By Doug Hewson

One of my favourite occupations is shunting the Yards either at Gilling East or on our own railway as I just revel in being surrounded by scale rolling stock with my Y4. However, the new Yard at Gilling has a headshunt which falls away from the yard on a gradient of 1 in 100 and the Y4 really struggles when trying to transfer a heavy coal train from the marshalling sidings to a departure road. I can manage to push about 16 loaded coal wagons up the bank and any more than that and I usually have to move the train in at least a couple of bites. It is not so bad on our own railway as we do not have the gradients and neither do we have the room to handle 40 wagon trains.

To this end I decided to look for a more powerful six coupled tank loco, preferably LMS or LNER but there are few designs available for such things and of course anything freelance is absolutely out of the question. In any event as a founder member of GL5 (The Ground Level 5” Gauge Main Line Association) they would not be allowed under our rules but neither would I thank you for one, anyway. 

Whatever I design or build has to be such that it can be photographed among the scale rolling stock and be indistinguishable from the real thing.  However, I was aiming to save time and as I couldn’t find a loco I was looking for I decided I would have to bite the bullet and advertise for a Pansy chassis as that seemed to fit the bill power-wise even though it was a product of the ‘Gas Works Railway’ and on our own railway we do at least have a GWR yard where it would be at home sometimes. 

I hardly dare mention it to our other Club members with my apparent change of allegiance but they will have to learn to live with it. I placed adverts in ME and EIM and I ended up with a reasonable chassis (or so I thought!) and then another very nice one came along so I decided it was too good to miss so I bought that one, too, with the intention of completing both in different guises and different liveries eventually. I was also offered a brand new one which I also bought as I knew one of our club members was on the look out for one so that one was passed on to him so then I wasn’t quite alone but Oh! the shame.

Anyway, the main reason for wanting a running chassis was that I thought I wasn’t going to be at all fussy about the lower works or the bits out of sight as long as the upper works looked right. I thought that I could then go to town on the cab, tanks and bunker to make a nice looking loco.

I decided to try the first chassis on air but on hooking it up to the airline there was no response in either gear. When I looked a little closer I realized that the reversing arm was an inch short hence an extension which consisted of two strips of 20swg strip and a couple of 1/16” copper rivets on the top.  I bought the second one from a friend and he told me that this one ran perfectly as long as you ran it upside down! 

I had already been forewarned about that problem by Rob King who has one and is one of our regular Gilling drivers.  He had already modified his at a very early stage to prevent the valves from dropping off their seats but more of that later.

At this stage I decided to do a bit more delving as I could see that both lots of coupling rods were wrong as they should be fish bellied and although nicely finished these were definitely straight so they had to go. Also, the buffer beams were adorned with the most hideously over scale rivets, a lot of which were in the wrong places so they were also consigned to the scrap bin, as were the clumsy brackets which held them in place. Furthermore, the quite visible brake hanger brackets were nothing like the full-size ones, so the bin was quickly filling up. 

As I piled more bits into the scrap bin I began to realize there were far too many holes in the frames apart from the ones which the builder had already drilled in the wrong places, apart from which the holes were all way over scale size so I decided I had better treat it to a new set of frame plates.  If I was going to do that then they may as well be right in the first place so they had to be hole for hole to the works drawings otherwise there would be no point in this exercise.

I rang Neville Evans to see if he had any drawings of the Pannier Tanks and although the answer was negative “he knew a man who did” so he passed me on to a friend of his, John Hill, who kindly offered me a loan of his copy.  This drawing arrived in very short order and once I saw that there was obviously nothing for it but to start again almost from scratch. I say almost as I will explain. 

I already knew of a number of people who were either trying to complete a Pansy and wanted to improve the detail and others who already had one and wanted to make improvements to theirs at the next rebuild. Also I had a perfectly good cylinder block and crank axle.(that was all that was left of the original chassis!) and I was sure that there would be others who would want to improve their existing models. I, therefore, decided to stick with the original frame spacing (which is 3/8” too narrow to be strictly to scale) and so that all the new castings could be used for either the original Pansy or the new one.  I also decided to stick with the original arrangement with the valves underneath for reasons which will become clear later.

The Prototype

The next major problem was the fact that I couldn’t find out much information on the prototype GWR 5700s until my good friend Steve Whittaker pointed me in the direction of the RCTS book on GWR locomotives in which there is a brief section devoted to them. This together with other information provided by Steve and others has helped enormously. There are also some papers to be published shortly on the Pannier Tank Classes by Irwell Press so watch out for their adverts.

