The EFIE is an okay device for altering closed loop mixtures a very small amount. However, once you start leaning things out too much you go beyond the intended use of the sensor and the resolution becomes horrible. Here is a chart that shows AFR vs O2 sensor voltage. As you can see, there is almost no voltage change at higher AFRs.
The $olution? Use a wideband O2 sensor. They allow proper resolution for higher AFRs. All the new systems can simulate a narrowband sensor's signal, and you can set the AFR at which it operate in closed loop mode. For example, I can tell the wideband system to say 17:1 is stoimetric and it'll keep me at 17:1 while in closed loop.