LMS 2F Experiences

[Bryan Cedar 1Participant @bryancedar1]

Has any ME member of the Forum any experience or comment on the "Twin Sisters" LMS 2F

I am considering acquiring one. The ash grate and tray are difficult to manage and I would like to know the best way to deal with this issue. The version is the fully detailed one.

Does any member run one ?

[Bryan Cedar 1Participant @bryancedar1]

“Twin Sisters” 0-6-0 Tank

[julian atkinsParticipant @julianatkins58923]

Hi Bryan,

Designed by Jack Austen Walton from the late 1940s over a very extended period described in ME. The boiler designed by Jack was a very poor dangerous design. Lots of features departed from LBSC standards just for the sake of it apparently, and you have already noted the problem with the grate and ashpan over the rear axle.

If made to drawings (the boiler design was quickly altered) I would give it a miss. If the builder has made substantial improvements then it might be worth a punt.

The Kennions/GLR 'Butch' is very very similar and there is one in my workshop at the moment for repair and is perhaps the most badly designed miniature loco I have ever encountered.

I have seen lots of 'Twin Sisters' over the years on display but never one running. I dont ever recall the design being held in high regard.

I have always taken the view that what is in essence a shunting loco is unsuitable for passenger hauling in miniature simply due to the boiler proportions (small grate area) and you end up having to force the firebed resulting in clinkering etc.

Just my personal opinion of course.

Cheers,

Julian

[Bryan Cedar 1Participant @bryancedar1]

Hello Julian

Many thanks for your reply. Not very encouraging but most helpful. The model I was considering IS to the published design. As such I gather that you advise AVOID. Can you elaborate on "(the boiler design was quickly altered)"

Interesting that Kennions continue to market castings, drawings etc.

There seem to be few of these locos coming to the market.

Thanks again, Regards

Bryan

[julian atkinsParticipant @julianatkins58923]

Hi Bryan,

I doubt the boiler would be built to the original design as it was seriously flawed and quickly altered.

Externally and superficially it is a nice compact loco.

But it is an old design by a long forgotten model engineer, and the construction series was not a great success.

GLR/Kennions continue to market it because it shares the same castings as their 'Butch'.

If the loco you intend to purchase is 'in ticket' then have a drive of same before purchase. Or else get someone experienced to have a drive while you watch on behind on the driving trolley.

If it doesnt have a current boiler ticket and paper trail of previous tickets then that opens a different can of worms.

Cheers,

Julian

[John Ockleshaw 1Participant @johnockleshaw1]

Hello Bryan,

I tried to build the original boiler, none of the inner firebox plates were flanged, as it cooled during silver soldering the butt joints opened up.

Subsequently I built a gas fired boiler with the same outside dimensions, which is very successful.

I filled the side tanks with steel ballast to give it more adhesion.

The four wheel riding car is made from four lengths of steel tube which act as a gas tank, it carries a water tank as well as the driver.

The loco drives and hauls well.

Regards, John

[Peter LayfieldParticipant @peterlayfield]

Yes the boiler plates were not flanged, but nobody would build it like that ,The model was built to a scale

of 1" to 1ft, not the usual 1 1/16" which is the norm for 5" gauge , but other wise a nice little model

[MEinThailandParticipant @meinthailand]

There is an interesting discussion about the need to flange copper boilers on this forum and the Twin Sisters flangless boiler design is referenced and discussed. Also on the same thread in that forum, calculations are presented that prove, from a structural strength perspective that flanges are not needed at silver soldered copper boiler joints.

In fact a T joint without flanges (such as at the smokebox tubeplate) only has to be 18% as strong as the copper it is joining.

Flangeless Boiler Design Vindicated?​

Well, to my surprise (April 2025) I recently came across and example of this (5″ Gauge LMS 2F dock tank Twin Sisters) boiler built to the original flangeless design for sale (as part of a loco) for sale (now sold) on a reputable website:-

https://www.steamworkshop.co.uk/portfolio/5-lms-2f-3/

Quote: The boiler is of course also brand new and made to the published design,….. and fully silver soldered copper construction.

