Bluebell Motive Power

That all makes eminent sense; but why then did the GWR and some others not see it that way? What was their rationale for preferring a single glass and try cocks?

 

I wish I knew! Googling photos of GWR Cabs (https://www.google.com/search?rlz=1...biw=1504&bih=852&dpr=1.5#imgrc=N72zf4VIdJ50FM) and it looks the space on the driver's side of the cab where one would put the 2nd gauge glass is occupied by the brake and piping for that and the hydrostatic lubricator. So maybe they just decided there wasn't room for 2 gauge glasses given how they were laying out driver's controls.

Of course that then begs the question of why not move/change the brake handle to create room, especially as the brake handle is so far inside the cab it's difficult to use it while also looking out of the window (for example, to check the position of the train relative to points / platform before bringing the train to a stand).

 

All I've ever seen from contemporary sources is that a second glass was never thought necessary. I've noted from GWR memoirs that some drivers would deliberately break a gauge glass to train firemen in what to do in the event of a failure.

It is sometimes said that engineering perfection is reached when there's nothing left to take away. It seems to me that a second glass adds an extra set of failure points. All else being equal if you double the number of gauge glasses you double the number of failures.

Seems to me it's just one of those engineering decisions where there are pros and cons either way. The GWR certainly had no problem with being the odd man out in such circumstances.

 

Incidentally I note from the report of the S160 incident on the LNER that the GW produced an indicator disc to show whether the valve was fully open, and that it was being fitted to all S160s. What are the arrangements on the ones in the UK now? Do they still have the screw down valves, and if so do they now have an indicator for fully open?

 

The one that visited the Bluebell at least had them replaced with a more conventional pattern:



Tom

 

In case anyone is interested, this is an annotated image of an S160 backhead as initially built (there were various developments over the construction period). The reflex gauge is obscured by its lamp. I think a contributory factor to the UK accidents was the assumption then any driver/fireman was automatically qualified to operate any steam loco. In "Over Here" for instance there is a contribution from a fireman at Colwick who was convinced that the reflex gauge contained mercury as a proxy for the water level.

 

The Weekly Round Up email newsletter shows 34059 now having had its boiler fitted again after having it removed for smokebox to be reattached properly as well as the ashpan.

 

Whilst other Western European countries got rid of their S160's fairly quickly the Italian FS had to keep them going longer due to a shortage of other motive power. Whilst the low water incidents were the cause of the firebox crown collapses a contributory cause was that the firebox sides, crown and backplate were only 3/8" thick and the coarsely threaded crown stays only engaged by 5 threads so, an overheated crown could pull over the stay heads. The FS lowered the B.P. from 225 to 200psi, fitted a better type of fusible plug and installed extra mudholes/washout plugs. They also fitted a plate over the bottom of the gauge glass so that there was a minimum 1" of water over the crown when the water was seen in the bottom of the glass. As the stays came up for renewal they were replaced with ones at a finer pitch so that 12 threads were engaged in the plate. Finally, as the fireboxes came up for renewal the plate thickness was increased from 3/8" to "just under 1/2" (12mm?). Ref: Rich Tourret's USATC locomotives.
Ray.

 

So by the same logic having two injectors is unnecessary

Peter

 

I’m afraid your logic is flawed. Doubling the number of any mechanical device does not double the instances of failure. Looking at it the other way, if you are on a two engined aircraft and one engine fails does that reduce your chances of landing safely by half? The answer is no, the chances of a failure are vanishingly small for both engines.
It’s just prudent to have a backup of important components if anything goes wrong, it’s why you have two kidneys and lungs.

 

False equivalence. The chances of an engine failure are doubled with two engines (assuming equal reliability), but provided the aircraft can remain airborne on one engine then the consequences of that failure are far smaller.

Same again. Reliability and consequences. AIUI injectors are not especially reliable and consequence of failure very great, hence two. We've already established that water gauge problems resulting in critical water shortage for UK designed locomotives was not a problem in the 20thC and there were always try cocks, so the situation is quite different to injectors.

Not every control or device on a typical steam locomotive is duplicated. There's only one reversing lever, one regulator and so on. Whether to duplicate or even triplicate equipment is an engineering decision, hopefully rationally based on reliability, consequences etc. For injectors I believe there was universal consensus by the 20thC that two was correct, for other equipment similar consensus that one was correct. Clearly when it comes to provision of gauge glasses there was no consensus as to whether two or one was the best option as is demonstrated by the single glass on the S160, so it's a reasonable presumption that there were rational arguments for choosing one or choosing two. I made my own guess as to what that might be. My guess may be complete nonsense. I have no problem with being told that. But to claim that one glass is definitively wrong and two glasses definitively right in every circumstance would seem to fly in the face of the evidence.

