Imaginary Locomotives

[DenysW]

3 hours ago, JimC said:

the wide firebox Atlantic solves or at least deals with a number of the ten wheel design problems

Total agreement.

 

But what isn't obvious is that the successful Pacific designs all seem to have required the learning from earlier designs within a single design office. The wonderful Bavarian S3/6 came after one-off 4-4-4 and a prior (Wurttemburg) mediocre Pacific. Same (ish) in France. And Gresley used the wide-grate Atlantic experience to get his Pacifics right.

[billbedford]

1 hour ago, DenysW said:

And Gresley used the wide-grate Atlantic experience to get his Pacifics right.

 

..but first, he put a long firebox between the frames giving a 7'3" x 9'0" driving wheelbase and making the whole loco too long for a bogie. 

[Rich Uncle Skeleton]

I’ve always liked the idea of a Terrier with 8 wheels. Like an engineer saw a Bavarian GTL4/4 and thought not a bad  idea that.

[Compound2632]

9 minutes ago, billbedford said:

..but first, he put a long firebox between the frames giving a 7'3" x 9'0" driving wheelbase and making the whole loco too long for a bogie. 

 

The 2-6-0s, you mean.

 

One way to achieve a larger grate area between the frames is the uncouple the driving wheels, avoiding the need for heavy-section connecting rods and any associated stiffness, if your manufacturing tolerances aren't exceptional. It worked like a charm for Webb:

 

 

Edited October 4, 2023 by Compound2632

[billbedford]

Yes, but Gresley didn't have Webb's out-of-the-box thought patterns. 

[AlfaZagato]

I think comparing firebox design begets comparing the A1 and A2 Pacifics of LNER.  If I remember my uncited reading, as-built the A2s were the better loco, but didn't (or weren't) develop much past what they had.   

 

Actually, better thought, how might have the LNER Pacific line developed if Raven's design won out?

[DenysW]

Is there a good reason other than standardisation why the early (1840s-1860s) railways didn't settle for small-wheeled 0-8-0s for mineral work? Reduces the axle-loading for big boilers for heavy trains and the small wheels improve the grunt at low speed. Giving three standard configurations (once the 2-2-2 singles were abandoned): 2-4-0 for passengers, 0-6-0 for merchandise, and 0-8-0 for minerals.

[rodent279]

On 04/10/2023 at 15:55, AlfaZagato said:

Actually, better thought, how might have the LNER Pacific line developed if Raven's design won out?

I think that has been explored before, and the answer was something like "not very well." Something to do with the positioning of the cylinders all driving the leading axle necessitating a long boiler and front heavy design.

[rodent279]

8 minutes ago, DenysW said:

Is there a good reason other than standardisation why the early (1840s-1860s) railways didn't settle for small-wheeled 0-8-0s for mineral work? Reduces the axle-loading for big boilers for heavy trains and the small wheels improve the grunt at low speed. Giving three standard configurations (once the 2-2-2 singles were abandoned): 2-4-0 for passengers, 0-6-0 for merchandise, and 0-8-0 for minerals.

I guess there just wasn't the need for the haulage capacity of an 0-8-0 until the early years of the C20th, and engines were not heavy enough to warrant 4 axles earlier than that.

[PhilJ W]

1 hour ago, DenysW said:

Is there a good reason other than standardisation why the early (1840s-1860s) railways didn't settle for small-wheeled 0-8-0s for mineral work? Reduces the axle-loading for big boilers for heavy trains and the small wheels improve the grunt at low speed. Giving three standard configurations (once the 2-2-2 singles were abandoned): 2-4-0 for passengers, 0-6-0 for merchandise, and 0-8-0 for minerals.

 

1 hour ago, rodent279 said:

I guess there just wasn't the need for the haulage capacity of an 0-8-0 until the early years of the C20th, and engines were not heavy enough to warrant 4 axles earlier than that.

An 0-6-0 was perfectly adequate even for heavy coal trains https://www.lner.info/locos/J/j20.php the one restriction was the strength of the underline structures such as bridges (and the GER wasn't noted for strong bridges).

