Abstruse CSB Theory

User avatar
Russ Elliott
Posts: 930
Joined: Thu Jun 02, 2011 6:38 pm

Re: Abstruse CSB Theory

Postby Russ Elliott » Fri Sep 16, 2011 12:20 pm

Will L wrote:The joy of Alan's spreadsheet is that it will allow you to specify a chassis with a CofG in any reasonable location, get it to sit level and and know what the weight distribution would be.

That's good news, Will, and when I can load Alan's spreadsheet properly, I'll want to do a few correlation checks, as you have done already, thanks.

Now, with Alan's spread sheet, if anybody badly wants a loco with the CofG well away from the chassis centre, we can now do a plot for that too.

Even though this thread title includes the words 'abstruse' and 'theory', let's try to stand back from this and question this stance from a working modeller's point of view.

Whilst it's great to have a range of spreadsheet CSB tools, I can't help wondering about the validity of promoting a constructional starting point that assumes a modeller has intentionally chosen to use a CofG well away from the chassis centre. I accept that there will be cases where constructed steam bodies might be a bit front or rear heavy, but the only obvious nasty examples that come to mind are 4-6-0s with heavy cast boilers, which are front heavy, or some 0-6-2Ts, which can be too heavy on the rear. Normally one would address this by adjusting the bogie or pony springing and support forces, a process which can be considered as quite separate from the mainframe (driving wheel) CSB. (See your previous posts on this matter, or 41.0, on applying a fairly simple principle of moments scenario.)

Let's take the example of an 0-4-2 tender loco with a CSB over the whole lot, and where the model loco body is perceived to be 'rear heavy'. The temptation of Alan's spreadsheet would (I think, not having played around with it yet!) be to intentionally slacken off the front drivers. Statically, or when running light engine, that won't matter much, but under the effect of drawbar pull, the resulting (effective) CofG will be moved back still further because of the effect of drawbar pull, and that will further unload the front driver and load up the rear driver. The situation is reversed for a 2-4-0, where the drawbar effect on CofG is likely to have a beneficial effect on the driver force for a 'front-overheavy' body, the downside being of course that the front carrying wheel could get a bit 'lively' (nodding donkey syndrome, not unknown on the prototype!). Such benefits or drawbacks are compounded, and may be exacerbated or diminished, in 0-4-2Ts and 2-4-0Ts, depending on their direction of travel. I do wonder whether all of these wildly different scenarios are likely to be thought about by a modeller who "badly wants a loco with the CofG well away from the chassis centre". The problem is with the word 'want' - are we talking about desire or necessity?

(This thorny CofG positioning and drawbar pull effect is why I personally would steer clear of a CSB that covers a non-driver. The simple springy equaliser is immune to all CofG movement, whether intentionally set in the 'wrong' place or by virtue of drawbar pull effects, and is immune to chassis horizontality. But let's leave aside my personal inclinations about the 2-axle case, and go back to the 3-driven-axle case, where the CSB is the killer technology.)

So let me ask a basic question relating to that difference between desire and necessity: why would anyone start building an 0-6-0T with the intention of putting or accepting a CofG in the 'wrong place'? (I can see the reason why the design CofG could/should be shifted to cope with drawbar pull on a tender 0-6-0 optimised for forward running.) Unlike those who can probably add symmetrical traction weighting to a near-finished loco with a 'conventional' CofG, a modeller choosing or accepting an enforced 'wrong place' CofG is unlikely to have any room for manoeuvre, because he probably won't be able to add symmetrical weight. He's burnt his boats. I think we need to reflect on this, sorry to use the jargon, 'audience-facing' question. I want to be able to sell Alan's spreadsheet (on the CLAG web page, for starters, with Alan's permission of course), but I'm kinda struggling to find the rationale, at least one that is predicated on good (best?) practice.

Please note that I am not knocking Alan's spreadsheet per se. On the contrary. It highlights the CofG issue, which is of course pertinent to any fulcrum plot. But I'm tending to think that Alan's spreadsheet might promote 'extreme CSB-ing', for want of a better phrase. Is this where we want to go?

Russ

P.S. For numbers context, typical drawbar pull CofG shift for a heavy train will be about 3mm to 4mm (on a conventional 14mm drawbar height). The CofG shift is proportional to the pull.