 

[noel shelleyParticipant @noelshelley55608]

Julian, if I may digress slightly, you mention the similarity with the kennions design Butch. You consider it to be badly designed please could you elaborate as I have one, I also have the drawings and would like to know what might be altered to improve it ? Noel.

[Nigel Graham 2Participant @nigelgraham2]

Examining the ‘Steam Workshop’ advertising, I ask…

Built to which published design: original or amended?

The boiler – was it built commercially or privately? Either way, was that locomotive sold with the boiler’s test papers? No mention in the ad. I searched the web-site further and still found no mention of test certificates or lack thereof, which I thought odd; but I can’t imagine a company run by such luminaries as Doug Hewson would ignore this vital point.

Presumably Steam Workshop either bought the locomotive for re-sale, or was selling it on behalf of the builder (or, we must consider, perhaps an inheriting family).

…

Turning to “this forum”…

Close examination of the drawing shows the joints are slightly better than first suggested, but not much. The barrel is turned to create little rebates that give positive locations and increase the joint areas very slightly. Nevertheless, it is a fundamentally weak  design.

The inner firebox plates have plenty of stays but are a mere 14swg (2mm) thick! The outer plates are 1/8” – still a bit thin for quite a sizeable boiler.

I could not understand the staying. None of the rods themselves are shown on the drawing, only their locations plus a mysterious instruction implying they pass through only the outer wrapper. Surely not?

Incorporating the blow-down valve in the foundation-ring? What – directly in the copper? Apart from the thread wearing rapidly, or evenn stripping, I don’t think it would weaken the ring significantly. However, because it creates an abrupt elbow, it would not be very effective as a blow-down, let alone wash-out point. It’s not clear where it discharges, either.

There are other puzzles. What are the two (tubes? rods?) arcing back from the regulator? What are the two hidden-detail (chain-dashed) lines presumably where the safety-valves would go, shown above the crown-plate?

.

Hmmm. Let us turn to the various responses below the quoted drawing from Model Engineer.

As for its strength, we now know that a properly brazed boiler has joints that are as strong as, or even stronger than the parent metal, and advantage is taken …. to ensure that the finished article will withstand at least four times its normal working pressure.

I think this is what has led our friend in Thailand astray, with his remark about the joint needing be only a fifth of the strength of the plates.

 

Four times…. well, clearly the boiler should have a good factor of safety so four times makes sense provided that is calculated-only, and no-one reads the F.O.S. as the shell test factor!

Is this why J Austen Walters’ boilers distort? His comment that tapping the metal back will “instantly” (really?) work-harden it so you can over-pressure it again, certainly ain’t in the Orange-On-White Good Book.

300 (psi I assume…) As a shell-test pressure that implies a working-pressure of 150psi: high for a copper boiler on a  5″g locomotive; and especially with such thin firebox walls. Yet that writer calmly talks of the test-pressure “reaching up to 300”, and repeating this “test”.

I think he has been quoted only partially there, leaving me wondering about “no more” noises from the boiler’s innards. Yet I don’t think he follows the M.E.L.G. Test Code, which seeks to protect innocent boilers from bad “testing” as much as protecting people from dangerous boilers.

 

Thin? Look closely at the photograph of the boiler on its own, with its front propped on the dome cover. Has the protruding rim of the barrel pushed in slightly where it rests on the dome, or is that an optical illusion?

 

I do not suppose I am alone in having serious misgivings about that particular locomotive; and I do hope the buyer will not find himself with an unuseable purchase whose provenance seems odd.

