far as revving safely, we have a 7k redline stock, which is pretty high for a traditional BMW I6 (normally it was like 6200rpm). Only a 10% bump is 7700rpm - but the issue here isn't the strength of the motor, it's the tuning of the whole induction system which is not designed beyond the stock redline. See the power dive on the N52 using a 3-stage manifold vs an N54 manifold (which while likely not ideal, doesn't have the high RPM resonance issues).

Anyway, all of that is likely irrelevant on a boosted motor, which is probably not going to have an 8k redline anyway - but that's for BPC to answer.

Que your independent throttle bodies and CSL style airbox... for NA pleasure

[CobraMarty]

There is no TQ sensor that I know of.

It can all be done with speed/acceleration off the speed sensor and known vehicle weight.
This is how a DynoJet dyno works. How fast it takes to accelerate a drum of known weight. Some fancy math and boom, calculated TQ and HP numbers based on acceleration, NO strain gauges.

3500 pounds requires some amount of TQ to be applied to accelerate any given amount/rate. I am sure BMW knows exactly how fast acceleration is for a given vehicle weight, delta speed/time, should take with 350Nm. If that time is faster than value in a table, Over TQ limp mode occurs.

Test is to accelerate say 0-80 on a downhill section of road. The ecu will not know that you are on a downhill which will increase your rate of acceleration.

So this issue only occurs with manual and steptronic transmissions? Not with automatic transmissions?
The BPC Alice car was originally an auto trans car and switched/converted to a manual. Making over 500HP and 400TQ, I'm sure it should have set off a over TQ limp mode but only if it originally was a MT. Being auto it does not have the 'acceleration' reference table.

IDK, just asking.

[Taskmaster]

Originally Posted by CobraMarty

There is no TQ sensor that I know of.

It can all be done with speed/acceleration off the speed sensor and known vehicle weight.
This is how a DynoJet dyno works. How fast it takes to accelerate a drum of known weight. Some fancy math and boom, calculated TQ and HP numbers based on acceleration, NO strain gauges.

3500 pounds requires some amount of TQ to be applied to accelerate any given amount/rate. I am sure BMW knows exactly how fast acceleration is for a given vehicle weight, delta speed/time, should take with 350Nm. If that time is faster than value in a table, Over TQ limp mode occurs.

Test is to accelerate say 0-80 on a downhill section of road. The ecu will not know that you are on a downhill which will increase your rate of acceleration.

So this issue only occurs with manual and steptronic transmissions? Not with automatic transmissions?
The BPC Alice car was originally an auto trans car and switched/converted to a manual. Making over 500HP and 400TQ, I'm sure it should have set off a over TQ limp mode but only if it originally was a MT. Being auto it does not have the 'acceleration' reference table.

IDK, just asking.

The torque limits can be adjusted, I’m sure BPC did it to my car.

[rjahl]

Originally Posted by CobraMarty

There is no TQ sensor that I know of.

It can all be done with speed/acceleration off the speed sensor and known vehicle weight.
This is how a DynoJet dyno works. How fast it takes to accelerate a drum of known weight. Some fancy math and boom, calculated TQ and HP numbers based on acceleration, NO strain gauges.

3500 pounds requires some amount of TQ to be applied to accelerate any given amount/rate. I am sure BMW knows exactly how fast acceleration is for a given vehicle weight, delta speed/time, should take with 350Nm. If that time is faster than value in a table, Over TQ limp mode occurs.

Test is to accelerate say 0-80 on a downhill section of road. The ecu will not know that you are on a downhill which will increase your rate of acceleration.

So this issue only occurs with manual and steptronic transmissions? Not with automatic transmissions?
The BPC Alice car was originally an auto trans car and switched/converted to a manual. Making over 500HP and 400TQ, I'm sure it should have set off a over TQ limp mode but only if it originally was a MT. Being auto it does not have the 'acceleration' reference table.

IDK, just asking.

I'm pretty sure torque is calculated from tables withing the DME. RPM, Airflow/mass, Lamda, actual ignition advance are compared against torque tables to calculate a gross torque then engine friction losses, alternator losses, air conditioner, power steering pumps etc are subtracted to provide a net Torque.

I don't think the DME is not going to calculate torque based on acceleration. It has no ability to measure incline or actual loaded vehicle weight In addition I don't recall seeing any tables for transmission losses, those would get really complex especially for an Auto with Torque converter.

I've also not seen any table for air resistance.

