Now that I have built several of the pioneer engines of the broad gauge GWR, including some of the more successful of the so-called ‘Brunel’s freaks’, and have also teased out the complexities of early valve gears, I have decided it’s time to tackle one of the two most freakish of the ‘freaks’.

 

I do not expect this to be a ‘quick build’, as there are lots of unknowns regarding the layout of these locomotives, as well as many complex parts to be designed and fitted but, following Mikkel's dictum, I shall tackle my elephant one bite at a time. First – a bit of background:

 

These two engines, named ‘Hurricane’ and ‘Thunderer’ were initially hailed as revolutionary solutions to some of the problems of the time but, sadly, the initial euphoria soon evaporated. Nevertheless, a record of the weekly activities of the engines that were running in November 1838 shows that ‘Thunderer’ was actually making a substantial contribution to operations between Paddington and Maidenhead – the end of the line at that time. At that time ‘Hurricane’ had only just been delivered and was not yet in service.

 

 

The original report, published in the GWR Magazine, April 1910, is not very legible but I have re-typed the relevant columns alongside. This shows that, whereas ‘North Star’ made 22 trips (11 round trips between Paddington and Maidenhead), ‘Thunderer’ made 21, so it was apparently heading as many trains as the ‘best engine’ of the time. Admittedly, these trains were extremely short – an average of 4 carriages for ‘Thunderer’ and these would have been small 4-wheel carriages at that time.

 

After reading the above, I felt the fact that this strange engine saw regular use in the early stages – until the railway extended to Reading and then beyond – was sufficient to tempt me into creating a model. The following illustration is taken from Sekon’s ‘The Evolution of the Steam Locomotive’, 2nd ed., 1899, p.78. As a drawing made long after the event, I must view it with appropriate caution.

 

Sekon’s drawing of ‘Thunderer’, published 1899

 

Once a tender has been added it’s easy to see why some people called it a ‘train in itself’ and, with a serious lack of adhesion weight, it had no chance of starting more than a very light additional train.

 

The illustration above shows that the ‘motion’ was fully visible and so must be featured in a model. For more information, there is a detailed description of the very visible ‘motion’ in Wood’s ‘A Practical Treatise on Rail-Roads’, 3rd ed, 1838. Note that, although Sekon’s caption above states the engine had 6 foot diameter driving wheels, Wood in his lengthy description of the engine states that they were 5 foot diameter, implying an effective 15 feet, as a result of 3:1 gearing. Since Wood was writing when the engine was new, I have more confidence in his description than Sekon’s.

 

 

Wood’s Figure 7, published 1838

 

In a detailed explanation of his Figure 7, above, Wood (1838) explains that the axle carrying the gear wheels was not fixed to the iron framing but could slide up and down with any movements of the axle of the driving wheels below. The two axle boxes were held apart by an upright pillar, supporting the upper axle box while resting on the top of the lower box. An iron strap, ran around both axle boxes to draw them together by driving in a tapered key. Thus, the teeth of the two gear wheel were kept in mesh during any movement of the wheels on their springs.

 

To prevent any play in the teeth, when the motion was reversed, the upper gear wheel was split into two halves, each with its own row of teeth. Only one half of the gear wheel was fixed to the axle while the other could be adjust by means of tapered keys. The offset could be adjusted to take up any back-lash as the teeth became worn during use. This arrangement of the gear wheel is illustrated in Figures 10, 11, and 12 of Plate XIII, shown below:

 

Wood’s drawings of the large gear wheel

 

Surprisingly, in view of the other innovative features, the illustration by Wood shows a very primitive, manually-operated reversing mechanism. From Wood’s description, the slide valves were worked by one eccentric for each cylinder, working upon the upper geared axle

 

I have written an account of this mechanism in an article published in the Broad Gauge Society (BGS) journal ‘Broadsheet’, No 78 (Spring 2018), which I summarise below.

 

The valve rods were attached to a rocking lever, mounted on a transverse shaft known as a 'weigh bar', which conveyed the movement from the slip eccentric. Slip-eccentrics allow the eccentric disk to 'slip' freely on the axle, for about half a turn of the wheels., determined by two ‘stops’.

 

My annotated sketch of ‘Thunderer’ valve gear

 

To reverse the running of the engine, the rod from the eccentrics had first to be disengaged from the rocking lever by means of the release lever.

 

The valves could then be adjusted manually into the correct positions to admit steam for reverse movement so that, when the pistons started to move and rotate the wheel in the reverse direction, the slip-eccentric rotated automatically to its reverse position on the drive-shaft. The release lever was then moved up, to re-engage the gab (notch) in the eccentric rod with the pin on the rocking lever, so that the engine continued to run in the new direction.

 

This mode of operation needed considerable manual intervention, which was made somewhat easier on this engine, since the parts were readily accessible to the engineman.

 

The cylinders and valves were conventional but, because the cylinders were mounted o a separate carriage from the boiler, flexible pipes were needed to carry steam to and from each cylinder. This was a severe technical challenge at the time and the design of these pipes was the subject of two more Figures in Wood’s book. I have annotated his Figure 8, below, to show the general arrangement of the connections between the ‘engine carriage’ and the ‘boiler carriage’:

 

My annotated extract from Wood (1838) Plate XIII

 

Wood provides a detailed explanation of the construction of the two flexible pipes that carried steam from the boiler carriage to the cylinders. He wrote that “the two carriages are fastened together by the bar … but, to compensate for the motion between the two carriages, a peculiar kind of joint is used for the steam pipe and the discharging pipe, Fig.9 shews the form of joint used, which is a hollow universal joint”

 

 

Short pipes ending in the outer parts of spherical joints were fitted by flanges to each of the two carriages. The inner parts of the spherical joints led into two pipes, one of larger diameter than the other, These pipe could slide over one another, to compensate for any relative horizontal motion of the carriages, with the space between them filled with hemp packing. A gland was screwed upon this packing to keep the joint steam-tight, while allowing horizontal movement when necessary.

