REF: Rail wheel profile and "to do" list update

This section allows guests to comment or ask questions. Posts from guests require explicit approval (which generally takes a day or so), before they appear, so that we can prevent unwanted spam.
Proto87Stores

REF: Rail wheel profile and "to do" list update

Postby Proto87Stores » Sun Apr 21, 2019 9:17 pm

An important point re the wagons on the tight oval video. The model trucks (bogies) have working suspension by equalization, not raw springing. It's also almost identical to that used on the same US prototypes.

As to the N7 (not N2), it's slightly behind the current work on a full "Grand Union" for the streetcar track. Ditto for the Airfix Mineral Wagons. No technical issues, just lack of time given to the next stages.

GU bases have been completed to the pre-production stage, and the current effort is installing the various girder rail sections, crossings, etc., so that a full working model can be tested.

Image

Image

The 2-rail wiring is quite complex as all crossings will be powered, and switched as needed. I expect to use something like an Arduino with a standard multi-relay shield to simplify things somewhat for the poor signal man ;)

The situation with the N7, etc is subject to some additional radical thinking, even for me :o . I'm seriously considering checking out using 2mm axles, rather the traditional 1/8", as it could make the chassis even more open spaced inside and likely less expensive still. The Isinglass Drawings seem to support that as being closer to their prototype drawings than 1/8". But I'd have to rely on you guys knowing what the real locos axle diameters are

There are several other happenings already in progress, but I'll save those for a later update.

Andy

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: REF: Rail wheel profile and "to do" list update

Postby Julian Roberts » Mon Apr 22, 2019 5:21 pm

Andy regarding the discussion on weight in the related thread, do you have any views?

Have you already described elsewhere how you make the crossings of the street track in your photo?

User avatar
grovenor-2685
Forum Team
Posts: 3918
Joined: Sun Jun 29, 2008 8:02 pm

Re: REF: Rail wheel profile and "to do" list update

Postby grovenor-2685 » Mon Apr 22, 2019 5:33 pm

Have you already described elsewhere how you make the crossings of the street track in your photo?


They are etched, full details on Andy's website http://www.proto87.com/
Rgds
Regards
Keith
Grovenor Sidings

Proto87Stores

Re: REF: Rail wheel profile further stuff

Postby Proto87Stores » Mon Apr 22, 2019 9:15 pm

Julian Roberts wrote:Andy regarding the discussion on weight in the related thread, do you have any views?


I'm not a knowledgeable expert on wheel climbing on the prototype, or on the model. So I can only re-iterate a few observations (mostly already known or obvious ) that might help draw some useful conclusions.

1. Wheels "roll" up the path of least resistance rather than "climb" willy-nilly. And the deepest part of the wheel flange/tyre rail contact is actually always stationary with respect to the rail at all speeds. So it's the next (teeny tiny) portion of the rotating wheel/flange to touch the rail that is just descending that will either prematurely grip the edge of the rail head and "climb", or slide down instead due to insufficient friction. If the former, then the next teeny tiny rotated rail/wheel interface point needs to be less likely to continue climbing, if the climb is to correct itself.

So I would look for clues in the shape of the flange approaching the rail, rather than worry about the shape directly at the bottom. That's going to be a mix of the horizontal cross-section of the wheel flange at rail height, as well as the vertical cross-section of the wheel. Not just the contact point curve.

2. The flatness of the vehicle chassis and its c of g symmetry. as well as the flatness of the track, can have huge effect on whether any particular wheel descends fully (and evenly weighted) to the rail at its lowest point. If a rigid chassis is not absolutely true, then it's going to dynamically rock across two wheels, leaving one higher than the rails sometimes and thus much more likely to climb. This leads to the issue that it's not easy to make a model chassis so true that all wheels bear fully down on flat rails all of the time. And of course making truly flat track "ain't easy either".

