| Still sounds to me like there could be a leak internally in the fuel tank. This is where I think my leak was. As I recall from my TT swap, there is a small piece of hose inside the fuel tank that runs from the pump to the hardline that then exits the top of the tank. The piece in the tank is only an inch or 2 long. If it leaks there, fuel just runs back into the tank. You'd never smell a leak since it is below the big o-ring seal on the top of the tank. Although, I don't see how that would allow you to have 38 psi just off the pump and 25 after the filter. Maybe you have more than one problem at the same time? Another idea (maybe remote, but you seem to have tried many things thus far) is that the FPCU is acting up once warm. Maybe a capacitor starts breaking down once it gets warm or something, then when it cools off it is OK again. You could take the FPCU out of the loop by hard wiring the pump as I described. Then drive and see what happens to your fuel pressure. The FPCU provides a varying voltage to the pump to alter its speed (again, from memory). If the FPCU is intermittently failing, perhaps it is letting the "steady state" or "cruise" voltage drop too low and the pump isn't pumping enough fuel. But again, that still doesn't explain 38 psi off the pump and 25 at the rails. I can’t see how anything in the hardline from the tank to the fuel filter could have a time-delayed issue. Unless somehow a line has so much crap in it that there is only has a narrow opening left, and when the line gets heated up (maybe at the firewall where it is closest to exhaust components?) it expands and closes off the opening even more. Then you’d still have higher pressure on the upstream side of the blockage (pump end), but lower pressure after the blockage from the reduced flow rate (rails). I’d say that’s a pretty remote possibility though. One more far out thought. I had an ’85 RX-7 that would start and run fine and then start loosing power after a while, to the point where it would barely run. If it sat for a while, it would start and run again until it started loosing power, just like before. Turns out there was a bunch of rust in the tank, and the rust particles would clog the mesh “sock” inlet to the pump after a while. After sitting I guess some of the rust would fall off the sock and there would be enough flow area again, until the pump sucked enough fuel/rust through the sock to block the flow again. Just some thinking out loud, but might give you some ideas,
Mike --------------------------------------------------------
   |
