Flying Scotsman

Modifying a design for new construction is a different matter from modifying an existing engine.

 

'xactly so. This is not the place for a lecture on poppet valve internal combustion engine design, but suffice to say that if you increase the bore significantly you will normally need to change the cylinder head and valve design in order to get the extra gas in and out efficiently. Its not unrelated to the A1/A3 history I suppose since long travel valve gear was also all about getting the gas (in this case steam) in and out efficiently. In either case you'll get reduced benefit from increasing the cylinder size unless the valves and valve gear are capable of supplying, and maybe even more importantly removing the extra gas capacity.

When you're boring out an engine for engineering reasons - to return the cylinders to an accurate cylindrical shape and remove the effects of wear, you aren't generally changing the swept volume enough that it will make a noticeable difference to the breathing, and in any case all you are seeking to do is to return the engine to mechanical efficiency. An extra few hundred pounds of tractive effort in around 40,000 won't really make any difference to anything except Mr Pole's publicity material! Its very unlikely that an engine would be so accurately designed that such a small change would be critical enough to overstress anything unless of course the design was already inadequate and a bad situation was made worse.

 

I think what this saga shows is the interconnected nature of designing a complex machine such as a steam locomotive. Steam supply, steam utilisation and mechanical strength to absorb the resultant forces are inter-related, and changing one may have unpredicatable effects on the others.

The historic companies had large design staffs to consider such issues, and generally design was an iterative process: increasing cylinder size in a new design showed up a weakness in steam production, so the next iteration had a bigger boiler, but that showed up frame weaknesses with the now heavier and more powerful locomotive, so the next design had strengthened frames and so on.It was only rare that there were genuinely transformational designs built in large numbers straight off the drawing board - and often such designs were fraught with difficulties.

In comparsion, in preservation we may have better analytical tools to look at stresses etc, but we don't have the large design staff to consider such matters, nor the benefit of an internal tradition of a hundred years of corporate memory about locomotive design.

So I am always wary about the siren calls from some to "improve" historic designs (most commonly to improve the draughting), since such "improvements" may have unexpected deleterious consequences that would take a lot of analysis and experience to work through. Far more sensible in my view to accept such machines for what they are, and just accept that sometimes they are flawed or otherwise have sub-optimal design. You can turn an A3 into a non-streamlined A4 - but at what mechanical cost?

Tom

 

That's precisely the point though - you can't.

The two locomotive classes might share a number of similar parts and the same basic wheelbase, but few other parts are directly interchangeable without modification. Not even the obvious parts such as the driving wheels and cylinders are interchangeable due to the differing balancing and the steam chest/steam circuit design which differs one to the other.

The A4 boilers share the same basic footprint but needed modifications to be fitted to an A3 and particularly where the cladding was concerned, for example (bearing in mind that the A4 boilers were intended for streamlining not the traditional cladding style of the A3s).

It's interesting to note that, for all of the criticism of Thompson's A1/1, by utilising Gresley's A4 frame design and boiler (the original frames went back into the pool of spares) 4470 did not suffer the same level of frame fatigue that it had done so as an A1 or A10. Peppercorn too in his designs beefed up the main stays as per Thompson on his machines. Where Thompson's Pacific design suffered most was at the steam circuit end, where on the A2/2s the frames were split into two sections and this probably exacerbated the overall problem of a high tractive effort through the cylinders, onto a shorter wheelbase with decreased adhesion and thus perhaps a lot more potential for damage at the front end. The A2/1 and A2/3 utilised new frames from the start and, as far as my research allows (accepted not perfect) never suffered quite as badly as the P2 rebuilds.

It's interesting to note that the classes which suffered the worst frame cracking seem to be Gresley's A1s/A10s/A3s and Thompson's A2/2s, for different reasons.

The Gresley locomotives had the distinct advantage of having a larger pool of spares by virtue of being a larger class, so perhaps their weaknesses are less known because in workshop terms, it didn't matter as much because they always had parts available.

The A4s never seemed - at least in the research I have done anyway - to have suffered the frame cracking problem anywhere like the other classes did, despite Doncaster's fondness for ample tolerances and a lack of optical alignment until the 50s, and this was down to a different arrangement of the main stays in the frames. It's the one major engineering different between the two Gresley Pacific classes so it must have had some advantages.

I find the whole thing fascinating, because it seems most of the problems could have been solved by optically aligning the frames in the first place. What was the problem with Scotsman's frames when examined this time around? Alignment amongst other details.

There's no substitute for studying history, applying the best engineering techniques and the knowledge of others experienced in these matters. Scotsman's finally had all of that benefit and hopefully will continue to now.