Symmetrical equalization can convert a non-true chassis to have all wheels fully and equally bearing down, whether on flat or not flat track. Springing can convert a non-true chassis to have all wheels bearing down, but with slight differences in wheel weights as it compensates for the chassis un-trueness. Ditto for springing adapting to not flat track. if both chassis and track are un-true, then springing differences on each wheel are due to the net of both un-true effects "summed" (i.e. taking into account adding both + and - values). And of course making the practical tiny, weak, springs on each wheel precisely the same value is possibly more difficult than making a chassis precisely true.

3. Hardened Ball bearings dropped onto a flat hard steel plate will "bounce" quite dramatically, due to the extreme rigidity of the elasticity of hard steel. The bump energy is reflected. Model wagons OTOH generally are made of much softer and lighter materials and will "give" and absorb the energy of impulses of bumpy track with much less bouncing (reflections). However adding springs to the wheels will cause much more temporary energy storage from track bumps and so increase the potential energy dissipated in bouncing back, once the impulse has ceased. This will also dynamically change the weight on each wheel, for short time after each bump.

4. Finally, I would think that the frictional force that enables flange climbing is applied more sideways, due to the centripetal(?) force from the rail curvature, than up. I.e the upward climbing force is only a small fraction of the sideways centripetal force. Which would infer that the wagon weight, acting fully downwards, works against flange climbing more, if the wagon is heavier.

Hope this provides some food for thought.

Andy

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: REF: Rail wheel profile and "to do" list update

Postby Julian Roberts » Wed Apr 24, 2019 6:21 am

Thank you very much Andy.  Regarding your video that Will put on the related thread - how much did the vehicles weigh, do you know?  I did dip into your long discussion with Will and others regarding the competing merits of compensation and springing but couldn't find what I seem to recall reading somewhere at the time - your approach to vehicle weighting.  So I wonder if you could share your thoughts again on that?  I focused my point in the related thread around 4 wheel wagons, and of course you don't have those as a prototype there, but it would be interesting to be reminded of your thoughts nonetheless regarding the prototypes you do model. And also to be reminded what is the flange depth of your wheels.

Proto87Stores

Re: REF: Rail wheel profile and "to do" list update

Postby Proto87Stores » Wed Apr 24, 2019 4:58 pm

Julian

The unqualified answer to your question is:

Video demo - Car weight breakdown

Plastic body 22gm, 12.5 cm long
Full length Mazak floor with original plastic rigid trucks and Alan Gibson Plastic centered wheels 54 gm

Total out of the box with no suspension 76 gm

As shown in the video with swapped out Kadee metal equalized trucks with NWSL solid NS wheels, added 12 gm net below the floor.

So Total with metal trucks, wheels and suspension, 88gm,

To equate that with say a P4 4 wheel wagon, there was 44 gm bearing down weight centrally per 4 wheel truck. Operating radius was 20 cm with a truck wheel base of 18 mm. Wheel dia 9.5 mm. P:87 flange depth is 0.33 mm. I'll leave it to you guys to calculate the attack angle for P4 equivalence.

My qualified answer is that the only thing I thought I'd "proved" with the demo was that P:87 wheel/rail interface, with proper working suspension, and moderate weight, does not climb, even under the most extreme conditions. And such conditions being far worse that would ever be encountered on any scale model railroad. I didn't try the experiment with either lighter or heavier cars. But I suspect cars that were too light weight would climb on the many small bumps in the track. And cars that were too heavy would roll over on entering the curves at high speed.

FWIW, I don't consider weight to be the only factor affecting wheel climbing unless all other suspension issues have been optimized first.

A couple of analogies worth thinking about are:

1. Hitting a light golf ball with a golf club will usually send it up in the air and along for a distance of a hundred feet or so. Try to do the same by hitting a football with same club, and you will be lucky to move the ball 10 feet along the ground and less than few inches above it.. Compare that concept with a light vs. a heavy model wagon hitting track bump.

2. Imagine a precision but wavy sideways shallow groove ( say 5 mm deep) on a slightly sloping down flat surface. Imagine then the result of rolling a light table tennis ball down the groove, with rolling a clearly much heavier steel ball of exactly the same diameter. As the rolling speed increases, which ball do you think would climb the edge of the groove first?

