I was looking into this fairly recently (like a month ago).
There are a few companies that manufacture them specifically in this location (
Autoexe do one, as does
RacingBeat), but it's not cheap. It definitely would look nicer than a DIY job, although if you just bought the plastic cover it'd look much neater.
The article about intake pipe length can be found
here. It actually had a hand in helping me decide on not going with a super short CAI.
I would say the technical reasons for less power/torque low down, but increases high up, would be due to air turbulence. At low speeds, the air entering the intake would be fairly turbulent as it enters the throttle body relatively quickly and doesn't have a chance to \"smooth out\" as such. Because the air isn't \"smooth\", the air flow isn't efficient. Also, I would think that having a longer intake length would mean there is some air within the piping itself, which the throttle body would suck in from stop. If there's less piping at standstill, that means less air, which would translate into a less efficient burn at first. Meanwhile, at high speeds, the air is coming in so quick it's pushing all the air that's already in there harder, creating more pressure and thus has a much smoother air flow.
I'm not against the mania CAI, but I just thought there could be a more efficient way of getting air to the throttle body that didn't require that extra length of piping that runs along the side of the engine and the drilling. It seems like the air has that extra distance to travel and would require more force to move it through all the piping thus losing velocity as it finally made it to the throttle body.
Luckily enough, one of the guys I work with is mates with the guy that runs/owns SuziTech, and he was keen to get in to MX5's, so it was the perfect opportunity to get a CAI made up for me and for him to have a test vehicle to work off.
The end result was something very similar to the Racing Beat example, but with completely new piping, a cover piece (although not as nice as the ones autoexe/racingbeat have, but made of metal rather than plastic) and missing the bend of the racing beat one (ie it's just the filter sitting in line on the end of the piping, if that makes sense). Cost was $550. I don't know whether SuziTech will adjust their prices after having done mine. The guy who runs the place said it was that price because they did more testing/tuning/engineering/whatever than he first thought they would do to get it right, so that added a bit of extra cost to it. It might be cheaper now that they've done it once before and have a working model/design to build off.
As for power/torque, I haven't had a noticeable gain/loss low down, but there's definitely increased throttle response. It seems that when I get on the gas, it picks up a lot sooner. I've found that when I do step on it, the power in the engine is more readily available compared to before.
As for high end power, my NB8A with only the CAI modification pulled an extra 2kw more than a purely stock NB8A at the MX5 Mania dyno day this Saturday just past. The stock NB8A had about 25,000 less kilometres on it than my car as well. The extra 2kw isn't guaranteed to be down to the CAI, but I like to think it has something to do with it
The induction note is great when you're on throttle too. For my CAI, if you take it easy on the gas, the induction note sounds almost stock, but with a bit more body in the note (if that makes sense). But once you get on it and open it up, you really hear it go.
Sorry about the long post, but I hope the info helps
