[Update 7 July 2010: The decoder failed by February. TCS advised that the CN-GP is a very delicate decoder; consequently, I cannot recommend this installation method. Currently looking for a new method of installation. Watch for future posts.]
Read Part 1.
It’s done. Well, almost. [Update 10 Jan 2010: Yep, it's done.] Anyway, it works, and that’s what counts. Here then I recount the two hours I spent last night making it work, and the year of effort that culminated in those two glorious hours. (Yes, it takes me about one year to install one decoder. No, I’m not that slow: it’s just a function of how much free time I have, and how I choose to spend it.)
I lost count of how many decoders had to die for this to work. I think four? Too many, anyway. I’ve learned some important lessons.
I tend to rush things, sometimes, and now I know to always, always check my work for unwanted shorts with a multimeter before putting it on the tracks. This simple check would have saved me at least two decoders.
I also learned to be careful soldering. Last night, I was sloppy with one of the wires I had to solder to the decoder. This sloppiness caused a nearly invisible short, as some minuscule whisker of wire jumped to a nearby solder pad. Reworking the solder joint properly solved that problem.
Finally, I learned that the CN-GP is not the panacea I thought it would be. The frame milling necessary to get the CN-GP to fit and seat is complex, and I have only clumsy handtools. Since last January, I moved from handfiles to cheap rotary tools to expensive rotary tools (thanks Johnny!), but even my fancy Dremel was too awkward and random for the precision needed to re-sculpt the frame to receive the CN-GP. In the end, it worked, but I had to add significant solder pillows to the pickups on the decoder to compensate for poor milling technique. Indeed, having sent the loco to TCS for their professional opinion, they responded that I had botched the milling job, and that I would not succeed in getting the CN-GP to work. They were ultimately wrong (sorry, Arthur!), but their reasoning was good. No more complex milling for me.
But even if I could manage the complex milling, the plastic retaining clips that Tomix uses needed modification to allow all the wires and components to fit. Cutting notches into these weakened the plastic, and one of them is now cracked and straining. This worries me. I’ve put a dot of superglue to help prevent further cracking, but I don’t know if it will hold. The clip is important, because it holds the decoder boards against the metal frame, ensuring electrical contact between the two. This particular conversion technique—using a CN-GP in place of the stock lightboards—seemed promising, but I worry that the design of Tomix locomotives is deceptive; the nature of the clips seems to be secret anathema to the technique.
I had originally planned to install markerlights, but accessing the solderpads on the decoder for extra functions would require further modification to the retaining clips. So no markerlights for this loco. Yet another reason to prefer a different technique for future conversions.

WIth the wires all spudged down, the shell will fit neatly over the whole affair. Notice the slight bulge in the plastic retaining clip on the left.
Anyway, the project isn’t quite done. To simplify testing (and to prevent damage to expensive parts), I pulled all of the clockwork. I need to tear the whole thing down, and reinstall and lubricate the clockwork. Things could still go wrong, but the likelihood is low since I won’t have to resolder anything. I could still break a clip, which could be a real problem. [Update 10 Jan 2010: No problems with reinstalling the clockwork. Still runs well!]
The upshot of all this work is that I will not use this technique again: Getting the milling just right is too complex, and the reliance on the fragile plastic clips is worrisome. I now have a definite strong preference for Kato locomotives (having the option of drop-in decoders) over Tomix locomotives for future purchases. But that’s not to say that Tomix makes bad locomotives—this thing was an amazing runner unmodified under DC power. Just that I don’t relish the thought of another conversion like this one.