[MEinThailandParticipant @meinthailand]

Nigel, referring to your comments on the boiler drawing for the ‘Twin Sisters’ 5″ LMS 2F Dock Tank by JI Austen Walters published in Model Engineer magazine 1949, It is reassuring to know that other model engineers review and comment on proposed designs to ensure the continued safety of the hobby. I should clarify that I have not myself reviewed the design at all – I merely posted the drawing and the reference to an actual boiler said to have been built to this design for sale at Steam Workshop, as a matter of interest.

Also I would like to clarify, for the benefit of our readers, that your reference to the four times working pressure as a design target was a statement made by JI Austen Walters himself.

Coming now to the “I think this is what has led our friend in Thailand astray, with his remark about the joint needing be only a fifth of the strength of the plates”, let me clarify.

Stress calculations for a silver soldered joint applied to a disk of 1/8″ thickness silver soldered into a 2″ internal diameter pipe to determine the maximum uniform distributed pressure that the joint can withstand in sheer were made.

https://modelengineeringinthailand.com/forum/index.php?threads/bassett-lowke-traction-engine.55/post-363

The answer was 752 psi. This was compared to the yield point of the copper disk being 4,347 psi for fully annealed copper as follows:-

752/4,347 X 100 = 17.3%.

i.e. a joint strength of 17.3% of the copper yield point is OK (in theory).

Therefore I believe “our friend in Thailand” has not been led astray but has simply stated the potential validity of flangeless endplates.

[DiogenesParticipant @diogenes]

Are tensile forces at the joint (discussed in the calculation referenced above) only derived from flexure of the circular end plate – is the pressure not trying to expand the barrel away from the disc as well?

As the end of the barrel is anchored by the soldered joint at the plate, expansion wouldn’t seem able to be even – does this not concentrate maximum force on the inner edge of the joint – i.e. stress isn’t shared evenly across the full width of the soldered joint ..?

[MEinThailandParticipant @meinthailand]

Diogenes, thank you for your comment, it’s great that people are actually taking an interest in this rather difficult and rarely discussed topic of strength calculations for model steam engine boilers.

I assume you are referring to the calculations and discussions on this web page:-

Your question is perfectly valid and was actually posed as a question (albeit worded slightly differently) and subsequently answered in the post you are referring to:-

https://modelengineeringinthailand.com/forum/index.php?threads/bassett-lowke-traction-engine.55/post-363

If you look for the following, your question should be answered:-

But that is not the whole story. Because the circular plate is bending there will be a force pulling the plate away from the pipe, effectively putting the silver soldered joint in tension as well as in shear?

If I may make a suggestion, Diogenes, it would be better if comments on topics on the forum mentioned https://modelengineeringinthailand.com/forum/ were posted on that Forum so that the context is readily available for other readers. Readers of this forum will see your question and my response but since all the discussion is on the other forum they will probably not understand unless, of course, the visit the other forum.

Posting comments on this forum about a topic on another forum can make things a bit confusing.

Thank you for your understanding and for your question.

 

 

[JasonBModerator @jasonb]

I think John’s post is the one to take note of. Although a flangeless joint can be shown by calculation to be upto the job. In practical terms a flanged one is more likely to be successful as when the boiler cools the flange can flex and move a bit so you won’t get the joints opening up.

It is also a lot easier to put a couple of loose rivits or screws through a flanged joint to hold the various plates in place while you solder. Not so easy to fix into the thin edge of a bit of 16g copper. So again the joint and overall boiler are likely to come out as you want rather than things moving during soldering, coming apart on subsequent heats and/or ending up twisted and out of line.

[SillyOldDufferModerator @sillyoldduffer]

The other forum contains this sentence, my bold:

Exactly! The shear stress at the silver-soldered joint (544 psi) is significantly higher than the bending stress in the plate, meaning the joint is stronger in shear than the plate is in bending. This confirms that the joint itself will not be the limiting factor in structural integrity—the plate material will yield first before the soldered connection fails.

