Guy Rixon

Side knees: what Spitfire said. Note also that the knees are substantial bits of forged metal, unlike the thin washer-plates on the outside. They would be 2 to 3 inches thick at the bend (where they disappear into the floor) and maybe an inch thick at the top. Below floor-level, they are bent through 90 degrees and have a leg pointing towards the centre of the wagon. These legs are bolted to the middle bearers and the floor planks are rebated on the underside to accommodate them. Nothing of the horizontal part shows above floor level. I shall experiment with the weathered red. I have a wagon to do in SER red that could be finished that way.

Can you remind us how the weathering was done, please? I ask because the weathered colour in the photos is looking much, much better than the basic primer-colour in the previous pics.

Nice that they're getting built at last. That's about £70-worth of kits at current (eBay) prices! If one is using the floor that comes with the kit, it rides a bit high if the headstocks are shimmed. I chose to reduce the height of the solebars. But your floor is thinner, so the shims are right. Will you simulate the side knees on the inside? It's trivial to do at the stage shown in your last photo.

LSWR axleboxes are back on sale. I have thickened the problematic detail along the axis parallel to the axle; I think this will not be very visible. I have not test-printed this version, so we still don't know if Shapeways will print it. I will try a test in my next batch of prints. The new prices for FUD and FXD are now in force. My buffers and axleboxes have come down in price by 35-40%.

The buffers and axleboxes in my shop are down 35-40%.

And the carriage (MSLR tri-composite) is now attractively priced and I shall get one in my next order. Thanks!

The advice about applying the solder directly to the joint is really important when using cored solder without liquid flux. If you carry the solder to the joint on the bit, then the flux from the core is gone, or already reacted, by the time the solder reaches the joint. If you put liquid flux on the joint first, then carrying the solder on the bit does work. Holding the work in one hand is not ideal. I find that if I jig an assembly (such that I could wield solder in one hand and the iron in the other, but I don't), then I get better joints. There exist devices that clamp to a soldering iron and dispense solder from a reel as the iron is applied. I saw one used in a demonstration but have never seen the like on sale.

More cylinders allow better balancing of the motion and less hammer-blow to the track. Hammer-blow increases significantly the load that a locomotive puts on the track and, from c. 1900 if not earlier, designers were keen to reduce it. Hence, Churchward designed the 4-series "Star" class for the fastest trains, even though most of his standard designs had two cylinders.

The wagon with rounded ends on the second standard-gauge road from the left possibly belongs to the LSWR. The LCDR had wagons with a similar end-profile, but those had a framing member across the end level with the top of the sides.

SER CCT of 1870 on HMRS site: http://www.hmrs.org.uk/drawings/drawinginfo.php?id=22395. HMRS 5482 looks like it might be a later drawing of the same vehicle.

Nickel silver is definitely easier. Ordinary brass quickly forms mucky oxides and these stop the solder from bonding. I've had cases where the joint mucked up before the solder melted! Nickel silver is less prone to this. Using the right flux helps. Years ago, I started building stock in 2FS where the wagon chassis are usually etched in nickel silver. I used cored solder and Carr's Red flux as that was all I had available. For the NS, this turned out to work really well. Later, I moved back to brass kits in 4mm scale and found that the same solder/flux combination was a bit pants. For ordinary brass, I think one needs a stronger flux. Carr's green works well for me. One of the problems with lead-free solder is that the replacements for 145-degree solder melt at a higher temperature and the leaded version and are not so good at filling gaps. If one gets a close fit in the metal to be joined then it works; if one tried to bodge it across gap then it won't play.

The C & W wheels are, IMHO, excellent, but they are hard to assemble into wheelsets unless one has the Exactoscale back-2back gauge; and those have been out of stock for ages. I predict that the wheels would sell much better if the gauges were made available again.

