PORSCHE 911 Turbo S on pump 93 with M1 Methanol injection Lambda TARG and Issues with load balance between R-Load vs Set-Load at WOT

Unfortunately, we are not the ones tuning this vehicle, we are assisting our customer with dyno ramp runs and data logging so they can be adjusted by the company that provided the ECU flash.

This vehicle also has a PDK TCU flash to help with the torque limits. We have two separate issues here from what we can tell.

Mods detail and tests:

Brand new 3071 VTG Turbochargers, 3" high-flow exhaust system, M1 METH injection kit with dual 250cc nozzles.

Meth spray starting point is at 12 PSI, and full meth injection is set to 24 PSI. The meth kit was tested, and it is working very smooth on the bench.

We have no fault codes and no boost leaks either. We also tested the turbochargers for bank-to-bank balance by logging, and it all checks out, so we do not have any mechanical issues.

We are having issues with drivability, mostly while shifting between 3,000 and 4,000 engine speed. The PDK feels like it is holding gears in this range; Porsche max torque is at 4k engine speed. When shifting gears between 3 to 4k engine speed, the PDK feels like it's limiting torque, as if the gear is being held. When the throttle is released at load, we get a really big fireball in the back and a loud backfire.

Looking at the Lambda reading at WOT, they lean out from starting at 2800 engine speed. At WOT, the reading starts at 0.90 lambda, dips to 1.00 lambda, and stabilizes down to Lambda 0.82 at 3400 engine speed. At 1 bar of boost, we see a jump to Lambda 1.

From my experience with P-Car DME/ECUs, they are pretty good with handling the transient enrichment on stock map. For the additional increase in the oxygen content while using 1M, I would like some input please from those that tune using m1. How much additional fuel enrichment is normally ideal as a start. I personally would want to see at least at a minimum lambda 0.80, and with M1 injection, even more fuel and especially at WOT/full load in the mid boost level.

Another issue when comparing R-load vs. Set Load, we have a large discrepancy from 2800 to 4200 engine speed. I have a feeling this is PDK limiting the torque, but not 100% sure until I can actually log the PDK TCU.

Any significant meth injection without true integration with a full time closed loop torque based ECU is going to cause issues like that.

You chose a lambda target, then attempt to reduce DME fuel injection to allow room for the externally controlled meth injection. You do that by hiding airflow which skews load and causes incorrect torque reporting to the TCM, and with incorrect torque calculation you will have issues on both the ECU and TCM sides.

With enough work you can skew all the torque related tables and reduce the issues, but ultimately because the external injection is boost based and the torque tables are based on load which isn't defined directly by boost, this cannot work in a proper way.

You can switch to water injection, adjust nozzle size(s) accordingly, and because it's not a fuel you won't have to skew anything on the ECU side, so both it and the TCM will operate properly.

Or if you really want external fuel injection, a standalone ECU i.e. Motec will provide an integrated solution where total mass of both fuels is known, fuel gets distributed properly, torque is correctly calculated, etc.

Hello Mike,

I really appreciate your reply; it makes total sense. I do agree about the standalone Motec, not available as of now for the 992. The interesting part is, I have tried turning off meth injection completely and logging with a less aggressive map, just to see what would happen, and I had no change in R load to Set load, which I thought was very weird. For example, in the log file I attached, starting boost 0.35 bars from row 994 and 0.78 bars at raw 997, the load difference is being affected without meth injection taking place. I know that's a short comparison. I will post a log with meth injection turned off as well.

What are your thoughts on how the Lambda readings between row 995 through 1006?

Thank you again for your time.

Looks like misfiring.

Hello Mike,

Performed some testing with my early map with meth injection completely off, and also monitored for misfires in this problem engine speed range. No misfires picked up between the engine speed with the unbalanced R load vs R set load, no change between testing meth on or off. I do get some misfires at idle and low RPM range, but very minor, however it is not in the load range that I am experiencing reduced power.

One thing I did not mention is that the larger MAP sensors are used, they were upgraded from 3 bar to 4 Bar. We do experience some un-even idle with Lambda fluctuation between .96 to 1.00. at idle. Idle Lambda target is 1.00.

I would expect that the MAP scaling is off and VE tables needing some work, including during transient fueling.

Any other subjecting's would be greatly appreciated.

Nick,

I don't know what means of misfire detection you're using, but the PCM won't generally give any indication when misfires are occurring at high load unless it's extremely severe. Any time lambda is varying back and forth massive amounts in successive data samples i.e. 0.1 lambda swings like you have, while fuel delivery and airflow are pretty stable, that's what misfiring looks like. The root cause could be related to tuning, or mechanical or electrical faults i.e. cracked spark plug.

Hey Mike,

I do appreciate your reply. I was using my factory PIWIS 3 Porsche computer to perform misfire logs, I will try with another software just to confirm.

Thanks you again Mike!

Yea that won't detect misfires unless it's a hard misfire and at that point you wouldn't be asking if it's misfiring or need a tool to tell you, you'd know for sure when it's that bad. :)

Hey Mike,

Confirmed that the Factory PIWIS 3 has a very accurate data log feature for the misfire counter, as I verified with another data logger with the same result. I did test and replicated your input regarding misfires and fluctuation in Lambda, and it is exactly as you described by cranking up early Meth injection; it does pick up misfires right away with great accuracy as you mentioned. However, I turned off the Meth injection completely and switched to a non-meth map. We are experiencing the same power reduction and load mismatch between the 3-4 engine speed range. After logging the PDK/DCT gearbox, we have come to a conclusion that the clutch clamping force is not sufficient; PDK clutches are being held too long, and there is what feels to be slip at max torque between 3,000 and 4,000. The ECU is limiting torque. Your point is, however, very valid, and not to discredit what you have mentioned, as I do appreciate it. However, it is evident that in our case, it appears to be directly related to the PDK clutch slip and power reduction.

Interesting how both of these scenarios can result in the similar results as far as what you suggested vs clutch slip.

Have you experienced this with DCT transmission before?