Boost loss at high RPM

We ghetto charged our lemons Civic as we got bumped to A class after winning B.

Aside from the new challenges of dealing with the extra heat that creates, we also have an issue with holding boost. The boost will spool quickly to 6psi from low RPM, which is what we want, and as you get up to redline, the boost gradually tapers off to 3.5psi.

The turbo is an internal wastegate journal bearing T25 from a Saab 9000, so it is sized correctly for this low boost application, and I just rebuilt it. We tried two different wastegate actuators and two different blow off valves, no change. We tried running with no intake air filter with no change. There are no apparent loose connections in the charge pipes for the boost to leak out of, especially this much.

Any ideas on how to keep the boost steady instead of falling off?

Have you tried pressure testing the intake? There may be a leak that increases as the boost comes up. Cap off the cold side turbo inlet, rotate to TDC, open the throttle all the way, and pressurize the system up to 6 psi with your compressor. Watch the boost gauge to see how fast the pressure bleeds out and locate any leaks. Leaks occur in unexpected places: intake manifold gasket, throttle shaft o-rings, t-bolt clamps that don't seal evenly, etc.

If the boost leak checks come up negative, I'd suggest disconnecting the wastegate and driving it. If you're having to back off to keep it from passing 6 psi at lower rpm but it still won't make 6psi at higher rpm at WOT, I'd suspect the turbo might be to small or you've got a poor flowing exhaust manifold or something that's choking it.

Is this motor built in any way? Cams for lower rpm would build boost lower and taper at the top (not sure why anyone would do this with a Honda engine)

Where are you sourcing your wastegate signal? If it's from the compressor outlet, then boost will taper as RPMs rise because of increasing back pressure from the intercooler. An overly-restrictive intercooler can produce a significant boost drop. You can address that by using a better intercooler, sourcing the boost signal from the pipes between it and the throttle body, or by using EBC.

I think that is the answer. The wastegate pressure source is on the compressor housing. The intercooler is a no name that I got in a pile of parts on Craigslist (remember, lemons car), so it wouldn't surprise me at all if it was restrictive, even at fairly low boost. We are going to add a port on the charge pipes just before the throttle body to see if that changes anything.

Personally, I'd start with rslifkin's suggestion. It's really quick and easy and cuts down the potential causes in a hurry.

Codrus and Armitage both give good advice.

Boost leaks can show in the oddest places. Pressurize the intake tract and spray with soapy water. If you see bubbles, you have a leak.

I noticed a significant change in boost behavior between using the turbo housing as a source for the wastegate versus using a pressure source between the intercooler and the throttle body. The latter position gave a more accurate and consistent reading since it avoided false reading boost created by any restriction between the turbo and the engine intake.