CSB; softer centre axle loading in 0-6-0

[essdee]

Help, please, CSB buffs!

I am submitting this topic in the hope that not only will it directly help me, but the nature of the question may encourage those who have yet to explore CSB.

I have been exploring the CLAG CSB thread and the use of Alan Turner's spreadsheet to define the fulcrum layout in an unequal 0-6-0 wheeelbase, and would like some advice on the degree of 'softness' of the centre wheelset which is desirable.

Wheelbase is 24mm + 26mm, ie. the CoG will be at 2/3 mm in rear of the centre axle, weight to be 250gm nominal.

Plugging these values into Alan's spreadsheet, and aiming for broadly equal axle loads A1, A2, A3, with softer centre axle loading to avoid 'porpoising', I arrived at a solution which gave a central CoG(the red triangle) and axle loadings:

Example 1: A1 33.46%, A2 32.66%, A3 33.89%. Spring span ratio 1.349

These seem satisfyingly similar, with a slightly softer A2. Possibly not soft enough?

For a softer A2 loading I then increased the centre axle's 'c' span by 0.5mm, needing a commensurate reduction in the outer 'f' span of 0.5mm to maintain a central CoG. This gave me:

Example 2: A1 34.41%, A2 31.12%, A3 34.47%. Spring span ratio 1.405

Certainly a softer centre axle - but now too soft?

Finally, I reverted the 'c' span as for Example 1 and this time increased span 'd' by 0.5mm, with a commensurate increase in span 'f 'of 0-5mm to maintain CoG. This gave me:

Example 3: A1 33.24%, A2 31.96%, A3 34.81%. Spring span ratio 1.372

An intermediate value to the first two, although less symmetric.

My gut feeling is that Example 3 is probably the one to use, but I would appreciate comments from those of you who have had far more practical experience than I have so far. I must say that, daunting though it appeared at first, Alan's spreadsheet and its iterative interaction now seem very much more user-friendly, and I look forward to further exploration.

Many thanks in anticipation,

Best wishes

Steve

[Will L]

Steve

To get to grips with a CSB plot what you really need is the fulcrum point positions. There isn't an easy way to recreate these from the weight distribution as you can get very similar weight distribution results from a whole series of different fulcrum points some of which don't give dynamically stable results. And I just don't understand were your spring span ration comes from as on a 3 coupled chassis there are 4 variable lengths of spring to consider which I would have said isn't amenable to a simple ratio.

But back to your basic point. I've come to the conclusion that the idea of making the centre axle softer has been overplayed. The issue of weight on the centre axle is not so much that it should be softer than the outer two, more that it shouldn't be much harder. Given that fulcrum positions are not normally specified with an accuracy greater than 0.5mm, and that an asymmetric chassis never works out exactly even, you go for font and back as equal as possible and the middle one as close to that as you can get it but erring on the slightly softer side.

It's hard to judge without the fulcrum points but assuming there isn't something odd about them, I would have been happy with the weight distribution given by your first set.

[billbedford]

There has been so much written about obtaining some sort of 'optimum' placement of CSB anchors that I'm not surprised people get confused. The idea that the centre axle should carry less weight than the others is a particularly pernicious form of obfuscation by precision. Look though any book that shows full size loco diagrams and check the axle weight of any six coupled loco and you will find that that majority have more weight on the centre axle than on either of the other two axles, sometime very much more. Part of this will be due to the unsprung weight of the crank axle and motion, but not all.

I suggest that if the OP builds three models, one with each of the CSB schemes he has outlined, he will not be able to see any difference in the running of his models.

[essdee]

Will, Bill,

Excellent, gents - thanks for the prompt and reassuring replies. You have actually confirmed what I hoped and 'instinctively' suspected, for a practical application,as opposed to theory.

I readily admit to being one of those 'confused and bewildered' by much of the more detailed CSB theory, and hence was pleasantly surprised by how Alan's spreadsheet responded to my initial probings. Ironically, Will, this was the reason I left out the fulcrum plots in my original posting. I wanted to establish a broad principle without too much detail, and thought that my three examples, with a single variation in each case from an initial position, provided sufficient figures for the more tentative CSB explorer, reading this, to handle at one go.

Here now, for those who would like to replicate - or extend - my experimentation, are the fulcrum spacings (mm) in my three examples:

a b c d e f Spring span ratio (c+d/eg. a+b)
Example 1: 10 11.5 12.5 16.5 9.5 12 29/21.5 = 1.349

Example 2: 10 11 13 16.5 9.5 11.5 29.5/21 = 1.405

Example 3: 10 11.5 12.5 17 9 12.5 29.5/21.5 = 1.372

Sorry about the sloppy term 'spring span ratio' Will, I should have defined it more clearly. This is the ratio of the centre axle spring span to (either of two equal) outer axle spring span, as used in Russ Elliot's worked examples on the CLAG website's 'Continuous Springy Beam section': "Effect of moving fulcrum points on springrate", to be found just over halfway down the topic. I found this a particularly useful illustration, thanks Russ.