Alan Turner
Posts: 559
Joined: Sun Jul 20, 2008 4:24 pm

Re: Abstruse CSB Theory

Postby Alan Turner » Fri Sep 16, 2011 3:18 pm

The reason for my spreadsheet and approach is because, as a matter of principle, wheel loads cannot be an input - they are necessarily an output.

The reason my spreadsheet gives the CoG of the loco and the CoA of the sprung axles is because my output is not valid unless the two are coincident. It was simply too much complication to do otherwise and as the two need to coincide for the loco to remain level then there seemed no point in it being otherwise.

However it means that if the CoG is not in the centre, for what ever reason, this can be catered for.

It was simply not the intention to go "extreme CGing" what ever that is.

As regards Will's comment "hiding the works" - what is hidden is the stiffness matrix and its inverse. It would be little point, in my opinion, to have that on show.

I promised a release version for Scale4um. There have not been any comments on the test version so here it is.

Regards

Alan
Attachments
CSB sheet - Combined version - v1.15.xls
(704 KiB) Downloaded 271 times

User avatar
Will L
Posts: 1770
Joined: Sun Jul 20, 2008 3:54 pm

Re: Abstruse CSB Theory

Postby Will L » Fri Sep 16, 2011 8:04 pm

Russ Elliott wrote:.. why would anyone start building an 0-6-0T with the intention of putting or accepting a CofG in the 'wrong place'? ...


Russ

I think if you consider my wording you will see I'm inclined to agree. If in your enforced absence you may have missed the thread CSBs a question of Gravity in which I tried to go over this in some detail. This followed on from the Abstruse CSB Theory -In defence of the CSB Spread Sheet which appeared in this thread in July.

None the less I find it much easier to defend a position of saying that you should be ware of the impact of the locating the CofG in a given position, and why a central CofG is preferable, than having nothing better than an assumption that putting it it central on the chassis was the right thing to do.

Will

User avatar
Russ Elliott
Posts: 930
Joined: Thu Jun 02, 2011 6:38 pm

Re: Abstruse CSB Theory

Postby Russ Elliott » Sat Sep 17, 2011 12:18 pm

Alan - taking the 3-axle case and where a required deflection has been input - when moving the fulcrum points about, which axle(s) does the 'actual deflection' value refer to?

Russ

P.S. Will, yes, I had read your question of gravity thread, which covered much the same ground as 41.0, and seemed to get sidetracked down the 'which axle slips' debate.

Alan Turner
Posts: 559
Joined: Sun Jul 20, 2008 4:24 pm

Re: Abstruse CSB Theory

Postby Alan Turner » Sat Sep 17, 2011 3:48 pm

[quote="Russ Elliott"]Alan - taking the 3-axle case and where a required deflection has been input - when moving the fulcrum points about, which axle(s) does the 'actual deflection' value refer to?

Rquote]

It applies to all axles, that's why you have to get the CoG coincident with the CoA. The deflection is the same because the rigid chassis imposes a given deflection to all axle positions. If the CoG and CoA are not coincident then the deflection is the "average" but the important thing to note is that, because of the rigidity of the chassis, the deflections at each axle are linearly related (i.e. lie on a straight line). In adjusting the fulcrum points there are two objectives: 1) to ensure coincidence of the CoG and the CoA and 2) to get the axle loading as you want it to be. However 1) overides 2) and the solution is invalid if 1) is not achieved.

The calculation of the deflections in a non-coincident situation is beyond the present spreadsheet and that is why the spreadsheet output is only valid when the CoG and CoA are coincident.

I must emphasis that, in the static condition, all deflections of all axles are the same on level track for both the actual model and the spreadsheet.

Alan

User avatar
Russ Elliott
Posts: 930
Joined: Thu Jun 02, 2011 6:38 pm

Re: Abstruse CSB Theory

Postby Russ Elliott » Sun Sep 18, 2011 1:44 pm

Alan - thanks for that, but before responding, I notice that there are horizontal and vertical scales shown on your 'centre of gravity location' graphic. I find the lettering very difficult to detect against the dark blue background. Could the background colour be changed please?

Russ

Alan Turner
Posts: 559
Joined: Sun Jul 20, 2008 4:24 pm

Re: Abstruse CSB Theory

Postby Alan Turner » Sun Sep 18, 2011 3:07 pm

Yes there are figures but you are not supposed to be able to read them as they are meaningless. I have altered the bachground colour in the attached so that they cannot be seen. I originaly got the shade of the font and the background slightly different (due to developing the spreadsheet on both Excel 2003 and 2007).