With my new found knowledge I thought it would be useful to include some of the details here which I have found for interests sake.  I will use the term 5700s as a generic term as that is how they are designated in the 1950s Ian Allan loco spotters books despite their different number classes, all of which had their differences.

The 5700s have a long pedigree and were actually an almost unaltered continuation of the rebuilt 2700s but had the earlier rectangular section coupling rods of the 1854 Class and prior to that the 1813 Class of 1882. Building commenced in 1928 with the first of the class being completed in 1929 and production continued on and off until 1950 well into BR days so there is a good range to choose from and if you don’t want a Swindon product (dare I say it!) there were quite a number built by outside contractors. These were all built during the late 1920s and early 1930s under a Government scheme to alleviate the effects of a trade depression.  Has a familiar ring to it doesn’t it!

The locos were built with quite distinctive cast iron chimneys so no polished copper ones please, and a lot were turned out with painted safety valve bonnets thought they were mostly polished up later.  If you wish to build a slightly more simple version then 5700-99, 7700-99 and 8700-49 were built without steam heating and for those who want an even more simple version No.s 6700-49 were built with steam brakes only and three-link couplings as they were built purely for shunting duties.  This also applied to 6750 to 79.

Beginning with 8750 in September 1933 the locos were fitted with an improved cab with the square windows and according to the works drawing the dome covers were 3” lower. The new cabs had sliding shutters and hinged doors and in 1937 a drawing was issued for these to be fitted to the older locos. 

A further modification was made with two steps being welded to the left hand side of the bunker and some additional handrails (added to the Fireman’s side). The first loco to be so fitted was 9795 in 1934 but these too were added to the earlier locos as they went through the shops.

In 1942 the first locos emerged with top feeds and the first of the older locos to be retrospectively fitted were 8770 and 5735.  The first loco to be fitted from new with top feeds was 4656 although it did not seem to be standard fitting on new locos until 1944.  The top feed arrangement was then fitted to the older locos but subsequently some lost theirs later again during boiler changes. 

The third sub class was the 9700s which were built with condensing gear and partly deepened side tanks and were all allocated to Old Oak Common to work the Metropolitan Lines to Smithfield. The end result was 863 locomotives being built.  Another little difference which people may like is that the locos 7700-24 built by Kerr Stuart had riveted baffle plates in the tanks resulting in several vertical rows of rivets in the tanks, some single rows, some double and rivets down each end.  Some of these were used on the London Underground and some preserved but sadly the one at Quainton Road lost its rivets when it was rebuilt with new tanks.

According to the RCTS book, under BR they were designated Class 4F and apart from 3656 and 3742 all had BR front number plates.  No.s 6760-79 and 9673-82 were built with BR front number plates from new.  However, the “4F” designation is open to question as they were always known as “3Fs” in the Ian Allan Loco Spotters books.

The following locos were sold to London Underground Ltd.:-

5775 (L89), 7711 (L90) (1), 7760 (L90) (2), 5752 (L91)(1), 5757 (L91) (2), 5786(L92), 7779(L93), 7752 (L94), 5764 (L95), 7741 (L96), 7749 (L97), 7739 (L98), 7715 (L99).  7760 and 5757 were bought as replacements for 7711 and 5752 at a later date and took their original “L” numbers.  The following London Underground locos are preserved as below.

5775 Keighley and Worth Valley

7760 Tyseley

5786 Worcester locomotive Society, South Devon Railway

7752 Tyseley

5764 Severn Valley Railway

7715 Quainton Road

Note, that all these locos had cut down cab roofs to give 2½” clearance in the tunnels which meant clipping the sides of the roof off flush with the cab sides and moving the rain strips inwards by 6” from the original edge so beware when copying these preserved ones.  This is quite obvious on 5764 which I photographed recently at Bridgnorth.

Other Preserved 5700s


Didcot Railway Centre



Didcot Railway Centre



Bodmin and Wenford Railway



Severn Valley Railway

Awaiting Restoration


Llangollen Railway

Undergoing Restoration






Undergoing Restoration


Glouceter and Warwicks’ Rly

Undergoing Overhaul


Dean Forest Railway



Chinor & Princes Risborough


The new GWR 5700 and 8750 Classes


As briefly mentioned I have already designed new frames and so that the rivets and bolts are all in the correct places so therefore I decided that everything needed to be changed so the frame stretchers are now all to the works drawings and a pattern is being made for a new motion plate This will be provided with double flanges and the correct ribbing around the openings as it is quite visible under the tanks. It will also be cast with double trunnions for the valve gear rocking levers to give support from both sides. 