I like Bill's suggestion that three otherwise identical models using the above examples would be indistinguishable in performance; knowing this, I could have performed exactly that experiment on a trio of Caley 0-6-0s recently built! A wasted opportunity....but at least I can now be confident of shifting fulcrum points to avoid brake hangers/motion brackets etc with more confidence.

Many thanks again, both (and Russ) for your clear explanations,

Best wishes

Steve

[essdee]

Aaaarghhhhh!

Sorry all, but my submitted last posting has somehow lost all the spacing of the figures and headings for the three examples and bunched them into strings of figures. I hope you are able to deduce what the strings of numbers mean; basically the fulcrum spacings a to f, with the working to get the ratio at the end.

John - for the future, how do I get the draft layout to translate to the submitted version, please?!

BW

Steve

[Russ Elliott]

Steve, I can't really add to what's already been said, but Bill is quite right in saying any of your three examples would provide an indistinguishable running result, and I also agree with Will about the risk of overplaying the softness of the middle axle. The small differences in the middle axle loadings in your examples are not worth worrying about. Perhaps I am in some way to blame for this emphasis on the CLAG page, which was borne out of observing a couple of early CSBs where the middle axle was far too strong - not an edifying sight! (Btw, in context, that 'far too strong' was of the order of more than 30%.) That experience led to throwing a bit of numerical investigation at the subject, which resulted later in the spreadsheets. Don't put too much store by the 'span ratio'; this is a bit of a throwback to pre-spreadsheet days, where I thought there might be some single optimum relationship between outer and inner spans - this proved not to be the case, although interestingly (at least it seemed at the time) it usually fell within the 1.3 to 1.4 range, which I think is still quite a useful 'marker'. There's probably far too much 'information' (or obfuscation, depending on one's inclination and familiarity with the subject) on the CLAG CSB page, but most of it does arise from real questions that were asked at the time, and subsequently, and are still being asked to this day. One of the original objectives of the page was to show, regardless of the subtleties and arcane discussions of middle axle loading differentials, that a chassis could be produced where the overall axle loading pattern was an order of magnitude better than a typical compensated loco.

Generalising from your original question though, there could be a difference of predeliction regarding middle axle softness, and hence pitch stability, between the following two things:

pitch-thinking.png

[Will L]

There has been so much written about obtaining some sort of 'optimum' placement of CSB anchors that I'm not surprised people get confused. The idea that the centre axle should carry less weight than the others is a particularly pernicious form of obfuscation by precision. Look though any book that shows full size loco diagrams and check the axle weight of any six coupled loco and you will find that that majority have more weight on the centre axle than on either of the other two axles, sometime very much more. Part of this will be due to the unsprung weight of the crank axle and motion, but not all.

I suggest that if the OP builds three models, one with each of the CSB schemes he has outlined, he will not be able to see any difference in the running of his models.

I certainly agree with Bill basic point. The fact that the spread sheets produce figure's with endless places of decimals shouldn't confuse anybody into believing that this level of accuracy is in any way meaningful. (see also Russ's post which turned up while I was typing this) For CSB be a useful and practical method there has to be some leeway either side of the theoretical position to allow for practicalities and a degree of variation between the plan and the actuality. E.g. when using handrail knobs (recommended) there has to be some doubt about the exact location of the fixed fulcrum point (i.e. somewhere in the approx 1mm length of the hole in a typical handrail knob).

However this does not mean we can safely just guess where they should go because the interaction between the adjacent spans isn't particularly intuitive and getting them significantly wrong can be detrimental to performance both in terms of road holding (insufficient weight on a wheel to ensure it doesn't derail easily) and puling power (optimum power demands an even weight distribution). Therefore we should see the tools as necessary to get us into right ballpark but not get too tied up with spurious claim for accuracy once you've got there.

Not sure I'm entirely with the relevance of the prototypes failure to achieve optimum axle loads, In many ways they were more constrained by practicalities than we are. E.g. they had much less control over where the loco's Centre of Gravity (CofG) ends up and equal load on each axle is entirely dependant on being able to get the CofG in the right place. Of course if you really want to subscribe to the "getting it all right" mantra to the point of trying to giving your loco something approaching the same weight distribution as the prototype, that is an option, but this is one occasion where I'm inclined to things that prototype fidelity is effectively invisible and you might as well go for the best available performance from the model.

[John McAleely]

John - for the future, how do I get the draft layout to translate to the submitted version, please?!

Tabulation is a weak spot in the BBCode used by phpBB (our forum software). We've tried to extend this in the past, but not met with success.

I think the only option now is the 'code' bbcode tag, and lots of spaces:

Code: Select all

1      2     3
a      b     c


This is not easy to edit in the provided window on the forum - it might be better done in notepad or other simple text editors, and cut & paste as needed.

I generally don't bother, and reformulate what I want to say to avoid tables. Not always possible or ideal, I'll agree.

