Wheel coning and trackholding

[James Moorhouse]

Sorry to reopen old wounds!

Just reading through this thread, it seems nobody mentioned the physics of how real trains stay on the track. For a concise explanation watch this Youtube clip.

I believe the same physics applies to our models, but contact between the wheel and railhead must be maintained and therefore springing is required. Martin's rationale for using EM wheels is simply to obviate the need for springing.

[Flymo748]

Sorry to reopen old wounds!

Just reading through this thread, it seems nobody mentioned the physics of how real trains stay on the track. For a concise explanation watch this Youtube clip.

I believe the same physics applies to our models, but contact between the wheel and railhead must be maintained and therefore springing is required. Martin's rationale for using EM wheels is simply to obviate the need for springing.

And if my understanding of Feynman's very clear exposition is right, it is the lack of the effectiveness of the coning due to the wheel not being in contact with the rail-head that means that the EM flanges come into play. Feynman rightly refers to the flanges as a safety mechanism, rather than the primary means to steering a vehicle around the curves.

It's a really clear explanation. If only all physics was explained in this way, I may have done better in my A Level than I actually did!

There is a bit of circular argument - lack of springing is permitted by EM size flanges that are required due to the lack of springing, and so on ad nauseum. Of course, we deal with different magnitudes of forces, including the centring effects of gravity, but my brayn would ache too much if I tried to think about that!

Flymo

[Russ Elliott]

Feynman rightly refers to the flanges as a safety mechanism, rather than the primary means to steering a vehicle around the curves.

Feynman was cool, but only up to a point. The really cool dude was called Brunel, who had the correct perspective in 1838:

"For small displacements from the centre of straight or slightly curved track the primary mode of guidance is conicity and it is on sharper curves, switches, and crossings that the flanges become the essential mode of guidance."

Toot! Toot!

[Jol Wilkinson]

Feynman rightly refers to the flanges as a safety mechanism, rather than the primary means to steering a vehicle around the curves.

The really cool dude was called Brunel,

Toot! Toot!

With that hat? I think there has been too much imbibing of the old Toot.

Jol

[martin goodall]

Richard Feynman's explanation of the function of coning on the wheels of railway vehicles is an excellent and entertainly presented exposition. But I have always felt that it is a fallacy to believe that our models behave in the same way. The masses are far smaller, and so our models cannot be gauaranteed to behave in the same way as full-size railway vehicles. Furthermore, most of us (in fact I suspect 99% of us) do not set our rails at the regulation 1 in 20 inward cant, so as to match the angle of the railhead with the angle of the coning of the wheels; we set the rails upright, so that the coned tyre treads run on the inner edge or corner of the rail-head. They are therefore much more prone to 'hunt' than their prototype counterparts.

For these reasons, I strongly suggest that the coning of our 4mm scale wheels cannot be relied upon to keep our vehicles on the track; we need those flanges. In fact, it is because I came to the conclusion that P4 flanges are too small to do the job with complete reliability that I resorted to using EM wheels re-set to the P4 back-to-back, so as to give the added 'insurance' of deeper flanges.

[Alan Turner]

Furthermore, most of us (in fact I suspect 99% of us) do not set our rails at the regulation 1 in 20 inward cant, so as to match the angle of the railhead with the angle of the coning of the wheels.

So do Swiss railways and their trains seem to run OK. American railways have a 1 in 40 inclination.


Alan

[nigelcliffe]

The masses are far smaller, and so our models cannot be gauaranteed to behave in the same way as full-size railway vehicles.

I think the mass, and the way mass scales is part of the issue (2010 Royal Institution Christmas Lectures recommended on this topic).
Another factor is wheel bearings. If wheels run in cylindrical bearings its probably fine. But many (most?) use pin-point bearings. In these, in theory a point on an axle runs in a conical hole in the bearing. In many cases the point doesn't actually sit in the dead-centre of the conical hole, and instead precesses around the bottom of the hole (which, under high magnification will be anything but conical!).

That said, for my small short trains, I'm running with "normal" P4 flanges. The problems only occur when the track is not good enough, or when the B-B is wrong.

