Maybe it is a something to do with BOV placement being too close to the MAFs on aftermarket kits.

100% Recirculated for sure. There is a LOT of piping and some of the bends are pretty harsh-

For years now everyone has been preaching the "Maximum Boost" principle that the increased internal volume of the massive amount of piping forces the turbo to take longer to pressurize the whole system. That makes sense, and I don't disagree that it plays a role. Over on the more technical ECU tuning boards, people have been logging and analyzing the issue to find that an untunable rich dip is present with FMICs. When I say untunable, that includes tip in or any other basic parameter you can imagine. If you try to tune it out, you get a lean spot at other times.

So people started experimenting with blow through MAFs and they were able to tune most of it out. Speed density doesn't have this problem at all.

Is the MAF directly after the filter in that picture? If it is that 90* bend immediately after is very suspicious if you ask me. Anyone else see what I mean?

What people are experiencing is the limitation of the stock ECU. It sounds like either there isn't a deceleration strategy in place for that rich condition (because the factory didn't need it), there aren't enough map breakpoints (because the factory didn't need them), the MAF metering strategy is limited, or it's just plain sloppy tuning. On some higher end standalone ECUs, you can tell it to ignore the MAF reading on deceleration via negative throttle actuation based on deg/sec movement. Then there's no more rich spike.

Maybe some open source nerd will implement that strategy into the stock ECU....after 42 tries.

Jere- that is the MAF location, but that pic was just an example. All of the different style intakes seem to have the problem however straight (and well scaled) they are.

Eric- so this strategy you refer to, would it essentially be a "tip out" logic? In other words, taking away injector pulse width based on how fast the throttle position is decreasing? If that's the case, it has been discussed a bit IIRC, and there isn't that logic in the stock ECU, and noone is making a patch AFAIK.  When you say map breakpoints, what map are you likely referring to?

Yes, deceleration enleanment you could say. Pectel uses a form of this strategy in the EVO 7/8 Group N rally ECUs because the 7/8 only came with a very low range MAP sensor. MAF was used for fuel metering in these models. The 9 was FIA homologated with a 3bar MAP sensor so the Pectel EVO 9 ECU uses speed density.

When I said map breakpoints, I was forgetting that the MAF voltage spike also causes the ECU to read a falsely high load. But maybe you can get the one of the open source nerds to map switch to a different main fuel table on decel. In that other main fuel table you would enter super high values of say for example 29.98 A/F so the ECU would shoot for 29.98 A/F on any deceleration by shutting off the injectors. That is assuming the stock ECU's fuel pulse algorithm doesn't include a minimum pulsewidth parameter. So you would be effectively turning a main fuel table into a main decel table.

When exactly does thisrich dip occur?