Quote:
Originally Posted by pokeybritches
So are you saying the duty cycle is what changes when AFR changes, and not the amplitude?
I could easily program a microcontroller to modify the duty cycle, but it would have to be based off of the primary O2 sensor. All I would need to do would be to splice a wire from primary O2 output and set it to an input on the microcontroller. From there the microcontroller could analyze the signal and output the appropriate duty cycle. The microcontroller's output signal would replace the output signal from the secondary O2 sensor, and it would be fed into the ECU. I guess I could do the same with an analog output if the signal is voltage-based and not based on duty cycle.
I need to dig into how O2 sensors work. How can they alter the duty cycle with changes in the AFR? I would think it would need some kind of processor onboard, and I don't think they have them.
|
An o2 sensor is basically a chemical comparator. It compares the amount of oxygen in the exhaust stream to the amount of oxygen ambient.
I know back in OBDI land, the duty cycle is what the ECU to control the AFR.. Basically it was like an ON/OFF switch, so if it was OFF it would read lean, and if it was ON it would read rich. Depending on how often the sensor switched from lean to rich would change the AFR. Again, this was on OBDI cars where the max the o2 sensor could change the AFR was maybe 10-20%, where I think on more modern cars (OBDII) it can vary it much more.
I'm not very familiar with electronics, so please excuse my ignorance if I use the wrong terminology... I think you're right, the ECU looks for the changes in amplitude not duty cycle..
http://www.mr2.com/TEXT/O2_Sensor.html
http://en.wikipedia.org/wiki/Oxygen_sensor
Bottom line it sounds like your methodology sounds correct. Since a catalytic converter essentially removes the same amount of pollution regardless of how lean or rich it is (not exactly true, but true enough for this discussion), it would seem to me the voltage drop (either positive or negative) would be fairly linear.