Scalefour Forum

The RAIB has published an accident report today which contains interesting information about the P1 wheel profile:

https://www.gov.uk/government/publicati ... ember-2018

Martin.

Oh dear. So any passenger-carrying track of radius less than 8.5 feet in 4mm scale needs a check rail! Best reach for that pack of rail I bought a few weeks back.

Bob

A few years ago, the American Passenger Transport Association were evaluating a proposal to adopt a different wheel profile for LRT trains, one with a greater flange-contact angle. The idea was to allow higher speeds on sharp curves. Makes me think that a less-prototypical profile might help our over-curved layouts.

Guy Rixon wrote:A few years ago, the American Passenger Transport Association were evaluating a proposal to adopt a different wheel profile for LRT trains, one with a greater flange-contact angle. The idea was to allow higher speeds on sharp curves. Makes me think that a less-prototypical profile might help our over-curved layouts.

Unnecessary Guy, the wonders of the scale effect on the physics mean that, in this respect at least, our models live in a much more favourable world than the prototype. You know this is true, consider the average toy train whizzing round impossibly tight curves.

The bogies of the class 37 locomotive each have three wheelsets, which increases the overall wheelbase compared to a two-axle bogie, and this is likely to have adversely affected the ability of the bogie to corner.

Yes....but.....they've been running since the 1950s!

The question that occurred to me was why isn't it suggested that the P8 profile be used on the Class 37?

The Class 66 also is a Co Co. Am I right that it has arrangements so that the individual wheelsets of the bogie adjust to curves in some way? Yet it has the P8 profile that is more resistant to flange climb. I wonder what other considerations there are between the two profiles. Like how they wear.

Also I don't get how the angle of attack increases with the gauge being 20mm wide in the case of a Co Co bogie, supposing the centre wheelset is pushing both other wheelsets to the outside rail.

And where the curve merits a checkrail the gauge widening starts to apply too which they don't mention; so it might not be 20mm wide to whatever the widening is nowadays for a 200m curve.

If you Google [or whatever your preferred search engine is] for "UK rail wheel profiles", there are several pdf files on the topic.

Will L wrote:

Guy Rixon wrote:A few years ago, the American Passenger Transport Association were evaluating a proposal to adopt a different wheel profile for LRT trains, one with a greater flange-contact angle. The idea was to allow higher speeds on sharp curves. Makes me think that a less-prototypical profile might help our over-curved layouts.

Unnecessary Guy, the wonders of the scale effect on the physics mean that, in this respect at least, our models live in a much more favourable world than the prototype. You know this is true, consider the average toy train whizzing round impossibly tight curves.

Empirically, yes. But I think those glued-to-the-rail toys have changed the wheel profile in useful ways, and this is more significant than scale factors. Not that I want Dublo wheels on my models, but if subtly reshaping the root of the flange helps then that might be a good thing.

Network Rail standard NR/L2/TRK/2102 specifies a nominal track gauge of 1438 mm on curves with a radius of between 200 and 178 m.

That's from the Ordsall Lane report which is much more thorough. So just 3mm widening is the first step below 200m radius, while at Doncaster it was 20mm wide.

Guy, the P1 profile wears into a flange angle more like the P8 say the reports. Isn't our wheel profile already designed to approximate a worn profile, which also has worn to a deeper flange?

I think I read the nearby Sheffield Supertram wheel profile developed for running on the heavy rail network has an ability to cope with the different check gauges that apply to each network. If that's right it means clever things can be done with the profile! - but doubt has much relevance at our scale...

Isn't our wheel profile already designed to approximate a worn profile, which also has worn to a deeper flange?

Well, no it isn't. see the P4 standards in the digest.

2. The tyre profile is based on BS 276 contour A.
3. All dimensions are in millimetres.
4. The tolerances on any contour tools used for the production of tyres or rail are recommended to be:
± 0.0125mm, for all flange dimensions (with the exception of the tyre front edge radius);
± 0.025mm, for railhead dimensions.