[CobraMarty]

So according to my simple math, and with a few assumptions for simplicity,

a stock 328 with 210lb-ft TQ (285Nm) can accelerate 0-60mph in 6 seconds. +10mph in 1 second or +1mph in 100ms.

With a TQ limit of 360Nm, which is 360/285 = about +25% more, our 328 could make +25% more TQ than stock before exceeding the 'over TQ threshold'. Stock 210TQ +25% = 262TQ (355NM).

For simplicity, If we assume it requires 2xTQ to half the 0-60 time, 210TQ=0-60mph in 6 seconds, then 2xTQ, 420TQ=0-60mph in 3 seconds,

then 360Nm (265TQ) would produce 0-60 in 5.21 seconds, (+10mph in 0.87 seconds or +1mph in 87ms).

SO, if the car accelerates FASTER than 1mph in 87ms, then it must be making more than 360Nm of TQ and ECU must go into 'over TQ limp mode'.

The ECU can make this calculation as a way to monitor TQ.

So if you made the car lighter and accelerated or accelerated downhill or with a strong tailwind or MORE ENGINE TQ, then the ECU would calculate the rate of acceleration or change in mph and if for 1mph change occurred faster than 87ms, then 'over TQ limit mode'.

In the case of our modified 328's, if they make more than 262TQ then they will exceed the 360Nm TQ maximum. It sounds like we are exceeding this TQ figure.

[CobraMarty]

BiginBoca has been testing and using the Dragy GPS timer and recording his acceleration and time required. This is just what the DME does with the vehicle speed sensor and it's clock.

He says he can set any speed range. 20-60, 60-90, etc. Even down to 1mph.
He can chart his rate of acceleration all along a curve from 10-100.

[CobraMarty]

Here is a sample of what the Dragy GPS can do and chart. It can calculate acceleration and chart it.

Now if it knew the rpm and the weight of the car it could calculate the actual TQ.
BMW knows the rpm and the 'minimum weight' of their car and could theoretically calculate the TQ it produces based on acceleration.

Draggy-

[rjahl]

Originally Posted by CobraMarty

Here is a sample of what the Dragy GPS can do and chart. It can calculate acceleration and chart it.

Now if it knew the rpm and the weight of the car it could calculate the actual TQ.
BMW knows the rpm and the 'minimum weight' of their car and could theoretically calculate the TQ it produces based on acceleration.

Draggy-

Sorry but no, I don't think so.

This is simply nowhere near accurate for BMW. The car weight can change like 25% between light load and fully laden. Then add transmission and other losses.

There are far too many tables within the DME that monitor expected losses for thing like power steering, alternator and friction to think that BMW would use such an inaccurate method.

If you look at the maps you can see plenty of effort to calculate the actual torque produced by using airflow / mass fuel and timing with plenty of correction values and no values /maps for acceleration. Think about it, not one map that defines vehicle weight or gearbox loses. An automatic would need a multitude of correction values based on the converter losses and TCC state.

[CobraMarty]

F=ma

The M, mass of the car is known. The 'best case' mass (which is lowest weight). It can never be lower, just higher.
The A, acceleration is known. The 'best case' acceleration. No matter what the variables, temp, altitude, etc, all these will only decrease the power and acceleration, never make it more than 'best case'.

So if we make more F, force/more TQ, then acceleration must go up for the same weight. And vice verse, if acceleration goes up then force must go up for the same mass/weight.

We only need to look at Maximum Force as far as the transmission maximum input is concerned. The trans doesn't really care about anything less than maximum.

[rjahl]

Originally Posted by CobraMarty

F=ma

The M, mass of the car is known. The 'best case' mass (which is lowest weight). It can never be lower, just higher.
The A, acceleration is known. The 'best case' acceleration. No matter what the variables, temp, altitude, etc, all these will only decrease the power and acceleration, never make it more than 'best case'.

So if we make more F, force/more TQ, then acceleration must go up for the same weight. And vice verse, if acceleration goes up then force must go up for the same mass/weight.

We only need to look at Maximum Force as far as the transmission maximum input is concerned. The trans doesn't really care about anything less than maximum.

I'm no longer sure what you are really getting at.

An automatic does care a lot about the actual engine torque output. So much depends on accurate torque values for shift pressures and times. Then you have internal torque control and estimation, particularly when there is a large difference between engine rpm and turbine speed resulting in high torque multiplication values. Without these, You can actually over torque these when the vehicle is stopped.