The new frames are also designed to take scale working leaf springs on the front two axles though for those who wish to keep it simple the existing axlebox design with the LBSC coil springs can be used if required. The new design includes the large coils springs for the rear axle which are a prominent feature in the cab. These are ready for production now.

New patterns have been produced for the correct spring hanger brackets and leaf spring pads and the new axle box patterns are on the way. Hopefully, the castings will be available in both gunmetal and cast iron for preference.

I have also made a new pattern for the horn keeps to go with the leaf springs. The keeps for the coil spring axleboxes will be laser cut with the frames. Along with the stretchers is a laser cut set of parts for the buffer beams and all the necessary outside brackets such as buffer beam brackets, brake hangers, tank supports, steps and platform brackets. 


The cylinder patterns are largely as LBSC designed but the bolting flanges have been raised slightly so that the smokebox wrapper can be correctly bolted to the sides. The one thing I have done is to alter the valve rods to nearer the scale size and added tail rod guides to reduce wear on the glands. The glands will also be fitted with O-rings both for the piston rods and the valve rods. 

As recommended by Rob King I have also modified the valves to take a couple of small coil springs to keep them on the port face when the steam is off. These are not strong enough to prevent trapped water being forced out but it will mean that the loco will run when the right way up!  I have also provided a full set of cylinder drain cocks. 

The main reason for leaving the valves underneath is to make it much easier for the average Model Engineer to set the valves and inspect them, and, the fact that the under scale frame spacing doesn’t really allow enough room to fit the valves between as they should be.  It also means that the large bore can be accommodated which will keep it well within the Class 3F category.

Wheels and Coupling Rods

On the Pansy design the wheels are wrong for no good reason as they have a spoke on the centre line of the crank boss.  This, therefore, means that the balance weights are in the wrong place and can’t be orientated correctly. The patterns for the new wheels are being made with a spoke either side but of course they can still be used for Pansy but will then be correct for either loco. 

On the full size loco the balance weights are loose steel plates riveted to either side of the wheel so all the wheels can be cast from the same pattern.  The balance weights will be supplied as a laser cut set with rivet holes ready cut.

The main problem with the coupling rods is that they should be fish bellied and much slimmer to be scale.  Again there is no good reason for them to be so over scale or the wrong shape or the chunky look beloved of Model Engineers who say to me “I decided to beef them up a bit”.  My question is, “Why”?  They are hardly going to bend in our scale. 

This also means that the crank pins need to be shortened.  However, you can still fit the new rods to the old crank pins but they will need a washer behind to make up the difference and they will look a lot better for it. This will also mean that the new CNC machined rods can be fitted to a Pansy.

Valve Gear

I decided to consult Don Ashton on the valve gear to see what his general opinion was on the existing Pansy valve gear and the answer was: “Not a lot”. In return I received a six page report on it but the gist of it was that there was a problem in both forward and reverse. In forward gear the cut off error is 38% and in reverse 42% so that explains why the locos do not notch up very well.

Don issued me with a new drawing detailing a couple of very simple alterations which reduces the error considerably.  These alterations can quite easily be carried out on existing Pansies.

The new valve gear incorporates these amendments and reduces the error to less than 7% in either gear. The two alterations mean lengthening the lifting arm by 3/8” and the lifting links to 2.41” centres.  Details are on the new drawing.  The other problem with the existing Pansy gear is that the radius of the expansion links is wrong and the length of the eccentric rods are too short. These are more major alterations but do not affect the valve gear as much as the rods described above. 

Apart from the items mentioned above I have also altered the eccentric straps so that they are divided at 45° which will mean that the rods can be dropped out without having to resort to removing the boiler.  The other alteration is to make the new eccentric rods bolt on to the front of the straps as per full size to eliminate the rivets across which tend to work loose.  Again, these improvements can be applied to Pansy to improve mileage and maintenance.  While trying to set out the LBSC valve gear on the CAD I found a misalignment of 1/16” so this explains why the rods on the ones I bought have bent eccentric rods.  This has been eliminated by correcting the fork joint offsets.

Steve Whittaker also rang me to say that it is virtually impossible to oil the little ends on the con rods as they are just not accessible. He usually overcomes the problem by standing the loco on its front buffers and squirting oil inside the crossheads.  Not an option when the loco is in steam! 

I have therefore altered the design with a bit of artistic licence so that the rods are similar to the 5100 class which has an oil reservoir behind the crosshead and a passageway to the little end. The big ends have been provided with a split bearing and cotters so that they can be taken up as required.

