On Feb 9, 2013, at 6:09 PM, Tracy wrote:

> I decided to experiment with a wideband sensor.   They have come down in price so even penny pinchers like me can afford them.  $150.00 range.  They typically put out a voltage between 0 & 5 volts instead of 0 to 1 volt like narrow band does.   The polarity is opposite as well so rich mixture is at low end of range instead of high.   With the latest version of the software (after Feb 2013) you can now setup the EC3 to use either  type sensor.   With the wideband sensor it is much easier to target the exact mixture you want and to achieve the best fuel economy in closed loop.

The biggest issue I've found in tuning closed loop is that the computer I'm using doesn't understand lean surge.

The rotary gets into a situation where the mixture is to lean to ignite.  The mixture doesn't exhaust as normal (because there is no expansion from heat), so a lot of the mixture gets recycled into the next combustion event.  The next event fire just fine (because it has double the fuel charge).  The O2 sensor sees some combination of the two events, and the combination is NOT a useful F/A mixture.  

Ive found that the most useful way to programmatically tune to a lean mixture is to move up to it from the rich side.