There are many older threads on this in general, but it takes a reading a while to find this issue. The early, 8/2003, tank leak test procedure suggested to use 66" of water in a manometer for the leak test. A later, circa 2006, changed that down to ~30" of water.
I got mine up to 48" and the tank pillowed greatly, a little scary, really. It finally stabilized at 42" after repair of some test hose leaks. Left it there for the test. Even then, the tank was visibly expanded. It is really a leak test not a pressure structural test, so 30" seems fine and much less stressful - pun intended.
Test protocol summary: I used the schrader valve with 40 psi line pressure to charge and it was a little difficult to get adjusted - a drop to 20 psi should be easier. Temperature is a big factor, as is barometric pressure. I left mine for 12 hrs overnight and for some reason the manometer zero changed, but the tank pressure (after zero offset adjustment) only .2 " Easily a temperature difference. 1 deg F is .8" of water change. If you use this method, get local barometer at beginning/end and use an accurate thermometer to measure tank temp. Tape an area to use the laser thermometer as the reflective surface is hard to read. Ratio the absolute temperature (Rankine or Kelvin) and absolute pressure for the correction factors. This will make the process much more definitive and accurate. No use guessing unless you like that sort of test.