Andy

Proto87Stores

Re: REF: Rail wheel profile and "to do" list update

Postby Proto87Stores » Sat Apr 27, 2019 3:09 pm

Post by davebradwell » Wed Apr 24, 2019 9:04 pm
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.


The Nadel formula refers to the partial upward sliding (in our case climbing) force generated by pushing horizontally against an upward sloping frictional surface interface. It doesn’t mention the optional separately applied downward countering force of some amount of added weight. Certainly substituting the mv^2/r centripetal effect for the horizontal force proportionally affects the upward sliding force due to the Nadel formaula. But that doesn’t at all affect the fact of the constant gravitational downward force due to the original static weight (not just mass) of the vehicle. The same weight is there even when v=0. So more static weight will in fact bear down harder on the rail, but as a constant, independently of the rotational speed.

So in fact, more weight is an answer. It has not been eliminated. Otherwise I agree with the independent conclusion that most derailments are first suspension/construction related.

In my particular case I would definitely defend the consistent operation of the same equalization for primary suspension, at all speeds and train lengths, as I published video evidence of the same design suspension running just as well at high speeds with long (and heavy) trains being both pulled and pushed.

e.g.


Andy

Proto87Stores

Weight matters!

Postby Proto87Stores » Fri May 03, 2019 3:31 pm

Re my previous post of around April 25.

Re Nadel's formula and Dave Bradwell's analysis, but correctly taking into account the added downward constant force due the vehicle weight, it is clear that rail climbing derailments are reduced by making the vehicle heavier, regardless of the suspension system.

This strikes me as being extremely important for the successful implementation of general P4 modelling. As far as I am aware, that was not included on the Digest Suspension Section. And it is frequently contradicted in many member posts.

Andy

Proto87Stores

Re: REF: Rail wheel profile and "to do" list update

Postby Proto87Stores » Sun Jun 09, 2019 5:39 pm

Julian,

I'm curious as to what conclusions you were able to draw from my answer on the weight of the free running box cars in my video?

My subsequent thinking about Nadel's formula leads me to the actually obvious conclusion that the major solution to preventing P4 and P87 derailments in otherwise compliant track work is to weight vehicles heavily, with as low a c of g as possible.

Unless the track work or suspension has a fault that can create a least resistance path to actually lead a wheel over the top of a rail, then one can start at at the extreme case and say that, Miss Piggy like, if you can't lift a model vehicle, not only should you not eat it, But more importantly there is no way it can lift itself over a rail without a huge external lifting force. Working backwards from there, it's clear that provided that each individual wheel weight always exceeds its instantaneous Nadel value, it can't derail.

And if the overall vehicle weight is only slightly over the Nadel value per wheel, then the least problems will occur if the instantaneous weight on each wheel is equal. And that probably explains why so many here have derailing problem steam loco bogie and pony trucks. Any usable suspension solution must include equalizing their wheel weights as well as, and to match, the weight of the drivers. Similarly for the centre wheels of six-wheel vehicles.

Andy

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Weight matters!

Postby Julian Roberts » Mon Jun 10, 2019 7:12 am

Proto87Stores wrote:Re my previous post of around April 25.

Re Nadel's formula and Dave Bradwell's analysis, but correctly taking into account the added downward constant force due the vehicle weight, it is clear that rail climbing derailments are reduced by making the vehicle heavier, regardless of the suspension system.

This strikes me as being extremely important for the successful implementation of general P4 modelling. As far as I am aware, that was not included on the Digest Suspension Section. And it is frequently contradicted in many member posts.

Andy


Andy I do apologise for not responding to your post here around Easter. Somehow I had not seen these subsequent ones. What you write here is exactly the realisation I came to before even starting P4 modelling. I don't know if you can access Snooze back numbers but I wrote about this general subject in No 199.

:idea: I was going to suggest on the related thread a formula that the smaller the flange the more the weight needs to be. :idea: :?:

I think that the scaling effect on the weight, i.e. that it reduces 76.2x76.2x76.2 acts to our DISadvantage. (I think the opposite was being argued on the related thread). The idea that an empty P4 wagon scaled down in weight to somewhere around 16g or 8g per axle ought to stay reliably on the track seems daft to me. A 25g per axle minimum seems essential.

