Steam Loco building & design in the UK and Abroad.

[Allegheny1600]

I’ve been interested in this sort of thing for some 20 plus years now, ever since I read “Express Steam Locomotive Development in Great Britain and France” by Colonel HCB Rogers in 1990.

I read once that the flow of high pressure steam from boiler, through the cylinders to exhaust was the most studied area of "gas dynamics" until the invention of the rocket motors developed for the NASA space program!
I have been contemplating many questions over the years so I may as well ask them here.
Why were British locos all built with plate frames when other countries railways used bar frames? Which was better, really? I’m guessing the plate frame was cheap and simple.
What was the most advanced British loco and what were it’s features?
How did this compare with advanced locos by Chapelon, Porta et al?
How did American “Super Power” locos compare against these designs?
What was the better cylinder arrangement? (for which job?).

 

I don't wish to start any 'wars' between proponents of the various companies, I just want to discuss the development of that most wonderful invention, the steam locomotive.

Thanks in advance for any contributions,

John Edge.
 

[Sheffield]

It used to be said the the science of thermo-dymamics owed more to the steam locomotive than the steam locomotive owed to the science of thermo-dymamics. A study of the steam circuits of a lot of British locomtives may lead one to think their designers were asleep when such things were being taught.

[34theletterbetweenB&D]

...Why were British locos all built with plate frames when other countries railways used bar frames? Which was better, really? I’m guessing the plate frame was cheap and simple...

 

  The plate frame was superior for the constraints the UK railway infrastructure imposed. Locomotive weight was always pushing at the limits of what the relatively weak constructions supporting the tracks could bear, and there were also 'set in stone' rather small radii to negotiate all over the system. For any given weight of steel you get more vertical stiffness and lateral flexibility out of a plate frame than a bar frame, and a well designed plate frame can have lightening holes too to position the maximum vertical stiffness where required and save weight where not  It is not that well appreciated that there was a real difficulty in getting a high output loco not only within the confines of the structural loading gauge, but also within the weight limits the UK infrastructure imposed. An advance to and acceptable 25 to 27 ton axleload instead of the 22 ton limit of the UK steam period would have made the design of pacifics - and potentially 4-6-4s and 4-8-4s - much simpler

 

An interesting historical perspective. Just four miles South of where I sit is an area called 'Marshmoor'. This was exactly what the name suggests, before it was drained and built up to allow the Great Northern Railway's main line to be constructed. On this construction, now rests the East Coast main line seventeen miles out from Kings Cross, on a lovely curve. About every thirty to forty years a fast train falls off the track at this location. These coincide with new faster traction being introduced, more weight, more lateral load; and the rails break, and off she come.

 

...What was the most advanced British loco and what were it’s features?... 

 You need to state your criteria for 'advanced'. If it's as much power as can be got per unit cost of construction and operational cost, then without doubt the Kylchap equipped A4. The speedworthiness of this design reveals a sustained power production per ton as high as any other UK design achieved, and they were cheap to build, and subsequently maintain and operate as BR's comparative data revealed. The turbomotive was a good effort, and if it had been developed and multiplied as an LMS standard would definitely be in with a shout.

 

The BR 9F introduced a notable feature to the plate frame, a horizontal diaphragm, and the 8P a good implementation of the poppet valve. A shame these came too late for general application.

 

 

...How did this compare with advanced locos by Chapelon, Porta et al?...

 

 Unadventurous! Little success with poppet valves and better gears, and no resuperheat compounding for example. But superheaters are heavy and this type of loco was pushing at the weight limits aready mentioned

 

...WHow did American “Super Power” locos compare against these designs?... 

 Against UK designs, so massively greater in weight and physical size. as not to be 'scaleable' to UK restrictions. It is well known that the first Gresley pacific (A1) was scaled from the PRR's very successful K4, roughly sixty percent of a K4 by mass, 41 sq ft of grate vs 70. As finally developed in the UK, 50 sq ft grate pacifics, little physically larger than the Gresley A1, but significantly heavier, could match a regular piston valve K4 in power output for as long as the fireman held out (no mechanical stoker, precluded by weight for a start!).

 

Chapelon would so loved to have had his hands on some typical US superpower. He was pretty confident he could have doubled the typical power output. Imagine what one of the lovely Nickel Plate Berkshires could have done with twice the grunt - half the running cost of diesel...

 

...What was the better cylinder arrangement? (for which job?)... 

