Number five injector was of course opening too early causing the early ignition knock I imagine. You have kind of proved it does not take very much to upset the individual cylinder timing. More reason than ever to at least check the injectors every 100k as recomended by mercedes.

I feel that old milli volt test would have picked up the hotter cylinder quite easily as well . Or perhaps just reading the glow plug temperatures with a laser thermometer with the engine warmed up and running.

At this point I am not aware of anyone trying that. Or if it would have any merit.

Since quite a few of us own one of those temperature readers I may look into it somewhat. Or someone else could and post if the plugs sit at different temperatures. It might be informative. I assume the sensitivity of the whole glow plug to internal individual pre combustion chamber temperatures is going to be the question.

What does it really cost to scan the operating surface temperature of the exposed glow plug end? A major temperature difference on any given plug might indicate something is unusual. Could be as simple as excess carbon buildup I imagine.

Needless to say the glow plugs should be wiped clean when cool as a different surface color would effect the laser temperature reading I think. I also would not want to see anyone burn themselves.

Who knows one day we may yet stumble on something quite useful that is cheap and fast to establish certain faults. In my limited knowledge on a well balanced warmed up engine all the external portions of the glow plugs should be at simular temperatures.

As long as all the glow plugs are the same brand regardless of internal element resistance . I think they should transmit heat equally well from the combustion pre chamber. Just a thought. Worth looking into?

I have this crazy compulsion to stir the pot periodically. The benifit from my perspective is at least it gets people thinking . Their contributions then tend to increase general knowledge in my humble opinion.

Sometimes for example we will just reach halfway or much less to a possible usable method. Yet even with that shortfall we come to understand applications of different troubleshooting approach possibilities better. Even more important is the possibility of one thing or a combination of them leading to a real usable senario at some point. That personally is my ultimate hope.