Encouraged by how smoothly the DCC conversion of 31 327 went, I decided to do 31 004. This one also ran very smoothly on DC so I thought it would go just as well as 31 327. WRONG!!

The actual hard-wiring went more smoothly because I knew exactly what I was doing this time. I pulled the Digitrax decoder out of 31 327 and plugged it in to 31 004. It ran like a dog! It stuttered and stopped and refused to complete even one circuit without stalling. I thoroughly cleaned the wheels, oiled the mech, cleaned the wheels again, making sure I got the backs as well, checked the pickups and tried again. It was better but not brilliant, so I tried a Gaugemaster decoder (demonstrating the wisdom of using the JST-compatible decoders!) with BEMF but that was no better (probably slightly worse), so I put the old Digitrax one back in. The Gaugemaster decoder worked a treat in 31 327 though.

After much persuasion (both physical and verbal!), 31 004 has been trundling around the layout for an hour and a half now. Hopefully this will free it up somewhat.

I took a few photos of the hard-wiring process this time so you can see what s involved.

Incidentally, I removed that capacitor between the brushes on both models after the photos were taken.

And here is 31 004 in service. I detailed it with the separate handrails on the fronts some years ago. 31 327 was easier in some ways because it only had two handrails at each end whereas 31 004 has four, albeit they are all straight ones. It now just needs the red circle coupling codes added above each buffer (these early 31s had electro-magnetic control), some pipework on the buffer beam at the end without a coupling, and a lot more weathering added.