[JimC]

Easy to get a circular argument I guess (can't have longer trains because the locos aren't big enough, don't need longer trains with the stock we have), but there must have been all sorts of technical limits its difficult for us to appreciate. I only have much data for the GW, but late 19thC Dean Goods were rated at power class A, which was allowed 29 loaded coal wagons on a 1:150 gradient. Aberdare 2-6-0s, which are probably roughly equivalent  to most pre group 0-8-0s, were allowed 48. That's a very considerable increase required in terms of length of refuge sidings, strength of couplings, train braking and all the other issues. And presumably the longer the unbraked mineral train is the more difficult it is to manage. I recall Cox pointing out that given a sufficiently long unbraked train it can stretch over two hills, making the problem of managing it extraordinarily complex. There were a handful of very early 8 coupled locomotives, but it seems evident they were not considered advantageous.

[Compound2632]

8 minutes ago, JimC said:

Easy to get a circular argument 

 

Turntables. A six wheeler could fit comfortably on a turntable of the 1860s/70s - 42 ft diameter on the Midland and I think typical generally - but an eight-wheeler would be a bit of a squeeze. 

 

A 4-4-0 (4-2-2) is a 2-4-0 (2-2-2) with the leading axle replaced by a bogie, so the layout and dimensions of boiler, cylinders, and motion can be unchanged (and are basically the same as you're using on your 0-6-0 or 0-4-2 tender engine and 2-4-2T, 0-4-4T, or 0-6-2T tank engine). A 0-8-0, if the cylinders are to be kept ahead of the leading axle, raises all sorts of problems - longer boiler, grate over one axle rather than between the axles, etc. 

 

Further, a 0-8-0 is only more powerful than a 0-6-0 if you can fit bigger cylinders, which was a problem for a two-cylinder engine, either inside or outside. That's why the first British 0-8-0s produced in quantity were Webb's three and four cylinder compounds for the LNWR - larger total cylinder volume. But the long boiler was still a problem.

 

Wilson Worsdell managed to fit 20" diameter outside cylinders - about the limit within the loading gauge - to his Class T 0-8-0s, which I think were the next built in quantity. 

[DenysW]

1 hour ago, Compound2632 said:

42 ft diameter on the Midland and I think typical generally - but an eight-wheeler would be a bit of a squeeze. 

I was more thinking 36" wheels, and the (much later) L&YR 0-8-0s had the same wheelbase as their 0-6-0s.

 

Perhaps another was to look at it is "If I was designing a mineral engine to drag wagons at 15 mph would I go small-wheel 0-8-0 or medium wheel 0-6-0?"

[Compound2632]

6 minutes ago, DenysW said:

I was more thinking 36" wheels

 

There was a balance to be stuck between tractive effort and piston speed. bearing lubrication, etc. Plodding along at 20 mph, your 3 ft 0-8-0 is going like a high-stepping 7 ft-er at 50 mph, in terms of piston speed and axle rotation rate, whereas a 4 ft 9 in 0-6-0 is equivalent to your high-stepper at 30 mph.

Edited October 20, 2023 by Compound2632

[The Johnster]

A Victorian 'goods' 0-6-0 was like Thomas, a Really Useful Engine.  'Goods' is not a particularly apt designation for them though, as they were very much the mixed-traffic locos of their day, happy on passenger, parcels, perishables, up to about 60mph.  Their driving wheels were of the order of 5'-5'6" diameter, and with the lower pitched boilers of 19th-century practice ran steadily at speed while being adequate for most freight work. 

 

The heavy haulers were 'mineral' engines, usually using components standard with the 'goods' types, but with smaller driving wheels for better T.E., more like 4'6".  It was these that 'informed' the 8-coupled designs of the Edwardian era, when loads increased and it was desirable to have long coal trains without double heading.  The GN 02, GW 28xx, GC 04, NE Q6, and others were fundamentally enlarged Victorian 'mineral' engines, some with leading ponies to help support longer boilers and cylinder blocks set ahead of the leading coupled axle. 