 

Of course there's another interesting development/upgrade example in your shed at the Bluebell. What's now 9017 started life with an unsuperheated round top box 160psi boiler and 18*26 cylinders with slide valves for a TE of around 17,000. By the 1930s she had a superheated, belpaire box, 180psi boiler, piston valves and TE in the 19,000 region, and much patched frames that were approaching end of life. She then got a superior design of slightly younger frames from a heavier locomotive that were still much reinforced, and even those have required a lot of work recently, although may well be inevitable corrosion and old age more than an overloaded design.

 

You would not expect me to agree with this don't you? Imho the design staff of the fiftees (or earlier) did not understand the fluid dynamics principles behind a proper exhaust system and choose to neglect available research. Ell for example was aware of the 1933 Young research as he used it in a lecture by Tuplin. As a consequence a large part of the now existing heritage fleet has flawed exhaust systems which do work of course but at the cost of an increased thermal load on the locomotive. I really don't see why a change resulting in less effort for the system to draught properly would have improper consequences elsewhere. An example would be appreciated.
Kind regards
Jos Koopmans

 

Looking at the frame arrangement drawings for the A1/A3 and the A4, the major difference appears to be the elimination of all, but one, of the lightening holes in the frame plates. The frame stretchers for both classes appear to be very similar, especially the cruciform section main stay.

 

The more important change was probably to the horn stays, being attached directly to the frame plates as well as to the horn blocks. A similar change was introduced on the Black Fives, which were well known for frame fractures as first built. Why this change wasn't gradually introduced to the A3s as and when new frame sets were built or damaged sets rebuilt, I don't know. Doncaster reputedly had a spare set of frames which would be sent out with the next A3 to come in, to give ample time for repairing whatever came in, so it would surely have been possible to gradually improve them as and when they came in for overhaul. I can't imagine the additional cost of the modifications would have been that significant, especially compared to the potential savings on future repairs.

There's a story that quite late in the day, somebody suggested that the Great Eastern main line was finally of a standard that it could take A10s but not A3s. Unfortunately, by the time the scheme got any momentum they'd all been rebuilt and were too heavy again.

 

When the original Gresley Pacifics were built some flexibility in the frames was desired since this was seen as useful in helping to negotiate curves. There was also the desire to save some weight, hence the four 1' 7" deep lightening holes, one on each side of the leading coupled axle gap (or horn gap if you prefer) and two between the driving and trailing gaps. There was also a 1' diameter hole just ahead of the cylinders. the first indication of trouble came in May 1925 and by March 1929 Gresley agreed that all lightening holes should be reduced to 1' diameter for new construction. The reinforcement of the gap between driven and trailing coupled axle began in May 1925 and the frame stay in this area was also strengthened. 1931 and cracks leading from the leading gap to the leading hole were becoming a problem and reinforcing plates were used around the leading gap and also partially covering the two leading holes. April 1932 and the practice of fitting new three quarter frames came about. Four sets of frames were produced making use of material from A1s that had received completely new frames. The salvaged frames were cut ahead of the trailing gap and new three quarter frames were welded to the sound rear portion. The inside face of the weld was reinforced with plates that were cut with holes to match the lightening holes in the frames proper. In the October of 1933 Gresley decided that the only lightening holes should be those ahead of the cylinders and this change was first seen in 2500 when it was new and older Pacifics were brought into line when their frames were repaired. With thanks to Part 2A.
The A4 had a single 1' lightening hole directly behind the driving horn gap. The rear section was modified in order to avoid having to dish the frames in order to clear the boss on the radial truck. The important change was the horn stay design which now bolted to both the hornblock and the frames.
The LNER was not a wealthy company. The acquisition of machinery that permitted locomotives to be assembled to the standard of accuracy that was desired had to wait until after WW2 and nationalisation.
You could improve an A3. The 94A was exchanged for the 107 but with the A4 boiler working at 220 psi. it needs some new cladding sheets and that is about it. Keep the working pressure at 250psi, fit the double Kylchap exhaust, enlarge the main steampipes and return the cylinder bore to 20". You get a very good class 8 but you have to do some work on the frames, A4 type hornstays, extra frame bracing (the leading hornblocks being produced as a single casting?).
You can go further, making an A3 equal an A4 is not the greatest task. 4472 came very close to equalling the best efforts of 46229.

 

You know far more about this stuff than I, so it's pretty foolish of me to argue, but improvements in draughting to some fairly recent designs in the 50s produced sometimes 50% or more increase in max steaming rate. It seems reasonable that at least similar improvements could be made to older designs.