Alan
Attachments
CSB sheet - Combined version - v1.16.xls
(697 KiB) Downloaded 267 times

Alan Turner
Posts: 559
Joined: Sun Jul 20, 2008 4:24 pm

Re: Abstruse CSB Theory

Postby Alan Turner » Sat Oct 08, 2011 9:15 am

At the request of Russ Elliott I attach a version for use of the CLAG group.

regards

Alan
CSB sheet - Combined version - v1.17.xls
(692.5 KiB) Downloaded 284 times

User avatar
Russ Elliott
Posts: 930
Joined: Thu Jun 02, 2011 6:38 pm

Re: Abstruse CSB Theory

Postby Russ Elliott » Mon Nov 21, 2011 2:38 pm

Having had some time in the last month or so do a lot of correlation checks between the different types of spreadsheets, I'm pleased to report I've yet to find any discrepancy between them in the results they give. I've also come to appreciate what Alan Turner was getting at in saying that wheel loads are not 'an input', but I suppose this was always the case in a sense with the moment-distribution method because wheel loads could be tweaked (as well as the fulcrum positions of course) to obtain equal axle deflections. Alan's approach sets the equal deflection and derives the loads according to the fulcrum positions, and it's certainly a quicker method, and a more rigorously correct one. Personally, I'd prefer if his fulcrum position gradation was reduced to 0.25mm increments rather than 0.5mm, but I say that not so much because that accuracy is necessary in implementation, but for purposes of giving the modeller or kit designer an idealised objective to aim for in marking out.

What has surprised me, in going through the correlation checks, is how consistent axle loadings are when fulcrum positions are 'more or less in the right place'. As with any figures connected with CSBs, the notions of 'consistency' and 'more or less' need to be qualified of course, and it depends on how much finessing of the equality of axle loads we feel is necessary or desirable. This brings into focus how much (unsprung) weight we are adding to a driven axle with gearbox and/or motor, and that's not something one can factor in to a generalised plot. Mind you, the prototype could do nothing about the weight of its inside motion, so we are perhaps no different in that respect.

What was more surprising to me was that a seemingly large deflection differential - what Will's spreadsheet calls the 'centre axle bias' - produces such a small difference in axle load when the wheels are, as on real track, deflected by an equal amount, at least in the static flat-track condition. This underlines the essence of Alan's approach. Accordingly, I've had something of a damascene conversion and dropped mention of (the potentially misleading) deflection differential percentages from the examples on the CLAG CSB page and translated everything into axle load percentages. It is, after all, the axle loads we're most interested in as the 'end product' of the CSB process.

The older style of spreadsheet does have some definite advantages though, and provides a route, albeit a tortuous one, in investigating what happens to axle loads as wheels traverse defined bumps and dips. This investigation is primarily an academic exercise of interest only to CSB nerds like Roger Wyatt and me, but I think something useful in terms of graphical animation might eventually result.

Changing subject; I've always thought the principle of unloading the middle axle slightly was 'a good thing' (and many of the examples on the CLAG page continue to implement that notion), but it's probably fair to say that we don't really know how valid an assumption this is - all we do feel, if only intuitively, is that the middle axle shouldn't be stronger than the outer ones. I arrived at that notion a long time ago as a sort of deliberately conservative approach, because of factors like the effect of drawbar pull, the unknown location of extra unsprung motor/gearbox weight, and perhaps most important of all, the difficulty in implementing a desired CofG position. In having a good natter with Will at Scaleforum, he did pose the question - To what extent is our wariness of middle axles stronger than the outer ones a valid point of view?

User avatar
Will L
Posts: 1770
Joined: Sun Jul 20, 2008 3:54 pm

Re: Abstruse CSB Theory

Postby Will L » Tue Nov 22, 2011 12:43 am

Russ Elliott wrote:...What has surprised me, in going through the correlation checks, is how consistent axle loadings are when fulcrum positions are 'more or less in the right place'. As with any figures connected with CSBs, the notions of 'consistency' and 'more or less' need to be qualified of course, and it depends on how much finessing of the equality of axle loads we feel is necessary or desirable. ..