 

A comparison to the as built S160 backhead is seen here

This is the S160 preserved in the Italian Railway Museum at Pietrarsa near Naples. The loco is FS 736.114 (ex USATC 3671). The US style water gauge has been changed for a more conventional one with a GW style linked shut off cocks. It does however retain the original tri cocks, seen here hiding behind the regulator quadrant. It has also been converted to burn oil, its tender being on the adjacent road.

Peter (with apologies for going so far off topic)

 

12 threads in 3/8" is 32tpi which would give a very shallow thread. As you know 11tpi was a global standard for boiler work. Are you sure Tourret is correct?

 

There is no doubt that these engines had a lot of faults but it must be remembered they were thrown together to do a job, help win a war and then be thrown away. There was not thought that they would last in service as long as they did or be working in the 21st century

 

12tpi was also common with stays. Is the statement in Tourret a mis-interpretation of this? 3/8” was also quite common in steel boxes. The ends of the stays were also riveted over so there was more holding power than just the threads.

 

C.H. Hewson in his book on boiler explosions gives the following examples of firebox collapses in Britain. He is notably critical of the LMS practice of linking the top and bottom cocks such that the two passageways can't be independently tested; several accidents in the list might have been avoided had they been fitted with separate cocks. There is also at least one accident (on the GWR) that can be partly ascribed to only having a single glass which made a fitter's mistake go undetected; and two examples on S160s where having only a single glass made crew mistakes go undetected.

Carstairs (LMS), 10 September 1940, loco 6224 (Streamlined Coronation)
This appears to have been crew error, trying to mortgage the boiler on a heavy train so as to avoid losing time. The explosion occurred on a descent after cresting a summit, when the water level would drop, and this appears to have been their undoing.

Honeybourne (GWR), 17 November 1943, loco 2403 (USA S160 class 2-8-0)
As alluded to earlier, these locos had a single gauge glass (and try cocks) in which the upper and lower passages were opened and closed using screw type valves. The upper valve was found to be only partly open, probably on account of crew unfamiliarity, the effect of which was to give a falsely high reading in the glass. (Although Hewson doesn't mention it, I can't help thinking that had there been two gauge glasses, incorrect operation of the valves would have been more likely to give a difference in levels which would have been a warning sign that something was not right).

Subsequently instructions were published to all four mainline railways to stress that the water valves on those locos had to be "fully open" (the previous instruction being that they should be "open"). Experiments indicated that the water level could read falsely high by up to 12 inches if the valves were not correctly opened.

Thurston (LNER, GER section), 12 January 1944, loco 2363 (USA S160 class)
Again, appears to have been caused by the steam valve to the gauge glass being insufficiently opened, leading to a falsely high reading of water level. The LNER's instructions were that the driver should have tested the try cocks when taking over the loco, but failed to do so.

Again, this is an example where a gauge glass was giving a false reading, with no independent check of the level being possible because it only had one glass.

South Harrow (LNER, GCR section), 30 August 1944, loco 1707 (USA S160 class)
This was more of a mystery, with the gauge glass cocks found to be in good order after the explosion. It is possible that the crew misread the glass in the dark, thinking the boiler was full to the top nut when in fact it was empty to the bottom nut. As built, the locos had a prismatic gauge glass that although it gave a very sharp distinction between steam and water, was of a type unfamiliar to British crews. Essentially in the absence of seeing the surface of the water (i.e. steam / water boundary) they may have mistaken a solid column of steam for a solid column of water.

Hinton Admiral (SR Western section), 23 April 1945, loco 854 (Lord Nelson class)
This appears to be crew error. The crew took over the loco and failed to test the gauge glasses; what they thought was a water level over the top nut was in fact below the bottom nut. The crew they had relieved also appear to have been lax in checking the water level.

(As an aside, in all the above cases I am loathe to be too critical of crews working under the strain of wartime conditions).

Hither Green loco (SR Eastern section), 15 October 1947, loco 1572 (Wainwright C class 0-6-0)
This one was poor maintenance practice at Ashford. On entering the works three years earlier, the boiler inspector had given instructions for the repairs to be carried out, in particular to the crown girder stays. Due to a somewhat casual system of logging and checking repairs, those repairs had not been carried out and the boiler was refitted with badly corroded crown stays. This hadn't been picked up in subsequent "on shed" boiler inspections, and eventually the crown stays simply gave way even though there was adequate water level.

