George Armstrong's Masterpiece

[MikeOxon]

In another thread,  I explored some of the earliest standard-gauge locomotive to run on the Great Western Railway.  The first of these were from the two Shrewsbury railways, which were absorbed by the GWR in 1854.  Not only did the GWR acquire a very miscellaneous collection of locomotives from these railways but they also gained a locomotive engineer by the name of Joseph Armstrong.  

Joseph Armstrong soon became responsible for all standard-gauge locomotives running north of Oxford, but he had little opportunity to design and build his own engines until the resignation of Daniel Gooch, ten years later in 1864.  Then Joseph moved to Swindon, to take on the immense burden of transforming the GWR from a broad-gauge to a standard-gauge railway.  When he moved, however, he left his younger brother, George Armstrong, in charge at Wolverhampton, with a high degree of independence, which included retaining the blue-green livery, originally used by the OW&WR.

Thus, it was George who took on the task of renewing the locomotive stock in the Northern Division of the GWR.  Initially, his designs followed the ideas of his brother very closely, but then he designed a small tank engine with an 0-4-2 wheel arrangement, which was to achieve a fame and longevity out of all proportion to its size.  If I mention that the first of these locomotives, built in 1868, was No.1040, it will probably ring few bells in anyone's mind but, if I mention that the first sixty were re-numbered in 1870 into a sequence starting at No.517, then these will be recognised as some of the true 'greats' in locomotive history - a greatness based not on 'flashy' deeds but on sheer competence and overall utility.  

For the modeller, the '517-class' presents both opportunities and problems.  Over the years, the class received so many modifications that it would be possible to build a diverse collection of models based solely on this class!  On the other hand, there were so many re-builds that applying a running number to a model is fraught with dangers, since the result could bear very little resemblance to any of the many incarnation of its prototype.

A few years ago, I made a rather simplistic conversion of a K's 'autotank' into a loose representation of a '517-class' locomotive. The original model had a Belpaire-type firebox and I wanted to include inside trailing-wheel bearings, so I hit on a photo of No.835, which showed both these features on the same engine.  More recently, however, I've started to look at how the prototype engines were developed over the years.  Although, there is much detailed information in the RCTS Part Six of 'The Locomotives of the GWR', it is quite difficult to extract the information most relevant to modellers, when it is spread over 12 pages!

For my own interest and, hopefully, of some use to others, I have made the following summary, starting with the original forms of the various stages of construction.  I have concentrated on visually obvious changes, such as wheelbase and tanks, rather then on the lot numbers.  I based my coloured illustrations of the different stages on original photographs. EDIT Since reading 'Great Western Way' more carefully, I have changed the splasher fronts on Wolverhampton engines to green, rather than brown, as originally posted.

Constant features

A few things remained constant throughout :  the mainframes, the layout of cylinders and motion, the principal boiler dimensions, and the coupled wheelbase : 7' 4" throughout.  The wheel diameters were 5' 0" and 3' 6" for the coupled and trailing wheels respectively, although both these dimensions grew by 2" over the years, as a result of fitting thicker tyres.

First Group - built between 1868 and 1870

The first 60 engines were saddle tanks, except for the final six, which had side tanks from new.  They were numbered from 1040 to 1087 and from 1100 to 1111 until July 1870, when they became Nos. 517 to 576.  (I'll use the later numbers in the following descriptions of the changes)

Even within this first group, the overall wheelbase started to increase; first from 13' 7" (Nos. 517 - 522), to 13' 8" (Nos. 523 - 552), and then a jump to 14' 8" (Nos. 553 - 576, which includes the first engines to be built with side tanks : Nos. 571 - 576)


No.517 in original form, with short 13' 7" wheelbase
shown in Wolverhampton livery


No.556 in original form, with longer 14' 8" wheelbase
also in Wolverhampton livery

Second Group - built between 1873 and 1885

This was the largest group, comprising 90 engines, mostly turned out in batches of a dozen at a stretch.  Numbers lay in various ranges: 826 - 849 (24 engines), 1154 - 1165 (12), 202 - 205 and 215 - 222 (12), 1421 - 1444 (24), and 1465 - 1482 (18).   All these engines had side tanks and a common overall wheelbase of 15 feet (7' 4" + 7' 8")  Other developments in boiler and tank details occurred during the build period.


No.1164 with side tanks and 15' wheelbase
shown in Wolverhampton livery

Third Group - built 1885

This was a small group of six engines, Nos. 1483 - 1488, which were the first to be built with outside axle-boxes to the trailing wheels, together with a further increase in the overall wheelbase to 15' 6".


No.1487 as built, with outside bearings and 15' 6" wheelbase
shown in Wolverhampton livery

Continuous Improvement

Set out like that, it all looks fairly simple but then the modifications and re-builds began, with all the improvements made in the later groups being applied to earlier engines.