There in lies what I like to think of as the robustness of the CSB method. While clearly you should aim to get the fulcrums as close to the optimum points as you can manage, disaster wont ensue if normal building tolerances aren't working in your favour.

Russ Elliott wrote:....What was more surprising to me was that a seemingly large deflection differential - what Will's spreadsheet calls the 'centre axle bias' - produces such a small difference in axle load when the wheels are, as on real track, deflected by an equal amount, at least in the static flat-track condition.


More evidence of that useful robustness. Investigation suggests that my spreadsheets "centre axle bias" figure is in the region of 3 times the size of the actual weight difference expressed as a percentage, using figures provided by Alan's Spread sheet. Quite why you'll have to ask Rodger. However, by magnifying the effect it actually improves the sensitivity of the measure, even if the units its expressed in remain a little obscure. I agree that calling it a "percentage difference in the weight carried" is not correct, and you will notice that I didn't, but it remains a useful indicator that the fulcrum positions are correct.

Russ Elliott wrote:....I've always thought the principle of unloading the middle axle slightly was 'a good thing' (and many of the examples on the CLAG page continue to implement that notion), but it's probably fair to say that we don't really know how valid an assumption this is - all we do feel, if only intuitively, is that the middle axle shouldn't be stronger than the outer ones. I arrived at that notion a long time ago as a sort of deliberately conservative approach, because of factors like the effect of drawbar pull, the unknown location of extra unsprung motor/gearbox weight, and perhaps most important of all, the difficulty in implementing a desired CofG position. In having a good natter with Will at Scaleforum, he did pose the question - To what extent is our wariness of middle axles stronger than the outer ones a valid point of view?


I was considering situations where some of the weight was being carried by none driving wheels, say a 2-4-2 or a 2-2-2 chassis where equal distribution of the weight across the axles would represent a big loss of adhesive weight. A recent discussion on E4um asked why a CSB chassis typically doesn't oscillate about its equilibrium position as some spring systems are inclined to do. Experience says that they don't, and I came to the conclusion that this was probably due to the damping effect of the friction between horn blocks and horn guides, and between CSB wire and fulcrum points. This again opened up in my mind the question of how far you could push your luck and overweight the centre axle before nodding does become an issue, and how far you could rely on the inbuilt damping to prevent an over lively ride.

I agree with Russ that for your average 0-6-0 chassis there is no point in pushing your luck on this, but faced with a 2-2-2 or a 2-4-2 I would be inclined to experiment to see what I could get away with. As a GER inclined person, 2-4-2 tanks are definitely on the agenda, if not at the top of the list for the moment. What I do know is that my C12, that is an effective 4-4-0 (but a 4-4-2 if you must count the wheels) is running very nicely with about 40% of the weight on each driving axle and only about 20% on the bogie.

Will

User avatar
Russ Elliott
Posts: 930
Joined: Thu Jun 02, 2011 6:38 pm

Re: Abstruse CSB Theory

Postby Russ Elliott » Tue Nov 22, 2011 2:37 pm

Will L wrote:Investigation suggests that my spreadsheets "centre axle bias" figure is in the region of 3 times the size of the actual weight difference expressed as a percentage, using figures provided by Alan's Spread sheet.

Yes, and that's what puzzled me at first and started me off in the 'correlation between spreadsheets' exercise a couple of months ago. What is happening is, with a chassis sitting on flat track (in distinction to its 'designed centre axle biased state'), the deflections become equal, so we are in effect 'resetting' the centre axle bias to zero - this alters the load relationships (and hence the axle loads) throughout the rest of the beam. (Alan Turner will be wagging his finger at this point saying "That's why I kept saying loads aren't inputs!") I suspect it's some mad cubic relationship, and we're looking at the quasi-linear long tail end of it as we approach our optimum fulcrum point positions. I updated the CLAG page a week or so to try to explain it, but I'm not sure I've got quite the right set of words yet!

2-4-2s, hmmm, I have some crazy thoughts on those. Here's some below. The conventional CSB approach would require a 'multi-take' with a Roger-style 4-axle spreadsheet, the first stage being to see whether a reasonable fulcrum set could be obtained to establish an inequality of design deflections between carrying and driving axles (2-4-2s tend to have long wheelbases, so the outer axles might need to be more accommodating of track irregularities), with the iterative stage being resetting d1 to be the same as d2 to confirm the final axle loads. All a bit tricky, I reckon. I understand there are stiffness matrix methods that can cope with vertical datum offset scenarios such as these, but they're probably mind-blowingly complicated.