Even if the peak power output doesn't change, wouldn't that mean that peak power could be sustained for very much longer? Isn't it feasible that such great improvements in power delivery could show up previously insignificant inadequacy in design and perhaps especially lubrication elsewhere in the locomotive with unfortunate results?

 

Improving draughting cannot physically damage a loco unless you start raising the boiler pressure or enlarging the cylinders and thereby produce more power. What it will do is improve the combustion and hence steaming rate and allow for poorer coal and or firing technique. Also a reduction in clag will save fuel and hence cost.

 

Improved draughting reduces back pressure and so allows more power to be developed in the cylinders. It also improves the maximum evaporation by more effectively driving combustion.

So you have situation where the locomotive delivers a far higher maximum power output than it could in pre modified state. When the A3 class was fitted with the double Kylchap exhaust system the increase in horsepower that was deliverable was around 500 and the frames began to pay the price. If only the hornstays had been attended to and maybe the wedges "Franklinised".

The higher vacuum levels achieved by the improved system increase carry over of unburned particles from the firebox and the tube ends pay their price. This can be dealt with by changing the combustion system.

An improved exhaust system can effect so much of an improvement on a locomotive that it can indeed damage itself.

 

yup

 

Which was the sort of point I was getting at. Any improvement to a locomotive that increases the steaming rate, or improves the distribution of steam, must almost by definition open up the possibility of increasing the mechanical strains on the locomotive. Such optimisation may well have been commercially justified in the 1950s, but I would suggest is not historically justified now just for its own sake.

I guess where I am probably at variance with some is that I can understand that there is an interesting academic and historical argument to be had about enhancing existing designs, as well as trying to understand why such enhancements were not necessarily made back in the day. But that doesn't to me automatically lead on to a desire to actually modify existing historical relics in practice just to prove the theories. As a simple device for efficiently moving people from A to B, steam locomotives are no longer the optimal choice: if you want to do that efficiently within the constraints of modern legislation and working practices, use an electric or diesel loco. So instead, I'd rather enjoy the locos we have for what they are: ancient historic relics that betray the flaws in their design but are all the more interesting for it. After all, no-one would seriously consider restoring a Mark 1 Spitfire and then bolting on a Griffon engine (even though such a machine would undoubtedly be faster than the original - while it lasted). So why do metaphorically the same thing with historic steam engines?

Tom

 

I think you are selectively quoting me - though I suspect we are possibly in violent agreement...

I wasn't meaning that you could literally turn an A3 into an A4 just with a bigger boiler running at higher pressure. Rather, that carrying out such a modification figuratively was taking the loco in that direction. If indeed you bore out the cylinders, add a larger and heavier boiler and increase the boiler pressure but without making all the other design modifications between the two classes (especially in respect of the frames) and then operate the loco up to the limits of its new enhanced form, then maybe it shouldn't be surprising when such modifications start to highlight previously dormant issues with mechanical reliability!

Tom

 

Oh - so should the double Kylchap now be removed in the best long term interest of the loco and it's frames? I live in hope

 

There you have it. Alan Pegler was right, returning the sole surviving A3 back to being a typical LNE version of the class in the 1930s was the best way of ensuring its long term working survival. Actually the engine never looked right to Mr Pegler in double chimney guise even though he admired the advantages of the system and this was the main driver for the decision to make the reversion. Unintended consequence - more life was extracted from the frame structure to be traded against a lower performance level. Actually the performance level cannot have been that low - the engine was sabotaged while running on the GW, things had moved on from 1925 and even then the LNE engine had a horsepower advantage. By the 1960s the old A3 was a formidable challenger to established performance records on the Western. Someone took action to ensure that these records continued to stand.

Fully returning the engine to 1930s condition, vacuum brakes and all, with air brake equipment being an extra interfering as little as possible with the the design of the original would be best. That could be next stage in the work done in 7 years or so. But I expect to be disappointed.

 

I seriously doubt if there is any thing left in 4472 which is "original" and if you want to return it to "1930's condition" (an A1 as it was until about 1947) it would necessitate a huge number of modifications including conversion to right hand drive. If you want an accurate A1 then you'll have to build a replica. I think the NRM has done the right thing in restoring it to its late B.R. condition - the only truly accurate form in which it can be. We owe a lot to Alan Pegler for saving 4472 but he did not return "... the sole surviving A3 back to being a typical LNE version of the class in the 1930's" - he simply had it converted to single chimney form and painted it apple green which did not make it an A1 - in fact it only made it a mongrel.
Ray.

 

Perhaps the fitting of effective spark arrestors might seriously impede a 1930's esque A3, that's not to say some form of single chimney might be adequate...