My Crab Comet Conversion thread was largely concerned with the suspension and concluded that the weight on the pony needs to be far more than most modellers put there. Indeed considerably more than on the front driving wheel. I don't know how many times I've seen pony trucks derailing when leading the direction of the loco. Where that leaves the adhesion is another question, all the more in a 2-6-2! My conclusion is that the more wheels there are the greater the problem. A tank loco can be stuffed with lead. A tender loco less so unless the motor is put in the tender with driveshaft to the loco.

My other related realisation was that it's the leading wheel which is most critical so I think weight needs to be optimised on it. As most stock and locos must generally go equally well in both directions, there is a 'leading' wheel at each end.

For a given radius curve and length of fixed wheelbase, the middle wheels of an 8 coupled loco have less sideways thrust (from the track on curves) than a symmetrical 0-6-0 (most aren't exactly symmetrical but are near enough that this still applies). The central wheels of a 6 wheel coach will be the next thing I find out about (much longer than a tender so more sidethrust) having just bought a kit for one. I reckon though that obviously there needs to be enough weight to keep them down on the track but still it's the outside wheels that really need the weight. At least there's plenty of room for weight and I'll start with the 25g per axle idea.

I cant follow such complex stuff as the suspension section of the Digest but I get the impression it makes out compensation to be full of flaws which I don't think is the case at all. Critical to the way I think is the optimisation of weight on individual wheels, and with compensation that weight distribution will stay constant whatever position the wheel takes as it responds to track level changes.
Last edited by Julian Roberts on Tue Jun 11, 2019 3:30 pm, edited 1 time in total.

proto87stores

Re: REF: Rail wheel profile and "to do" list update

Postby proto87stores » Mon Jun 10, 2019 11:13 pm

Julian,

Thanks for the quick reply.

My position is that Nadel's formula is proven and the consequences are that models should be heavy (or at least heavier) compared with the worst expected impulses from bumpy track, and evenly and equally loaded per wheel. That latter reasoning supports the running requirement that the first occurring wheel derailment is then a whole vehicle derailment. It would seem obvious that given any track deformity that is just slightly less that would result in rail climbing for average wheel weight, then any lighter loaded wheel will then climb and the lightest most easily before any other.

I see no reason to scale weight. Nor to design suspension systems that break the equal weight per wheel requirement.

Andy.

User avatar
Guy Rixon
Posts: 909
Joined: Sun Mar 27, 2011 6:40 pm

Re: Weight matters!

Postby Guy Rixon » Tue Jun 11, 2019 7:12 am

Proto87Stores wrote:Re my previous post of around April 25.
Re Nadel's formula and Dave Bradwell's analysis, but correctly taking into account the added downward constant force due the vehicle weight, it is clear that rail climbing derailments are reduced by making the vehicle heavier, regardless of the suspension system.
Andy

This will only hold true if the added weight does not increase the lateral forces.

User avatar
Noel
Posts: 1975
Joined: Wed Jun 23, 2010 1:04 pm

Re: REF: Rail wheel profile and "to do" list update

Postby Noel » Tue Jun 11, 2019 10:07 am

Andy, I have read your last post several times, and I'm afraid I am baffled by it.

proto87stores wrote:My position is that Nadel's formula is proven


The Nadal [not Nadel] formula does not define the forces involved. What it does is use the expected vertical and lateral forces to define the minimum requirement for the relationship between wheel and rail. Equally it can start with that relationship and tell you if the relationship between resulting vertical and lateral forces is acceptable or not. It implies that a vehicle which is satisfactory on one rail profile may not be on another unless the wheels are changed.

proto87stores wrote:and the consequences are that models should be heavy (or at least heavier) compared with the worst expected impulses from bumpy track,


As indicated above, I am not convinced that the conclusion you appear to draw follows from the formula [although I don't necessarily disagree with the conclusion]. Also you fail to define, or indicate how a reader might define, "the worst expected impulses from bumpy track", or how they might be measured in any practical application.

proto87stores wrote:and evenly and equally loaded per wheel


So far as a static analysis is concerned, I would accept this, but a moving vehicle is a dynamic system, and the loadings will change over time, for a variety of reasons, such as hunting, spring reaction time, vehicles not parallel with the track [particularly in models given the amount of built-in slop], etc.

proto87stores wrote:That latter reasoning supports the running requirement that the first occurring wheel derailment is then a whole vehicle derailment.


