I am bringing this thread back from the dead, since recent testing has shown that AFM testing is valuable, and maybe important to keep you engine running. I also got a question on this today, so I figured I would post my reply here for all to see.
First, some back ground the car tested. It finished 3rd at Nationals after heroic week of driving by Dan Pina, making 133.5 HP on the dyno. That was after fixing a vacuum leak and loose cam wheel (it made it less power before that). It is a low compression motor with a shaved head, late cam, and with about 3 seasons of racing on it. Early ECU & AFM. Seemed like a fairly average motor.
We took it to the dyno yesterday. This dyno showed air fuel ratios in the 15's (!). That's frighteningly lean. Still, it made 136 HP on this dyno. The 2.5 extra HP may be dyno variation, or the switch from 20W50 to 10W40 oil for our last event (temps in 30's!).
So, we cracked the previously unopened AFM and started moving the wheel counterclockwise. It took 8 clicks to get it right (that's huge, don't try that w/o a dyno). This resulted in a gain of an even 5HP. This is with a stock airbox, and short straight exhaust, and a modified collector. We tested a lot of things on the dyno, and this tuning made the biggest difference, by far. More on the rest soon, but this was the most interesting finding.
Tune tune the AFM (on a dyno, with air fuel monitoring, please!), lift up the spring on the black geared wheel (pictured below). Move it CCW to make the car run more rich, and CW to make it run more lean. Estimate 1/3 to 1/2 point of change per click (it's not always consistant).
Here is the before and after dyno:
These results may not be typical, but they illustrate what can be gained. They also illustrate why it's important to get your car checked out - you don't want to run that lean!!!