I can tell you with 95+% certainty that this is a tune issue. The symptoms you describe are something that I fix for customers on a weekly basis. Idle dipping, near stalling when you come to a stop. It's almost always in the tune, and it's due to extremely high fueling error and LTFT being turned off in closed loop. The only other thing that I've ever seen cause this is when the vacuum lines pop under the manifold, but usually then it's way more exaggerated, and the car will stall all the time (not to mention you can hear it clear as day).
You just need a new tune. This has nothing at all to do with a BOV or your MAF. I'd bet my paycheck on it.
EDIT: @
nelsmar pretty much got it, and it's really pretty simple: latency is wrong so you have massive fueling error on the low end port side. without LTFT there to expand the compensation range (ltft and stft stack, so giving up ltft halves the operable correction range) and stabilize things the car simply can't correct itself in time to keep from stalling when it falls back to port-only injection.
This is why port and direct injection balancing is so important to drivability, and why it's so important that fueling error be as low as possible in every case. Even if you had LTFT turned on, that doesn't necessarily solve the problem. Trims have no knowledge of injection ratio. The common case is this (when LTFT is left on in closed loop):
- Car starts and idles on ports only.. builds up a massive trim (often because latency is wrong)
- You drive away... car warms up while you're driving
- You pull up to a stop, car is now warm and attempting to idle on DI only
- Car stalls or idle dips because the error on the port side (for which the low bucket LTFT was calculated when you were idling there warming up before) is wildly different than the error on the DI side.
In other words, the car has two independent 'fueling error amounts' that it needs to compensate for with trims, but only one mechanism (set of trims) with which to do it. It never says 'oh I'm on DI now... i need to trim differently than a while ago when I was in the same maf bucket but was on port only'. This is why it is absolutely crucial that you have extremely low trims independently on both the port and di side of the equation. they have to both be minimal or else you'll have wild afr swings when the injection ratio changes that cause issues like you describe.