I thought I had done the Brunel ‘freaks’ to death and was thinking about what I might try making next. Then, I started reading the series of articles about Broad Gauge engines, published over several issues of the Locomotive Magazine. Starting with the January 1901 issue (Vol.6), a monthly series of article described in considerable detail ‘The BG Engines of the Great Western Ry’.

 

In fact, this series, which continued until February 1903, has been used ever since as a major source for many publications, including the ‘RCTS Part Two ’ (Broad Gauge) booklet and Brian Arman’s ongoing series on ‘BG Engines of the GWR ’

 

While there are errors in the series, some of which later authors have corrected in the light of further information being discovered, there are also snippets that have subsequently been ‘lost’. For example, I found that the ‘teeth’ on the large gear wheels of Harrison’s engine ‘Thunderer’ were actually strips of wood – quote: “The engine consisted of two cylinders, each 16-in. in diameter with a stroke of 20-in., driving a large fly-wheel which had broad wooden teeth, engaging in a pinion of similar construction, this latter being on the axle of one pair of the 6-ft. driving wheels.” This explains the size of the ‘drum’ which I reproduced in my own model but have not seen elsewhere.

 

My attention was also drawn to a mystery surrounding two engines ‘Snake’ and ‘Viper’ According to the Locomotive Magazine : “… two geared engines were delivered to the railway company by their builders, the Haigh Foundry Co. , of Wigan, in the following order :-”

 

 

The Magazine stated that very little was known about these engines. Later, in the same series of articles, the Magazine dealt with two other engines that it considered to be of unknown origin – ‘Teign ’ and ‘Exe ’. Of these, ‘Teign ’ became the subject of one of E.T.Lane’s beautiful drawings, made at Swindon and dated 11th November 1848.

 

 

Further examination of the records has led historians to determine that the above two pairs of names actually belong to the same two engines and hence, a rather complex picture of re-builds and transfers of ownership has emerged.

 

According to MacDermot/Clinker [1964] “ … correspondence during March - May 1839 shows that alterations were to be made, probably in 1839 and 1840 and both ‘Viper’ and ‘Snake’ appear to have been back in service by June 1843 and July 1844 respectively.” These alterations involved new cylinders and removal of the gearing.

 

Following the crisis caused by the failure of the Atmospheric system on the South Devon Railway, the GWR supplied steam locomotives to the SDR, starting in 1846. These appear to be the same two engines after re-building, when they were re-named ‘Exe’ and ‘Teign’, which were appropriate for the South Devon location. Their time in Devon must have been short, however, probably because their single-driver design made them unable to tackle the steep banks of the line

 

It is clear that ‘Teign’ alias ‘Viper’ must have been back at Swindon in 1849, when Lane made his drawing. Lane also made a dimensioned sketch of ‘Exe’, presumably at around the same time so, taken together, these illustrations provide sufficient information to construct a 3D model based on their appearance after the first re-building.

 

 

The original names were re-instated when they returned to service on the Windsor branch from Slough, where they are known to been allocated in 1850. By then, the two engines had been rebuilt again as saddle tanks and, fortuitously, ‘Snake’ was photographed in this form at Oxford (Grandpont) Station in about 1865.

 

‘Snake’ at Oxford c.1865

 

 

Planning a Model

 

While it would be fairly straightforward to create a model based on one of the rebuilt forms, I thought it would be interesting to try and re-create the layout of the original geared engine, assuming that the main external features of the engine were left unchanged. Thus, I started to search for whatever information I could find out about the original geared locomotive.

 

An obvious place to start was Daniel Gooch’s own ‘diary’. Unfortunately, this contains a collection of reminiscences, made long after the events to which they refer. It is now well known that Gooch’s memory was incorrect on many points and his description of these engines falls into this category. Gooch wrote: “The same plan of gearing was used in the two engines built by the Haigh Foundry ; their wheels were 6 feet diameter and the gearing 2 to 1, but the cylinders were small. I felt very uneasy about the working of these machines, feeling sure they would have enough to do to drive themselves along the road.”  Sadly, these ‘facts’ are contradicted by other evidence discovered more recently.

 

According to the revised edition of Mac Dermot/Clinker [1964] “… motion was communicated to the driving wheels by means of toothed gearing, … What they were originally like or how the gearing was arranged is not known. The cylinders are shown in the list prepared at the end of 1838 as 14¾ inch diameter by 18 inch stroke, the diameter of the driving wheels as 6 feet 4 inches, and the ratio of the gearing as 3 to 2.”