- Nigel

[Alan Turner]

The masses are far smaller, and so our models cannot be gauaranteed to behave in the same way as full-size railway vehicles.

I think the mass, and the way mass scales is part of the issue (2010 Royal Institution Christmas Lectures recommended on this topic).
Another factor is wheel bearings. If wheels run in cylindrical bearings its probably fine. But many (most?) use pin-point bearings. In these, in theory a point on an axle runs in a conical hole in the bearing. In many cases the point doesn't actually sit in the dead-centre of the conical hole, and instead precesses around the bottom of the hole (which, under high magnification will be anything but conical!).

That said, for my small short trains, I'm running with "normal" P4 flanges. The problems only occur when the track is not good enough, or when the B-B is wrong.

- Nigel

Mass is not the problem it's the inability to scale time that's the problem. Because our models move then their dynamics are not properly scaled simply by scaling their linear dimensions.

I have posted on this aspect many time before - I was ignored then and I will probably be ignored now.

Alan

[Russ Elliott]

I don't see that the presence or lack of a 1:20 rail inclination has much effect on the situation. Redtenbacher's 1855 formulation of the rolling of a coned wheelset on a flat curve proved it was not dependent on mass, and applies regardless of scale.

Also, I would strongly dispute whether wheelset hunting is present on our models like it is on the prototype, where it is a significant problem requiring sophisticated yaw dampening on bogies. Notwithstanding the effects of longitudinal wheel creep (the major step in identifying the cause of hunting), yaw dampening on model bogies should be avoided in my view, i.e. we prefer the conicity for our curve guidance rather than flange guidance, even if we are thwarted in that objective because of the prevalence of our sharp curves. (Bogie yaw inertia moments are speed and mass-related, so our model considerations in this respect will be radically different to that on the prototype.) If bogie hunting was observable in 4mm models, one would expect to see it most of all in OO. I've seen all sorts of wobbling in OO (arising from other causes), but not bogie hunting.

[LesGros]

Alan Turner wrote:

...I have posted on this aspect many time before - I was ignored then and I will probably be ignored now.

Alan,
Be of good cheer, I, for one, do not ignore your postings; the phrase "silence implies acquiessance" comes to mind. There is, perhaps, a reluctance to risk breathing too much life back into this long standing EM wheels topic. Readers may choose to believe in legal minded Martin G, or believe in the Engineering Science that you and other learned practitioners espouse on these matters.

Personally I incline to the latter, primarily because the term "EM wheels" is about as usefully descriptive as the term "OO wheels". As we have seen, any technical discussion which is not crystal clear about the differences within the detail is always likely to become contentious.

Happy New Year

[Will L]

.. it's the inability to scale time that's the problem.

I'm with you Alan, there is no such thing as scale time. I learned this building a scale brick hut out of scale plastic bricks. Took about as long as building a real hut with real bricks. It was even more exact than you expect as, just like with real bricks, you couldn't build up more than a few courses a day as you needed to let the solvent dry out!

Will

[Russ Elliott]

I agree with Les, Alan - you weren't ignored then and you're not being ignored now. (Although I will plead guilty to not understanding your postings, sometimes, but that is my fault.) Anyway, since we've drifted way off topic, here's another drift (apropos to this thread topic, actually, but that's another story) - I can't find the EM track standards page on the EMGS website - can someone post it?

[James Moorhouse]

Richard Feynman's explanation of the function of coning on the wheels of railway vehicles is an excellent and entertainly presented exposition. But I have always felt that it is a fallacy to believe that our models behave in the same way. The masses are far smaller, and so our models cannot be guaranteed to behave in the same way as full-size railway vehicles.

Although you are quite right in saying that the masses of our models are small compared with those on the full size railway, and indeed the mass to momentum relationships in the prototype do not scale linearly to models, I don’t see what relevance mass has in the context of how the coning of wheels keeps trains on the track. The effect of coning on the wheels means that a train is steered into a curve: there is no gravitational centring effect in play. This explains how a train can move very slowly on super-elevated track, where its centre of gravity would be moved away from the centre line of the track, without riding off the rails.

