I saw a thread in another forum that linked an article about MB diagnostics:
CEL article
Fascinating, but technical--it's over my head for the most part, but I've been dealing with a CEL problem for a few months, so most of the codes and jargon are becoming familiar. To recap: I just tried my third "new" AMM (I say "new" meaning "new to me"). The first one was the wrong model, but my independant mechanic had it on his bench (from a ML320) and suggested I give it a try. I successfully drove to a very nearby office park, where I floored it and it quickly went into limp mode. We put the old one back in. Fine. Then I ordered one from ebay, put it in at home, quickly went someplace I could floor it: limp mode again. At this time I figured that I had tried a wrong AMM and a bad AMM, and that maybe mine wasn't the problem at all (OBSII codes were P0170 and P0173, of course). But after months of driving with CEL on, I tried another AMM from ebay (different seller, one with lots of AMM's). This time I read the codes after replacing the AMM, and several P030x misfire codes in addition to the fuel bank malf's. I reset the light, and took it for a slow, easy test drive; no flooring it immediately this time. After gradually bringing it up to highway speed, I did some spirited passing and, lo and behold, no CEL, no limp mode. This would appear to be consistant with what he said on page 3 of his article, that when the 1.32 adaptation falls to 1.0, random misfire codes ensue (I got five: P0300-P0304B). My questions (finally; sorry for the long post), (1) is it possible that I was too quick to stomp on the earlier AMM replacements, and that had I been gentler, they may have worked? (2) Am I even wise to believe that this AMM has solved my problem? (3) And am I correct when I infer that the whole time my CEL has been on, I've been burning 15-32% more fuel than I'm supposed to? Thanks for the article, stevebfl, wherever you are!