If I were an experimental physicist who had access to a data logger, I would be tempted to rig up a system to ground out my second spark plug....
If you were an experimental physicist you wouldn't be
tempted, you'd be
compelled. Going off topic for a moment, when I was in graduate school in Southern California I left the lab late one foggy night and as I left the parking lot on my Triumph it started missing on one cylinder (because of the moisture). I reached down to wiggle the spark plug wire and my memory is still painfully fresh of the feeling caused by three or four pulses of 5000 V going through my elbow grounding out the spark plug.
A mini heat wave has us already at 96 oF at 1pm and headed for a high of 98 oF so I only was able to make one ride today. OK, I
could make more, but did I mention it's going to be 98 oF?
A few notes in reading the attached two graphs from my data logger. The values for AFR (magenta), rpm (black), and voltage from the throttle position sensor (red) at the upper right are from the time into the run (11 min. 29.33 sec. for the first graph) where I had the cursor set when I took the screen shot. The rpm values shown in black are 2x what they should be, apparently because the inductive clamp I have laying directly on the magneto is picking up a signal from there as well as the wire. The voltage values are 1.5/5.0 of what they should be because I incorrectly tricked the program when it imported the data -- I should have used a factor of 5.0/1.5.
During the interval between 11:22 and 11:48 on the first graph I had the throttle at a fairly constant ~0.1 so it was mostly running on the pilot circuit. The AFR can be seen to be a very rich 11 or so. The four sharp spikes in this time interval are because the engine missed, allowing raw fuel and oxygen to pass through to the exhaust. Engine rpm was ~7800/2 = 3900 so each combustion cycle was (3900/2)-1 = 0.5 msec. The fact these spikes are seen tells us the response time of the sensor is at least this fast.
The three regions near 11:12, 11:20 and 11:52 show that when I completely rolled off the throttle the exhaust mixture got very lean because only air was passing through under that circumstance.
Turning to the second graph, I had the cursor placed at 8 min. 19.3 sec. for the screen shot, at which point the throttle was wide open. Note that the AFR is a
very rich 9.35.
Starting at the far left of this graph it can be seen that in a time of ~1¾ sec. I increased it to ~½ throttle. Up to ~¼ throttle the AFR stayed around 13, but then started dropping as the throttle opened wider, dropping to ~10. I briefly let off the throttle, then rolled it back on to full throttle. Again up to partial throttle the AFR went back to ~13, but again it dropped to below 10 by full throttle. Note that these are "transient" behaviors, not steady state, so there's a lag between changing the throttle position and the engine catching up with the change.
It's not worthwhile analyzing this data any further. The bike was
very rich and, as a result, it would load up at low throttle settings and I'd have to rev the engine to clean it out enough that it would again run smoothly. So, there's a lot of bad behavior reflected in various time intervals. Despite this, I thought some of you might like to see examples of what's possible with a data-logging AF sensor. Everything I discussed above went by way too fast to assimilate while on the road dodging cars and with the sun glinting off the display, but afterwards in the air-conditioned light of the living room much was revealed.
This morning I replaced the #30 pilot with a #25 but did not adjust it before today's run for reasons not worth going into. Had I adjusted it I assume the AFR at low throttle settings would have been more reasonable than the super-rich <10 that caused the bike to load up today. I'll drop the main jet by a couple of sizes and adjust the idle mixture before setting off on a similar run tomorrow.