Isn’t this the wrong way round?  The shear stress being higher at the joint doesn’t mean the joint is stronger, it means it’s experiencing higher forces trying to tear it apart. That’s backed up by full-size experience – boilers often failed at the joints, for which reason they are beefed up.

Applying the formula to a simplified case can also be misleading.  Instead, I suggest FEA with a CAD package.   Finite Element Analysis treats the object as a whole and colour codes stress concentrations.  And can do so over a range of forces, as when the pressure in a boiler starts to rise.

Doing that shows:

• At Standard Temperature and Pressure, everything in the boiler’s structure is equally stressed, and the stresses are low.
• Initially, as pressure inside starts to rise, the boiler shell inflates like a balloon, bending out unless constrained by stays or other structure.   At first the stress at the end-plate joints will be similar to that on the plates.
• But, as the plates bend further under pressure, the bends start to crowbar the joint apart, and the stress on the joints rises faster than that on the plates.   Note the crowbar, leverage is at work, not just straight pulling.

Although the maths applied during FEA isn’t particularly difficult, it requires millions of calculations.  Briefly, the software triangulates the structure into a mesh and then calculates the stress in the sides of all the triangles – thousands of them.  The tighter the mesh, the more accurate the result, but the number of calculations needed rises rapidly.    The sheer number of sums mean humans can’t do FEA with paper and pencil – life is too short!

The computers results are usually shown graphically by overlaying the model with the colour coded mesh, so it can be seen where the model is stressed and by how much.  Bright red means high stress, down through yellow and green to blue.    Red doesn’t necessarily mean the model will break, it only identifies the most highly-stressed areas.  (* some do identify breaking – if the material is fully specified.)   Anyway, the engineer looks at red areas, and strengthens the design at that point if the stress is of concern.  He also reduces weight and material costs by slimming the design where it is over-strong.

Reinforcing isn’t the whole story.   The design is likely to include a significant safety factor, and, there’s probably a need to allow for metal fatigue too.   Another opportunity is applying FEA to the thermal characteristics of a model, an important, an non-obvious factor in boiler design.

Unfortunately, FEA has a steep learning curve.   The object has to be modelled accurately in 3D-CAD, and then the FEA analysis has to be set up.   FEA assumes the engineer understands stress, it’s skilled work.   I’ve tried 3 or 4 times and barely dented the surface.   Nonetheless, I’d recommend anyone interested in modernising boilers to put the effort in.

Flanges provide useful reinforcement and ease construction.  It suggests builders would need a very good reason for not using them.

Dave

[duncan webster 1Participant @duncanwebster1]

Modern industrial steel boilers have corner welded joints, no flanging, and I know of several loco boilers with the same. There is a book about the design of modern steel boilers by a chap called Green (I think) which is well worth a read. He dismisses flanged plates as ancient technology, and he was chief designer for a large boilermaking company, so we can take if that he knew a bit about it. Delving into FEA is not for the novice, especially with ductile materials, which give a bit under initial pressure test to dissipate local high stress. As Jason says, it is a lot easier to hold a flanged joint together with a few bronze screws or copper rivets for silver soldering than with a corner joint. We must recognise that the throat plate to barrel joint on hundreds if not thousands of copper boilers is a corner joint. I’m not aware of any that have failed

Even though I know corner joints can be OK, I’ve been involved in designing and making 3 copper boilers, they all have flanged joints because on balance I reckon it’s easier. If I had the kit and skill to do TIG welding I daresay I’d change my mind.

[JasonBModerator @jasonb]

Not just modern commercial boilers. The likes of Haining was describing butt and corner joined steel boilers for models 45-50 years ago. Though Welded steel and with proper joint prep. And we have certainly had tig welded copper model boilers about for 20years again no flanges.

Back to practicalities. the difference is they are welded, you can do a few quick tacks and that will hold things together or get as carried away as you like with chunky jigs as they will not act like heatsinks which draw the heat out of silver soldered work and previous joints are unaffected by additional welding.