The first step was the conversion of saddle tanks to side tanks, which was (mostly) accompanied by increasing the wheelbase to 15 feet. (Occasionally the two 'operations' were staggered).  This applied to almost all of the first group between 1876 and 1886, with most being converted at Wolverhampton, although four were done at Swindon.  I guess that the Swindon conversion may have resulted in a livery change to the Swindon style - they certainly acquired polished brass dome covers.

 

No.551 as re-built at Swindon, with outside bearings and 15' 6" wheelbase
shown in Swindon livery

By the end of the 19th century, there were only two wheelbases, 15 feet and 15' 6", with all the latter having outside bearings for the trailing wheels.  Reconstruction continued at Swindon until 1915, providing new frames and the longer wheelbase, with outside bearings.  Meanwhile, at Wolverhampton, lots of changes were also being made, to improve bunker capacity, for example.  This involved lengthening the frames by 6 inches, without any change to wheelbase or bearings.  The most noticeable differences, below the footplate, between Swindon and Wolverhampton engines were in the depth of the valances and the shape of the footsteps.

Above the footplate, cabs and bunker designs were constantly evolving.  The variations are really too complex for a brief summary here so, I fear, one is forced to look at information for each individual locomotive - much of which can be found by carefully scouring the pages of the RCTS Part Six of 'The Locomotives of the GWR'

In the case of my own model of No.835, I now know that this started out in December 1873 as a side tank with 15' wheelbase.  It was re-boilered in 1894, again in 1910, and then received a Belpaire boiler in 1916.  For some reason, it retained its original inside bearings and wheelbase throughout, being finally withdrawn in 1935.
 

No.835 after receiving a Belpaire firebox in 1916

Shown with black frames (post 1906)

So, my model is quite wrong in combining Indian Red frames with a Belpaire boiler and it certainly should not have wing plates on the smokebox. I still enjoy it as a model, though, and perhaps one day, I shall be inspired to build a more authentic version ....... but, which to choose? 
 

[Mikkel]

Interesting summary, Mike, thanks for that. It would be interesting to do the second group type. And not too difficult, perhaps?

[MikeOxon]

...................It would be interesting to do the second group type. And not too difficult, perhaps?

I agree that the second group represent the 'quintessential' 517-class : side tanks and inside bearings on the trailing wheels. 

 

Remember too that the earlier saddle tanks had all been converted to side tanks by 1895 (just No.544 continued with a saddle until 1899)

 

If you are thinking in terms of converting a Collett 14XX model, then it is not too difficult to get a 'fair' representation but there are some serious hurdles if accuracy is required.  Those which affect the basic structure include the change from a wrapper-type smoke box to the drum-type, used by Collett, who also raised the boiler pitch by around 6".  The 14XX have the longer 15' 6"  wheelbase.

 

I have summarised the main dimensional differences in the following table:

 

 

Mike

[buffalo]

In addition to the information in the RCTS volume, no one modelling a 517 in any of its many forms should be without copies of GWRJ 74 and 75. John Copsey's article covers more than forty pages and provides numerous photos.

 

Nick

[pointstaken]

They must have been tough old blokes in those days - 1164 in second group photo only has a spectacle plate, third group has almost a proper cab.

 

Dennis

[Welchester]

I'm sure I've seen a photo of a Metro Tank with only a spectacle plate taken in the mid-1930s.

[buffalo]

I'm sure I've seen a photo of a Metro Tank with only a spectacle plate taken in the mid-1930s.

 

That would have been a late survivor of the ones with condensing apparatus, most of which only had spectacle plates.

 

Nick

[MikeOxon]

I'm sure I've seen a photo of a Metro Tank with only a spectacle plate taken in the mid-1930s.

According to John Gibson,where a photo appears in his book 'Great Western Locomotive Design'. these engines, fitted with condensing gear were nick-named 'Get Wets'.  He suggests that sooty drips off tunnel roofs were possibly more injurious than clean country rain!

 

Mike

[MikeOxon]

They must have been tough old blokes in those days - 1164 in second group photo only has a spectacle plate, third group has almost a proper cab.

 

Dennis

According to Harry Holcroft in 'GW Locomotive Practice', the Armstrongs held that, in a confined space, the products of combustion ... were more injurious to health than the weather could be.  This was probably true in the days when coke was used as fuel, creating carbon monoxide as a source of drowsiness and potential death. 

 

The Armstrongs had been drivers in their younger days and were therefore implementing what they themselves were used to.

 

Mike

[MikeOxon]

Since posting my paintings of the early engines, I have been reading 'Great Western Way' more carefully and, accordingly, I have changed the splasher fronts on Wolverhampton engines to green, rather than brown, as originally posted.

 

Mike

[Mikkel]

I didn't realise you had done the colouring. Very useful, thanks Mike.

[MikeOxon]

I didn't realise you had done the colouring. Very useful, thanks Mike.