[essdee]

Russ, Will, John,

Many thanks for the further feedback, guys - very heartening indeed. I take the point about over-worrying on theoretical accuracy when workbench accuracy will not be up to the same mark, even with great care. Russ's warning about differing pitch scenarios is a valuable reminder, although I will be OK in this case.

The last 24 hours, with the aid of Alan's spreadsheet and the prompt, clear advice from this forum, have made an enormous difference to my confidence in tackling CSB planning. I would recommend anyone half-minded to explore this suspension method, to give it a go. The water is not half as cold as it may at first appear......

Thanks for the layout tip John, I will give that a whirl next time I need to arrange data.

Best wishes

Steve

[David Thorpe]

Just a word of warning. I've just about completed an 0-6-0. Initially I used CSBs and the chassis ran beautifully. Then I added the body...... Stupidly, I hadn't got my weighting right and I'd put too much weight inside the boiler, particularly towards the front. As a result, once the body was on the chassis, the loco was front heavy, with the front drooping very noticeably lower then the rear. Unfortunately, I found it impossible to remove the weight I'd carefully stuck in the boiler, and eventually I had to abandon the CSBs and rebuild the chassis with beam compensation. Fortunately, it still seems to work well and in spite of the weight in the front the loco is now level.

DT

[Alan Turner]

As the author of one of the CSB programmes I can assure you that the figures following the decimal place are meaningless as you won't know the input parameters to that precision.

I gat really despondent of young engineers dashing of to computers and reporting factors of Safety to 3 decimal places for analyses that they had done when they didn't even know the input parameters to within 20%. Just indicated to me that they hadn't grasped the fundamental issues they were dealing with.

regards

Alan

[dal-t]

As the author of one of the CSB programmes I can assure you that the figures following the decimal place are meaningless as you won't know the input parameters to that precision.

Fairly easy matter to ensure a spreadsheet only returns (well, at least displays) whole numbers. I used to get equally frustrated with financial forecasts to the last £ when single payments could amount to 5 if not 6 figures - to me, that demonstrated a fundamental failure to grasp the nature of probability.

[billbedford]

Just a word of warning. I've just about completed an 0-6-0. Initially I used CSBs and the chassis ran beautifully. Then I added the body...... Stupidly, I hadn't got my weighting right and I'd put too much weight inside the boiler, particularly towards the front. As a result, once the body was on the chassis, the loco was front heavy, with the front drooping very noticeably lower then the rear. Unfortunately, I found it impossible to remove the weight I'd carefully stuck in the boiler, and eventually I had to abandon the CSBs and rebuild the chassis with beam compensation. Fortunately, it still seems to work well and in spite of the weight in the front the loco is now level.

While I can understand why you went to such lengths to correction loco, it might have been simpler just to stiffen the springs on the leading axle buy adding a secondary wire between the first and second anchor points.

[Will L]

Just a word of warning. I've just about completed an 0-6-0. Initially I used CSBs and the chassis ran beautifully. Then I added the body...... Stupidly, I hadn't got my weighting right and I'd put too much weight inside the boiler, particularly towards the front. As a result, once the body was on the chassis, the loco was front heavy, with the front drooping very noticeably lower then the rear. Unfortunately, I found it impossible to remove the weight I'd carefully stuck in the boiler, and eventually I had to abandon the CSBs and rebuild the chassis with beam compensation. Fortunately, it still seems to work well and in spite of the weight in the front the loco is now level.

Given that rebuilding a chassis to use a different form of suspension is hardly a trivial job, I’m surprised that worked out easier than removing some of the misplaced weight but circumstance and cases and all that sort of thing. Of course I assume your warning is meant to highlight the need to apply CSBs properly, rather than anything else.

While I can understand why you went to such lengths to correction loco, it might have been simpler just to stiffen the springs on the leading axle buy adding a secondary wire between the first and second anchor points.

Alternatively, while indulging in major chassis surgery, you could have measured where the CofG really is and used Alan spreadsheet to rework the fulcrum points to suit what you actually have. Either way you have lost the haulage performance advantages of a well balance chassis and while it may run quite nicely pulling itself, how much else will it pull?

[David Thorpe]

Given that rebuilding a chassis to use a different form of suspension is hardly a trivial job, I’m surprised that worked out easier than removing some of the misplaced weight but circumstance and cases and all that sort of thing. Of course I assume your warning is meant to highlight the need to apply CSBs properly, rather than anything else.................. Either way you have lost the haulage performance advantages of a well balance chassis and while it may run quite nicely pulling itself, how much else will it pull?

Rebuilding the chassis wasn't really very difficult. The hornblocks were in position and the chassis jig was still set up for this loco. Removing the weight, on the other hand, would have been well nigh impossible as I'd firmly epoxied some rolled up lead into the boiler, the smokebox door had been secured, and the cut out for the motor didn't leave enough room to go poking inside the boiler.

As to your other comments, yes, the purpose of my post was to warn anyone else off making the same mistake as me and to ensure that when building a CSB chassis they remember to give proper consideration to weight distribution. And while I appreciate what you say about haulage, on my small layout that is unlikely to pose a problem.

DT