 

I believe these dimensions are more likely to be accurate than Gooch’s reminiscences from later years, since they are based on more recently discovered contemporary correspondence.

 

In a letter to the makers, dated 16th July 1838, Brunel stated that in the original drawing submitted to him the lowest point of the gearing was shown 17 inches above the rails, and that he was surprised to find it very different, the teeth of the driving wheel being within 10 inches of the rails, so that when the springs yielded this would probably be reduced to 8½ inches. For this reason he did not consider them safe.

 

A Directors' Report of the 12th August reported in the following terms : "The directors are under the necessity of declining to receive two engines made for them in consequence of a material variation in the plan of them since it was submitted to and approved by their engineer." In October 1838, however, they were accepted under limited conditions.

 

In this condition Whishaw [1842] quotes these engines as having the following dimensions : cylinders 14¾ -in. in diameter with a stroke of 18-in., diameter of driving wheels 6-ft. 4-in., and of leading and trailing wheels 3-ft. 6-in.; boiler barrel : length 9-ft. ; diameter 3-ft. 3-in. ; containing 111 tubes measuring 9-ft. 4-in. long with a diameter of 15/8-in. ; firebox 2-ft. 6-in. long by 3-ft. 6-in. wide by 3-ft. 5-in. deep ; heating surface : firebox 47.17 sq.ft., tubes 439.0 sq.ft., total 486.17 sq.ft.

 

Designing the gearing

 

Since we are told that the piston stroke was 18in, this set the minimum diameter for the primary gear wheel. We also know that the lowest point of this part of the mechanism was 10 inches above the rails and was sprung, with a potential vertical movement of about 3 inches.

 

In his book, Arman suggested that a chain drive might have been used but this would isolate the primary gear from the suspension so that vertical movement would not be expected. It seems more likely that the primary gear meshed directly with the driving wheel axle and moved with it during suspension movement.

 

With these constraints, I started to set out possible arrangements for the gearing. I found that there was sufficient space below the driving axle for the gears to be placed immediately below a pinion on this axle. Fiddling with the known dimensions, I arrived at the drawing shown below:

 

My conjectural arrangement of Haigh Foundry gearing

 

If my hypothesis is correct, the arrangement of the gears would have been very similar to that used by Harrison on his geared engines ‘Thunderer’ and ‘Hurricane’, in which the two axle boxes on each side were arranged vertically between a common pair of guides. Of course, in the present case, the primary gear had to be placed below the driving axle, because of the boiler, and this led to the ground-clearance problem noted by Brunel. It is possible that the gears took the form of elongated cylinders with wooden ‘teeth’, as in Harrison’s engines.

 

With this arrangement, it would not be a very difficult job to remove the gearing, as was done within a year of the initial delivery. Lane’s sketches of ‘Exe’ (no gearing) shows a four-eccentric type of valve gear that appears to have been similar to the Stephenson arrangement used on North Star. In a letter dated 25th October 1838, however, Brunel informed the builders that the Company had not yet succeeded in making the two engines 'perform the regular work of the trains, so as to be used at all for this purpose', and added that Gooch reported the regulator work so weak as to require renewing and that the valve gear had also to be altered. This suggests that a more basic form of valve gear was used on the original engines – perhaps of the Carmichael single-eccentric type as described in Clarke’s Railway machinery.

 

This choice is made more plausible because the Carmichael gear, with its characteristic X-shape forks, is visible in the drawing by E.T.Lane of the contemporary GWR engine ‘Ajax’. In addition, Gooch’s report to the Directors concerning the early engines made similar comments about the weakness of the valve gear on both ‘Ajax’ and ‘Viper’.

 

 

Carmichael’s single-eccentric Reverser

 

 

Constructing the chassis

 

I decided to start with the chassis. I created a pair of inside frames, based on those shown in Lane’s drawing of ‘Teign’ but with lengthened horns to carry both the upper and lower drive shafts. I assumed that the vertically-paired axle boxes on each side were linked by pillars, as was done on Harrison’s geared engines.

 

 

 

From now on, my model must be regarded as a purely conjectural reconstruction of ‘Viper’, as originally delivered from the Haigh Foundry in 1838.

 

The driving wheels had a very ‘spare’ appearance, when compared with other designs of the period and seemed to have carried domed polished ‘hub caps’. I made my model by my usual method of extruding the rim, hub, and one spoke as individual ‘bodies’, then using the ‘Circular Pattern’ tool to complete the wheel.