 For general purposes, Stephenson got it right: two. Only advance to three if two cannot deliver enough power. Under no circumstances have four or more (simple) in a single engine frame, unnecessary expense for a start, four plays three.

[PenrithBeacon]

In the 1820-30 period most British locos had bar frames too; see Rocket.

 

As locos got bigger plate frames became essential in UK but industry had facilities tp make these, even then the plates had to be welded together.

 

The engineering industry in the US wasn't as sophisticated so they had to develop the bar frame to cope.

 

There are advantages to both.

 

Regards

[rogerzilla]

GWR practice was to have low superheat and higher boiler pressure, as they found this was equally efficient.  Superheating is ineffective unless you have extended running at reasonable speeds, and it reduces the amount of evaporative surface in your boiler because you could have had more smoke tubes instead of the big superheater flues.  Churchward also revolutionised valve lap and timing, which had a big impact on efficiency as well as steaming.

 

Some of the "advanced" designs were actually pretty terrible; Bulleid's Pacifics were much better after they were rebuilt by BR (they were apparently going to be scrapped as they were so troublesome, but the boilers steamed well and this just tipped the balance in favour of a rebuild programme) but they still used far too much coal.  Likewise, Gresley's No. 10000 looked brilliant on paper but it idn't work as well as had been hoped.

 

The best advanced design was probably Turbomotive, which really did work well and was only rebuilt because of the cost of fixing the turbine after many years of good service.

 

ISTR that a steam locomotive with a free exhaust can't be more than about 10% thermally efficient and is generally much, much worse.

[Allegheny1600]

  You need to state your criteria for 'advanced'.

 

 For general purposes, Stephenson got it right: two. Only advance to three if two cannot deliver enough power. Under no circumstances have four or more (simple) in a single engine frame, unnecessary expense for a start, four plays three.

Wow!

Many thanks (Paul?) for such a comprehensive and informative reply.

For "Advanced" I guess I really meant "Loaded with features" like high superheat, thermic syphons, "gas flow" grate (a la Porta), even things like mechanical lubricators, rocking grates and so forth.

I never thought about REALLY advanced machines like 'Turbomotive' or 10,000. I guess I am a classic case of 'having a little knowledge is dangerous'!

I thought 'we' got it pretty much right with the Standards wrt cylinder provision, it was much the same across the USA and in both Germanies - where their post war new build steam programs essentially mirrored ours (just more emphasis on the larger locos).

Cheers,

[Allegheny1600]

  Against UK designs, so massively greater in weight and physical size. as not to be 'scaleable' to UK restrictions. It is well known that the first Gresley pacific (A1) was scaled from the PRR's very successful K4, roughly sixty percent of a K4 by mass, 41 sq ft of grate vs 70. As finally developed in the UK, 50 sq ft grate pacifics, little physically larger than the Gresley A1, but significantly heavier, could match a regular piston valve K4 in power output for as long as the fireman held out (no mechanical stoker, precluded by weight for a start!).

 

Chapelon would so loved to have had his hands on some typical US superpower. He was pretty confident he could have doubled the typical power output. Imagine what one of the lovely Nickel Plate Berkshires could have done with twice the grunt - half the running cost of diesel...

I was surprised when I first learned, not that many years ago, that the Pennsy K4 led to the A1 - I'd spent many years in the dark about that one!

The K4 wasn't even classed as a "Superpower" locomotive, I think the first such machine was the Lima 2-8-4 of the mid 1920's and onwards and had a lot to do with high superheat and massive grate area.

 

How I should have loved to have seen what Chapelon could have done with a Berkshire!

Cheers,

[Allegheny1600]

GWR practice was to have low superheat and higher boiler pressure, as they found this was equally efficient.  Superheating is ineffective unless you have extended running at reasonable speeds, and it reduces the amount of evaporative surface in your boiler because you could have had more smoke tubes instead of the big superheater flues.  Churchward also revolutionised valve lap and timing, which had a big impact on efficiency as well as steaming.

 

Some of the "advanced" designs were actually pretty terrible; Bulleid's Pacifics were much better after they were rebuilt by BR (they were apparently going to be scrapped as they were so troublesome, but the boilers steamed well and this just tipped the balance in favour of a rebuild programme) but they still used far too much coal.  Likewise, Gresley's No. 10000 looked brilliant on paper but it idn't work as well as had been hoped.

 

The best advanced design was probably Turbomotive, which really did work well and was only rebuilt because of the cost of fixing the turbine after many years of good service.

