A discussion on water-methanol injection encompassed the subject of check-valves and their use with nozzles. This further expanded to ponder the question of how much, if any, vacuum was present in the bipipes where WMI nozzles are most commonly located.
My personal experience has been that the flow rate through the Flow Sensor is different when nozzles have check-valves or do not. What I have observed is a brief spike in the flow rate through the sensor upon system activation if the WMI system has not recently been operated. I’ve attributed the spike to the system pump filling the lines, which results in a higher flow rate than is sustained once adequate line pressure is achieved, followed immediately by the WMI spraying into the intake pipes.
My assumption has been that when the system is not active and the car is running there is vacuum in the bipipes that slowly draws out the water from the WMI lines. The discussion about check-valves raised this subject, and I decided to take a pressure reading of the bipipe with the car idling to find out if my assumption was correct.
I’ve got a Dwyer magnehelic differential pressure gage that is well suited to this experiment. Drawing slightly on the low pressure port by mouth causes sufficient vacuum to move the needle, thus if there is vacuum in the bipipes this gage should be able to register it.
I decided to attach the gage to the car at the WMI bung that is located on the top end tank of the intercooler. The position is near enough to that on the bipipes and would be easier to access since I did not already have a WMI line running to this location.
I used play-doh to seal the hose to the WMI bung.
Then I started the car and checked the pressure gage:
Nothing. No movement from the needle at all. I confirmed the hoses were sealed well and still no difference. Inside the car the boost gauge, that is fed from the intake manifold plenum, was reading:
So much for my theory that vacuum in the bipipes is what was causing the WMI flow spikes that I recorded without the check-valves.
My new theory is that airflow past the nozzle creates a venturi effect that draws the fluid out.