Started the day fiddling with my Bachmann Class 37 254 in a BR blue. Did a brief test on a DC test track to find it rather sluggish and then failing. Weird part was the centre of the roof was hot. Took the body off to find it has a Bachmann (ESU) 21pin DCC decoder. Hmmm. Faffed about and fitted a blank plug and it worked sluggishly. So stripped the bogie frames off to find one bogie was full of grease. Cleaned that out and cleaned the wheelset faces and treads, plus the pick ups. Some light lubrication when the wheelset were refitted, and on it was drawing much less current. So tried the decoder and it was dead, hence guess it’s fired through an overload condition.  So a bit disappointing. 
 

no obvious signs of failure. How do you tell though what failed? 
 

All the diodes seem to be ok. 
 

I guess they aren’t repairable?

I then faffed about my Sprog 2 Version 3, Decoder Pro on my Mac. This is a hateful system which seems to be rather temperamental, but it’s the only DCC controller I have at the moment. Upgraded to version 5.6 of Decoder Pro and the Java 11 environment. It’s weird how the Mac knows the Sprog is there. But I don’t think Decoder Pro does? In the end I gave up with it, because it wouldn’t talk to the either the Class 37 or a Hornby 56 nor a Hornby 50 all with working decoders. 

Finally I had a look at the 56 further. This is a bit of a rat bag model. It is an acquired 56 sound chassis that was in a sorry state.  Managed to fix it up and it was working last time I test it.  Anyway it was dead. After much investigation find a wire of an inductor coil had broke away from it pad, hence open circuit to the motor. Soldered this back on and with 21 pin DCC blank plug, it worked. So slightly pleasing. 
 

If anyone has compressor details or the outer sand box mouldings or a 3D print or them, let me know because I will gladly buy them to detail up this chassis and move it along into a finished off model.