 

ISTR that a steam locomotive with a free exhaust can't be more than about 10% thermally efficient and is generally much, much worse.

Thanks Rogerzilla,

I must read up on GWR locomotives!

Weren't the 'Spam cans' rebuilt due to features like the oil bath chain driven valve gear and the casing itself? Wasn't the boiler and firebox always good, even before the rebuild?

Cheers,

John.

[Arthur]

For general purposes, Stephenson got it right: two. Only advance to three if two cannot deliver enough power. Under no circumstances have four or more (simple) in a single engine frame, unnecessary expense for a start, four plays three.

Expense aside, why? What is the problem with four cylinders?

[asmay2002]

One of the problems with trying to decide which was the most "advanced" British locomotive is that almost no one type had all of the features that it could have had.   The last generation of British steam had things like roller bearings, Poppet valve gear,  boilers proportioned to steam well, thermic syphons, mechanical stokers, drop grates, hopper ashpans, etc but many of these features had been built into British locos for export since the 1920s.  Cast steel bed frames only featured in locos built for export (BR considered then but had no way of making them) and feed water heaters had been tried and discarded before WW2 as a maintenance burden. The pressure for reliability and simplicity tended to work against some of the proprietary developments like feedwater heaters. Subtle things like water treatment, the use of manganese steel to reduce wear on axleboxes and optical frame alignment made major improvements to mileage between repairs and could be applied to many types although application in the UK was patchy.

 

Probably the most advanced British built locomotives were the class 25 condensing locos built by NBL for South Africa.   In Britain itself the only ones to combine most of the features in the later days of steam would be the Caprotti standard fives or the Duke of Gloucester but that sadly didn't realize its potential until preservation as BR lost interest because of dieselisation before its steaming problems were sorted out. Even without that its documented test performance showed it as the most efficient conventional engine built in Britain in terms of lowest steam consumption per HP (overall it wasn't quite so good when built because of the steaming issues but those are fixed now -sadly too late).  On the other hand if you just want sustained high power unlimited by a human fireman at any normal speed with anything passenger or freight on the back it would be hard to look past the stoker fitted 9Fs. 

[JeremyC]

.....  Cast steel bed frames only featured in locos built for export (BR considered then but had no way of making them) . ...

Was any builder in Britain capable of making them? I'm sure I've read somewhere that when North British Loco built engines using cast steel bed frames the castings came from America.

[Arthur]

No locomotive builder in the UK had a foundry capable of casting a large bed frame. The English Steel Corporations River Don works certainly had the capability but I don't believe that they were ever contracted to cast one. There were possibly one or two other large steel founders with the capacity.

 

I wouldn't be surprised that even in the US the casting of such large and complex products was contracted out, there were certainly commercial foundries making them.

 

I've a photograph somewhere of a large cast bed frame and if I can remember where it is I'll post it. It is a remarkable example of the foundrymans art.

[jamie92208]

No locomotive builder in the UK had a foundry capable of casting a large bed frame. The English Steel Corporations River Don works certainly had the capability but I don't believe that they were ever contracted to cast one. There were possibly one or two other large iron founders with the capacity.

 

I wouldn't be surprised that even in the US the casting of such large and complex products was contracted out, there were certainly commercial foundries making them.

 

I've a photograph somewhere of a large cast bed frame and if I can remember where it is I'll post it. It is a remarkable example of the foundrymans art.

 

I can't remember which foundry it was in the states but I think you are spot on there Arthur.  The cast bed frames that allowed locos like the Big UP and other Super power machines were fabulous examples of the foundryman's art.  IIRC there were only one or two firms that could make them and I believe that they worked very closely with the railroads. In the UK  we had to go to Czechoslovakia's Skoda works to get the central axle casting done for the London Eye.

 

Jamie

[bike2steam]

 In the UK  we had to go to Czechoslovakia's Skoda works to get the central axle casting done for the London Eye.

 

Jamie

Yep - so sad, we were once the world leaders in heavy engineering ( I did my apprenticeship in one such company), now near enough extinct - due to whoevers propaganda you believe in.

[34theletterbetweenB&D]

Expense aside, why? What is the problem with four cylinders?

 It is expense, in both first cost and maintenance: simply unnecessary unless there are some other requirements to be met which mean that there cannot be enough cylinder volume from three: not a constraint that I believe ever need have affected a UK design. (The big problem with four simple in one frame wasn't found in the UK, the famous Duplex designs of the PRR were the 'trouble'.  Well worth a read, for any interested.)