Regarding model rails not being set at a 1 in 20 inclination, I think this is certainly true amongst the ply and rivet brigade, but more and more modellers are supporting their rails solely by plastic chairs which do hold the rail to the correct inclination. However, it is not always entirely prototypical to incline the rail especially through point and crossing work. Also, I believe that for a number of years BR specified that flat bottom rail be held vertically, but have since reverted back to the 1 in 20 inclination.

Regarding the hunting of wheels on track, on the prototype one of the main factors attributed to hunting is the elasticity between wheels and track where they are in contact with one another. On our models the same degree of elasticity would not be encountered.

For these reasons, I strongly suggest that the coning of our 4mm scale wheels cannot be relied upon to keep our vehicles on the track; we need those flanges. In fact, it is because I came to the conclusion that P4 flanges are too small to do the job with complete reliability that I resorted to using EM wheels re-set to the P4 back-to-back, so as to give the added 'insurance' of deeper flanges.

As I said earlier I believe that the physics of how a real train stays on the track also applies to our models. Of course, as on the prototype flanges do play a role some of the time, i.e. when traversing switches, crossing and sharp curves, in keeping model trains on the track, and this is why the shape of our flanges in P4 is important. Again to ensure our flanges work correctly contact needs to be maintained between wheel and track and for this reason suspension is required. In terms of practical experience, models that are sprung perform much better than those that are not. During a running session on Steve Hall's Drighlington layout my own stock, which is fully sprung, ran with zero derailments, whilst some of Steve's stock which was not sprung did have the occasional derailment.

[grovenor-2685]

Also, I believe that for a number of years BR specified that flat bottom rail be held vertically,

This only applied to pointwork. However not all railways incline the rails and it does not have any part to play in the steering effect, the contact patch will move across the wheel tread to give differential diameters whether the rails are inclined or not.
Keith

[Chris Mitton]

It's been a key principle of science for several hundred years now that experiment always trumps theory; a theory is never accepted unless it successfully predicts the result of an experiment. So would any brave soul like to sacrifice a wagon by turning the flanges off its wheels completely, then let us know if it stays on the track or not? Sorry I've neither a lathe nor the necessary skill......yours only slightly facetiously,
Chris

[Terry Bendall]

Picking up on a number of points on this thread ...

However, it is not always entirely prototypical to incline the rail especially through point and crossing work. Also, I believe that for a number of years BR specified that flat bottom rail be held vertically, but have since reverted back to the 1 in 20 inclination.

As Keith has pointed out, on British railways, the rail is vertical through point work. For those who want to replicate this, Colin Craig's flat bottom track is designed to be inclined at 1 in 20 for plain track, but his turnouts have the rails vertical.

Again to ensure our flanges work correctly contact needs to be maintained between wheel and track and for this reason suspension is required.

Many people have shown that suspension of some sort is NOT always required. My philosophy is that is if I can get stock to run without suspension, and not fall off, then that is what I will do. If things fall off, then I fit some sort of springing/compensation, depending on what the prototype is and what works best on the model. I have just converted 12 HEA coal hoppers by fitting some of Colin Criag's wheels direct to the existing bearings and they all run perfectly on Elcot Road without falling off.

As I said earlier I believe that the physics of how a real train stays on the track also applies to our models.

I don't see how this can apply since the mass of the model is not scaled down from the prototype. As the model engineering fraternity have found with things like steam ports, you cannot scale nature.

In my view the MRSG did a super job in 1967 and produced a set of workable standards that the trade and the average person could implement.

They did indeed and they do work. If things fall off, there is always a reason although finding it will often take quite a lot of time. I would prefer to find the reason and then correct the error, rather than correct the error by making a fundamental change such as fitting EM wheels of whatever profile. If I did that I would not be working to P4 standards.

but it just seems like a frightful waste of viewing time to be re-inventing this subject over and over.

Well some people get excited by all sorts of strange things, but if that is what rings your bell then fine. Personally I prefer to get on with building a layout and getting it to run successfully.

Terry Bendall

[Re6/6]

..................... Personally I prefer to get on with building a layout and getting it to run successfully.

Terry Bendall

I do so agree Terry.

[grovenor-2685]

As Keith has pointed out, on British railways, the rail is vertical through point work.