Southampton had the well-known lines from Canute Road to the town pier (also, earlier, serving the royal pier) and on into the Western docks, as noted above. As a child, I remember watching diesel shunters (07s?) trundling along this stretch. It also had some lesser-known tramways from the main line to the quays and small docks on the river Itchen; these were upstream from the Eastern docks. If memory serves, there was one that ran along Chapel Road and another, further north that ran to the quays near Belvedere Roads. These, apparently, were still running up to the early '80s.

I do, but increasingly I find it's too blunt an instrument for the fine work. Haven't yet tried whittling the stick to a finer point. I'm going to try the dispenser needles in the post above; got to be worth £2.85 to try them out.

Between 20 and 30, I would guess. Microsoft are reputed to pay their developers well, and I would be surprised if total cost of employment was below £100,000 p.a. Back-porting patches to an "obsolete" OS is not going to be a popular job within their company, and can't safely be done by inexperienced or second-rate staff, so (a) they may need to offer incentives, (b) they need a decent proportion of senior developers and © they need a larger testing and QA contingent than might otherwise be expected for the volume of software. 20 staff is possibly more than they need but not ridiculously more. I'd guess that a team of 10 might be about right, including management, QA and somebody looking after the internal paperwork. So yes, they're taking a large profit, but not, I believe, in the millions of percent. You can't do this work with a couple of interns and a dog. Also, £5.5 million is a tiny amount compared with the NHS budget (and with Microsoft's turnover). Choosing not to spend that extra money on the NHS computers was a ridiculously bad decision, whoever made it.

Organizations running Windows XP are able to get some continuing support (= security patches) by paying MicroSoft for a special contract. After the end of general support for XP in 2014, the NHS had such a contract. The price goes up each year. When the renewal cost reached £5.5 million, the support was not renewed. I don't know whether this was decided inside the NHS or by the treasury.

In fact, main-line 6-wheelers generally had softer springs on the central axle. One GER example I measured recently has: J-hangers and long suspension links for the central axle and plain spring-shackles for the outer two; Spencer's patent rubber pads as secondary springing on the centre axle only; nine spring leaves on the outer-axle springs but only eight on the centre-axle springs. When H. F. Stevens started buying stock, c. 1900, the 4-wheeled coaches of the larger companies would have been mainly in two classes: ancient and worn out, or recently renewed and not for sale. The stock available for sale and in good condition would mainly have been mainly 6-wheeled. He could have had a job lot of District coaches after 1905 but perhaps they were too far gone.

I suggest that the distant would be under the outer home signal of the terminus, not on the platform starters, and the the terminus would have an outer home somewhere in the covered way so that trains could shunt out of the platforms while remaining within station limits. However, if the branch is only a quarter mile long, that puts the outer home nearly halfway to the junction, so I wonder if the station limits of the terminus and junction could actually coincide; and, if so, whether there would be any need for a distant? Note that this branch is actually shorter than the normal distance from a home signal to that signal's clearing point!

Why not do it the other way round? Buy a cheap kit for a narrow gauge coach and rebuild it for standard gauge. It would have to be one of the large loading-gauge kinds, not an FR bug-box.

Not sure what you mean by "two rows of vents". Did you mean the roof-mounted ventilators (e.g. Laycock a.k.a. "torpedo" style)? They weren't specific to gas-lit coaches, oil-lit and electrically-lit coaches had them too. If the Hornby moulding doesn't have roof vents (can't see any in your photos) they're easy to add. +1 for upper footboards, specially on a light railway where platforms might be a bit low. BoT would probably insist. If you want to try a guard's compartment, prints of side lookouts are available on Shapeways.

Controllers with variable pulse-width tend to break coreless motors if the pulse frequency is low. The motors slow down and nearly stop after each pulse, then spin up again with the next pulse. Starting and stopping wears the motors out very quickly. Controllers based on variable pulse height, or with a high pulse frequency (or both) should be OK.