Note that BS 276 is a standard for steam locos and gives the new profile, of course how well this is followed by the wheel suppliers is a big unknown.
Rgds

Guy Rixon wrote:

Will L wrote:

Guy Rixon wrote:A few years ago, the American Passenger Transport Association were evaluating a proposal to adopt a different wheel profile for LRT trains, one with a greater flange-contact angle. The idea was to allow higher speeds on sharp curves. Makes me think that a less-prototypical profile might help our over-curved layouts.

Unnecessary Guy, the wonders of the scale effect on the physics mean that, in this respect at least, our models live in a much more favourable world than the prototype. You know this is true, consider the average toy train whizzing round impossibly tight curves.

Empirically, yes. But I think those glued-to-the-rail toys have changed the wheel profile in useful ways, and this is more significant than scale factors. Not that I want Dublo wheels on my models, but if subtly reshaping the root of the flange helps then that might be a good thing.

And your evidence for that?
I was irresistibly reminded of the following.

For those who haven't seen this before it comes from is Andy Reichert who is the US champion of Proto87scale which is the HO equivalent of P4.
His wheels and track standards are very much the equivalent of ours, and while not a member he is a regular contributor to our Guest book
If you eliminate any significant track and suspension defects you can do this in P4. Where Martin Goodall and our OO/EM/ friends score is that the extra depth of their flange compensates for a failure in the suspension department to keep the wheel tread on the tack. In these circumstance the flange rout is little to do with anything as very little if any of it will have been in contact with the rail.

P.S. if your not familiar with Andy's web site there is lots of interesting stuff on there some of which you may find comes in useful.
http://www.proto87.com/
(Hi Andy how's that N2 getting on?)

Will L wrote:

Guy Rixon wrote:A few years ago, the American Passenger Transport Association were evaluating a proposal to adopt a different wheel profile for LRT trains, one with a greater flange-contact angle. The idea was to allow higher speeds on sharp curves. Makes me think that a less-prototypical profile might help our over-curved layouts.

Unnecessary Guy, the wonders of the scale effect on the physics mean that, in this respect at least, our models live in a much more favourable world than the prototype. You know this is true, consider the average toy train whizzing round impossibly tight curves.

An Easter Day debate revolves in my mind Will. I don't see how the scaling effect of physics favours us particularly. A 16 ton wagon scaled down 76.2 x 3 times gives us 32g (if I have done the maths correctly) which as nearly everyone agrees is an inadequate weight even in 00 and EM.

As Martin pointed out here viewtopic.php?f=96&t=5030&start=50
one of us can easily blow a model wagon off the track while Russ estimated it would take 33 million people to blow a real one over.

Martin Wynne wrote:

Russ Elliott wrote:but I find that very slight difference doesn't really explain why our wheels still like to jump off the track so much more than the prototype does.

Loading. The presence of a few tons of load on a wheel is very effective at persuading it not to jump for joy. P4 trucks would hold the track just fine if stuffed full with lead, but then nothing locomotive-looking we could build would pull them along. Perhaps a motor on the axle of each one?

Martin.

quote="Martin Wynne"]

Russ Elliott wrote:I've struggled with some scaling sums trying to compare 10g on a model wagon wheel with 2.5t

Forget scaling the load, it is the rolling pressure which needs to match the prototype (load per contact area). Prototype wheels and rails actually deform under load from being a point contact to create a finite contact area. Models do the same (if it was really a point contact of zero area, the pressure would be infinity), but the 1:76.2 ratio destroys any similarity in contact area or behaviour. (Area changes as the square of scale, load changes as the cube of scale.)

To get a model wheel to create a scale-size contact area, the wagon would need to be almost self-destructive in weight, certainly not made of plastic. Back to Hornby-Dublo 3-rail, anyone? That didn't jump off the track.