:?:

proto87stores wrote: It would seem obvious that given any track deformity that is just slightly less that would result in rail climbing for average wheel weight, then any lighter loaded wheel will then climb and the lightest most easily before any other.


Slightly less than what? It may not be the lightest loaded wheel that climbs, if you mean the lightest vertical loading, as you appear to do, as it depends on the ratio between vertical and lateral forces acting on the wheel.
Regards
Noel

billbedford

Re: REF: Rail wheel profile and "to do" list update

Postby billbedford » Tue Jun 11, 2019 12:53 pm

proto87stores wrote:My position is that Nadel's formula is proven and the consequences are that models should be heavy (or at least heavier) compared with the worst expected impulses from bumpy track, and evenly and equally loaded per wheel.


Looks like confirmation bias -- We like heavy stock, therefore heavy stock works better...

Proto87stores

Re: Weight matters!

Postby Proto87stores » Tue Jun 11, 2019 4:23 pm

Guy Rixon wrote:
Proto87Stores wrote:Re my previous post of around April 25.
Re Nadel's formula and Dave Bradwell's analysis, but correctly taking into account the added downward constant force due the vehicle weight, it is clear that rail climbing derailments are reduced by making the vehicle heavier, regardless of the suspension system.
Andy

This will only hold true if the added weight does not increase the lateral forces.


Nadel's formula calculates the proportion of upward climbing force due to the anticipated sideways force and friction, contact angle, etc. The same added weight also increases the downward force and in the reverse same proportion. Provided your contact angle is steeper than 45 degrees (approx with friction,etc), The added weight has a greater effect downwards than sideways.

I think it was Bill Bedford who suggested earlier that he could just blow his ideal weight of a P4 model wagon sideways off the track. But Nadel and I are pretty sure he couldn't do that to a heavy wagon. ;)

Andy

User avatar
Guy Rixon
Posts: 909
Joined: Sun Mar 27, 2011 6:40 pm

Re: REF: Rail wheel profile and "to do" list update

Postby Guy Rixon » Tue Jun 11, 2019 5:27 pm

A thought about models with moving suspensions: the vertical force needed to defeat rail-climbing, as in the Nadal analysis, can be more easily defeated by friction if the vehicle is light.

I suggest that the frictional force in the suspension, which opposes downward movement of each wheel, goes approximately as

F = aW + S

where W is the weight on that wheel, a is a coefficient that is linear with the coefficient of friction, and S is a stickiness term due to imperfections in the structure; I'm guessing that S is constant for each wheel but might vary a lot between wheels on the same vehicle.

In a perfect model, S goes to zero: no rough edges or tight spots anywhere in the suspension travel. In an imperfect but heavy vehicle, aW can be much greater than S, so S doesn't matter. In a light and imperfect model, if S might become large compared to aW for one wheel, meaning that the wheel becomes unloaded. Then it climbs and comes off.

Edit: a is not the coefficient of friction but is probably linearly related to it.

Proto87Stores

Re: REF: Rail wheel profile and "to do" list update

Postby Proto87Stores » Tue Jun 11, 2019 6:26 pm

Hopefully I haven't posted this twice. I was interrupted for quite a while before sending it initially

Noel wrote:Andy, I have read your last post several times, and I'm afraid I am baffled by it.