Lamington, BR(M), 7 March 1948, loco 46224 (Coronation class)
Interestingly, this is the same loco involved in a firebox collapse in 1940.

This is a catalogue of errors between fitters and crew, but the short story is that a minor issue on one gauge glass was incorrectly diagnosed and the crew were recommended to run with it shut as being potentially misleading. What was missed was that the real problem was on the right hand gauge glass, which had been incorrectly assembled such that, with the LMS pattern of linked cocks, when the bottom cock was open, the top one was shut, giving a misleading high water level.

Hewson (who I believe had been an LNER shed master before writing the book) was very critical of the design of LMS gauge glasses, and in particular the inability to independently test the top and bottom passages to the gauge glass - had that been possible, the misassembly of the cocks would have been easily discovered before it became a significant issue. (The reason the left hand glass had been diagnosed as faulty was also due to the linked cocks, in which the linking rod was slightly too short, such that when opened, both valves didn't quite fully open. The effect of that was that the water level was correct, but slow to move when tested, and that had been seen as a primary problem, diverting attention from the more serious problem on the other glass).

Bevois Park sidings (BR(S) Western section), 6 April 1949, loco 2557 (LBSCR E4 0-6-2T)
Another one caused by the poor state of the boiler causing a a group of stays on the right hand side sheet to fail and the firebox side to collapse even though there was adequate water. The stays had been badly corroded by poor water quality at the loco's home shed of Horsham. The loco had recently been inspected by a boiler inspector who signed it off for 6 months more use; it didn't make it.

Wheatsheaf Junction, BR(W), 22 November 1952, loco 6859 (GWR "Grange" class 4-6-0)
Another fitter error, but exacerbated by only having a single glass. The fitter had overtightened the glands when reassembling the gauge glass, with the result that - unseen - the top of the glass broke, and the rubber packings then squeezed into the gap, constricting the top steam passage. The result was as expected when there is a constriction in the top passage: the water level reads falsely high. Previous crews had reported that the return of water level was sluggish, but the cause remained undiagnosed. The crew who suffered the accident appear not to have used the try cocks, but in addition the combination of linked top and bottom cocks, and having only one glass, made what should be a simple diagnosis of a partially blocked steam passage impossible.

Bletchley, BR(M), 24 January 1962, loco 46238 (LMS Coronation)
In this case, it appears that the glass on the fireman's side was filthy and almost unreadable. The fireman therefore relied on the driver's side gauge. It seems he was deceived into thinking a glass that was completely empty was in fact completely full. LNER practice was for the gauge glass backplate to have diagonal painted lines, which refract through water and make the difference between steam and water more obvious. 46238 had plain white back plates and seemingly the fireman was deceived with no obvious reminder that what he saw as a solid column of water was in fact a solid column of steam.

My reading of the above is that though there are many different causes, for those caused by low water (i.e. ignoring the two SR examples of poor boiler maintenance) having a single glass (on the S160s and the GWR Grange); and linked cocks (on the LMS and GWR examples) made diagnosis of errors by fitters or crew much harder, and therefore contributed to the accidents.

Tom

 

Just to observe that aero engines had twin magnetos from a very early date. There was no spirit of '' we've always done it this way" to justify only using one.

 

Checked the Lima spec for the S160s the staybolts were 12 tpi (which is more nearly 5 per 3/8").

 

The second engine might get you to the crash site quicker.

 

One of the sad, but interesting, details of the Lamington incident was that the crew had reported a blow in the firebox at Carstairs and requested that a fitter from the nearby shed investigate this. He could not find the cause so the train went on its way. Hindsight suggests that the fusible plugs had already let go but this was not detected. Clearly this could have been obvious had the crew been suspicious that the water level was getting critical, but being oblivious of this, and no doubt with a heavy fire in a large box, this wasn't apparent.

Fortunately such incidents were extremely rare in the latter part of the steam era but this one does show two things. One is the limitation of fusible plugs as a warning device. The other is how long the crown can be uncovered (Lamington is about 10 miles south of Carstairs) before catastrophic failure occurs.

EDIT. I think the first incident (1940) involving 6224 occured somewhere near Lockerbie and not as far north as Carstairs.

Peter