Martin.[/quote]

Russ Elliott wrote:Yes, Martin, but if one compares sideways impact forces (e.g. the front of a flange being shoved over abruptly* to a railhead shoulder, which is what will produce a vertical component of reaction force) between prototype and model, then, unless I've dropped an almighty clanger somewhere (always possible!), the sums seem to indicate our model wheels should be far more stable than the prototype. That's what I can't get my head around.

* by abruptly, I mean in a finite amount of time, which is necessary to calculate an impact force, taking vehicle speed into account as well of course

Martin Wynne wrote:How are you calculating the sideways force on the flange? And allowing for the forward momentum of the vehicle mass?

If I take a moving P4 wagon and blow hard at it I can derail it. Do you think 76 people could blow a full-size coal truck off the track?

Scaling the prototype exactly can make a model look like the prototype. But I never did believe it could make it behave like the prototype. CJF was not daft.

Martin.

Russ Elliott wrote:Energy absorption. Hmmm. Does that scale to the fifth power of scale?

Martin Wynne wrote:How are you calculating the sideways force on the flange?

I'll dig out some sums later.

Martin Wynne wrote:If I take a moving P4 wagon and blow hard at it I can derail it. Do you think 76 people could blow a full-size coal truck off the track?

I think that would take 33 million people. (assuming fourth power scaling for force).

It seems to me that what really helps is getting more weight than the scaling would suggest.

Hi Julian,

I'm wondering if pin-point axles play any part in this. Is there any empirical evidence on the relative derailability of wagons with pin-point versus parallel needle bearings? Note that the prototype doesn't use pin-point bearings. Nor traditional tinplate train sets.

If there is any end-float on a pin-point axle, it means that only one end is ever rotating on a fixed axis. If it becomes unloaded, the other pin-point can process around the inside of the bearing as the axle rotates. This suggests that any rolling tendency to flange-climb could be a very complex interaction.

Just asking. There has to be some explanation for the jumpiness of P4 wagons.

cheers,

Martin.

Julian Roberts wrote:An Easter Day debate revolves in my mind Will. I don't see how the scaling effect of physics favours us particularly.

Even though we all know that toy model trains comer unrealistic fast? And that Andy's video clip demonstrates this can also be true for scale wheels?
I would have thought that the evidence of your eye was likely to be more convincing, but if you want the math:-
For real railway engineers, this tendency for wheels to climb the outside rail can be quantified by the Nadal Formula (see Wikipedia). Do a dimensional analysis on Lateral and Vertical forces element of that and it shows that including a scaling factor of less than 1 (in our case roughly 0.0131233596…) does in deed make it much less Ilkley our creations will climb the outside rail.

A 16 ton wagon scaled down 76.2 x 3 times gives us 32g (if I have done the maths correctly)

you have

which as nearly everyone agrees is an inadequate weight even in 00 and EM.

But I would have said surprisingly close to our accepted 25gm an axle given how big the devisor is that got us there, and that there other factors which will influence the desirable weight, like a tendency for our wagons to bounce more than the full size prototype when they hit a track irregularity.

Not that I'm trying to say our vehicles don't seem a bit more inclined to derail than we'd like, but I do think you need to look else where for a reason than any particular tendency climb the outside rail on corners .

As Martin pointed out here viewtopic.php?f=96&t=5030&start=50
one of us can easily blow a model wagon off the track while Russ estimated it would take 33 million people to blow a real one over.

I'd have said 442450 people ought to be enough, but, which ever figure you accept getting them all stood close enough to the wagon to blow directly on it is going to be tricky.

Julian Roberts wrote:An Easter Day debate revolves in my mind Will. I don't see how the scaling effect of physics favours us particularly. A 16 ton wagon scaled down 76.2 x 3 times gives us 32g

16 tons was the permitted load. The tare was 7-8 tons. So a mineral wagon could weigh anything between 7 and 26 tons, and the model, if you want to use a proportional weighing system* of say 4 grams per ton, between 28 and 104 grams.

which as nearly everyone agrees is an inadequate weight even in 00 and EM.