The pairs of asymmetric equalisers route is easier design wise, but the deflection relationship at the ends of the cantilevers (a conventional principle of moments attribution, but where d is proportional to the cube of the cantilever span), and is therefore set by the chosen load differentials. The outer axles could therefore be a bit too lively, and on discrete equalisers, the fulcrum point becomes more of a true roller support, so some of the primary sources of friction (as on a conventional CSB) are no longer present. It's an adventurous approach. (Btw, on the damping issue, I've also put up a section on the CLAG page on frequency response, if only to counter the uninformed suggestions that CSBs will 'oscillate'. They won't. There's absolutely no chance of it, although the situation on cantilever equalisers is perhaps less clear - might do some sums later.)

The final suggestion is to have conventional 2-fulcrum thingies on the outer axles, set normally - this will be good for both pitch and roll stability. The tricky part is getting enough downforce on the long-travel equaliser over the drivers, and I suspect it will require a vertical adjuster on the middle roller to get that deflection/load relationship in the right place.

2-4-2.gif
2-4-2.gif (9.56 KiB) Viewed 9641 times


Edit: The last bit of the above diagram as originally posted was rubbish, so I've changed it. Will try to draw something better later.

Chris Mitton
Posts: 206
Joined: Sun Aug 31, 2008 1:18 pm

Re: Abstruse CSB Theory

Postby Chris Mitton » Sat Jun 22, 2013 3:38 pm

Hi all (especially Alan, Russ and Will!)

I've now built two CSB'd chassis using Alan's spreadsheet plus a fair bit of belt and braces. One is an 0-6-0 which is actually a 2-4-0 (ex-GER E4 which has been seen on this forum http://www.scalefour.org/forum/viewtopic.php?f=20&t=1305#p20848), and it runs well so far as I can see - however I haven't tried it under power yet, the collectors are on my bench as i write. [The other is a mogul, sprung as a 0-6-0 with separate pony truck, but that's waiting for enough Round Tuits to tackle Walschaerts! :twisted: ]

There are three more sets of frames on my bench (the bodies are largely done), and while unpacking a supply of enough RT's to contemplate having them running by year end, I fell to thinking about the optimum strategy for setting up the CSBs. At risk of re-opening old debates, I'd welcome any views on whether I'm talking sense or nonsense......

It seems to me that my overall strategy in getting the best out of Alan's spreadsheet needs to be:
1. Determine the weight of the engine.
2. Determine where I want the CofG to be.
3. Use the spreadsheet to work out where the fulcrums need to be to achieve that CofG and what spring wire to use.
4. Build the bloody thing!
5. Find what the engine really weighs.
6. Find where its CofG really is.
7. Add strategically placed weight to move the actual CoG to its desired position.
8. Add a bit more weight wherever I can at the back, and sufficient balancing weight in the smokebox or front frames, to bring it to the desired weight without disturbing the CoG.
9. If it's too bouncy, or not bouncy enough, change the spring wire, which so far as I can see makes no material difference to the weight distribution but only affects the spring deflection.
10. Enjoy a beer or three....

Now, I don't know step 1 as I haven't built the loco yet, but I can set a lower bound on it (the weight of the parts excluding unsprung stuff such as wheels), so I can use any high enough figure I like. For my J15 I've decided arbitrarily on a target weight of 180 grams.

For step 2, previous posts in this thread suggest that the CoG needs to be somewhere near the middle but some variation - how much? isn't fatal. So it struck me that I could aim for it to be somewhere near the position on the prototype. I don't suppose Mr Worsdell had access to any spreadsheets on his slide rule (remember them?) but he would have done some serious maths to achieve his design, and it did work - one is still working successfully 130 years later!
The J15 (according to my Isinglass drawing) carried 13 tons 1 hundredweight (35.2%) on its front axle, 13t 10c (36.4%) on the centre and only 10t 11c (28.4%) on the rear axle. With a wheelbase of 7' 7" (front) plus 8' 6", a simple calculation of moments tells me that the CoG on the real thing was 7' 4" behind the front axle, or 8 1/2 inches ahead of the centre-line of the coupled wheelbase.
Using this figure in Alan's spreadsheet, I can get spring positions as
front fulcrum - 10.5mm - axle - 15.33mm - fulcrum - 15.0mm - axle - 16.5mm - fulcrum - 17.5mm - axle - 10.5mm - rear fulcrum
which gives a weight distribution of 35.0% front, 36.4% centre, 28.6% rear axle. This is very close to the prototype's (and not much different from Bill Bedford's plot), but violates a theme that recurs in this thread, of not having more weight in the middle than the outer axles. It's not a huge violation, and the real thing certainly didn't bounce about too much (mind you, I was only ten when I rode on one! :D ), so does this matter?

