mk e wrote:
Pizzaman wrote:
You're going to have 12 o2 sensors??
No...14
My thought is to put WBO2 in the collectors for normal mixture tuning and normal use. Then WAY less expensive NBO2 ($25 each on amazon) in the pipes that I'll use for checking cylinder to cylinder mixture at or near stoich to make sure everything's matched and also since they react in a msec or so it should be possible to see any cycle to cycle variation if it exists at any mixture (but not measure actual mixtures away from stoich). I've never tried this before but I thing it will give some interesting data....the AEM/enginelab ECU logs at 1000hz (as does the M150 motec I;ve have in case the enginelab doesn't work the way I think it should) which will give me at least a couple readings per cycle even at redline.....and if the data is not what I expect I have caps for the bung holes and I'll go back to just 2 sensors
Got my answer to the last question!
I've seen systems which sample a single narrowband per bank many times per cycle to determine the individual cylinder fueling errors. The average of a cycle is used for the normal closed-loop control, and the variance of specific samples to the average of the cycle is attributed to specific cylinders. The end result trims each cylinder individually and can detect failing parts which aren't bad enough to be caught as misfire. So you could do with as few as 4 sensors, if they are narrowbands (which are really fast).
I'd expect the individual cylinder errors can be trimmed pretty well with 14 sensors. Depending on how you wire it you can probably estimate rough EGT from the narrowband heater resistance.
I've thought for awhile about the possible response of using a WB as an offset NB sensor. Instead of controlling pump current to hold the Nernst cell at stoich (control loop #1) and using the pump current to control fueling (control loop #2), it would set the pump current to the target lambda and do closed-loop fueling based on the Nernst cell feedback (1 loop instead of 2). I'm curious if this would be faster response, especially near stoich where the WB sensor keeps flipping polarity on the pump. This would allow individual cylinder measurement with the fast NB sensor and still the off-stoich operation of the WB sensor.
Login
Register
FAQ
Members
