I'm getting a little skittish about Rotax's 20Nm lubricated thread spec for spark plug torque. I cannot find a single source which supports that. NGK and Denso publish their accepted torques for 12mm plugs, e.g. https://ngksparkplugs.com/en/resources/spark-plug-installation, https://am.denso.com/plug-global/basic/torque/, and those are spec'ed dry. (See screenshot from https://www.youtube.com/watch?v=XtzpzsPfsYE). However, our plugs are installed wet, because they're lubricated with silicone grease (the thermal paste).
In general, thread lubrication reduces torque values needed for proper preload by 10-30%, and I notice Rotax's new spark plug is spec'ed for 16N-m lubricated. This is an almost exact 25% reduction in torque, as one would expect for lubricated threads.
While there's no danger to the plug by over-torquing, since they are designed to withstand the forces for installation into a cast-iron head, it certainly seems like there is global agreement that plugs into aluminum heads shouldn't go past 20Nm dry. Other spark plug manufacturers, such as Champion, publish almost identical stats.
So what's the smart money here?
1) Trust that Rotax knows what it's talking about because people have been doing it like this for decades?
--or--
2) Use the available data, combined with the knowledge of how engine manufacturers are extremely reluctant to admit fault or change, and that spark plug torque is almost always a range and not a specific number, to aim for a more conservative 16-18Nm, thereby reducing risk of damaging the threads or the plug?