The heat is also more localised so you will not see say a boiler barrel expand as much as it would if heated for soldering which reduces the risk of joints failing as the boiler cools. Even if the joints don’t fail on a silver soldered boiler it would be wise to allow some factor of safety into butt joints as stressing the cooling joint won’t do a lot of good for it’s strength.

[noel shelleyParticipant @noelshelley55608]

Any joining of metal by either soldering or welding HAS to be done with equipment suitable and powerful enough to make the joint by someone who has the skills to set up and use the equipment ! That the job was done commercially does NOT guarantee it is well done . I have found plumbing installed by professional plumbers where insufficient heat was used to flow the solder for capillary action to take place – the joint failed ! I was also involved in the catastrophic failure of a welded commercially made vehicle part. Looking at some of the questions asked on here one realises that older members of the forum grew up in a VERY different world. That we are still allowed to make a pressure vessel I find some what surprising. Hey ho, it’s dinner time. Noel.

[MEinThailandParticipant @meinthailand]

This is my response to SillyOldDuffer’s comment that he highlighted in bold:-

The shear stress at the silver-soldered joint (544 psi) is significantly higher than the bending stress in the plate, meaning the joint is stronger in shear than the plate is in bending. This confirms that the joint itself will not be the limiting factor in structural integrity—the plate material will yield first before the soldered connection fails.

That comment was put to AI (MS Copilot) who provided the original calculations and this is Copilot’s response:-

Let’s rigorously assess whether the original statement is correct.

 

Step 1: Understanding the Statement

The statement in question: “The shear stress at the silver-soldered joint (544 psi) is significantly higher than the bending stress in the plate, meaning the joint is stronger in shear than the plate is in bending.”

 

This implies that because the shear stress is numerically higher than the bending stress, the joint is stronger in shear than the plate is in bending. However, stress alone does not determine strength—failure depends on material properties and how close each component is to its failure limit.

 

Step 2: Comparing Shear and Bending Strength

Shear Strength of the Joint

The shear strength of silver solder is 6,900 psi.
The maximum shear stress in the joint is 544 psi.
The joint operates at 544 / 6,900 = 7.9% of its shear strength.

Bending Strength of the Plate

The yield strength of annealed copper is 4,830 psi.
The bending stress in the plate (using the formula) is 3P R² / 4t².
The plate operates at a fraction of its yield strength, but the exact percentage depends on the calculated bending stress.

Step 3: Evaluating the Statement

The shear stress (544 psi) is numerically higher than the bending stress, but this does not mean the joint is stronger in shear than the plate is in bending.

The joint is operating at only 7.9% of its shear strength, while the plate’s bending stress must be compared to 90% of its yield strength.

If the bending stress in the plate is closer to its failure limit than the shear stress in the joint, then the plate is actually weaker in bending than the joint is in shear.

Final Conclusion

The original statement is misleading because it compares stress values rather than failure limits. A more accurate statement would be: “Although the shear stress at the silver-soldered joint (544 psi) is numerically higher than the bending stress in the plate, the joint is operating at a much lower percentage of its shear strength compared to the plate’s bending stress relative to its yield strength. Therefore, the plate may be more vulnerable to bending failure than the joint is to shear failure.”

As I have mentioned in a previous post above, it’s a bit silly to post questions and replies on this forum relating to a discussion on another website. The question and answer are taken out of context and will be confusing for readers to follow.

I will be pleased to answer further queries and to discuss further, if queries are posed at the place where the original discussion is posted.

 

[JasonBModerator @jasonb]

Better here than for a forum with just a couple of members as it will get seen by far more people.

It was also you that brought it up in a 8 year old thread on here which has started the discussion again.

A read of ME4651 may be useful, this image taken from it. Note that the endplates are soldered onto the tube not into it.

[JasonBModerator @jasonb]

[duncan webster 1Participant @duncanwebster1]

Well that’s pretty convincing.

[Author]

Posts