My apologies for name spelling errors

proto87stores wrote:My position is that Nadel's formula is proven


The Nadal [not Nadel] formula does not define the forces involved. What it does is use the expected vertical and lateral forces to define the minimum requirement for the relationship between wheel and rail. Equally it can start with that relationship and tell you if the relationship between resulting vertical and lateral forces is acceptable or not. It implies that a vehicle which is satisfactory on one rail profile may not be on another unless the wheels are changed.

Nadal's formula does nothing except specifically calculate the upward wheel climbing force for a given sideways force, contact angle and friction.

proto87stores wrote:and the consequences are that models should be heavy (or at least heavier) compared with the worst expected impulses from bumpy track,


As indicated above, I am not convinced that the conclusion you appear to draw follows from the formula [although I don't necessarily disagree with the conclusion]. Also you fail to define, or indicate how a reader might define, "the worst expected impulses from bumpy track", or how they might be measured in any practical application.

One simple practical way would be to make a spiral test track and run an equalized wagon down it, with a range test weights. Note at which radius and weight it derails. - Otherwise you have to know the exact contact angle between the rail and wheel profiles you are using.

proto87stores wrote:and evenly and equally loaded per wheel


So far as a static analysis is concerned, I would accept this, but a moving vehicle is a dynamic system, and the loadings will change over time, for a variety of reasons, such as hunting, spring reaction time, vehicles not parallel with the track [particularly in models given the amount of built-in slop], etc.
Precisely. Nadal's formula has no time components, so it is valid at all instants, static and dynamic. So whatever per wheel weight changes (plus or minus) your suspension design applies, affect your track holding. This means that for springing, Hooke's Law is only beneficial 50% of the time

proto87stores wrote:That latter reasoning supports the running requirement that the first occurring wheel derailment is then a whole vehicle derailment.


:?:

I thought it was pretty obvious that only one wheel derailing is enough to consider the vehicle derailed.

proto87stores wrote: It would seem obvious that given any track deformity that is just slightly less that would result in rail climbing for average wheel weight, then any lighter loaded wheel will then climb and the lightest most easily before any other.


Slightly less than what? It may not be the lightest loaded wheel that climbs, if you mean the lightest vertical loading, as you appear to do, as it depends on the ratio between vertical and lateral forces acting on the wheel.


Just less than the minimum weight need to counterbalance the Nadal formula's calculated up ward force as per your un-enumerated example situation above. See all above to figure what that weight should be.

Andy

davebradwell
Posts: 1174
Joined: Fri Jan 18, 2019 3:48 pm

Re: REF: Rail wheel profile and "to do" list update

Postby davebradwell » Tue Jun 11, 2019 7:17 pm

Oh no, we're back to this stuff again.

There have been a few efforts recently in the Forum and magazine to prove that a wheel will just throw itself off the track (flange climbing if you like) and that weight is a factor. Can we just agree that such conclusions are fundamentally flawed and we get derailments because of imperfect workmanship in translating the Utopian drawing into the real world.

It should be obvious that a simple analysis is completely inadequate as soon as the thing moves (we know weight does matter) but we now enter the world of complex maths and differential equations to investigate what happens when a wheel hits a rail joint or other discontinuity. As no-one on the Forum seems about to try we'll just keep guessing, so why not try a methodical experimental approach. The only attempt at this that I've seen was back in MRJs 101 and 104 when Peter Kirmond rolled 0 gauge wagons through a couple of obstacles and recorded the results. It's worth a look.

If you're still determined to study the theory, then Google train suspension design but unless you've done a maths course at university and can remember it you're stuffed after line 1.

DaveB

Proto87stores

Re: REF: Rail wheel profile and "to do" list update

Postby Proto87stores » Tue Jun 11, 2019 9:38 pm

davebradwell wrote:Oh no, we're back to this stuff again.

There have been a few efforts recently in the Forum and magazine to prove that a wheel will just throw itself off the track (flange climbing if you like) and that weight is a factor. Can we just agree that such conclusions are fundamentally flawed and we get derailments because of imperfect workmanship in translating the Utopian drawing into the real world.

