[Nezil]

Originally Posted by ZPrime

Hmm. I would think that closing the throttle plate would also potentially be bad for the turbo, as it might cause a reverse shockwave back up the system as pressurized air "bounces off" the throttle plate. But there's a big intercooler in the way that I suppose would absorb most of that shock before it could reach the turbo itself. So maybe not as bad as I thought originally.

I have considered this... the boost spike is not that high, just a tad higher than our TMAP sensor can read at about 21 psi, so I don't think it's a huge problem. Additionally, the diverter valve can help reduce shock, though I'm not sure what that's doing in this case because there is no log for DV.

When I first started reading your research into all of this, what popped into my head was, "Is this part of the auto-revmatching?" Per BMW's documentation & statements, the ECU on 6MT cars does "magic" to help smooth performance during upshifts. My first thought was that this could be the root cause of this behavior.

I very much doubt that's the issue, but it could be. If you shift slowly, and take anything more than > 2 seconds over the shift, there is no problem at all, everything works as expected except you've lost absolutely all boost and have to build it up again. It does build up much faster, as I showed a few posts back, but it's overall still slower to shift slower without the lag than faster with it. Of course, shifting fast with no lag is way way faster.

To put this in to perspective, I was getting 0~60 times of just over 5 seconds in the past, and that was with the BM3 Stage 2 93 map. I'm now getting consistent 4.8s (= 4.45s with 1ft roll-out) times without trying that hard, and on the street not a sticky drag strip. The map I'm running at the moment is the starting point for a 91 octane tune from the tuner I'm working with, and is therefore quite a bit down on power from the OTS 93 map I was running previously as well.

I think it's absolutely fair to say that the DME induced shift lag adds between 0.3 ~ 0.4s in a 0~60 run.

Since we know that revmatching is disabled (on an OG car like Nezil's) when you disable traction control entirely, I was wondering if your tests were being performed with ESP entirely off, or whether that has any change in behavior? It seems like some people have tested with it off and say they are still feeling the lag though...

All of the logs I captured initially were in DSC off mode, as is typical for capturing logs without DSC interference. My tuner has been optimising Sport+ in the maps I've been given, so I've been capturing logs in that mode since starting work with them.

Also, have you tweaked anything else, like "Euro MDM" or other traction or DME-related stuff?

Although I don't think it's relevant, I am running "Euro MDM" and "M2GTS" for the DSC coding. I haven't coded the diff at this point, but might in the future.