< Safety Valves & Water Unit >

The safety valve shown in the picture is not from William's drawings. It comes from 'Mild Pop Safety Valve' presented by Gordon Smith in "Engineering in Miniature" in 2001. It has high venting capacity, low hysteresis and mild action. Stuart Hardy, a model engineer in N.S.M.E.E, introduced the valve to me. He has already completed the William with this type of safety valves.

The valve body is made from hexagonal brass bar. First, the bottom part was turned and threaded. Then it was screwed into a mandrel, turned, drilled and finished with a D-bit. Note the mandrel was chucked in a collet instead of three jaws. It means the mandrel can reproduce true center in any future use.

The valve stem is also made from brass. Note a round bar with a center hole hold in the tail stock, which supported the long and thin material when turning.

The safety valves exploded (left hand) and assembled (right hand). The spring was coiled by myself. The top ring is a lock nut which secures the adjusting screw made from phosphor bronze bar.

Venting pressure was adjusted with air-compressor. The top screw was adjusted until the pressure gauge of the compressor indicates constant 0.6 MPa. Final adjustment should be done in the first steaming up, because venting pressure of steam is different from that of air.

I designed the water unit at the rear part of the loco, due to following demands.

- Abolish coal banker and rear water tank in order to ease driver's operation.
- Mount 2 KG dead weight at the rear end of the frames to improve axle weight balance.
- The hand pump should be mounted as rigidly as possible.
- Driver can confirm side tank waterline at a glance.
- Driver can confirm water flow from the axle driven pump at a glance.
- Make available supplying water from a container in the trolley.
- Test running without the side tank can be done.

And the result is shown in the photo.

As a dead weight, two brass blocks were mounted on the rear buffer beam. The blocks have inside water manhole and each block will be connected to each side tank. The hand pump is fixed rigidly onto the right hand block and sucks water from the block. The two blocks themselves were connected each other with a balance pipe. The pipe has two blanch to the pair of axle driven pumps. The left hand block has a nipple at the rear corner, which will be connected to a water reservoir in the trolley. At the rear end of the left hand block, there is a fat brass pipe which works for three purpose as follows.

1) Side tank water line indicator
2) Air venting hole of the water unit
3) Return flow catcher from the axle driven pump

As a filter, a brass ring with fine brass gauge is sunk into the tube (see small photo). The blocks and connections will be hidden between the floor and the running board.

Free hand tube bending takes much time and resulted in bad looking. I designed all of tube bending in CAD and bent them before silver soldering. The photo shows how to bend a tube in required radius and angle.

For more precise bending, I employed full size drawings printed out from CAD. I don't like annealing because it makes copper tube too soft to bear rough handling in practice.

The lowest part of the loco except wheels is the suction fitting of the axle driven pumps. I changed the design of them in order to keep reasonable height from the truck.

Some fitting works were done with the loco backside up in the carrying case. If smokebox can support the boiler weight, same trick can be done for completed loco and it will make maintenance much easier.

The lubricator is driven by one of expansion links. This is standard of Japanese locomotive.

The photo shows all of the oil check valve components. Ball valve is push against valve seat with a coil spring via a cup. Note a pair of holes in the cup for oil path.

Branch tubes from the oil check valve were connected to both steam pipes from the smokebox. The two tubes should be the same length and height, otherwise the oil always drops into one side cylinder.

All of the components for driving loco were completed. In the next time, I will report the first test running !