Sir Nigel Gresley - The L.N.E.R.’s First C.M.E.

You've just explained for me why William Brown in his book on the W1 was convinced that the W1 could do what 242A1 did.

But you've missed one thing (which I had also missed, before reading today's posts) - crucially the receiver on the W1 was not as large as that on 242A1.

With a larger receiver, the W1 could have operated very similarly to the Chapelon loco.

So the reality is that Chapelon must have realised that Gresley was onto something with the W1 - and when he went to build his compound with a similar HP/LP setup, included a much larger receiver. That's speculative on my part but it cannot be pure coincidence, surely? He learned from Gresley's perceived failure.

 

Sorry, I was getting my terminology mixed up. I should have said that the inside valves lack their own eccentrics and eccentric rods but have all the rest of the components of a normal Walschaerts gear.

 

And your primary evidence is?

 

Hence the words “speculative on my part” in the section above.

 

I did add a smiley to indicate it was a light hearted post.

 

When you say "a similar HP/LP setup", what were the similarities? (forgive me, I know little about 242A1 apart from it being a three cylinder loco with a single HP cylinder). Gresley noted in his paper that he did not like only one HP cylinder as he preferred a "more even draw" on the boiler and a more even pressure in the receiver. Gresley otherwise says little about the receiver on 10000 apart from its being of 6 1/2 cu ft volume. Incidentally, the HP cylinders had the valves set to give different cut-offs at the front and rear ports to offset the effect of the reduced rear area of the piston due to the diameter of the piston (10", originally 12") being unusually small in relation to the diameter of the piston rod.

Gresley admitted in the paper that he was feeling his way by making the HP cut-off separately adjustable, as he did not know what the optimum ratio was (for a fixed relationship through a single reverser, which was the longer term objective). Wardale incidentally commented (The Red Devil p.518) that he thought compounding was poorly understood, even by Chapelon (although Chapelon presumably had a good feel for the practical issues of sizing cylinders and receivers).

 

Firstly my apologies, I have been undertaking much work for the MNLPS accounts this weekend. As I understand it, quoting from the William Brown book (page 100):

I don't know enough about compounding to say if Wardale is right or not regarding Chapelon, but this particular locomotive (10,000) remains Gresley's only four cylinder compound locomotive designed and built under the L.N.E.R. and it probably isn't wrong to say that he perhaps didn't understand it himself in the way that would have led to a more refined outcome for No.10,000. As it stands, looking at 242A1 by comparison, 10,000 could probably have achieved similar results had its receiver been to the same or similar proportions that 242A1 had.

 

I remember a well known designer in a very different field commenting to me on the difficulties of innovation. Went something like this. Supposing the status quo is 90% optimised and you have an idea for an innovation that's theoretically 10% better. 10% would seem worthwhile, but for your first prototype to be considered a success the first prototype would need to be better than 80% optimised out of the box, which is one hell of an ask!

 

I understand what a vector sum is, but not how you can have a vector sum of power.

 

I may have miswritten the quote, I will check in the morning against the book.

 

I don't think comparison with 242 A1 is particularly relevant, as a receiver with significant volume is essential for a 3-(or 2-) cylinder compound because the phasing of the steam leaving the high pressure cylinder differs from that entering the low pressure cylinder(s) whereas in a 4-cylinder compound the steam basically passes straight from the high pressure cylinder to the low pressure set at 180 degrees. Indeed, many 4-cylinder compounds operated successfully with no receiver at all, though some did not consider these as true compounds. The key factors would seem to be the point of compression in the high pressure, i.e. when the exit valve closes, and the cutoff in the low pressure; the greater the difference between these the more important the size of the receiver. The receiver volume of 10000 would effectively have been increased when the intermediate superheater was installed.
Incidentally according to the RCTS green book Gresley was originally intending 10000 to be a 3-cylinder compound (despite the quote above); the book quotes a calculation showing that to even out work between the three cylinders the low pressure cylinders would have to be over 25" diameter, surely too large for the loading gauge.

 

But that's not what seems to be described on 10,000:

That's interesting and may explain why there was a perceptible, but small, increase in overall performance when that was done.

Yes, you are correct there, that's why the three cylinder compound was ditched quite early on in favour of a four cylinder compound arrangement.

 

On 10000 the inside high pressure crank was at 180 degrees to the outside low pressure cylinder on the same side, hence the inside half expansion ling was set at 180 degrees to the drive from the outside radius rod as stated in an earlier post.

 

Ah, I understand now. Thanks Bill.

 

In other words, the same relative positions as on a GWR Star etc, good for balancing. I think I half understand now about the 135 degrees business. The vector sum of the phases of the HP cranks would be 135 degrees from each of the LP cranks; but with double acting cylinders that doesn't tell you anything useful about the variation of instantaneous power output with revolution of the wheels. With HP and LP in antiphase, the variation of power would be essentially the same as in a conventional 2-cylinder loco.