[Arthur]

 

The US foundry names which spring to mind are General Steel Castings Corp, Commomwealth and I'm pretty sure that Bethlehem Steel had a large jobbing foundry.

 

We do still have heavy forging capacity at Sheffield Forgemasters though I gather that they are having difficult times. Just what capacity for making large castings, or casting ingots for forging, resides in our remaining steelworks I have no idea.

[Arthur]

It is expense, in both first cost and maintenance: simply unnecessary unless there are some other requirements to be met which mean that there cannot be enough cylinder volume from three: not a constraint that I believe ever need have affected a UK design. (The big problem with four simple in one frame wasn't found in the UK, the famous Duplex designs of the PRR were the 'trouble'.  Well worth a read, for any interested.)

Presumably then, the GW and LMS believed that they couldn't get the required cylinder volume from three cylinders or were they wrong?

 

Would the Princess Coronations have developed such high dbhp from three cylinders? Open question, I don't know the answer.

[Michael Edge]

Beyer Peacock built a good number of Garratts for South Africa with cast steel frames - does anyone know where these were cast?

[asmay2002]

Beyer Peacock built a good number of Garratts for South Africa with cast steel frames - does anyone know where these were cast?

US I think.  The AD60s for New South Wales were done by General Steel Castings in the US.

[jamie92208]

US I think.  The AD60s for New South Wales were done by General Steel Castings in the US.

That's the name that now springs to mind for me with regard to the big UP engines.

 

Jamie

[34theletterbetweenB&D]

Would the Princess Coronations have developed such high dbhp from three cylinders? Open question, I don't know the answer.

 On equivalent cylinder volumes, a little extra. Every 'engine': cylinder, piston, valve and valve gear, piston rod, crosshead, connecting rod and crank weighs what it weighs, and then imposes mechanical friction in motion, and because the moving parts are oscillating draws energy continuously for acceleration. There's less of that in a three cylinder assembly doing the same work as a four cylinder assembly, so near a third more energy absorbtion from these causes in the four cylinder assembly, which is subtracted from the engine output before it can do work at the drawbar. This is one of those 'less is more' situations.

 

These advantages were not lost on UK designers: as multicylinder designs became necessary because insufficient cylinder volume was available from a pair of the largest possible cylinders that would fit, four cylinder, 2 x 2 was the 'obvious' choice and initially more common. But as the art developed, three cylinder designs became the more common choice, and the end result was BR had significantly more three cylinder machines in service than four cylinder.

[flubrush]

Was any builder in Britain capable of making them? I'm sure I've read somewhere that when North British Loco built engines using cast steel bed frames the castings came from America.

 

Some of the later Beyer-Garratt designs had cast frames and they came for the USA.

 

Jim.

 

[later]

 

Sorry - repeated what someone had stated earlier - must update threads before answering.

 

But the cast frames for the Garratt classes were produced by General Steel Castings Inc. in USA.  There is a picture of one on Page 29 of Durrant's Garratt book.

 

Jim.

[runs as required]

I've always been interested in the human side of train running, since I grew up with railway men close to, then later within the family.

 

I found British railway men even more partisan than enthusiasts about their particular line.

Stepfather-in-law at Buxton though he began work on the LMS was always pro Derby (Midland) and dismissive of Crewe (North Western).

On the other hand, however partisan, all railwaymen were nevertheless intrinsically suspicious of any change from the established working pattern.

An 'uncle' who was a Stratford driver first told me the story about the most efficient loco being one with a yellow painted funnel - since everyone believed it was the one the guys with the O&M clipboards were watching.

 

The French mécanicien had the reputation of being better theoretically prepared to address complexity - and one might say the same of Southern African and Indian sub-continent engine drivers.

 

By and large I reckon Britain got the steam locos it could run - the sexier stuff went for export.

 

dh

[Arthur]

On equivalent cylinder volumes, a little extra. Every 'engine': cylinder, piston, valve and valve gear, piston rod, crosshead, connecting rod and crank weighs what it weighs, and then imposes mechanical friction in motion, and because the moving parts are oscillating draws energy continuously for acceleration. There's less of that in a three cylinder assembly doing the same work as a four cylinder assembly, so near a third more energy absorbtion from these causes in the four cylinder assembly, which is subtracted from the engine output before it can do work at the drawbar. This is one of those 'less is more' situations.

