ODI directed the Vehicle Research and Test Center (VRTC) to purchase and test 30 subject regulators.The regulators were marked 2011 and 2012 and were purchased from three separate sources.Externally, the regulators looked similar.Internally, there were distinct differences.The 2011 regulators had a burr on the upper gasket seat edge and a casting flaw that formed a sharp edge on the lower seat.Both of these cut into the diaphragm gasket.The plating on the inside surface had a dull finish with loose, flaky particles that stuck to the diaphragm when it was removed.The 2012 regulators were the same design but the plating was of a higher quality.There were no loose particles and the seating surface was smooth.VRTC found about half of the 30 units had at least one loose housing screw when received.The first 10 regulators were tested with water at the recommended maximum inlet pressure of 8 psig.Pressure was gradually increased to 100 psig.This resulted in two regulators exhibiting a slow drip at the flanges; one at 70 psi and another at 76 psi.To test the effect of loose screws on a regulator, screws were slowly loosened with the regulator at 80 psig.The screws were completely removed and no water leaked.The modulating valve had closed from the inlet pressure and no water could flow to the outlet side.Additional tests with loose screws indicated the modulating valve would close at about 40 psig inlet pressure with the adjusting knob set at 5.5, and close at about 20 psig with the knob at 0.5.After finding that higher pressures reduced the chance of major leaks, VRTC changed the test strategy to use lower pressures.A diaphragm was punctured in one regulator.This resulted in water flowing out below the adjusting knob.In two consumer videos posted on the internet, fuel sprayed out forcefully between the flanges.This meant the diaphragm was intact and fuel was leaking below the diaphragm gasket and between the flanges.Tests were conducted with grooves cut in the lower side o