Brake Gear

The brake gear on Pansy is about as near to scale as you can get with the steam brake cylinder in exactly the correct place so all I have done is to provide the necessary detail for the alternative drop forged hangers and amended the cylinder design slightly so that the steam is admitted through the inner spindle as per full size. This is by no means an essential alteration, but it just means that the full size pipe layout can be copied. 

As for the vacuum brake, I am just working on the design for a scale working steam/vacuum brake valve to see if I can design one which not only looks right but will work right.  However, it may not work on an existing Pansy boiler as the hollow stay is too small and creates too much back pressure for the ejector exhaust. 

See later but I have designed a new boiler with a larger bore hollow stay but still don’t know yet if this new arrangement will work. If not I have an alternative plan up my sleeve to house the ejector in the smokebox and use the hollow stay as the steam pipe to feed it. This will then mean that the exhaust can go straight up the petticoat and should give the distinctive hollow roar up the chimney when the ejector is on. 

If this is the case the valve will then fit directly on to an existing Pansy boiler with a very slight modification to the bush. It will just require a small hole drilling in the bush face and a passageway drilling upwards at as steep an angle as possible to connect into the steam space above. This will provide steam into the back of the brake valve. The brake valve will operate the same as the full size one with brake “On” to the left and “Off” to the right.

The brake hangers on the Pansy design are correct for the early locos but then some were changed for the later drop forged steel ones but when this occurred I don’t know.  The original ones consisted of a pair of  plates with the brake block between. We already have the castings in stock for the drop forged type. 

While on Holiday I ventured into the Dean Forest Railway Shop and there staring at me was a magnificent book The Red Panniers  (Lightmoor Press) This recent book is of course the history of the GWR Pannier tanks which were sold to London Underground for maintenance work on their network. I found it an excellent book with many interesting photos and a lot of good detail. It also gives a fascinating account of the work which goes on largely unseen and in which the Panniers played a major part. 

However, I digress, the main point is that at the top of page 159 there is a photo of 5775 with double plate brake hangers but on the same page there is another photo of the same loco pictured in 1960 at Pontypool Road with the later type hangers. Other photos in the book show that quite a number of locos retained their double plate hangers until they were withdrawn in 1971. This illustrates another very good reason for finding as many good photos as you can at different dates of the loco you want to build so that you can get the detail and livery correct at that time.

The Boiler

I have to say that the boiler for Pansy is a very good design and quite near to scale with the hollow stay almost in exactly the right place for the brake valve. However, the main problem is that it just won’t go within the scale outline of the cleading.  If you have already got your boiler then slightly over-scale cleading will have to do but like a lot of LBSC locos there is not really enough room for any lagging as well.  I have therefore redesigned the boiler slightly so that it will fit and there is room for some good insulation and it will well make up for the deficit of the slightly smaller boiler.

One of the main problems is that the gauge glass is far too long and as built the bottom fitting is below the firebox crown.  There is absolutely no advantage whatsoever in having a gauge glass too long as the only portion that you want to see is the maximum and minimum working levels.  In any event the bottom nut needs to be above crown level just purely for safety reasons. 

The top nut should be positioned to give about 5/8” range which is about right for a boiler of this type.  If the water level is allowed to go above that there is a danger of priming which is not a disaster in our scale but more of an embarrassment. I have been on the footplate of numerous locos where the fireman has the water out the top of the glass so he hasn’t a clue where the water level really is and then they wonder why the valve gear gets bent or worse still blow the cylinder end cover off.

I would, therefore, suggest that if you are using a standard Pansy boiler you take a measurement inside the firebox to measure the height of the crown and then fit a loose brass sleeve over the glass. On the Winson Britannias we have to use a sleeve 1/2” long to give a minimum safe reading as the bottom bush is 1/2 too low. I have designed the boiler back plate as per full size with the alternative bushes for backhead clacks so that either configuration can be built, i.e. with or without top feeds.


The full size GWR 5700s only have two superheater flues and they are only there to prevent cracking in the top corners of the firebox tube plate. From my own experience I find that superheating a small loco in our size is not only essential but can transform a loco. 

LBSC obviously had the right idea as he included four flues so I had no reason to change this. The only change I have made is to use our stainless steel weldless elements. Each element is made from a single piece of 5/32” x 24 swg tube in a “U” tube which doubles back the same as other full size elements.  This means that there is 64 inches of element in the new design of superheater, a large proportion of which is directly over the fire. 