It should be obvious that a simple analysis is completely inadequate as soon as the thing moves (we know weight does matter) but we now enter the world of complex maths and differential equations to investigate what happens when a wheel hits a rail joint or other discontinuity. As no-one on the Forum seems about to try we'll just keep guessing, so why not try a methodical experimental approach. The only attempt at this that I've seen was back in MRJs 101 and 104 when Peter Kirmond rolled 0 gauge wagons through a couple of obstacles and recorded the results. It's worth a look.

If you're still determined to study the theory, then Google train suspension design but unless you've done a maths course at university and can remember it you're stuffed after line 1.

DaveB


I suppose the surprising conclusion you are making is that 00 and EM Chaps must be far more skilled at putting chassis together correctly than P4 Chaps? Because their derailment rates overall are far lower that P4. Nothing to do with their 2 x deeper flange depth?

Andy

User avatar
Will L
Posts: 2516
Joined: Sun Jul 20, 2008 3:54 pm

Re: REF: Rail wheel profile and "to do" list update

Postby Will L » Tue Jun 11, 2019 10:18 pm

I think there is little doubt that sufficient weight is necessary to stop the leading wheel from climbing the rail on corners, This is the implication of Nadal's formula, not to mention the real world accident investigations that Julian is fond of quoting. None the less, to assume that just because enough is necessary adding more than enough will produce any further improvement is a mistake, and what I think Dave is, rightly, objecting to. Surely, once somebody has demonstrated that they can get reliable running with a given weight over a given radius, anybody who can't get the same reliable running should realise they need to look at track and suspension faults, rather than thinking that adding weight is the right answer. So, thank you Andy, I think you have successfully demonstrated that it can be done with reasonable weights and ridiculous sharp curves.

Examples.
1. Suspension faults

Stock with no suspension or with a suspension fault may well have too little (or no) weight on one wheel, which may well then climb the rail at a corner as a result (probably on random occasions depending on several other factors). A bigger flange may save the day, as it may lift the wheel sufficiently to take enough weight to stop it climbing further before it tops the rail head. Adding extra weight may make this less likely, but fixing the underlying problem is a better answer, requiring no additional weight (or for that matter deeper flanges).

2. Bouncing

Problems with stock bouncing enough to derail when they hit a track defect may be ameliorated by adding weight, but the underlying problem is the track fault, so fixing that one problem is a better answer than adding weight to lots of vehicles.

3. Julian's observations about the tendency for leading pony trucks and bogies to derail.

Theory, and my own experience, says that a pony or bogie which really is free to find its own path and that carries sufficient weight to stop it climbing, wont derail without other provocation (see previous examples). And sufficient weight isn't actually all that much, when they really are free to find their own way. Nadal's rule of thumb suggests that if the lateral force between wheel and rail is more than half the weight on the wheel, it is likely to climb. These lateral forces come from the radius of the curve, the speed round it and the weight of the thing doing the cornering, for an unconstrained pony or bogie, these lateral forces will be very low. They must, of course, be allowed sufficient clearance to move relative to the rest of the loco, as contact with some other part of the loco is likely to cause a rail climbing derailment, as jamming the wheel against the front footsteps, or whatever, will produce a sudden sharp jump in lateral forces. As will any lack of free movement in whatever method is chosen to attach pony/bogie to the loco. The true fix is to provide the necessary clearance, or freedom of movement. Added weight might help but then again it might not.

If you now consider applying side control springing, you are immediately going to introduce increased lateral forces for which an increase in the weight on the wheels will be necessary. One thing that playing with CSBs has taught me is that judging the right stiffness of functional springs in our models is tricky*, and more often than not such springs need to be a lot softer than you might expect. I suppose you can win this one by piling huge amounts of weight on to the pony/bogie until it overcomes the lateral forces cause by an arbitrarily chosen bit of spring wire, but I'd suggest better practice would be to settle for a reasonable weight (e.g. what experience has shown works on a wagon) and tuning the spring till you get reliable running. Either way you are in trial and error territory, which is why I've given up on side control springs. I would also note that I have never ever seen a model run in a way that suggested that adding side control springs would produce a visible improvement.