 

These advantages were not lost on UK designers: as multicylinder designs became necessary because insufficient cylinder volume was available from a pair of the largest possible cylinders that would fit, four cylinder, 2 x 2 was the 'obvious' choice and initially more common. But as the art developed, three cylinder designs became the more common choice, and the end result was BR had significantly more three cylinder machines in service than four cylinder.

Right, so there's nothing inherently wrong with a four cylinder layout.

 

Three became common in the UK because it was able to provide sufficient 'power' in most cases. Yes, with four you add the extra costs of build and maintenance and the extra cylinder brings with it a set of frictional losses. And, as with the BR standards, you wouldn't use four if three would do, nor three if two would do. Neither would you adopt four smaller cylinders of the same total capacity as three larger ones.

 

However, if you need, within the constraints of the loading gauge, and presuming that the boiler can meet the demand, as much 'power' as possible then a fourth cylinder surely makes sense? Cheaper than double heading.

 

Comparing a Princess Coronation to an A1,

 

PC 4 cylinders, 16.5" x 28", volume of all cylinders 2900 cubic inches with a total piston area 103.6 square inches,

 

A1 3 cylinders, 19" x 26", volume of all cylinders 2328 cubic inches with a total piston area of 89.4 square inches.

 

That is a big difference in potential cylinder output, the extra frictional loses pale into insignificance. Could Peppercorn have fitted even larger cylinders to the A1 or had the outer two reached the loading gauge? Had he 'maxed' out the potential of a three, equal, cylinder layout within the loading gauge?

 

So I'm not seeing anything inherently wrong with a four cylinder layout if the extra complexity can be justified by the need for power, is that the case?

[EddieB]

 

One of the problems with trying to decide which was the most "advanced" British locomotive is that almost no one type had all of the features that it could have had.   The last generation of British steam had things like roller bearings, Poppet valve gear,  boilers proportioned to steam well, thermic syphons, mechanical stokers, drop grates, hopper ashpans, etc but many of these features had been built into British locos for export since the 1920s.  Cast steel bed frames only featured in locos built for export (BR considered then but had no way of making them) and feed water heaters had been tried and discarded before WW2 as a maintenance burden. The pressure for reliability and simplicity tended to work against some of the proprietary developments like feedwater heaters. Subtle things like water treatment, the use of manganese steel to reduce wear on axleboxes and optical frame alignment made major improvements to mileage between repairs and could be applied to many types although application in the UK was patchy.

 

Probably the most advanced British built locomotives were the class 25 condensing locos built by NBL for South Africa.   In Britain itself the only ones to combine most of the features in the later days of steam would be the Caprotti standard fives or the Duke of Gloucester but that sadly didn't realize its potential until preservation as BR lost interest because of dieselisation before its steaming problems were sorted out. Even without that its documented test performance showed it as the most efficient conventional engine built in Britain in terms of lowest steam consumption per HP (overall it wasn't quite so good when built because of the steaming issues but those are fixed now -sadly too late).  On the other hand if you just want sustained high power unlimited by a human fireman at any normal speed with anything passenger or freight on the back it would be hard to look past the stoker fitted 9Fs. 

 

Probably the most advanced British built locomotives were the class 25 condensing locos built by NBL for South Africa.   In Britain itself the only ones to combine most of the features in the later days of steam would be the Caprotti standard fives or the Duke of Gloucester but that sadly didn't realize its potential until preservation as BR lost interest because of dieselisation before its steaming problems were sorted out. Even without that its documented test performance showed it as the most efficient conventional engine built in Britain in terms of lowest steam consumption per HP (overall it wasn't quite so good when built because of the steaming issues but those are fixed now -sadly too late).  On the other hand if you just want sustained high power unlimited by a human fireman at any normal speed with anything passenger or freight on the back it would be hard to look past the stoker fitted 9Fs. 

 

I was going to make a very similar point, Andy.  Some of the most advanced locomotives to be built in Britain were indeed those exported to South Africa.  Largely built by North British, they were designed by South African Chief Mechanical Engineers whose names will be virtually unknown to most of us parochial Brits, but deserve to be accorded recognition alongside their better known British peers.

 

That David Wardale later chose examples of classes 19D and 25NC for his modifications did little to endear him to the South Africans, who considered them already to be "best of breed" (an opinion shared by no less than Gresley when it came to the 19D).  However, the Wardale-modified locos did show improved thermal efficiency and fuel economy when driven correctly (which was something of an Achilles' Heel even for Chapelon).  These, along with two modified R class 4-6-4 on Victorian Railways, prerhaps represent the furthest reach of British heritage in steam development.