On what basis exactly? 50gm or 1 ounce per axle, has always seemed to me to be based on what some guy wrote back in the mists of time and a lot of confirmation bias.

*OK I can't be arsed to do the scaling calculation.

Julian Roberts wrote:

Will L wrote:

Guy Rixon wrote:A few years ago, the American Passenger Transport Association were evaluating a proposal to adopt a different wheel profile for LRT trains, one with a greater flange-contact angle. The idea was to allow higher speeds on sharp curves. Makes me think that a less-prototypical profile might help our over-curved layouts.

Unnecessary Guy, the wonders of the scale effect on the physics mean that, in this respect at least, our models live in a much more favourable world than the prototype. You know this is true, consider the average toy train whizzing round impossibly tight curves.

An Easter Day debate revolves in my mind Will. I don't see how the scaling effect of physics favours us particularly. A 16 ton wagon scaled down 76.2 x 3 times gives us 32g (if I have done the maths correctly) which as nearly everyone agrees is an inadequate weight even in 00 and EM. ...

Consider the horizontal force needed to push a vehicle back into line if it hunts and the leading flanges strike a rail. Using F=ma, the mass decreases (clearly) as the cube of the scale factor and the required acceleration as ... the square of the scale factor? Because the acceleration needed is linear both with the velocity along the track and the required velocity across the track to correct the hunting?

The limit for track-holding is on the ratio of vertical to horizontal force*, and the vertical force scales as the cube of the scale factor, while the horizontal force looks like it scales as the fifth power of the scale factor, so a P4 model ought to have 76.2^2 times the ability to resist flange climbing. But they don't, and the Dublo does.

Speculation: the really coarse-scale stuff has extended flanges that are nearly vertical. Because the chassis are solid, the wheels spend most of their time with the treads off the rails and the vertical bit of the flange pressed against a rail. This is highly resistant to flange climbing.

Further speculation: P4 trains sometimes have massively out-of-scale accelerations along the track when shunting, because the locos can accelerate so sharply and the buffers are often solid. If full-size trains were propelled jerkily like the models then perhaps they'd fall off more often.

Final speculation (and I'll shut up now): the physics arguments only apply if the weight distribution of a wagon is fairly even. If a P4 model has a sticky or distorted suspension, then wheels get unloaded and it comes off ... just like the full-size train at Lewisham a few years ago, for which Martin posted the RAIB report. I think that our suspensions are not always as freely-moving as we'd like.




*Strictly, it's the component of the horizontal force tangent to the contact of the wheel and rail, which is why the flange profile matters.

billbedford wrote:
On what basis exactly? 50gm or 1 ounce per axle, has always seemed to me to be based on what some guy wrote back in the mists of time and a lot of confirmation bias.

*OK I can't be arsed to do the scaling calculation.

Agreed. That figure has always had an element of ‘plucked from thin air’ about it. I weight my wagons to about half that based on trial and error

Cheers

Jim

billbedford wrote:16 tons was the permitted load. The tare was 7-8 tons. So a mineral wagon could weigh anything between 7 and 26 tons

If you include 21T minerals [tare 9T +/- depending on the type of brake] and 24.5T minerals [tare around 10.5T to 11T], then the maximum goes up to 35T + on two axles. Up to 45T on two axles was permitted by the mid-1960s.

Re the 50gms/1oz weighting, I have pointed out before (too many years spent perusing 'Model Railroader' from cover to cover) that it was based on fairly extensive empirical research, but all with vehicles on US-style 4-wheel trucks. So applicability to UK rigid-wheelbase wagons is certainly open to question, and it would be nice to see further comparative testing (if not necessarily on the scale 9(sic!) the NMRA resorted to). Personally, I find with (small) pre-Grouping wagons 50gms is a comfortable amount of weight without needing undue effort to fit/hide it, but if I were ever tempted towards modern high-capacity tanks or hoppers I would probably go for more. I'm currently building a 'towable' Deltic prototype as a display model , and for various reasons it has already reached around 25gms per axle, so I shall be interested eventually to see how well it may ride (on a total lack of suspension and P4 wheelsets).

