| Author | Message | ||
     S21125r |
Hey all, Was trying to find a good way to objectively evaluate 02 sensor data. I originally set up a history table to display average voltage for a given load/RPM. But I didn't think the data was that useful since it was averaging voltages outside of it's usable range of .4 - .8 volts. So I came up with a better way that I figured I'd share. With in the ADX file there is the ability to set up a lookup table. Within the look up table I set up entries so that anything less than .415 volts gets a -1 output, anything over .562 volts get a +1 output, and anything in between gets a 0 output. BTW - those lean/rich voltage numbers are in the scalars section as "O2 sensor lean voltage" and "O2 sensor rich voltage". After the look up table is set up, you then have to go back to the "values" definition for the O2 you are watching and set the tick mark for "send converted data to lookup table". It will then prompt you for the lookup table you created. Now when viewing the history table for that O2 sensor, it will throw -1's, 0's, and +1's instead of the raw voltages. As expected, anything in the closed loop region of the map should hover very close to an average of 0, but for regions on the fringe of closed loop it could still be useful. For example, I had a couple of full throttle high RPM cell that had 6 samples with an average of .667. This tells me that 4 samples were richer than stoic, and 2 were stoic. I would have expected all of those samples to register richer and therefor the average should have been 1. Also found that several cells in the 13TPS row (low speed in town below closed loop...) are running very close to an average of 1. Probably running a little too much fuel there. Couple of other observations... When I ran my logs back at 2X speed I started seeing averages above 1.0. That shouldn't be mathematically possible if I'm only averaging a series of of 1's, 0's, and -1's. After dropping the playback speed to 1X speed, these quirks disappeared. Also I bumped my sample count up to 50 from the default of 10. 10 samples in logging mode is not very many - maybe a couple of seconds worth - and doesn't give you a good sampling count to assess. Anyway - thought I'd share for anybody that is tinkering with TP. Let me know if you've found anything cool in TP to exploit. |
