Try zeroing the proportional table. It uses a PI controller. P being proportional learnt data. You can use a very slight correction factor in integral tables. And moderate timing correction. The e38 is so fast it will try and adjust timing to maintain idle speed. If the rpm dips way to fast it is a sign the min idle airflow is to low. If it hangs or timing is very low try removing some timing. The reason why I change B1652 and 51 is they have a big effect on drivability. But I run B1652 high enough that it will allow enough air on cold start. Usually 1.20 B1652 0.7 B1651 is a good combination with around 7 g/s across the board in min idle airflow. I have tried every combination possible with idle tuning I always end up around here. There is a very high possibility that the "learned" data will still be there, ie the Long term idle trim. You have to zero this out in the tune to get it right. 6g/s-7g/s is about right. Some people have had good results not drilling the TB aswell. But there min idle air will be much higher. I personally get them idling better with a hole but everyone has there own methods. Not zeroing learnt data will give you a world of trouble. The main method I use will allow you to find best idle air without worrying about the learnt data but when max idle area is raised the idle lifts back up. Thats why if I drove around for a few hours and then raised the max idle area value the idle would not hang as it had learnt correctly.

As simon and I discussed a few months ago. Zeroing proportion is the right thing to do. (All proportion tables) Then you will be very close.

My email is [email protected] if you want me to look over your tune feel free to email me or Simon. We don't charge here lol , just help each other out.!