 

Better bearing materials and lubricants enabled some of these, particularly the 02s and 28xx, to reliably run fast enough to be used on faster goods work, which led to the later development of the Stanier 8F and Thompson 01. 

 

 

The GW might provide a general example of how the transition from 19th to 20th century traffic was handled.  Heavy late Victorian freight work was done by the Dean Goods, a typical Victorian 'goods' loco, with backup from the 2721 and 1884 saddle tanks, but the turn of the century suggests an element of casting about for suitable bigger replacements.  The late Dean/early Churchward era saw experimenting with the Kruger and Mrs.Kruger prototypes for heavy mineral work, until the 'Aberdare' was settled on; all these engines had 'mineral' sized driving wheels.  Churchward designed the 3150 large prairies for heavy main line goods work, though they don't look like that sort of engine to us and in fact developed a niche in banking and piloting, but they did haul some lengthy transfer freights in South Wales.  Churchward sometime defied conventional wisdom in matters of driving wheel size; his initial attempt at the very successful small prairie was the 4'1" diametered 44xx, only a few built before larger drivers were brought in.  Another success, the 43xx, was an early example of a 20th century British mixed traffic engine, but could often be seen being thrashed cruelly on long freights.

 

It would be fair, I think, to say that Churchward designed the 28xx with heavy mineral haulage in mind, especially as he later produced the 47xx with 5'8" drivers for faster jobs, but that by the time Collett signed off on the 2884s, he was considering them as general purpose heavy goods engines.  The 47xx were a failure in their original form with the no.1 boiler, which couldn't provide enough steam to keep up with the demand, and I wouldn't describe the bigger boiler as a success; they had enough steam on tap now, but were barred from about half of the GW's trunk routes.  That's why so few were built.  The role that Churchward originally designed the 47xx for was successfully taken on in 1929 by Collett's rebuilt Saint, the Halls, and the even better Granges.  Those engines could pull just about any main line job on the GW.  But when it came to replacing life-expired Dean Goods, Collett came up with another 0-6-0, the 2251, which looked more modern but wasn't really much better, and had inferior route availablity. 

 

All the railways had trouble adapting their locomotive policies to the changed conditions of the 20th century, which can be summarised as major increases in the loads that had to be hauled in all classes of main line work, and some did better than others.  It can be argued, and often is, interminably, by it's fans, that the GW did better than most, and certainly Churchward produced some very forward-looking designs that were inspirational to successful later engines, but the above comments suggest that not all was plain sailing even at Swindon; mistakes were made, some never properly rectified. 

[DenysW]

32 minutes ago, Compound2632 said:

Plodding along at 20 mph, your 3 ft 0-8-0 is going like a high-stepping 7 ft-er at 50 mph,

That's actually the point. Reciprocal steam engines with Victorian-style valves had a maximum power output as actuations/second increased but effective-pressure decreased (as p 40 of the Midland thread). For a 36" wheel engine it's between 18 and 25 mph - ideal for slow, heavy trains. A 5'3" wheel gives you a broader maximum, but it's between 25 and 50 mph, better for faster work, and taking a significant hit on the power available to accelerate its loads up to its optimum working speed. Still suitable for goods, but the merchandise that paid a lot more per ton than minerals.

[DenysW]

12 minutes ago, The Johnster said:

the changed conditions of the 20th century, which can be summarised as major increases in the loads that had to be hauled in all classes of main line work

Passengers: Yes, absolutely, much heavier carriages as measured in tons/passenger due to technology/comfort changes around 1895-1900 (ish).

 

But I think that the 'changed conditions' for merchandise and minerals were actually more about efficiency as times got tighter and 'good' railways shares no longer paid much higher dividends than government bonds*. That definitely pushes the companies to trying to run the same loads in fewer, bigger trains to reduce the cost per ton-mile. There's also the impact of the working day legislation making faster freight a partial solution to legal compliance - part of the reason for, and success of, the Midland's Train Control strategy was to improve compliance with this legislation by making lengths of working days (much) more predictable.'