Before I get Chris Gibbon's jig and the Dremel out, is this a sensible way to proceed?

Regards
Chris

PS Alan - how about adding prototype weights as inputs and prototype CoG as a result on your spreadsheet, just in case it's helpful as something to aim at? I tried it but the cell locking defeated me.... :(

User avatar
Russ Elliott
Posts: 930
Joined: Thu Jun 02, 2011 6:38 pm

Re: Abstruse CSB Theory

Postby Russ Elliott » Sat Jun 22, 2013 5:53 pm

I like your 10 steps, Chris! (Although maybe somewhere up front should be a judgement stage/loop on whether the desired CofG will be attainable at the end of the build for a given motor location.)

For our model purposes, I'm not convinced prototype axle weights should be taken too seriously - prototype driven axles usually carried a typically 3 ton excess because of (predominantly unsprung) inside motion, which the designer couldn't do anything to change, but I suppose one could argue we have the equivalent somewhere because of gearbox weight. In some ways though, we perhaps have more flexibility than the prototype designer did, and can play more games with our spring weights accordingly. Don't forget though prototype CofGs were placed in the knowledge that drawbar pull would bring it backwards. See here for the thinking behind a 2-4-0 plot, where drawbar pull will equalise the driver loads.

Concerning the 'violation' (!) of not having more weight in the middle than the outer axles, this is I think an area where we can have one up on the prototype, namely that we can have the advantage of all axles giving a reasonably equal contribution to the pull, but where we can lessen any porpoising propensity. No one knows how much of a 'lessening' this will be, it's merely intuition, and my guess is that a jury would probably still be out if assessing the performance of two otherwise identical locos with different CSB plots. Nevertheless, it's worth five minutes of spreadsheet play on Will's version to see how much changing the CofG position affects the chassis slope, if only as a tick-the-box confidence check.

Alan Turner
Posts: 559
Joined: Sun Jul 20, 2008 4:24 pm

Re: Abstruse CSB Theory

Postby Alan Turner » Sun Jun 23, 2013 8:31 pm

I thought I felt my ears burning!

OK some comments.

Prototype weight is meaningless. We only build models that reproduce the linear scaled dimensions. We cannot therefore reproduce the dynamic characteristics of the prototype because, simply, we can not scale time and therefore gravity.

What we have to do therefore is to build a chassis that does what we need it to do in our model. In effect the chassis is at a scale 1:1 for that purpose.

My spreadsheet does provide % of weight distribution of the model and there is no reason why that should not resemble the prototype weight distribution.

As regards your dilemma about having to build it first; that of course is an issue that confronts all designers - even the designers of locomotives. However as I guess the largest part of the model's weight is made up of lead placed about the body then you should be able to influence the final weight and distribution after construction. If you end up with a different weight than initially estimated then that can be corrected by changing the suspension wire size. You should however be able to achieve your desired distribution/CogG.

The prototype CofG might well be a good place to start and easily calculated from the prototype weight diagram. Remember however that the prototype had individual springs rather than the CSB.

Obviously the spreadsheet has a lot more going on than you can see. What you see is probably only 10% of the complete sheet.

Regards

Alan

User avatar
Will L
Posts: 1770
Joined: Sun Jul 20, 2008 3:54 pm

Re: Abstruse CSB Theory

Postby Will L » Sun Jun 23, 2013 11:13 pm

Chris

I have no real problems with your approach, and having modified my spread sheet a touch to allow the entry of a prototype weight distribution, I have plugged in your values. It comes out well balanced, and it will certainly sit flat, so long as the loco CofG is pretty close to the design position.

