Mercedes (and BMW) straight sixes have a rough idle "by design"?!

[sanchezc]

My car (a 1991 W126 M103 300SE) has had a rather rough and irregular idle since I acquired it from an MB dealer (with approx. 95,000 miles in August of 2002; it now has approx. 115,000 miles).

My current mechanic tells me that that idle is typical of Mercedes (and BMW) straight sixes. However my car's idle was PERFECTLY smooth once after being serviced (by Ditmire's team), although only temporarily. Which proves that the idle can be as smooth as I otherwise always heard and read a straight six's should be. Well, at least temporarily... Techron also helps, although not perfectly and also only temporarily.

So here is my question: Does any of you have a perfectly, and permanently smoothly idling straight six, and in particular an M103?

Among all the possible causes I have gathered by reading this (most useful and helpful!) newsgroup but haven't addressed yet are:
+ vacuum leaks: can they be tested for?
+ spark plug wires: but some claim that they are not a maintenance item?
+ timing chain: I've read that its state (stretch) can also affect the idle?


=========================
PS: Here is what I have changed so far and recently, with the main reason:

+ fuel injection:
---- fuel filter, because I wasn't sure when it had been changed before;
---- EHA, because it was leaking;
---- injectors, because they were leaking too, resulting in difficult starts;
---- O2 sensor, because it was original but is supposed to be changed every 60k;

+ combustion air:
---- idle control valve, because the car tended to stall when coming to a stop;

+ ignition:
---- rotor and cap, because the contacts were in bad shape;
---- spark plug caps, because there were blue sparks (i.e. leaks) around them in the dark;
---- spark plugs, because there was some serious knocking at idle due to a faulty set;

+ engine suspension:
---- engine mounts, because they were worn out, although surprisingly the new ones, installed just a couple of days ago, are not making any difference in the vibrations perceived at idle: do they need to settle down?