Guy Rixon wrote:Consider the horizontal force needed to push a vehicle back into line if it hunts and the leading flanges strike a rail. Using F=ma, the mass decreases (clearly) as the cube of the scale factor and the required acceleration as ... the square of the scale factor? Because the acceleration needed is linear both with the velocity along the track and the required velocity across the track to correct the hunting?

I'm not at all sure hunting is ever a problem for us, a theoretical possibility I agree, but it doesn't occur until you hit the resonant frequency of the vehicle/bogie. Given our vehicles are small and light I think that frequency is going to be high and the speeds its achieved at very high even by model railway standards. I have sat in a full size mark 1 having my teeth shaken out by a hunting bogie, and I've sat on Bletchley station watching a first generation electric coming north doing a ton and hunting so much I did wonder if it was likely to join me on the platform, but I don't think I've ever seen a model do it. Crab when being pushed round corners yes, but no hunting.

The limit for track-holding is on the ratio of vertical to horizontal force*, and the vertical force scales as the cube of the scale factor, while the horizontal force looks like it scales as the fifth power of the scale factor, so a P4 model ought to have 76.2^2 times the ability to resist flange climbing. But they don't, and the Dublo does.

I'm with you on the Vertical force (V), the lateral force (L) is governed by the radius of the curve and the speed (both linear), taken together the square of the Scaling factor. As the key ratio is V/l, I reckon we get just 76.2 times the ability. but I would have though that was sufficient to explain scale trains cornering ability. And I think this is shared by P4 stock given nothing nasty in the track work and, as you point out below, fully functional suspension. Did you not believe Andy's video clip?

Speculation: the really coarse-scale stuff has extended flanges that are nearly vertical. Because the chassis are solid, the wheels spend most of their time with the treads off the rails and the vertical bit of the flange pressed against a rail. This is highly resistant to flange climbing.

I don't think the current RTR products, with quite respectable flange profiles, are any less capable of high speed cornering than the old steamroller profile with near vertical flanges (which, in any event, disappeared with Triang didn't they?)

Further speculation: P4 trains sometimes have massively out-of-scale accelerations along the track when shunting, because the locos can accelerate so sharply and the buffers are often solid. If full-size trains were propelled jerkily like the models then perhaps they'd fall off more often.

Now this is I think worth following. I suspect that part of the reason we are apparently so much more inclined to get derailments than the prototype is because the force we have available to us to pull and push our stock about are much larger in realtionship to the mass of our vehicles than on the prototype. On the Welsh Highland recently they were trying to push a very rusty NG15 2-8-2 round a corner, but the pony truck rubbing plates were corroded nearly solid so push as they might it wouldn't move. If we tried pushing a chassis like that with jammed pony trucks on a model, instant derailment. That's the difference

Final speculation (and I'll shut up now): the physics arguments only apply if the weight distribution of a wagon is fairly even. If a P4 model has a sticky or distorted suspension, then wheels get unloaded and it comes off ... just like the full-size train at Lewisham a few years ago, for which Martin posted the RAIB report. I think that our suspensions are not always as freely-moving as we'd like. *Strictly, it's the component of the horizontal force tangent to the contact of the wheel and rail, which is why the flange profile matters.

I think we can agree that faulty suspension can and does cause derailments, but just the lack of weight on the wheel is enough to cause rail climb with the flange profiles being no more or less important than at any other time?

Bill pointed out that the weight of an empty 16 ton wagon was 7 to 8 tons. So just 16g for us. Experience of coal wagons at exhibitions where the 50g weight is mostly in the detachable coal load is that the empty wagons run considerably more unreliably.

Keith corrected me. Our profile is not of the worn flange but the unworn one. But would it seem from the accident report that the P1 profile is for older locomotives and is designed to wear into the steeper angle, while on newer locomotives because they "steer" the steeper angle will not wear away so can be created from the start in the P8 profile?