However because the location of the natural CofG is not known till after you’ve built it, and you will inevitably have to weigh the loco to give the desired CofG, my view is that there is little or no point in planning to have the finished loco CofG anywhere but the optimum point at the chassis centre, and equal weight distribution on all axles, so as to optimised adhesion/pulling power. I know the prototype didn’t achieve this but they were much more constrained as to where the CofG of the loco turns out to be than we are. While using prototype weight distribution looks like a nice example of the "getting it all right" approach in theory, in practice so you might as well do what's best for model performance.

That said, if you do decide to go with your current design, I wouldn’t be too concerned about the fact that you’ve apparently got more weight on the centre axle. I have come to the view that the proposing threat is more theoretical than real. There is sufficient friction in the horn blocks to damp out any inclinations to oscillate unless you go for something really extreme.

No matter what design you use, I think you should try to get the CofG to the design place with a reasonable degree of accuracy, as a deviation of multiple millimetres can have a significant effect. E.g. the CofG for your design is less than 2mm from the standard all axles with equal weight position. Some may say that needing to get the loco balanced right is an unnecessary complication of going CSB. I would rather look at this as bringing a proper focus to an important factor in chassis performance common to all chassis designs. The fact that it has been ignored in the past has led to the actually pulling power of finished locos being a bit of a lottery.

Will

billbedford
Posts: 708
Joined: Tue Oct 02, 2012 2:40 pm

Re: Abstruse CSB Theory

Postby billbedford » Wed Jun 26, 2013 10:31 am

What we need is some sort of competition, lets call it the 'Deputy Chairman's Cup', where all those that pontificate on CSB theory can show off their latest creations and prove the worth of their theorising…………


……….or not.
Bill Bedford
Mousa Models
http://www.mousa.biz

User avatar
Will L
Posts: 1770
Joined: Sun Jul 20, 2008 3:54 pm

Re: Abstruse CSB Theory

Postby Will L » Wed Jun 26, 2013 1:39 pm

billbedford wrote:What we need is some sort of competition, lets call it the 'Deputy Chairman's Cup', where all those that pontificate on CSB theory can show off their latest creations and prove the worth of their theorising…………

……….or not.


I agree

Will

David Catton
Posts: 77
Joined: Fri Mar 13, 2009 7:23 pm

Re: Abstruse CSB Theory

Postby David Catton » Wed Jun 26, 2013 4:48 pm

Hang on - I was the Deputy Chairman who instituted THAT cup. I was less than thrilled when its purpose was changed without even the courtesy of "Would it be alright with you?"

If there is ever a serious proposal to determine the rationale for awarding this trophy, I wonder if the (then) committee might see its way to dropping me a note with their proposal before proceeding . . .

Thank you!

David C

User avatar
dcockling
Posts: 638
Joined: Mon Jun 30, 2008 10:11 am

Re: Abstruse CSB Theory

Postby dcockling » Wed Jun 26, 2013 6:10 pm

Hi Dave, had you not noticed that when it was reintroduced in 2010 it was changed back?

http://www.scalefour.org/history/depchaircup.html

You were, as you know, but others may not, Deputy Chairman from 1978 until 1981, the first one, as the post didn't exist prior to that. The competition was changed in 1982 and the Committee then was: Ray Hammond, Ken York, Robin Barr, Martin Setchell, Clive Evans, Brian Popkin, Mike Peascod, Richard Johnson and Beth Evans, so none of the present incumbents are guilty.

The revised format ran until 2001. When it was reintroduced in 2010 it was in it's original format, which we didn't think you'd mind, although it is the Deputy Chairman's Cup not the Dave Catton Cup. The problem now as then though is getting any entries.

All the Best
Danny

David Catton
Posts: 77
Joined: Fri Mar 13, 2009 7:23 pm

Re: Abstruse CSB Theory

Postby David Catton » Thu Jun 27, 2013 9:30 am

Hi Danny,

Yes, I had seen it was returned to its original guise but I was responding to Bill B's suggestion and wondering out loud if any changes were envisaged in future whether at that stage I might be asked my opinion.

The list of committee members who made the change who have since departed this mortal coil makes depressing reading . . .

I guess if I continue to stalk the world naked in this chilly summer, I might not last long enough to have an opinion that can be accessed by currently available technology.