I feel that it is easy to forget that 19th-century railways were often very colourful, when we only look at the old B/W photos.  Adding colour gives a truer impression and I find it inspiring for modelling 

 

Mike

[Quarryscapes]

 

Ah yes the 517s. Above is as far as I've got with my 3D printing of some. So many things differ, even with a kit of a few standard bits you can create so much variation, as shown above, all the parts are colour coded to track them across the different locos! 

 

None of this takes into consideration the several different boilers that were fitted which has an impact not only on the firebox, but also dome position. Add to that further variation in smokeboxes and you can get almost infinite variety.

 

Lets see, so in terms of what may change you can alter:

 

Wheelbase

wheel diameter

running plate shape and width

valance shape and width

rear buffer beam

bunkers

tanks (not just saddle or side but size of side tanks!) 

firbox

smokebox

dome position

chimneys

smokebox doors

wing plates

sand boxes

toolboxes

cab profile

roof profile

push pull equipment

 

I may have missed a few....

 

I can highly recommend the GW Journals noted above, so many high quality photos illustrating loads of potential variations. I must try and get hold of the RCTS book too. 

 

[MikeOxon]

 

............... So many things differ .........................

 

Indeed!  On the same theme, I made a couple of overlays of 517 and 14XX drawings, which bring out just a few of the changes quite well.

 

My first illustration overlays a 14XX on a 517 drawing claimed by Russell (Pictorial Record vol.1) to be Official Diagram B, although it doesn't agree with his description of 15' 6" wheelbase (the drawing shows 15' 0"). It has the shorter frames, typical of earlier engines.  This 517 looks very small in comparison to the 14XX, mainly due to the short cab and bunker..

 

(517 in red)

 

My second illustration is of a late 517, after many re-builds, based, in this case, on a Col. Templar drawing, also in Russell.  Here, the two types are relatively similar.

 

(517 in red)

 

Mike

[CKPR]

There was a series of articles in Model Railways in 1980 (IIRC) by Iain Rice on building various 517 variants from the Mallard kit.

[ChrisN]

Thank you I have found this very interesting and useful.  I may well want to build a 517 at some point so something like this is a good starting point, (my main research has been Cambrian.)

 

How do you obtain the GWR Journals, or do you have to be a member of a society?

[MikeOxon]

You can get back issues from this website: http://titfield.co.uk/WildSwan/WSR_GWRJ.htm

 

Mike

[ChrisN]

It's a Wild Swan publication. They have no website, but you can find back numbers on their stand, at various book dealers and on eBay.

 

 

You can get back issues from this website: http://titfield.co.uk/WildSwan/WSR_GWRJ.htm

 

Mike

 

Simon and Mike,

Thank you both very much.  I shall look at that website when I am back from work later this week and if that fails I will put another search on eBay.

[Quarryscapes]

I got both mine from ebay. 

 

Looking through the journal, it would actually be fairly easy to do an RTR 517. The vast majority of them Ended up with the same frames, footplate, tanks and cab front, that would only leave the need for a range of smokeboxes, boiler tops and bunkers. Now who do I have to bribe to get the wheels in motion... 

[corneliuslundie]

No good for back numbers but it is stocked in Smiths here in  Newtown.

[TheCuckoo]

Thanks guys!

 

Just ordered the two journals - they'll be something for me to read on Christmas morning when I've finished my "Beano". I wonder what else I'll find under the tree

[Brassey]

There was a series of articles in Model Railways in 1980 (IIRC) by Iain Rice on building various 517 variants from the Mallard kit.

 

I have had these articles and a Mallard kit for quite some time and just getting round to building it.  What Ian Rice never picked up on was that the Mallard kit has the 15' 6" wheelbase whilst both his "variants" he demonstrated would have had the 15" wheelbase which is evidenced on the drawings with the articles.

 

Maybe he just chose to ignore that.

 

I need to build a 15" version.  My question is, when the wheelbase was lengthened to 15' 6", where the frames and footplate also lengthened at the same time i.e. is the Mallard footplate/valance etc too long for a 15" 517?

[Miss Prism]

I think the Mallard frames, footplate, valence are too long for the 15'. More importantly, the Mallard is also the wide tank version, so is applicable only to the wide-tank locos, all of which were 15'6".

[Brassey]

I think the Mallard frames, footplate, valence are too long for the 15'. More importantly, the Mallard is also the wide tank version, so is applicable only to the wide-tank locos, all of which were 15'6".

 

Thanks Miss Prism

 

I have built an M&L whitemetal kit which has the 15' wheelbase and shorter footplate but is the same width as the Mallard kit.  I did read of the narrower versions in one of the Ian Rice articles but found nothing else about them.  Any ideas where I might find more info on the narrow-tank locos.  I do have Russell and the RCTS volume 6 but might have missed it?

 

Cheers

[Penlan]

I think the variations over time of the prototype,

might be to much for 'Kernow Models' to

even contemplate doing a model of.  Just a thought.