Not quite what I said, not all pointwork has the rails vertical. Bullhead designs were inclined (I don't know of exceptions but there could be some). The flat bottom designs followed suit until the late '60s when a redesign introduced the vertical rails. IIRC the main reason was to reduce costs by eliminating the need for right hand and left hand designs for a lot of the components. The vertical designs were then dominant until the introduction of the UIC60 rail a few years ago. The new designs using UIC60 rail have reverted to the use of inclined rail.
Keith

[James Moorhouse]

As I said earlier I believe that the physics of how a real train stays on the track also applies to our models.

I don't see how this can apply since the mass of the model is not scaled down from the prototype. As the model engineering fraternity have found with things like steam ports, you cannot scale nature.

Once again, in the case of how trains stay on the track, i.e. the effect of the conicity of the wheels on the track, mass is not relevant.

Many modellers do scale mass down from the prototype when building their stock, e.g. the four grams per tonne rule. The problem with mass on our models is its relationship with momentum, this does not scale linearly from that of the prototype.

[Will L]

As I said earlier I believe that the physics of how a real train stays on the track also applies to our models.

I don't see how this can apply since the mass of the model is not scaled down from the prototype. As the model engineering fraternity have found with things like steam ports, you cannot scale nature.

Because it is not a mater of mass, more a mater of linear dimensions. If we can't scale them then we really are in trouble.

Will

[grovenor-2685]

If we can't scale them then we really are in trouble.

We have been successfully using the P4 standard for over 40 years now, why should there be any doubt.
IMHO, if it doesn't work for you then the problem can be seen in the mirror, make whatever compromise you want but don't suggest that a 40 year old successful standard needs to be changed to suit.
Keith

[LesGros]

James Moorhouse wrote on 5 Jan 2012

Sorry to reopen old wounds! ...

Request to moderator(s)
Given that the content of this interesting U-tube link and the discussion which follows has little to do with EM wheels, and everything to do with the physics of wheel coning and the dynamics of wheel and track, how about transferring the relevant postings to a new topic with an appropriate title eg "Wheel and track dynamics".

Such a move would not only break this irritating obsession with EM, it would also make the material easier to reference in the forum index.
regards

[CliveLincs]

So would any brave soul like to sacrifice a wagon by turning the flanges off its wheels completely, then let us know if it stays on the track or not?

I don't consider myself particularly brave but a few years ago,just as an experiment, I turned up a pair of SM32 (16mm to the foot) wheels with a 1 in 20 coning but no flanges. This wheelset will run down a length of straight track without derailing even coping with a small amount of side to side movement. Don't ask if it will go round a curve because I don't know.

Those of you who want deeper flanges for P4 trackwork should read Ray Hammond's article in S4 News Number 5 on machining Sharman wheels to near P4 standards. At the time of this article the only Sharman wheels available were for 00/EM.

CliveLincs

[jf2682]

James Moorhouse wrote on 5 Jan 2012

Sorry to reopen old wounds! ...

Request to moderator(s)
Given that the content of this interesting U-tube link and the discussion which follows has little to do with EM wheels, and everything to do with the physics of wheel coning and the dynamics of wheel and track, how about transferring the relevant postings to a new topic with an appropriate title eg "Wheel and track dynamics".

Such a move would not only break this irritating obsession with EM, it would also make the material easier to reference in the forum index.
regards

I wholeheartedly agree with Les! In closing, have any of us observed OO models derailing over turnouts at all? No? I thought not.....

John Fitton

[HowardGWR]

Is there not an issue of wheels changing direction over a switch? Most derailments I observe at S4 exhibitions (not a few I am afraid) seem to happen mostly on switches.

The progenitor of this discussion once proudly wrote of a line of his wagons successfully being reversed over a switch on Heckmondwike layout, their suspensions being only made flexible by having the axles loosely floating their pinpoint ends in the bearings of the axleboxes, the latter themselves not being sprung.

I was impressed with this tale, but I did assume that S4 wheels were used.

By the way if someone wants to pick me up on my S4 and P4 nomenclature don't bother as I know I get it wrong and always will.