DC

User avatar
dcockling
Posts: 638
Joined: Mon Jun 30, 2008 10:11 am

Re: Abstruse CSB Theory

Postby dcockling » Thu Jun 27, 2013 2:10 pm

David Catton wrote:I guess if I continue to stalk the world naked in this chilly summer, I might not last long enough to have an opinion that can be accessed by currently available technology.

DC


A cloak! A cloak for the poor Emperor!

All the Best
Danny

User avatar
Horsetan
Posts: 1063
Joined: Tue Jul 29, 2008 9:24 am

Re: Abstruse CSB Theory

Postby Horsetan » Thu Jun 27, 2013 8:51 pm

David Catton wrote:.....I guess if I continue to stalk the world naked in this chilly summer, .....


I'm tempted to say that I'd pay to watch this....... :mrgreen:
That would be an ecumenical matter.

David Catton
Posts: 77
Joined: Fri Mar 13, 2009 7:23 pm

Re: Abstruse CSB Theory

Postby David Catton » Fri Jun 28, 2013 7:22 am

Ivan,

I'm trying to decide if you are depraved or merely deprived.

And there we had better leave it or the serious minded folk who compose the erudite notes we read on this topic will shortly be as frustrated as you appear to be . . .

Cheers,

David C

Chris Mitton
Posts: 206
Joined: Sun Aug 31, 2008 1:18 pm

Re: Abstruse CSB Theory

Postby Chris Mitton » Sun Jun 30, 2013 9:07 pm

Before minds get too boggled, to get back on topic.....

Many thanks to Russ, Will and Alan for your very helpful comments....

I had actually forgotten to take into account the movement of CoG from the drawbar pull. On paper this would work nicely for my proposed plot, as it would pull the CoG nearer to the middle of the loco, but for one snag. As on many a bucolic East Anglian branch, J15 no.7567 will be pulling its trains tender first as much as chimney first (Stowe Fen won't have a turntable), and that would clearly pull the CoG even nearer the front. Not that it will be overly-taxed, three coaches or a dozen (ish) wagons will be the most it will be called on to move (should I ever get that many!).

So on balance I think I'll rework the numbers to put the theoretical CoG a tad abaft of the centre axle so that it doesn't move too far off-centre in either direction. I'll report progress on the My Workbench thread - but not for a while, a couple of weeks in the Dordogne with three granddaughters won't leave any modelling time! ;)

Regards
Chris

PS Will - some of your comments imply aiming for an equal weight distribution across all axles - but this is impossible unless the wheelbase is symmetrical (the J15 isn't) - simple consideration of moments about the centre axle will show that.
PS Alan - I did try and see what your spreadsheet was doing by looking at the off-screen bits and I have to take my hat off to you ...... :thumb

User avatar
Will L
Posts: 1770
Joined: Sun Jul 20, 2008 3:54 pm

Re: Abstruse CSB Theory

Postby Will L » Sun Jun 30, 2013 10:13 pm

Chris Mitton wrote: Will - some of your comments imply aiming for an equal weight distribution across all axles - but this is impossible unless the wheelbase is symmetrical (the J15 isn't) - simple consideration of moments about the centre axle will show that.


Sorry Chris but this is not true
There is always a solution that will allow equal weight distribution on all wheels. To get equal weight distribution, the CofG will be over the centre axle on a symmetrical chassis but otherwise it is displaced slightly to the direction of the longer segment of the wheelbase. If you look at my spread sheet it will tell you where it is. To get your J15 to perfection you will need to deal in very small parts of a millimetre but you can get close enough to an accuracy of half a millimetre, 31.5 mm back from the front axle.

The CofG shift on picking up a heavy load is real but I have to say it only becomes a factor of any significance with a loco like a Pacific with a load carrying pony truck. For such a loco picking up a very heavy train causes the transfer of weight off the driving wheels onto the pony truck and hence the well know light footed tendency of spam cans. For other loco's the slight lift at the front will only happen sufficiently to approach visibility when you loco is near its adhesion limit and will be so small you most unlikely to spot it.

Will

P.s. while playing with Chris's figures I realised that my spreadsheet's "required CofG location" calulation only works when looking for an equal weight distribution. I now have a version which will allow you to enter a prototypical loco's weight distribution in percentage form and tells you exactly where the CofG works out to be. I'm disinclined to move away from the existing version because many people already think it a bit complicated but if anybody is interested....


Return to “Chassis and Suspensions”

Who is online

Users browsing this forum: No registered users and 4 guests