 

Well, all of this is quite interesting too. But I think the LNER might have had some misconceptions about how a 3-cylinder compound should be operated if they thought they'd need bigger outside cylinders for one. Ironically, normally I would argue that the advantage of the classic de Glehn/du Bousquet arrangement in a British context is that you can put the HP cylinders outside, which is precisely what was not done with 10000, because with a 450 psi boiler you've got a lot of room to play with on cylinder sizes.

I've borrowed a copy of La Locomotive a Vapeur and been trawling through it for compound indicator diagrams (like this one). And fundamentally, you run a 4-cylinder compound differently to a 3-cylinder one. With both systems the aim is roughly equal distribution of power between cylinders, which inherently means that a 4-cylinder does half its work on the HP end and the 3-cylinder only does a third. But despite this, nearly all French compounds including 242A1 came in with an LP:HP volume ratio somewhere between 2 and 2.5:1

You can see from the diagram that (at least on the 'transformed' locomotive, which is 3566 of 1929, the very first Chapelon rebuild), the receiver pressure is essentially static, so definitely big enough even at that time. Chapelon had plenty of other compounds to look at when deciding what to do, after all. But also, the receiver pressure is about a third of boiler pressure. This makes good sense; the LP cylinders have about twice the volume, so they only need half the pressure drop to do similar amounts of work at similar cut-off (which is how the loco is set).

Now here are indicator details for 242A1 (sorry, no diagram). Notice how the receiver pressure is about half boiler pressure. This also makes sense, each of three cylinders of similar individual volumes need to do the same amount of work, so they each need about the same pressure drop across them. It's having to be run with quite a long HP cut-off compared to the LP cut-off to make this happen, but what this also means is your two big LP cylinders have more pressure across them than in a 4-cylinder compound and do more work. Basically, if you take a volume ratio of 2:1 and assume pressure drops inversely proportional to volume, a 3-cylinder compound does work proportional to 0.5 x pressure x total cylinder volume, where a 4-cylinder does 0.444 x pressure x total cylinder volume.


I'd have to play around with steam tables and make some educated guesses to get exactly how theoretically efficient these two approaches should be, but 242A1 was not inefficient when compared to the most efficient of the lot, the 4-cylinder 141P with fixed equal cut-offs. There's also an indicator diagram in the book of a 141P being run at 10% cut-off on HP and LP cylinders...


As a final note, the reason 10000 probably benefited from the intermediate superheater was that with very high pressures, if your maximum steam temperature is limited (usually to about 400 C for lubrication and materials reasons) then increasing the pressure actually reduces total superheat because the boiling point is higher, which means expansion is more likely to bring the steam into the saturated region too early. So with a high pressure compound you split the expansion and re-heat between to get further away from the saturation region without ever having to get too hot.

 

Right - time to put something to bed. Confirmed in writing today to me by another author: the W1 was intended to be streamlined. I have now also seen primary photographic evidence to support this.

I will therefore be amending my chapter on the book and making this point clear and making a relevant citation to the parties in question.

 

Pretty important breakthrough in research, it doesn't exactly look streamlined, with the exposed motion, wheels and cylinders like that. Did the windtunnel testing influence the design of the corridor tender fitted to the A1s?

 

Something like that in my memory bank too. And an idea that wind resistance rises as the square so at 90 rather than 75 mph was much more than 3oohp - very useful as at these sort of speeds the actual effective cylinder/wheel rim power output was tailing off and at best requiring more steam per horse power on top of the higher steam rate required by the higher speed ie the advantage of streamlining is cutting in just when you need it.

There appears to be no doubt that the A4s were reckoned to use such speeds in normal running from the reaction after Bill Hoole's famous exploit with a BR standard 2-10-0 when it was proposed to limit all locomotives to the the same number in miles per hour as the wheel diameter in inches which would have meant the A4s were limited to 80 mph.

 

I think this is a really common misconception where the W1 is concerned. No idea why. Look at the locomotive carefully.

It in fact does look streamlined and is quite obviously so in a number of areas.

It is functionally streamlined. The main aims for the design were:

• To streamline air passages between the cladding to improve draughting - note the front ducts for this purpose
• Streamline the casing at the front end to improve smoke deflection (including smoke deflectors and a teardrop profile shaped chimney together with tapered sides to the smokebox and a definite curved cut out around the chimney)
• Streamline the sides of the casing and the cab for drag reduction purposes and for additional smoke deflection properties

It was Gresleys first streamlined design, and was done really early on in the development of streamlining in a scientific manner by way of wind tunnel testing.

So naturally the W1 is not as refined as later designs, where a better understanding of air flow and drag reduction as a result of these sort of developments was happening. It’s almost a book in itself.

But you can have a streamlined form that does not fully encase the locomotive and its mechanical workings in a streamlined casing.

Case in point - Staniers Coronation locos were not covered over the valve gear or driving wheels either.

No, but later testing and evaluation of the shapes did affect the design of the corridor tenders fitted to the A4s, which were smoothed in a variety of ways including removing the beading on the side sheets, curving the end of the tenders around the corridor connector and adding streamlined fairings on the top.

These are more likely to have been done as a result of the overall design process and I need to check if they were included in the wind tunnel testing (current knowledge says not).