Bachmann K3


Another loco aimed mainly at our extended "Little Histon" layout, to add to our LNER period rolling stock operating in the station area.

Hattons photo of 1304

Likely configuration:

The DCC socket in this locomotive is located under the boiler. There is only room for a small decoder and possibly the smallest size of sugar cube speaker. Lighting options are also limited, with a metal forward cab bulkhead very close to the rear of the motor and no obvious routing path for wiring front oil lamps along the upper body shell, back to the cab rear. (Although the wires could be taken down externally between the wheels under the loco, but this would preclude future separation of chassis and body shell.

So for the initial fit, I have removed the DCC socket board, extended the motor wires, using heat shrink sleeving to insulate the solder joints and taken the motor wires down the tunnel in the metal chassis previously used to run the wheel contact wires up to the DCC socket, and then through holes drilled in the chassis bottom plate. The original driving wheel contact wires were removed and a new pair were also taken down through similar holes. These wires will be secured and connected to the tender. There is room in the tender for decoder, speaker and stay alive capacitor.

I'm going to use a Zimo MX648 decoder with Paul Chetter's A3 three cylinder sound project on board. This will drive an 8 ohm Zimo Dumbo speaker. Holes have been drilled in the tender bottom for the sound waves to escape. An AVX 6800uF supercap with zener diode based protection circuit will also go in to the tender. Some form of ballast weight will be needed to replace the original plates which have been removed to facilitate the speaker vent holes.

A footplate crew will also be added.

This locomotive is intended for wartime service, so the red lining has been removed and the "LNER" tender markings have been changed to "NE". 


DC Testing:

The locomotive running in DC configuration, straight out of the box, is not a complete disaster, but its not the smoothest I have seen. Additional lubrication of the motor gear chain helped to improve slow speed performance. Using my bench power supply which provides pure DC, the motion is very quiet.......However, initially the DC test was only at low I discovered later, as the speed was increased, running became increasingly rough.


Showing the locomotive in 1944 style, with the red lining removed (or painted over) while decoder installation is under way.


Zimo speaker glued to the tender chassis and the Zimo MX648 ready to be inserted through

the hole into the coal hopper. (The white plasticard insulator  covers the stay-alive supercap and

its circuit, to prevent a disaster if the glue ever fails.)


The stay-alive super-cap and its zener diode protection circuit, glued to the tender top & hidden by the insulator.


Decoder CV Adjustments:   TBA

e.g.   Some motor control adjustments:   CV3 to 60 & CV4 to 25 (acceleration and deceleration rates), CV9 to 87, CV57 to 120.

chuff rate about right (6 per driving wheel revolution) with CV267 dropped to 66.

Digitrains MX648 Function Details

Key Number

Function Description



F1 Sound ON
F2 Brake Key
F3 Long Whistle
F4 Short Whistle
F5 Heavy Train (off) / Light Engine (on) Selection
F6 Coal Shovelling (This sound is also a random sound)
F7 Live Steam Injector (This sound is also a random sound)
F8 Blower
F9 Wheel Flange
F10 Safety valves lifted (This sound is also a random sound)
F11 Westinghouse pump
F12 Handbrake
F13 Buffering up
F14 Coupling
F15 Cylinder Drains Opened
F16 Water filling
F17 Station ambience
F18 Guard’s Whistle
F19 Fade All Sounds

Shunt Mode

F21 -
F22 -
F23 -
F24 -
F25 -
F26 -
F27 Overall Volume Down
F28 Overall Volume Up


Adding the Hornby Detailed parts :

The loco and tender brake hose parts were glued in position using super-glue and the brake linkage was sprung back into its slots on the loco and  underside.


Poor high speed running:

The loco unusually exhibits jerky behaviour in the forward direction, as the speed is increased from a crawl..... and the uneven movement continues across the rest of the speed range. Now knowing what to look for, a DC retest was carried out with similar, although not quite such extreme results.

The loco was stripped down and the motion studied. The centre driving wheel pair wobble from side to side as speed is increased. This may be the cause of jerky movement. The motor and then the motor gearbox tower were removed so that the chassis could be rolled along the tracks. Movement was a little stiff, but there was no obvious cyclical resistance to movement. Axles, gear box and motor worm gear were all lubricated with high viscosity oil. All linkages in the rods and valve gear were lubricated with lighter oil. The locomotive was re-assembled and re-tested with little or no real improvement.

Curiously, there is little sign of the problem when the loco travels in the reverse direction.

The problem clearly has a mechanical dimension as it can be seen in both DC and DCC operating mode.......

Drastic plan B:  The chassis assembly from our original "Little Histon" DC mode K3 locomotive was exchanged with the problem new chassis. This behaves much better, although there is still more than a hint of the problem behaviour, suggesting that it is probably design related.

Showing the chassis with the worm gear cover removed & the motor wires routed down through the bottom cover plate.


With the chassis swapped, we now have slightly wobbly running as the speed increases, but its not too obvious. (The DC loco was re-built with the later chassis and still gives good lower speed DC behaviour but rather hesitant higher speed behaviour.)


K3 on the loft layout.


Hauling M10 Achilles armoured vehicles on Warwell wagons.


Further Investigation re the poor high speed running:

Running the loco fast enough to reproduce the hesitation effect, it seemed to be slowing very briefly when the right hand rods were in the horizontal position. The rear connecting rod seemed a bit tight when the wheels were in this position, so I unhitched the rod from the rear wheel and filed the rod end hole to give a little more play on the rod, cleaned and lubricated the joint then re-assembled it.

Now if the BEMF is disabled (CV58=0) The loco runs smoothly at speed. (Although starting behaviour is of course unacceptable with no BEMF active). Using the Zimo 3 point load compensation system involving CVs 10, 58 and 113, I set the BEMF compensation to operate only over the lower quarter of the total speed range. This provided a partial solution, but there was now a slight jitter noticeable at speed steps 1/28 and 2/28. Lowering CV58 a little largely eliminated this problem. 

The overall speed range was excessive with CV57 set at 120, so this was reduced eventually to 100, expanding the number of useable speed steps within the actual operating range of the loco. CV3 & 4 were reduced proportionally. 

Experimenting with CV56 and CV9 one and tens digits slightly reduced the tendency for low speed jitter, but the main influence on behaviour was the load compensation adjustments (CVs 10,58 and 113).

I got some interesting effects under load when the loco was moved on to the loft layout and coupled to the heavy tank train.... with my initial light engine settings, movement at around a quarter of max speed through the transition zone between BEMF on and off resulted in the loco repetitively slowing then speeding up until the throttle was further advanced, when normal running resumed.     Moving the transition point up a bit via CV10 and compensating by dropping the zero speed BEMF effect down to around 50% via CV58, restored behaviour under load to smooth operation across the entire speed range, with light engine behaviour still very acceptable, but with just a hint of a wobble at around a quarter speed.

Final Relevant CV values are as follows:

CV # Value
2 1
3 55
4 15
9 98
10 80
56 22
57 100
58 130
113 30




Click this line to watch the K3 in action on You Tube

Supplier website links:
Rails of Sheffield    The K3 was sourced from Rails of Sheffield.
Digitrains    Source for decoder, speaker and stay-alive super-cap.


Click to move back to Model Rail Index Page