Friday, 18 October 2013

Fitting DCC Decoder to a Bachmann Collett 0-6-0 with Cast Metal Tender

Going through my Dad's old model railway bits, I found a K's white-metal kit for a Dean Goods 0-6-0. The loco was incomplete, but one of the Collett 0-6-0's on our local branch line in the 1960s was fitted with the tender from a Dean Goods, and I had a spare Bachmann Collett 0-6-0. This model's destiny suddenly changed to one with a future as 2218.

This particular Bachmann loco was unusually a very poor runner, moving fitfully and stalling frequently. Adjusting the pick-up contacts had not resulted in any improvement, so I experimentally (ie as an unmaintainable lash-up) fitted pick-ups in the tender and two sprung electrical contacts between the tender and loco. But it was cumbersome and difficult to wire, and didn't make it run any  better, so it was removed. And it has sat in a box for a couple of years waiting for another idea.

Now that the garden railway uses digital control (DCC), I have been gradually fitting the appropriate locos with decoders. The older Bachmann Colletts have no space for a decoder in the loco without cutting metal, but it was clear that the floorless white-metal tender could easily be made to house a decoder with four linking wires to the loco. To alleviate the inconvenience of having them permanently coupled, I would use a connector. I produced a design and proceeded with an experimental assembly. Wiring was again extremely difficult, and the force required to separate the connector was impossible to apply where it was needed. So back to the drawing board again.

It then struck me that if the decoder was fitted to the loco with wires of the right length and slid into the tender during coupling, the difficulties would all be resolved. Wiring would be more straightforward (there would be none permanently in the tender), the loco and tender would be easily separable, and no connector would be needed.

First I made an insulated box to take the decoder. I used 30 thou plastikard. It must be wide and high enough to easily fit the decoder, and reasonably long. About twice the length of the decoder (if possible) seems about right. The bar across the bottom end (end A in the first photo) is to stop the decoder falling all the way through while still allowing air flow, and the packing under the entry end (end B) is as thick as is necessary to get the entry angle correct. To clear the wheels, the overall width should not be more than about 12mm for 00, which restricts the choice of decoders. I used a TCS M1, which comes fitted with a protective insulating sleeve. If you use an unsleeved decoder, you should fit a sleeve for protection.
The decoder box before installation in the tender.
The box was then fitted in the tender and held in place with a suitable adhesive. Check that the sleeved decoder and cables slide in easily, and the wheels turn freely.
The decoder box in place under the tender.
Now remove the loco body by undoing the two screws at front and rear of the chassis. Then loosen the central chassis screw, and carefully remove the keeper plate and wheelset from the main chassis block. To fully separate them, you need to cut the red and black wires leading to the motor. 
The three main chassis sub-assemblies.

Quickly and carefully (to avoid melting plastic), unsolder the red and black wires both at the motor and at the metal contact strip attached inside the keeper plate at C in the picture below. Drill two tiny holes each side of the tender hook, each just big enough to take one decoder wire. Cut short and sleeve the unused decoder wires (blue, yellow, white). Fit any protective sleeving you think necessary to protect the wires from chafing. Feed the decoder wires through the holes, with red and black closest to the metal contact strips inside, to which they will be soldered, and red on the right from behind when the wheels are down. Allow them to slide in and out while trying the position of the decoder until it is about half-way down its box with he tender coupled. When you are happy that it is just right, unthread the orange and grey wires to avoid damage, then trim the red and black wires to length and solder carefully to the contact bars.
Part of the keeper plate showing solder terminals C and tender hook. 

Re-thread the orange and grey wires into the keeper plate, then thread them through the square hole in the main chassis block, carefully re-fit the wheels and keeper plate ensuring the six contacts are correctly located behind the wheel rims, and tighten the middle screw. Run the orange and grey wires up to the motor terminals using the same route as the old red and black wires, trim and solder in place. Note that space is very tight, so correct routing of the wires is critical to avoid damage.
The finished "decoder tail".

You can now check there are no short circuits, especially from red to black or orange to grey. If ok, test on the track, and if still ok, refit the body and tighten the front and rear screws.
Loco and tender coupled together. The wires will be painted black.

A similar principle could be applied to other locos with limited space for a decoder, providing (a) it has a tender and (b) the tender has a suitably located space for a decoder box. Unfortunately, this rules out the majority of limited-space locos, which are tank engines.
The following day, the engine was given a test run in the garden. The heavy rain showers  meant that the trackbeds were saturated, but she still managed 8 coaches up the 1 in 50, with only occasional short bouts of slipping. The actual total weight of the coaches used was 1020g, which on a 1 in 50 gradient requires, ignoring friction losses, a pulling force (or tractive effort) of 20g. Scaled up 76 times to full size (i.e. multiplied by 76 cubed because weight or force is proportional to volume) gives a figure of about 20,000lb. For comparison, the quoted theoretical tractive effort for a real engine of this class is 20,155lb - a pretty good correlation!
2218 under main line test. She still needs some painting, detailing and renumbering.
A bonus result of this task was that when I removed the keeper plate, I found that the gears, axles, and sprung centre wheelset bearing were all packed with a thick grease, most of which had hardened to the consistency of candle-wax. It's not surprising it was a poor runner! As much as possible of the grease was scraped out, a little light oil applied, and it is now one of the sweetest of runners.