 

Finally, it's my belief that it took a couple of decades for the change from iron to steel boilers to trickle through to enabling much more powerful boilers in both grate and heating area and operating pressure. That then hits the tracks 1890-1910 onwards, just when needed.

 

No needs to digress into 'bad' railway shares here. Pick your own example.

[The Johnster]

Good points.  There were a lot of factors that demanded bigger engines to pull heavier loads at higher speeds around the turn of the 20th century; the population was increasing and cities were spreading rapidly into the countryside (they still are), bogie gangwayed stock eliminated the need for refreshment stops, better quality steel found it's way into wagon underframes and couplings, better lubricants allowed higher piston and wheel revolution speeds, and big marshalling yards appeared primarily for coal traffic at assembly point locations to be repeated in big coal depots in the cities.  Coal traffic was the bread-and-butter job of the railways, and had increased to the point where it was starting to be hard to find paths for the slow drags, leading to quadrupling schemes as well as the bigger engines.

 

This in an industry that prided itself on being at the cutting edge of technology, but was in fact extremely small c conservative in it's actual practice when it came to designing wagons and locos.  I worked trains in the 1970s that Stephenson would have immedieately recognised behind the loco drawhook, the coal wagons of which were not even braked from the engine.

 

4 hours ago, DenysW said:

That's actually the point. Reciprocal steam engines with Victorian-style valves had a maximum power output as actuations/second increased but effective-pressure decreased (as p 40 of the Midland thread). For a 36" wheel engine it's between 18 and 25 mph - ideal for slow, heavy trains. A 5'3" wheel gives you a broader maximum, but it's between 25 and 50 mph, better for faster work, and taking a significant hit on the power available to accelerate its loads up to its optimum working speed. Still suitable for goods, but the merchandise that paid a lot more per ton than minerals.

 

'Range' is a factor in this as well.  Let us assume a 36" diameter driving wheel locomotive with the same boiler and engine as a 63", your example.  The 3-footer uses the same amount of steam per piston stroke, but the distance it travels during that piston stroke is much less.  So it cannot manage the same distance as the 5'3"-wheeled loco before it needs to replenish it's tanks or take on more coal.  This is a significant matter when it comes to rostering locos for jobs, and a significant reason for the popularity of the Victorian 'goods' engine.

[DenysW]

29 minutes ago, The Johnster said:

'Range' is a factor in this as well.  Let us assume a 36" diameter driving wheel locomotive with the same boiler and engine as a 63", your example.  The 3-footer uses the same amount of steam per piston stroke, but the distance it travels during that piston stroke is much less.  So it cannot manage the same distance as the 5'3"-wheeled loco before it needs to replenish it's tanks or take on more coal.  This is a significant matter when it comes to rostering locos for jobs, and a significant reason for the popularity of the Victorian 'goods' engine.

At the risk of starting a discussion/argument I'll lose, range is probably a problem for express passenger services but only express passenger services. In the 1860s (the numbers are in Bradshaw's Shareholders Guides) the average passenger train travelled 21.5 miles, the average goods/mineral train travelled 33.5 miles (total miles divided by number of trains, as declared to the Board of Trade). The average Midland passenger paid 12-16 d (1868-1912), so travelled 12-16 miles/journey at 1d/mile in 3rd class. Presumably the same again to get back. Less if you allow for some of the revenue going in 1st class.

 

I've also looked at the LYR passenger scheduling for 1922 (Bob Mills, published by the L&YR Society), and almost none of their locomotives exceeded 200 miles in a day, and there were often re-coaling stops included in a locomotive's daily turn if it was towards the higher end of this.

 

At the other end of the time spectrum, LMS Garratts achieved about 13 mph start-destination, in part because they had several stops for water and/or coal and/or engine checks, just between Toton and Brent.