The brake rodding on the chassis makes no sense unless there are handbrake levers. As moulded, it seems to be derived from the LNER vacuum-brake arrangement, where there was a lever on each side. I suggest scrapping the moulded brakes and adding a single pair of V-hangers with a centre crank on the brake shaft. There should be steps on the coach end, so that the railwaymen can tend to the gas lamps. A gas-lit coach would need a lighting-control bar on the end opposite the steps, and pipes running up the end to the roof. Alternatively, you could remove the moulded gas lamps and fit oil-lamp tops. If the coach is to represent a vehicle running after 1889, you'll need brake connections, vacuum or air.

Any surface in contact with the support wax gets a poorer finish. Usually, you'd want you loco bodies printed right-way up so that the good finish was on the most-visible parts. If you print it with smokebox down, then the whole of the boiler and smokebox and running plate is in the wax to support the overhang of the cab. Your cab fittings will print better but all the rest suffers. If you print it smokebox up, then the inside of the cab is in the wax. There would be some support material next to the boiler, supporting the smokebox where it overhangs the boiler cladding, and supporting the chimney and the dome and the safety valves. The smokebox would be mainly clean, except aft of the chimney. Running plate and any side tanks should be cleanish, but any detail on top of tanks would entail support on the tank top behind that detail. This orientation would probably add a lot of wax (=cost) to the print and it's tall so the machine-space charges would be high. I would not go this way; I think it would cost more and give a poorer result. IMHO, I think your only lower-cost route for printing steam-engine bodies in FUD is to do them as kits.

Detection bars lay normally down, I understand, and were raised by the point rodding when moving the points. When a train wheel was above a bar, then that bar could not rise, thereby locking the points in their current position. Treadles, IIUC, normally lay in the raised position and were actively pushed down by a wheel flange. A treadle might, for example, be made to sound a gong to warn shunters of an approaching train. My hazy understanding is that a turnout would be locked by a detection bar if it could be changed under a running move of a train. That, I think means if a train could be moving over the turnout after relevant signals have been put back to danger and the locking freed to move the points. E.g., consider a train running into a platform and passing over points in the station throat and later the engine-release cross-over. While the train is approaching the station, the points are locked by the home signal governing the approach. When the engine has passed that signal, the signaller can put it back to danger, and that, absent the detection bars, unlocks the points. The bars keep the train safe. Another use for detection bars is to lock signals. Once a train is stopped within station limits - e.g. standing at a platform - and the signal that allowed it in has been returned to danger, nothing mechanical stops a signaller from clearing that signal to allow another train onto the same track. This is a significant risk where the view of the line from the signal box is poor. A detection bar locking the signal mitigates the risk. To be fully effective, a detection bar has to be at least as long as the greatest gap between wheels in a train - i.e. the distance between the inner wheels of the two bogies of a coach. Since your station is approached through covered way, it might well have had a treadle-operated gong to warn staff of trains. The gong itself would be visible on the layout, but the treadle would be out of sight on the approach. I think you need detection bars on the throat point-work leading to and from your platform roads, and on your engine-release cross-over; these protect against point movements under the train. You probably don't need detection bars to tell the signallers which platforms are occupied as they can see that well from the box. If trains ever shunt onto the running lines in the covered way they may be hard to see from the box and then you should have detection bars to remind the signallers. One should also consider when the full suite of safety measures came into force. Bear in mind that train brakes only became mandatory in the late 1880s. Interlocking of points and signals started in the 1860s, but was not applied everywhere until much later. EDIT: while googling for some images, I discovered that the proper term for what I described above might be "fouling bar" rather than "detection bar". Whatever the name, you still need to have them. There is a rather different device called a "facing-point-lock detector" (google it for a picture) that connects the signalbox locking to the position of a FPL. "Detection bar" might be another name for one of these. FPL detectors also became mandatory on facing points in passenger lines, but possible not in 1880.

Your biggest problem will be the cost of the support material inside your shells. Work out the volume inside and add $0.50 per cubic centimetre to the current cost. Then knock off $2.50 for the drop in per-print charge. Sounds like your prices would go up a bit.