Charlotte the Starlet


The purchase

On 4th June 2017 I received an e-mail showing a loco that was for sale in Orange. The loco was part of a deceased estate and was built by Bob McKinney who was a member of Orange Society of Model Engineers. I rang Matt (the seller) and he told me some details and sent me some more pictures. The following Friday we drove to Orange and bought the loco and the riding carriage.

The loco is an 0-6-0 saddle tank (Sweet Pea Metre Maid) and looks like a model of a small (possibly 2' gauge) loco. It is unusual because it has a Bagnall type boiler and Hackworth valve gear. The saddle tank connects to an axle pump and the riding car has a tank for the injector. Both the loco and the riding car have manual brakes. The cab valves are for the blower, whistle, drains, and injector.

The loco seems peaceful but we're keeping it in its cage because it might have a mischievous or destructive streak. Some saddle tanks can be nasty particularly when in a new environment.

Initial testing

The first test was a boiler leak test using the garden hose as the pressure source. The mains water pressure here is close to 80psig which is the working pressure for the boiler. We put the loco on blocks on a stand, swapped a safety valve for a test connector, filled the boiler, and finally pressurised the boiler. The only leaks were dribbles from some of the valves. Having the loco on blocks means we can spin the wheels and check the motion.

The next test was a proper squeeze to twice working pressure. For this we swapped both safety valves for a plug and the test connector. Also, the pressure gauge was replaced with a broken one because the test pressure here (160psig) wrecks the gauge. As before, the boiler showed no leaks. Notice the correct pressure reading on the big gauge and also that the wrecked small gauge is way off the dial. I wrecked this gauge doing a test years ago and I'm glad I kept it for future testing like today.

Now we could do the initial steam test. Much more fun than the water tests. We lit a fire and opened the injector steam valve to let the excess water out as the boiler heated. Once we had steam (30psig) we used the loco blower to keep the fire alight and opened the regulator. After spewing out all the oil we'd previously put in the steam chests the wheels started turning and we started testing other things. The injector worked well but we couldn't get the hand pump or axle pump to work. Other problems included a steam pipe leak in the smokebox and that the blast pipe is far too low below the petticoat.

Water pumps

The loco has both a hand pump and an axle pump. The unusual design feature is that both pumps are in series. The hand pump kept jamming and when it did move water just leaked out the cylinder. I found that the outlet check valve was jammed closed. Freeing the check valve and replacing the O-ring fixed this pump.

After fixing the hand pump there was a terrible leak from the back of the axle pump. This pump is mounted on a frame stretcher and this stretcher forms the back of the pump. There was a rubber gasket between the pump and the stretcher and this gasket had torn. I soft soldered a brass plate to the back of the pump to seal the end of the cylinder. Replacing the gasket wasn't an option because this part of the stretcher is rusted and pitted. Both check valves were leaking so these were cleaned and the piston O-ring replaced. Access to this pump is very limited but luckily I managed to remove the stretcher and pump as one unit.

Clack valves

The boiler has two clack valves - one for the hand and axle pumps, and one for the injector. The pumps' valve developed a leak and started blowing water into the saddle tank. When dismantled I found the valve had a dislodged O-ring. I removed the O-ring and cleaned the seat and refitted the valve (minus O-ring) after successful testing.

Later, after filling the boiler via the injector drain, the injector clack had a bad leak. Again, the solution was to remove the dislodged O-ring, clean the seat, and whack the stainless ball. The valve now works well without the O-ring.

Drain cocks

The cylinder drain cocks were steam powered and most stuck closed. The drain cock piston has an O-ring and the piston also pushes against an O-ring to seal the cylinder outlet. Testing showed that when "stuck closed" it's not the piston that jams but that when the piston moves back to uncover the drain hole the front O-ring also moves back and covers this hole.

The solution here was to replace the steam-powered drain cocks with manual ones. This required drilling a 4mm hole in each frame for the actuating rod. The first photo shows the jig used to guide the drill. The jig is located by a bolt in a frame stretcher hole, aligned by the square, and locked by the clamp. The second photo shows the crude levers and rods that link everything together.

Blast pipe and blower

The blast pipe didn't create a proper draught and the loco needed the blower on to keep the fire alive. The blast pipe is 80mm below the petticoat which is 40mm ID. To achieve the recommended 1 in 3 gradient from the blast pipe to the petticoat I fitted a 22mm high extension to the blast pipe. This improved the draught but the loco still struggled on the first test run. I reduced the blast pipe size from 6mm to 5mm and also cleared the chimney which was heavily coated with a sticky mess of ash and oil.

The blower is a piece of 1/8" tube pointing up the chimney. To improve its efficiency I fitted a restrictor to the end of the pipe. Without this most of the steam pressure is lost as the steam flows from the blower valve to the smokebox end of the pipe. The blower jet(s) area must be less than the area of the pipe to reduce this loss.

Boiler inspection

I rang Ben de Gabriel at Bathurst to ask for contact details for the boiler inspector at Orange. He gave me Roger Kershaw's number and when I spoke to Roger he said he remembered the loco and would post the boiler history. Sure enough, the required documents turned up a few days later. These documents record that the boiler was approved and I needed this record for a renewal inspection. I then rang Brian Day (our inspector) to request an inspection.

On 24th June 2017 I took the loco to Galston Valley Railway (Hornsby Model Engineers, my club) for the boiler inspection. The squeeze (cold hydraulic test) went for 20 minutes and no leaks were found. The steam test after took a while because the boiler was left full of water and again all was deemed okay.