 

So yes, if you are trying to get a fast steam locomotive from St. Pancras to Trent as its first stop, range is an issue. If you are scheduling a normal train's turn for an average distance, not so.

[Compound2632]

1 hour ago, The Johnster said:

'Range' is a factor in this as well.  Let us assume a 36" diameter driving wheel locomotive with the same boiler and engine as a 63", your example.  The 3-footer uses the same amount of steam per piston stroke, but the distance it travels during that piston stroke is much less.  So it cannot manage the same distance as the 5'3"-wheeled loco before it needs to replenish it's tanks or take on more coal. 

 

Not quite true. The same amount of energy is needed to move the train from A to B in a given time. Because the 36" engine moves a shorter distance for each cycle of the cylinders than does the 63" engine, it uses less energy from the steam each cycle - i.e. less steam is admitted to the cylinders. So in principle it should use the same amount of coal and water.

[DenysW]

I think I agree with both of you, as follows: At 15 mph the 3-footer will be putting out more power because its cylinders are being used more per second. So it will consume more coal and water because it can use that extra power to haul a heavier train. So it will have less range because it's doing more work. At 50 mph the 3-footer can't get steam to its cylinders because the valves are too slow and the effective pressure drops and now the 63" wheels let you pull a heavier train.

 

I've just checked and the Edwardian L&YR 0-8-0s had 54" wheels and squeezed them into a 16'5" rigid wheelbase. So they were thinking turntables, track curvature, etc., as per @Compound2632's comment.

 

Edited October 21, 2023 by DenysW

[Jeremy Cumberland]

1 hour ago, DenysW said:

I think I agree with both of you, as follows: At 15 mph the 3-footer will be putting out more power because its cylinders are being used more per second. So it will consume more coal and water because it can use that extra power to haul a heavier train. So it will have less range because it's doing more work. At 50 mph the 3-footer can't get steam to its cylinders because the valves are too slow and the effective pressure drops and now the 63" wheels let you pull a heavier train.

 

I've just checked and the Edwardian L&YR 0-8-0s had 54" wheels and squeezed them into a 16'5" rigid wheelbase. So they were thinking turntables, track curvature, etc., as per @Compound2632's comment.

 

The 3-footer can put out more power at the same speed. That is, it can haul more, or acccelerate quicker, or go faster up hills. If the load and speed is the same for the two locomotives then both locomotives put out the same power, and the only difference in coal consumed comes from internal losses. This includes moving the pistons faster, but I imagine this is dwarfed by the front end design and how the engine is driven, which could favour either engine. Of course, if one locomotive is heavier than the other, then it will need more coal simply to move itself around.

 

One thing worth remembering is that although the 3-footer can put out more power, if it only has the same boiler as the larger-wheeled locomotive, it won't be able to do so for long. You can have as powerful as you like (within reason) if you only need it for short bursts, but if you need it over a sustained period of time, you need a bigger boiler and firebox. It is this, primarily, that drove the need for more wheels at the end of the nineteenth century - locomotives became heavier and longer.

 

If 0-6-0s could manage the trains that ran in the 1850s, what need was there for 0-8-0s, with their increased mechanical losses and maintenance costs? Was there really a need for longer/heavier trains? As has already been said, small wheels won't get you faster trains, well, not unless they had previously been very slow indeed. If a railway had been struggling to get a mineral train up to 10 mph with a larger-wheeled 0-6-0, then they might well have looked at a smaller wheeled 0-6-0, but there seems no need to go to an 0-8-0 unless they were going to use a bigger boiler.

[DenysW]

9 hours ago, Jeremy Cumberland said:

If the load and speed is the same for the two locomotives then both locomotives put out the same power,

Not so, I'm afraid, although it took me a long time to realise this - at the maximum power vs. speed a locomotive can deliver.

 

Possibility 1: the cylinders are limiting. Coming back to basics, the amount of work done by a reciprocating steam engine is set by the volume of the cylinders, the effective pressure achieved in the cylinders, and the number of times the cylinders are actuated per time (from the speed & wheel diameter).

