[bestclimb]

A scan gauge or similar will NOT provide usable data.

It derives it's fuel consumption data from air flow. The premise of this mod is that it is doing things with airflow which means data that the scan gauge/ultra gauge or similar is suspect. Something that monitors the quantity of fuel injected directly is the only data source that is even close to reliable. That means graduated cylinder (both feed and return), super accurate fuel flow gauges comparing feed and return, or an MPguino or similar that measures fuel injector pulse width and frequency.

edited to add
There is WAYtoo much variation in tank to tank data for it to be valid. I have gone from 25-28 mpg just in the last couple months. Had I put some sort of mod on the car I could easily attribute that change to the mod when it is environmental. In tank to tank comparisons driver technique has a huge role to play. Did it rain slightly more this tank than last, was it cooler, did I get stuck in traffic, did I happen to hit a lucky bunch of green lights, etc all would invalidate a tanks usability as data.

I laid out a test in This post that could provide usable data without using a rolling road.