I then ran the loco for a few hours. One job was to take a wandering lunatic back to the funny farm. The loco pulled well but I had trouble keeping the fire active and the steam pressure up. Firing certainly is a "little and often" affair on this loco. I'd like a longer regulator handle for better control. And maybe footpegs on the loco so I can stretch my legs.

Safety Valves

The safety valves opened a little early (about 70psig) and kept leaking even once the pressure dropped. I dismantled the valves and found that the stainless balls were badly encrusted. Replacing the balls and cleaning the seats helped here.

Before re-assembly I made and fitted some collars to keep the springs central on the spindles. Previously they could drift sideways a bit and apply uneven force the the spindles and seats,


Charlotte came with a riding car containing a coal tray and a water tank for the injector. This car is good because it has brakes but is overwidth (500mm) for Galston (my track) and the ride is hard because the bogies are unsprung. We had a spare tender which was unusable because the wheels were badly rusted and in fact some of the flanges were missing. So we removed the wheelsets and checked the ball races.

I ordered some wheels from Dinki Di Engineering in Adelaide. They advertise 80mm steel wheels with AALS profile flanges which is just what we needed here. The machining quality is excellent. Ian Thomas (Dinki Di) also offered axles and supplying the wheelsets complete but I decided to make my own axles. The difficult bit was machining the seats for the wheels to 0.6894" so the wheels would be a good press fit but still removable if required. Six of the eight seats turned out well, one was a bit tight, and one was loose and required Loctite to secure the wheel.

The tender has a hook coupling at the back and it's welded to the frame. Rather than cutting it out I left it and fitted an AALS standard clevis to the underside of the frame. The coupling pocket at the front was too high and the holes for the pin were too large. I enlarged the pocket hole and fitted sleeves to suit a standard 1/4" coupling pin. My brother made new footpegs to replace the 3/8" carriage bolts that were previously used and later made some coupling pins which gave him an excuse to try his new knurling tool.

For the initial test I sprayed the body with KBS Grey paint and sealed some leaks in the tank. The test run at Galston went well and the loco ran for hours until it was full of ash and most tubes were blocked. It steamed better than last time which was amazing because it had a steam leak in the smokebox which we discovered during cleaning the next day. Of course there is a slight scale difference here because the loco is 2.5" scale and the tender is 1.125" scale. Hopefully most people won't be bothered by this and the tender will look better when properly painted.


There was a leak at the steam header in the smokebox and the only way to to remove the header was to undo the bolts some of which were obscured by the petticoat. The chimney was secured by four 6BA bolts fitted from inside the smokebox. These bolts were perpendicular to the smokebox and angled in towards the petticoat. Removing these bolts (especially the back ones) was extremely difficult and refitting them would be even worse.

The solution here is to drill vertical holes in the chimney skirt and fit studs that extend down past the bottom of the petticoat. Spacers (not shown) are required above the nuts when fitting the chimney. These spacers have a slanted top to match the smokebox. To drill the vertical holes I machined a piece of rod so it had a point at one end (to align the chimney) and a drill protruding 9mm at the other end. Then I mounted the chimney on a vertical table on the lathe and aligned one of the angled holes using the point then drilled the new hole using the drill end. Finally I tapped the new hole. Repeat three times then fit the studs.

After enlarging and aligning the holes in the smokebox the modified chimney was positioned in place and the spacers and nuts fitted. The spacers are just short pieces of brass tube. Assembly is still a bit fiddly but it will be much easier to remove and refit the chimney if required. Notice the blast pipe extension and sophisticated blower arrangement in the second photo.

Test run 3

After fixing the steam leaks in the smokebox and doing a test steam I took Charlotte out to Galston for another test. It ran well hauling two carriages with three people for hours until finally all the tubes blocked. The tender holds enough water and coal for such an extended run. The loco does need an ejector to work the vacuum brakes in the carriages. This is high on the to do list. Another suggestion is an angled plate at the front of the grate to reduce the amount of ash that falls forward and blocks the tubes. This run has been the best so far and the loco is looking better all the time.

Grate assembly

The Bagnall boiler has a circular firebox with a removable grate assembly. The grate is higher than some of the firetubes and there is a vertical plate at the front of the grate to restrain the fire. But ash that gets over this plate blocks the firetubes. The change here is to fit an angle plate to the front plate. The original plate is badly burnt and the new plate should protect the old one and also (hopefully) reduce the amount of ash that gets carried forward and blocks the firetubes.

Brake ejector

The locomotive has a wind on handbrake and the original riding car has a good manual brake. But since the riding car has been replaced by a tender with no brakes all we have left is the wind on brake on the engine. This is not sufficient when hauling passenger carriages so the solution is to fit an ejector and piping so we can use the carriage brakes. All club carriages must have at least one bogie with vacuum brakes.

Since I replaced the steam-controlled drain cocks with mechanical cocks we had one valve spare on the turret in the cab. So I used this valve to supply the ejector which has been fitted under the footplate. This ejector has a check valve underneath and a pipe to the connector on the drag beam. The tender has piping from the front to the back buffer beam and also a leak valve which is used to release the vacuum. A proper setup on the engine would also have a vacuum pipe to the front buffer beam and also a connection for a vacuum gauge in the cab. These extras are a future job.

In the picture here the ejector and piping are partially obscured by the injector water valve and its piping. The ejector is just under the footplate and has its check valve underneath. Then we have the bent pipe to the connector on the drag beam.

To be continued

Last modified 2017-09-06