AEM Water Injection Failsafe Device: Part 2 – Catching Our Mistakes and Upgrading Our Water Injection System

AEM Water Injection Failsafe Device: Part 2 – Catching Our Mistakes and Upgrading Our Water Injection System

By Mike Kojima

Over the past few months in-between other projects we have been tinkering around with AEM’s water methanol injection kit on our in-house Evo IX project car.  You can read about the kit here.   A few weeks ago we installed AEM’s new injection monitoring system which you can read about here.

In looking at our data, something didn’t look right; there was too much scatter at higher flow rates for the volume of our pump.  The flow would fall off drastically, especially at high rpm and there was a lot of variation in the flow. Even considering throttle modulation and the fast responding Evo IX turbo, the data didn’t look quite right. The system was functioning correctly but we felt that the flow should be greater under high loads.

The kind of logs we were getting from the system didn't look right to us.  Full throttle runs should leave consistent flow traces and ours were scattered like this.  The flow would rapidly drop as well and the lines would converge at a low flow point.  This was weird.  We decided to look into why this was happening.

This led us to look for problems.  We started by looking for leaks or restrictions in flow before the flow sensor.  We found no obvious obstructions like kinked lines in the system at all.  When we tested the pump's flow by pressing the test button under no load the flow was fine.  We next decided to check the water filter.  When we were about to open the filter case, we found that while we wiggled our suction side lines, they would leak a few drops.  We had found found our elusive problem, when installing the system we tried to make the water lines short and sanitary.  Unfortunately we made them a little too short and some of the suction lines were not fully seated in the push fittings.

The scattered data forced us to run the control limits very far apart to prevent annoying alarms.  The wide limits hampered the effectiveness of the monitoring system as well.