 

Modelling driving wheels for ‘Viper’

 

The carrying wheels followed by the same method. For the paired gear wheels I used the same methods as on my previous model of ‘Thunderer’ , starting by extruding a cylinder and then adding one tooth as a strip and repeating the pattern, using the ‘Circular Pattern’ tool, around the rim of the cylinder.  In my current model, the larger gear has 24 teeth and the smaller has 16, so providing the required 3:2 ratio.

 

I now had all the parts needed to create a ‘rolling’ chassis. In my illustration, I have removed one driving wheel so that the gears and driving cranks can be seen clearly.

 

My model of a geared chassis for 'Viper'

 

I realised at this point that the model has many similarities to my previous model of ‘Eagle’, so I created the ‘bicycle’-type splashers using similar methods to those used before.

 

Now it was time to re-create my conjectural ‘motion’ and valve gear. Since the crank axle was set below the driving wheel axle, it was clear that the cylinders would have to be inclined. The eccentrics for the valve gear would also have to be located on this low-set axle, with just one eccentric for each cylinder to operate the ‘Carmichael’ valve gear, described above. I placed the eccentrics outside the drive cranks.

 

I imported the drawing of the Carmichael gear, shown above, into ‘Fusion’ as a canvas and scaled it appropriately for my model. I then traced the main components – forks, valve rocker levers and slotted lifting link – and extruded them into bodies that I could then arrange in appropriate locations on my model. I adopted an iterative procedure to work towards a ‘best fit’ that provides a plausible arrangement for my conjectural model. One side of my 3D model of the gear is shown below:

 

My 3D model of Carmichael’s reversing gear

 

To reverse the engine, the operating lever moved the lifting link between the positions where the forked gabs engage with either the forward or reverse pins on the valve rocker.

 

I duplicated the gear for the opposite side of the engine and then added connecting rods and cross heads to drive the piston rods. To expedite creating these parts, I copied several items from my previous model of the running gear of ‘Eagle’.

 

I can now show the complete motion within the context of the rest of the chassis:

 

3D model of my conjectural geared motion for ‘Viper’

 

Boiler Assembly

 

I could now add the boiler assembly over this chassis. Again I copied several parts from my previous model of ‘Eagle’, modifying dimensions where necessary. The reason for the very small diameter of ‘Viper’s’ boiler (only 3’ 3” on the prototype) became clear, because it has to clear the upper gear-wheel.

 

Some of the boiler fittings were unusual – in particular the safety valve cover was a design apparently unique to the Haigh Foundry. I created my model by sketching the profile and then using the ‘Revolve’ tool to create the cylindrical body. I made one hole through the upper part of the body and the used the ‘Circular Pattern’ tool to create a ring of holes, so creating the ‘pepper-pot’ appearance of the prototype.

 

My 3D model of the Haigh Foundry safety-valve cover

 

The 1865 photograph (shown above) of the re-built ‘Snake’ shows wooden lagging around the large dome over the firebox, so I have replicated this feature on my model.

 

These additional fittings completed all the major features of my model, leaving one ‘sting in the tail’.  According to the Lane drawing of ‘Teign’ there were rather ornate rails on each side of the footplate. Fortunately I realised that ‘Fusion’ provides tools to facilitate creating such structures, including ‘Create Torus’ and ‘Cylinder’ tools. With the aid of these tools, I soon had a model of the frame:

 

My 3D model of Haigh Foundry handrail

 

With the addition of these rails, I can now show overall views of my conjectural model of the geared ‘Viper’. There are few visible signs of the gearing, which is largely hidden behind the driving wheels:

 

Two views of my conjectural model of ‘Viper’ as first delivered

 

For me, this has been a ‘fun’ project, in which I tried to re-think the methods by which the Haigh Foundry attempted to meet Brunel’s very restrictive design parameters for the original broad gauge engines of the GWR.

 

There is a reminiscence reported in the April 1901 issue of the Locomotive Magazine. A correspondent apparently wrote in August 1838: “I have just returned from witnessing the performance of an engine on the Great Western Railway, built by the Haigh Company, upon somewhat of a new principle, which combines what the writer deems to be essential to the perfectibility of the locomotive engine - namely, slower motion of piston with increased speed of engine. The experiment was completely successful, and, although Mr. Harrison has abandoned his plan, the principle of increased speed by the application of tooth and pinion gear is fully established by this experiment. The engine started from Paddington with five carriages, to Maidenhead, and returned with five carriages and two wagons loaded with iron, and frequently travelled at the rate of 40 miles an hour. The engine then took the five o'clock train with passengers, to Maidenhead , and performed the journey at the rate of 36 miles an hour with from 120 to 150 passengers.”

 

Mike