I have several friends who have already fitted these elements and they say they have transformed their locos merely by changing the type of element. One of our members has just three in his K1 and he won IMLEC with his. The only other change I have made is to the header where I found it useful on the Y4 to be able to change individual elements so I have copied this on the Pannier so that the elements on the new design are attached to the header with unions. With these elements it is also possible to push a flue brush down the flues to clean them out, another distinct advantage.


The smokebox has been re-designed so that the door opening is in the correct place 3/16” above the boiler centre line and the wrapper has been shortened by 9/32 so that it is the correct length.  I did say that the chimney is too short but on re-checking it is only 1/16” too short so the new pattern will be the correct height. The rivet size for the wrapper has been altered to suit my own measurements taken off 9681 and they are now at the scale spacing.  Other details have been added such as the row of set screws in the front just below the smokebox door.  Hopefully the chimneys will also be available CNC machined along with a matching petticoat pipe.

Boiler Fittings

The new turret which I have designed especially for the 5700 fits exactly in the correct place on the LBSC boiler and incorporates the regulator bush. This has to be added to the casting as an additional bush because of the way the turret is cored out.  The new castings are already in stock. 

The other fittings such as injector steam valves are all built up on to the turret as per full size with all standard GWR fittings.  The two injector steam valves are one either side of the centre line and then raised up at the right hand end is the blower valve and opposite is the isolating valve for the steam heating.  Below and to the left is the ‘Masons’ valve which controls the steam heating pressure and has a coiled spring type handle. In our case of course this is a dummy unless of course you want a steam heated driving truck seat. 

The blower valve is quite a tiny little valve perched on the top right hand end of the turret and the pipe from this goes through the spectacle plate and under the tank cover plate and turns into the smokebox with a standard blower elbow. On the front of the spectacle plate is a nice little fairing which covers the blower pipe where it dips down under the boiler casing. All the new valves have captive spindles to comply with the new boiler code. I have also re-designed the safety valves as there is now a pair as per full size to comply with the current boiler regulations and they just fit in the correctly sized polished bonnet.

Cab, Bunker, platforms and platework

On one of the chassis I bought, the side platforms had been completed but once again they were adorned with the most enormous over scale rivets and the line along the platform edge was a token sprinkling to say the least.  For the new locos I have already drawn up a complete set of laser parts for the platforms, splashers, bunker, cab and step plates for the later type 8750 Class and the 5700 will follow shortly.  In this case all the joints in the platforms and valances have been copied from the works drawing and my own measurements from 9681 and 7754 so that they are rivet for rivet as the full size loco. 

Yes, I am an unashamed rivet counter and very proud of the fact. 

What is the point of doing it any other way, you may as well build a Simplex otherwise! Included in the kit are all the fixing angles for inside the cab and all of the rivet holes come ready cut. The only ones omitted are some down the rear of the bunker which need to be drilled after the corners have been flanged over a former. To this end I had some formers made and carried out tests to see if the rear corners of the bunker could be flanged out of 20swg steel plate. This proved to be quite easy with a little patience and care so they will be supplied ready formed in the kit of parts. The remaining holes can then be drilled using the lap strips as templates by the builder. I also intend to supply the brass window frames, hinges and catches as a lost wax cast set.

Hopefully this will give a bit of an insight into my thinking and planning.  I have not got any of the drawings plotted out yet and neither are the new patterns ready but I am hoping that the full set of drawings will be ready before Christmas (2020), apart from maybe the vacuum brake valve which still needs a bit of thinking about. 

Drawings Available are:

  1. 1.General Arrangement LHS with end views and plan (top of page)

  2. 2.Frame details, horns, axleboxes and motion plate

  3. 3.Buffer beams, brake hanger brackets, frame details and leaf springs

  4. 4.Frame and Platform plan, part brake gear, couplings and buffers

  5. 5.Sanding gear, toolboxes, lubrication details

  6. 6.Cylinders, wheels, axles, brake gear, coupling and connecting rods

  7. 7.Motion layout, details and reverser

  8. 8.Smokebox arrangement and details, regulator and superheaters, dome covers, safety valves

  9. 9.Tanks, fillers and cover plating

  10. 10.Boiler, ash pan, grate and top feed details

  11. 11.Cab and bunker details 8750 Class

  12. 12.Cab assembly details

  13. 13.Cab layout and details 8750 Class

  14. 14.Drg No. 8750_01, 8750 Class GA (Below)

the only free and the only weekly magazine for model engineers. 

Editor: David Carpenter