 

To allow circular motion, the spherical ends of the pipes had vertical flanges, to which were screwed rings, leaving a space of about two inches in between. Metallic 'ring packing was laid in this space, which was screwed down to form a steam-tight joint, while allowing motion in any direction.

 

I should mention another drawing of ‘Thunderer’ which appeared in Colburn’s ‘Locomotive Engineering’ p.48 published in 1871. Thus drawing diverges considerably from the description given by Wood, most notable in the layout of the steam pipes between the engine and boiler carriages.

 

Colburn’s drawing of ‘Thunderer’, published 1871

 

Of course, it is possible that alterations were made after Wood’s 1838 account, since the engine continued to be used for experimental purposes until December 1839. It may indicate that the regulator was moved to the engine carriage at some point. I have decided to discount this version overall, although the boiler fittings look much more plausible for the period than Sekon’s ‘Dean’ type of safety valve cover.

 

First Steps to a Model

 

Having worked through the various peculiarities of the design, I now had to decide how to capture the external appearance of these workings in a 3D model. The boiler carriage and tender were conventional so should not need any ‘difficult’ treatment’ but where to start with the rest … ?

 

I have not found any plan view of this engine, so my first step was to work out a possible layout for the various components across the width of the engine carriage. For this, I needed to determine a few dimensions. Sekon (followed by many others) stated the driving wheel diameter to have been 6’ but the scale attached to Wood’s Plate XIII clearly indicates a wheel diameter of 5 feet and an overall length of the engine carriage as 12 feet. I have decided to adopt Wood’s figure because his is a contemporary reference. This means that the scale on the drawing reproduced in Arman’s ‘Broad gauge Engines of the GWR, Part One’, p.55 is incorrect.

 

After that, I thought I should do some ‘tooth counting’ and again the result differed from the received wisdom. I feel less confident about this, as the drawing may well have been stylised in this respect – who knows?  In addition, engines at that time rarely corresponded to their drawings.

 

For what it’s worth, the calculation from this drawing indicates a gearing ratio of 74:22 or 3.3 : 1, which doesn’t seem to agree with anyone else!

 

Colburn and Ahrons both state 3:1, whereas the RCTS Part two states 27:10. All these sources give the driving wheel diameter as 6 feet, not 5 feet, which I am using.

 

It may be a coincidence but the RCTS ratio of 2.7:1 with 6’ wheels corresponds quite closely to my 3.3:1 with 5’ wheels, both yielding an ‘equivalent’ wheel diameter of about 16 feet.

 

 

 

 

 

 

Counting Teeth

 

 

 

When I imported Wood’s Plate XIII into the 3D modelling software ‘Fusion’ as a ‘canvas’, the story took another twist – I measured the diameters of the two gear wheels and found that these were in a ratio of exactly 3:1. It seems that, rather sloppily, someone just looked at the diameters, without considering the numbers of teeth.

 

After assuming a wheel diameter of 5 feet, ‘Fusion’ indicated that the overall length of the engine carriage was exactly 12 feet. I could now start laying out my model in earnest, assuming a length of 12 feet and a width of 8 feet, across the frames.

 

 

 

 

 

 

I decided that creating a 3D model of the gear wheel would be a good place to start. One reason for this is that Wood’s Plate XIII includes both side and end views of the large gear wheel which, as explained above, was actually in two parts. I decided to fabricate my model wheel from several individual components before bringing these together to create the overall assembly.

 

As usual, my method was to extrude 3D ‘bodies’ from Wood’s Plate XIII, which I imported into Autodesk Fusion as a ‘canvas’. Initially, I extruded a simple cylinder on which I drew and cut out a single slot to start the process of cutting all 74 teeth.

 

 

 

From that single slot, I could use the 'Circular Pattern' tool in ‘Fusion’ to repeat the first slot all around the circumference, so producing 74 teeth. Next, I split the toothed cylinder cross-wise into two bodies and placed a smooth (no Teeth) section between them, after which I could ‘Combine’ them all into a single cylindrical body. I drew a circle on one end of the cylinder and ‘Pushed’ this through to create a hollow centre into which the spokes could be fitted. I also chamfered the inner edges at each end of the cylinders, as indicated on Wood’s drawing. The result at this stage is shown below:

 

I then went back to the canvas and extruded the pattern of the spokes from Plate XIII. Again, I drew a single spoke in detail and used the Circular Pattern tool to create six evenly spaced spokes. This resulted in a ‘spider-like’ body as shown below;

 

 

I made a copy of this part and fitted two sets of spokes inside the hollow centre of the toothed cylinder assembly. Now I had a complete structure that visibly represents the gear wheel shown on Wood’s Plate.

 

 

Note that I have not (and do not intend to) replicated the anti-backlash adjustment, which slightly offset the teeth on one half of the gear wheel relative to the other.

 

Seeing this wheel in place over Wood’s Plate reveals what a large structure it actually was, which was not apparent from side elevation drawings alone.

 

Well, that has made a start – one bite – with a very long way to go but it’s good to feel something is under way …

 

Mike