Taken in 1952, this picture shows part of  the Rolls-Royce Experimental Instrument Department workshop. The lathe in the immediate foreground, at the right-hand end of the bench, is the second Rolls Royce built by "JY "and "GJH" and to its left the first example, JJ's work, also with  8 mm collets. On the next bench is a pair of Pultra lathes - that on the left a standard 15/50 and on the right a capstan-equipped  15/50 fitted with the maker's raiser blocks. On the third bench is a standard Rivett 608 - in its distinctive "war" finish - with painted instead of polished headstock and tailstock stock castings -  and mounted on its original wooden underdrive cabinet stand. Beyond the Rivett is a long-bed South Bend Model A, with screwcutting gearbox and power cross feed, and the lathe at the end is a Mk. 1 Smart and Brown Model M manufactured circa 1947.

Continued:
Only one of the lathes is known to have survived, that shown in the colour pictures below and designated by R-R as their Type "GJH" - the appearance of these initials on a surviving general-arrangement drawing being in conflict with previous claims that it was known as the "CY". The lathe was tiny, but crafted like a larger machine with a noticeably robust headstock and carriage assembly. The overall length from collet handwheel to the tip of the leadscrew-feed handle was 25 inches, with the bed accounting for 19 of those. The centre height was 23/8" with 7" available between centres. A letter from Gordon Hudson revealed that the bed was machined from a length of case-hardened nickel steel that was ground and then lapped - with the headstock and tailstock cast from close-grained iron. To the writer's eye, the iron castings look to have been form-milled and hand scraped, in the manner used on Rivett lathes; however, bearing in mind the skill of the builders, it is entirely possible that the fine profile of these units could have been achieved by handwork alone. As a final touch, and to give an exemplary finish, the headstock and tailstock were satin-chrome plated.
Although later used in a modified form by the Japanese Toyo Company on their precision ML1 lathe, the bed and saddle design of the GJH was, at the time, probably unique - and very interesting. The top was flat and edged with narrow 45-degree V-shaped ways that guided the tailstock and aligned the headstock. The flat front and back faces of the bed sloped outwards (forming a trapezoidal shape) and it was on these, and the top surface, that the duplicated front and rear saddles ran. For such a small lathe the resulting contact area, at around 22 square inches, was enormous and, with the ways extending to the bottom edge of the bed, tremendous support was given to the cutting tool. On the underside of the bed the saddles (formed from castings with bolt-on top and bottom plates made to a dead fit), appear to have been located against an inner vertical surface by one face of the bronze feedscrew nut. The carriage was moved by a graduated handwheel at the tailstock end that drove a feedscrew running centrally down the bed and gripped by the bronze nut set immediately below the toolpost. Because of this construction, and doubtless very careful assembly and fitting, the carriage moved with the slightest touch on the handle, yet had absolutely no play.
Running in traditional, adjustable, hardened steel plain bearings - conical at the front with a steep taper and at the rear with a parallel bore but sitting in a tapered housing - the hardened, ground and lapped headstock spindle carried a threaded nose and accepted 8 mm draw-in collets. Simple and reliable wick-feed oilers provided lubrication. The 3-step pulley took a round belt and was arranged with its largest diameter to the left, with the outer flange pierced by a circle of 60 indexing holes; these were engaged by a particularly strong, spring-loaded pin that passed through the left-hand wall of the headstock.
Conventional in appearance, though beautifully finished in polished steel, the cross and top slides were broad and with sufficient gib-strip adjustment screws to ensure that the setting would could be absolutely precise over their full length. The cross slide had a travel of 2 inches, the top slide 1.5 inches with their feed screws made from hardened nickel steel. The 0.001" graduated micrometer collars were also hardened and could be locked by small, knurled-edged finger screws that ran through the centre of the handles. By causing the dial to travel axially (and not tending to tilt it over, or move it sideways, as conventional locking screw are prone to do), this arrangement avoiding any alteration in the setting. Even the balanced ball handles were special and, reflecting Rivett practice, were retained not by a crude taper pin but by screws that passed neatly down the shank of each ball-ended shaft. A form of "Norman Patent" toolpost was used, (identical in concept to that used for many years on the post-1924 Drummond M-Type) but much improved by Mr. Hudson. Like the ordinary "Norman" it could be both quickly altered in height and swivelled around its mounting boss - but with the added sophistication that the whole unit was self contained and could be instantly released and slid out of its T-slot holder. Releasing the lock-down handle enabled four coil springs to lift the assembly, the height setting then being finely adjustable by a knurled ring at the top. It's possible that, like many precision lathes, two tailstocks would have been provided, each with 1.5 inches of barrel travel: one took a Morse taper centre the other being adapted for 8 mm (draw-in) collets. Both tailstocks were designed so that their centres could meet that of the headstock, yet still leave sufficient room to use a range of slide-rest movements.
As an example of what the lathe could produce, a sample was retained of a minute bush: just 0.025" long with a diameter of 0.008" and a 0.004" bore. Still in existence, it requires a powerful magnifying glass to appreciate..

16th April, 1943 and a young Mr Hudson works at a bench in the foreground. Sitting at the desk is Mr. W.H.R. Clay ("CY"), designer of the first R-R lathe. The watchmaker's lathe in the middle of the photograph looks to be a pre-war Lorch

Possibly 1953 - and a more senior Mr Hudson at his desk in the strictly functional office used by the Experimental Instrument Department. On a shelf to the right can be seen a canon, used to conduct early tests of compressor blades. At the time, jet engines were being developed to use centrifugal instead of axial compressors but concerns were raised about the effect that a detached blade might have. Fearing that such a loss could cause the whole assembly to fail, it was decided to conduct some destructive testing and Mr. Hudson was asked to make a suitable cannon to fire an assortment of 2BA nuts and bolts into a compressor disc blade running at 29,000 r.p.m. Far from causing the blades to detach, the bolts simply embedded themselves in the disc material - and development proceeded. Of course, it would have been possible to make just a simple "test-rig" gun but, this being Rolls Royce, a proper job was done that both worked while also being aesthetically pleasing.

A form of "Norman Patent" toolpost was used, (identical in concept to that used for many years on the post-1924 Drummond M-Type) but much improved by Mr. Hudson. Like the ordinary "Norman" it could be both quickly altered in height and swivelled around its mounting boss - but with the added sophistication that the whole unit was self contained and could be instantly released and slid out of its T-slot holder. Releasing the lock-down handle enabled four coil springs to lift the assembly, the height setting then being finely adjustable by knurled ring at the top.