Just too many variables for the DME to use another module ,DSC to accurately calculate engine output. It's not making a rough guess,

Then there is the whole concept of what's happening on a dyno when the car is not moving. Does torque management just fail?

[CobraMarty]

My point is that on these cars, 328 with 6MT and 128 with steptronic, they are going into 'over TQ limp mode' when they exceed 358-360Nm of TQ.

How does the DME know when the engine is exceeding 358-360Nm? I am saying the DME knows by the acceleration table. Time vs speed/distance. If this number is exceeded, an 'over TQ limp mode' is triggered.

Automatics so far have not had this 'over TQ limp mode' thrown. I'm not saying automatics don't monitor TQ and function related to TQ, I'm just saying that so far NO automatics have experienced a 'over TQ limp mode'.
IDK why, maybe they just don't accelerate as fast as the manuals and steptronics, and exceeded the 358-360Nm acceleration threshold, yet.

[hassmaschine]

Originally Posted by CobraMarty

My point is that on these cars, 328 with 6MT and 128 with steptronic, they are going into 'over TQ limp mode' when they exceed 358-360Nm of TQ.

How does the DME know when the engine is exceeding 358-360Nm? I am saying the DME knows by the acceleration table. Time vs speed/distance. If this number is exceeded, an 'over TQ limp mode' is triggered.

Automatics so far have not had this 'over TQ limp mode' thrown. I'm not saying automatics don't monitor TQ and function related to TQ, I'm just saying that so far NO automatics have experienced a 'over TQ limp mode'.
IDK why, maybe they just don't accelerate as fast as the manuals and steptronics, and exceeded the 358-360Nm acceleration threshold, yet.

I keep telling you guys, there's no 360nm torque limit. You're confusing the actual torque reported from the expected torque threshold that was exceeded (268nm). This table can be adjusted so it's a moot point.

and as Rjahl says, the transmission has nothing to do with it. 2D59 is entirely triggered by the DME. Otherwise, it would be impossible to trigger this error on a manual car, but we know that isn't true. And autos absolutely have triggered this error. I could name several people I know who have had the same issue.

as far as how the DME knows - because it was calibrated that way. That's why if you increase torque too much without recalibrating it properly, you get a torque overflow error. It's all tracked through a very complex model - it's not a simple rate of acceleration curve at all.

Anyway, this is waaay outside BPC's thread topic - it is an interesting discussion though. If there were any mods around I'd ask them to split it off into a separate thread..

[tom.gnade]

I think there is an expected upper limit of 268nm and a +90nm maximum allowed difference between actual and expected, resulting in an effective 358nm limp mode threshold. In any case, the only thing that matters is dealing with it effectively, and hopefully reasonable mods won't blow the transmission.

Back to BPC's thread now...

[hassmaschine]

You're still confusing the actual value with the threshold. 360nm is just the DME calculated instantaneous torque value which is stored when it triggered the error - it's completely coincidental. You could have triggered it at 300nm or 400nm or 365nm. The point is it's higher than the expected threshold of 268nm.

there is hysteresis, but it's more like a counter that increments every time you're outside the threshold than a torque value. Once this counter increments past a certain value (24 times, actually) then 2D59 is triggered and it goes into limp mode. Also if you trigger it 23 times and restart the car, the the counter goes back to 0 and you never see an error.

I'll leave it at that - perhaps another thread dedicated to 2D59 and torque monitoring would be worthwhile.

[Phil@BPC]

Making progress!

[hassmaschine]

that's freakin awesome! I'm sure it adds some weight, but what does something like that cost?

[Phil@BPC]

Once everything is completed we should be able to give you a quote.

[Torgus]

I like how the sleeves make it basically a closed deck.

Why did Darton not machine the sleeves ahead of time? It does not look like there would be nearly enough coolant flow without the machining you did...

Hass: FWIW Darton sleeves on the N54 retail for ~2k. I'm guessing these are equal or more given the volume for the N52. A fully built engine I would guess would be roughly 10-12k assuming you want a fully worked head and all the extras etc.

It's cool BPC is still moving forward with this project. Not many shops give the N52 much love. Especially given the market for things like built engines and turbo kits.

Can't wait to see the progress!

[Noir]

wow :O

[rothwem]

Originally Posted by Torgus

Why did Darton not machine the sleeves ahead of time? It does not look like there would be nearly enough coolant flow without the machining you did...

My guess is that they expected some distortion from the press-fitting. Also, its probably helpful to have the bore centers exactly right, and I'm thinking the best way to do that is to machine in-situ.