> ...but a four direction speed test shows that my airspeed is within one knot of the gps.

This is a common misconception. Airspeed tests do not show that the static port is free of error. It only shows that the static port and pitot port have the same pressure errors at the tested airspeed.

To test static port accuracy, you need to fly different airspeeds at the same known altitude and check for any variation.

I used the horizontal roofline of the tower at my local airport as the altitude reference and flew successive laps of the landing pattern at speeds ranging from 60 kias to 200 kias. The faster I flew, the higher the indicated altitude (+300' at 200 kias initially). The errors were parabolic (depending on the square of the speed) as expected. Based on that information, I was able to reposition the static port until the altitude error was reduced to about 30 feet over the usual speed range. 

I am using a Piper style combined pitot-static blade on the bottom of the wing. The adjustment that I made was to remachine the end of the blade to a different angle to make it more forward-facing.

After the static port is calibrated, THEN check the pitot pressure by comparison of ASI with GPS speed.

 

bob mackey