Okay, I am a little disturbed by the fact that nobody has every really explained this in full to anyone. I hear newbies say... If i run xx pounds of boost on a t3, how many hp will I get? and then they get responses like.... psi doesnt mean shit...it is how large the turbo is. Yes, that is true...but only under certain circumstances. The rule is this: 8psi is 8psi coming from any turbo (big or small), as long as there is no pressure drop at the high end due to your engine being able to suck in more air than the turbo can blow. If you had a guage in your h22a prelude that showed how many psi the turbo system was running at, and you were running 7 psi on a t3 ( or some small turbo) the guage/rpm would look something like this: remember I am just using these as a visual...these are not actual numbers from a dyno..they are just out of my brain. 1000 rpms: 0 psi 2000 rpms: 0 psi 2500 rpms: 4 psi (+ 25 percent more tq theoretically) 3000 rpms: 7 psi (+47% more tq..theoretically) 3500 rpms: 7 psi (+47% more tq..theoretically) 4000 rpms: 6 psi (+40 % more tq..theoretically) 4500 rpms: 5.5 psi (+ 37% more tq..theoretically) 5000 rpms: 4.5 psi (+30% more tq..theoretically) 5500 rpms: 3 psi (+20 % more tq) 6000 rpms: 3 psi ( +20 percent more tq) 6500 rpms: 2 psi (+13% more power) 7000 rpms: 2 psi (+13% more power) so...if the power peak for the h22a engine was at 7000 rpm, you would have 170 ft/lbs instead of 150 ft/lbs...but your average torque from 2500 to 6000 rpms would be something like 35 percent higher. that same small turbo on a d15b7 at 7 psi might look like this: 2000 rpms - 0 psi 3000 rpms - 0 psi 3500 rpms - 3psi 4000 rpms - 6 psi 4500 rpms - 7 psi 5000 rpms - 7 psi 5500 rpms - 7 psi (47 percent increase in torque from 4500 rpms to red line) 6000 rpms - 7 psi what if you put a big ass turbo on a d15b7? 2000 rpms - 0 psi 3000 rpms - 0 psi 4000 rpms - 0 psi 4500 rpms - 0 psi 5000 rpms - 3 psi 20 percent more tq 5500 rpms - 5 psi 35 percent more tq 6000 rpms - 7 psi 47 percent increase in tq notice how the average torque gain is a lot lower than the smaller turbo, and the peak torque gain is the same? that same large turbo on a h22a might look like this: 2000 rpms 0 psi 2500 rpms 0 psi 3000 rpms 2 psi 3500 rpms 5 psi 4000 rpms 7 psi 4500 rpms 7 psi 5000 rpms 7 psi 47 percent increase in tq 5500 rpms 7 psi 6000 rpms 6.5 psi 44 percent increase in tq 6500 rpms 6 psi 40 percent increase in tq Horsepower = torque x rpm/5252 If the h22a had 150 ft/lbs of torque w/o boost, with 7 lbs, it would have close to 220 ft/lbs of tq. the h22a @ 7k rpms w/o boost would have 199 hp/ 150 ft/lbs the h22a @ 7k rpms w/big turbo would have 293 hp/ 220 ft/lbs the h22a @ 7k rpms w/small turbo would have 226 hp/ 170 ft/lbs the d15b7 @ 6k rpms w/o boost has 102 hp / 98 ft/lbs the d15b7 @ 6k rpms w/big turbo would have 150 hp and 140 ft/lbs the d15b7 @ 6k rpms w/small turbo would have 150 hp/ 140 ft/lbs, but it would have one hell of a lot more power throughout the whole power band...making it a much better daily drivable car A b16/b18 would be somewhere inbetween the d15/h22a this is my 2 cents on turbos...hope it helps someone

If you had 7 psi at the throttle body coming from a t3, it would have the exact same increase in torque/power as 7 psi at the throttle body coming from the biggest ass turbo you can find. the only difference is when you step on the gas, there might be a drop in pressure from the small turbo because the engine can suck in more cf/m than the lil turbo can produce. the big turbo, however, would continue to stay at 7 psi.

make sure you get the right size of turbo for the application. If you are going for more power for daily driving, go with a smaller turbo. If you are going for drag racing, get a larger turbo, but dont get one larger than it needs to be, because you will actually lose peak horsepower. This is because it does actually take energy to push the turbo, and if you are pushing it harder than it needs to be, you are just wasting horsepower to push air through the blow off valve

Hmm. I've got a clearer explanation using some real numbers.. but I have to write it up real quick. I was going to link to Brian's article, but it's somewhat confusing still.... I'll post here as soon as I rewrite the article.