As the train negotiated the curve, the first two right-hand wheels of the leading bogie of the class 37 locomotive climbed the outside rail of the curve and derailed to the outside of the curve. The third wheelset on the leading bogie stayed on the rails until the bogie reached the crossing of a set of trailing points, at which point it also derailed.

The driver of the train reported feeling a jolt and hearing a loud bang. He looked out of his window, saw sparks from the class 37’s front bogie and immediately applied the brakes. As the train came to rest with the class 37’s leading bogie over a set of trailing points, the first wheelset of the leading bogie rerailed and the second and third wheelsets became trapped between the switch and stock rails. The points were damaged and required a half set of switches, numerous sleepers and other components to be replaced.

(My italics)

Does anyone else find it "interesting" that the photo shows the second wheel of the bogie derailed inside the rail? It's difficult to see but it's clearly not outside it. That bogie did quite a gymnastic feat on the crossing - the front wheel that was derailed outside rerailed, the second wheel that was derailed outside managed to get derailed inside, and the third wheel which had not derailed already managed to get derailed, we are not told which side. Amazing the whole crossing didn't need replacing too!?

There is no evidence that the actions or driving style of the driver contributed to this accident.

Might not the speed of the train be a question? After the previous Ordsall Road accident where it became clear newly profiled P1 wheels could flange climb, and the lack of a flange lubricator on this curve for moves in this reverse direction, I'd have thought a simple recommendation could be for similar trains to go dead slow in that location and similar ones.

Might not the speed of the train be a question? After the previous Ordsall Road accident where it became clear newly profiled P1 wheels could flange climb, and the lack of a flange lubricator on this curve for moves in this reverse direction, I'd have thought a simple recommendation could be for similar trains to go dead slow in that location and similar ones.

Perhaps, but in the absence of any such instruction there was no fault in the driver's actions. And the move must have been fairly slow anyway.
Rgds

Now my maths is rather rusty and my serious maths forgotten but if I take Will's Nadel formula L/V (lateral and vertical forces) where vertical must be twice lateral and replace lateral by mv^2/r (tension in string when whirling a rock) we find that the tendency for the flange to climb is independent of the mass m and proportional to the square of the velocity v. Nothing to do with scaling at all (except we are, conveniently, moving 76 times slower) which is understandable because the formula doesn't know how big the wheel is - it looks solely at flange angle and friction. It also gives us a whopping margin of safety which would suggest that derailments are solely the fault of the builder of track and vehicle.

I think it's dangerous to use scaling factors like squares and cubes because they are an over-simplification. I suggest we have trouble because we can't scale the little things like misalignment at rail ends and, in particular, friction is going to behave very differently under our very light contact forces. Rough turning marks on a model wheel acting on the drawing marks on new rail will easily outdo the slightly rough finish on the doomed class 37 - they will be more like a ratchet on a tiny scale and don't have to lift many tons to cause trouble.

Andy on the other thread suggested that weight was the answer which I have eliminated so far. It's certainly required to get compensation working or get a spring into it's comfort zone but perhaps he's explained it himself and that it brings the materials from which our models are made into the range where they are resilient and can absorb shocks that would otherwise spell disaster. I thought we have 50g wagons because that is the weight of a whitemetal wagon and we want them all the same.

Somebody had, of course, to mention hunting and I was relieved when it was swiftly swept aside. The maths predicts that coned wheels will aalways oscillate but our track isn't straight enough, our axles insufficiently parallel and our wheels too wobbly and eccentric for it to stand a chance. Nevertheless, all these factors are just queueing up to to help tip our latest creation onto the ballast if we let it.

My final comment relates to our own conflicting experiences of different suspension systems. I suggest that small layouts with low speeds and short trains are less sensitive to technique or design than large systems where, in particular, speeds may be much higher or trains longer and it would be helpful if folk described the circumstances when commenting on things like suspension design. There's been a number of reports in the model press of the increased challenge when moving to a larger layout.

Cheers,

DaveB