 

As the speed increases the number of actuations increases linearly, but the effective pressure decreases non-linearly as a function of the piston speed.

 

So a 36" wheel will deliver twice as many actuations as a 72" wheel and, potentially, deliver twice as much power as a result - at low speed. BUT it suffers more drastically from a loss in effective pressure, so delivers less power at high speed.

 

Possibility 2: the boiler is limiting. In this case the maximum hp that can be put out is set by the ability to convert heat into steam, and the locomotive will slow down as the pressure drops until steady state is achieved. In this case wheel size doesn't matter much because the cylinders are starved in both cases.

9 hours ago, Jeremy Cumberland said:

 

If 0-6-0s could manage the trains that ran in the 1850s, what need was there for 0-8-0s, with their increased mechanical losses and maintenance costs?

I'm confused by this assertion of increased costs. Please support it. Mechanical losses: one extra set of bearings on the extra axle?

 

I actually agree for the case that 36" wheel 0-6-0s could do the work required. But if you need a bigger/heavier boiler to supply bigger cylinders to do more work and yet still want to keep at 14 tons/axle (a typical 1850/60s value), then 0-8-0s (warts and all) look like a good possibility for slow-speed minerals work.

[Jeremy Cumberland]

2 hours ago, DenysW said:

Not so, I'm afraid, although it took me a long time to realise this - at the maximum power vs. speed a locomotive can deliver.

I wasn't talking about maximum power, but the power output of two locomotives pulling the same load at the same speed.

For a train on level track at constant speed, the power requirement is simply the sum of all the drag forces multiplied by the speed. In SI units you don't even need a conversion factor: the drag in kilonewtons multiplied by the speed in metres per second equals the power in kilowatts.

 

If the locomotive puts out more power than this, then the train will accelerate. Both the small-wheeled and the large-wheeled locomotive will be capable of accelerating the train, and the small-wheeled locomotive will, if pushed to its limit, be capable of accelerating the train faster than the large-wheeled locomotive, but how much acceleration do you actually want?

 

The power output of a steam locomotive depends on many things. In the long term, meaning anything much over 15 to 30 minutes, the power output is almost always limited by the boiler and the quality of the fuel (and the fireman), but for short durations you can mortgage the boiler, and the absolute limit depends on the boiler pressure, the steam circuit design (which affects the steam chest pressure), the number of pistons and piston diameter, the piston speed and the wheel/rail adhesion. A smaller-wheeled locomotive will, at a given speed, have a faster piston speed and therefore a higher maximum power output (until it slips, at any rate), but what are you going to do with that power?

 

Going back to our two identical trains travelling at the same speed, both locomotives will be using steam at much the same rate, in terms of kg per second (or however you care to measure it), but the smaller-wheeled locomotive will be using less per piston stroke. It will therefore be working at either a lower steam chest pressure or at a shorter cut-off. Or perhaps not, for it will need more steam to accelerate the mass of the pistons and connecting rods than the larger-wheeled locomotive.

 

2 hours ago, DenysW said:

I'm confused by this assertion of increased costs. Please support it. Mechanical losses: one extra set of bearings on the extra axle?

It's more the imperfect alignment than the six extra bearings, as such (two axleboxes, two coupling rod ends and two knuckle joints). Each wheel is free to move up and down, and the knuckle joints allow for this to some extent, but they don't allow the coupling rods to stretch, which is what the geometry requires. Just as in our models, there needs to be some sloppiness to prevent binding, and this in turn leads to increased wear. The more coupled wheels there are, the stiffer an engine is to move, and this is one reason for the renaissance of singles in the late nineteenth century, because they were "free running".

[Compound2632]

32 minutes ago, Jeremy Cumberland said:

the renaissance of singles in the late nineteenth century, because they were "free running".

 

Single drivers plus piston valves plus 170 psi boiler pressure = 90 mph. 

 

 

(Plus a long 1:200 down-grade!)

 

Edited October 21, 2023 by Compound2632