As long as you see some cross counts (from lean to rich) on a narrowband O2 during light throttle, likely it's working properly. TBI data speed is incredibly slow even by that era's standards (TBI is 160 baud, TPI is 8192) so you miss a LOT of information due to how fast things change in and on an engine. Coupled with the narrowbands terrible resolution, that leaves a lot of guesswork.
Knock count will climb as the sensor detects knock (I'm probably improperly interchanging knock and ping), so you should be seeing the total count of knock as a number. If it sticks at 10 (IIRC the ECM will only count to 255 and start back over at zero) then you probably aren't seeing significant knock. That's one reason I like the .csv. I can see the conditions a millisecond before knock happens, and have a good idea where timing needs subtracted, or fuel added, to prevent that knock. My experience has been that the ECM will do a pretty good job of pulling timing if knock happens, and do so more aggressively as knock gets worse, but if that is the case, you'll hear the knock, and you'll feel the "bog" as the timing is cut. Even when my timing and fueling was off by a good margin, knock only happened in small bursts, so not huge knock numbers, but they were spread out over a long period of time as they ramped up then ramped back down.
All this is one reason I generally don't like seeing people modify an engine from base, if they aren't all in on tuning. You simply cannot change something on an injected engine (at least until the OEM's started using wideband O2s) without causing a conflict with the stock tune. When someone tells you they changed something and "it runs fine", you can look back at your datalog and realize that isn't the case. Ignorance of fact doesn't make a presumption reality. Running and not knowing any different, is not the same as running as good as it can or should.
Knock count will climb as the sensor detects knock (I'm probably improperly interchanging knock and ping), so you should be seeing the total count of knock as a number. If it sticks at 10 (IIRC the ECM will only count to 255 and start back over at zero) then you probably aren't seeing significant knock. That's one reason I like the .csv. I can see the conditions a millisecond before knock happens, and have a good idea where timing needs subtracted, or fuel added, to prevent that knock. My experience has been that the ECM will do a pretty good job of pulling timing if knock happens, and do so more aggressively as knock gets worse, but if that is the case, you'll hear the knock, and you'll feel the "bog" as the timing is cut. Even when my timing and fueling was off by a good margin, knock only happened in small bursts, so not huge knock numbers, but they were spread out over a long period of time as they ramped up then ramped back down.
All this is one reason I generally don't like seeing people modify an engine from base, if they aren't all in on tuning. You simply cannot change something on an injected engine (at least until the OEM's started using wideband O2s) without causing a conflict with the stock tune. When someone tells you they changed something and "it runs fine", you can look back at your datalog and realize that isn't the case. Ignorance of fact doesn't make a presumption reality. Running and not knowing any different, is not the same as running as good as it can or should.
