jaa55 - have you had a wide band air/fuel ratio on the car to see what it's doing across the rpm/load spectrum? Unless you have, then 'seems to run perfect' and actually running perfect may be two different things. It may be that the change you made affected things within the computer's adaptive strategy range to 'fix' different maf readings it's seeing.
Plotting voltage vs. flow would be a great thing to do. Then you'll know what's going on. The ecu is expecting certain voltages proportional to certain flows - it's called the transfer function. Hopefully you can see that moving the electronics to a different sized housing, even with the same sample tube, is going to alter the flows that move across the hot wire --- and that results in a different transfer function. Now, if the change is small enough for the ecu to adapt - things may 'seem' perfect. This is what happens when one uses a 70mm maf (housing and electronics) from a 94/5 HO Stang on a fox body A9L processor - it's close enough to work. If you use the 93 Cobra maf - and look at the transfer function, you'll see that there's as much as a 20% difference (engine goes lean) in the voltage vs. flow numbers at high rpm/high flow. Idle, around town, cruise - everything 'seems' perfect. But at higher rpm, the motor's going to run lean because the adaptive strategies can't compensate for that 20% difference. Unless you flow it/measure voltage - you really don't know what's happening when you start moving electronics around.
But ideally - you don't move electronics around, and you try to give the ecu the transfer function it's expecting.
Michael Yount - Charlotte, NC - 82 Volvo 242 - 6.2L; '15 Fit; '16 CrossTrek