If you want to transfer loco weight onto a pony truck or bogie, you immediately get more weight on the wheels. This helps those determined to fit side control springs, but you inevitably constrain the ability of the pony truck/bogie to move freely, so if it derails, as like as not, it is the result of these constraints not any failure to meet the Nadal rule of thumb. Finding and fixing the issues produced by these constraints will be more likely to solve any problems than trying to fix it by carrying a disproportionate percentage of the loco weight. Fotunately carrying loco weight on a bogie isn't usually difficult to achieve, particularly with wheel arrangements where proper balance demands it (e.g. any 4-4-0 or 0-4-4). Achieving this with pony trucks is a lot harder, mostly because the pivot points and the weight transfer arrangement don't coincide. That said I am currently building a LNER F6 2-4-2 which I intend to fit with CSBs acting on all wheels and radial trucks front and back. We'll see how I get on but I’m hoping radial trucks will prove easier to achieve than weight carrying pony trucks would be.

(* unless you use CSB which takes the guesswork out of providing springs that are right for your loco)

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: REF: Rail wheel profile and "to do" list update/weight matters!

Postby Julian Roberts » Wed Jun 12, 2019 7:30 am

I had the following written before I saw your post Will. Lots to think about there.


From the current Snooze No 212 The Bardfield Chronicles
20190612_074126-1.jpg

I think what my approach (disproportionately weighting the leading wheels) does is compensate for all-too-possible track imperfections. I get my kick not from owning a layout but making stock (and occasionally some trackwork) for a shared one, so I've got little control over the environment this stock will run on.

I described on the Crab thread how I increased the weight on the front pony (by decreasing the spring force in the front driving wheel) till it would not derail even on this point on the straight road. Looking at it closely like this it's clear the straight route blade and diverging stockrail are not up to standard (which is why I'm remaking it, illustrated another thread).
20190514_195509-1.jpg

By increasing the Crab pony weight till it coped even with this I possibly have an OTT degree of optimisation. But my other locos didn't derail there and I think a pony wheel particularly has what I can only describe as a twitchiness to it when leading.

Here is another example of a point giving trouble. My loco with optimised weight on the front wheels sails through this point (made using undercut blades). The other one derails there. This was not a random derailment - the same thing happened to both locos on repeated efforts to understand the problem.



I'm not saying that without optimisation all locos fall off at this potentially tricky location, merely that optimisation optimises the likelihood of it not falling off - at places where our workshop skills don't rise to the Utopian drawing!
You do not have the required permissions to view the files attached to this post.
Last edited by Julian Roberts on Wed Jun 12, 2019 8:33 am, edited 1 time in total.

davebradwell
Posts: 1174
Joined: Fri Jan 18, 2019 3:48 pm

Re: REF: Rail wheel profile and "to do" list update

Postby davebradwell » Wed Jun 12, 2019 8:27 am

I suspect if you just file an angle on the gauge face of those blade tips, Julian, in line with the other correspondence, all will be much improved. Some of my early efforts weren't wonderful but they've been poked and filed until they work. Replacement would be a last, desperate measure when 00 wheels were seen as the only alternative.

DaveB

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: REF: Rail wheel profile and "to do" list update

Postby Julian Roberts » Wed Jun 12, 2019 8:44 am

Too late Dave I'm afraid! It was felt that other people had problems there too so was worth another go. Nothing wrong with the crossing however.
You do not have the required permissions to view the files attached to this post.

User avatar
Paul Townsend
Posts: 964
Joined: Mon Sep 14, 2009 6:09 pm

Re: REF: Rail wheel profile and "to do" list update

Postby Paul Townsend » Wed Jun 12, 2019 10:15 am

I reckon you would have got away with just replacing the switches and a very small repositioning of the stock rail on existing rivets.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: REF: Rail wheel profile and "to do" list update

Postby Julian Roberts » Wed Jun 12, 2019 4:17 pm

Yes Paul possibly but access to layout is limited timewise and it's a lot easier to make from new at home than stoop over the layout fiddling around making tiny adjustments. (By the way this point was not the one featured in the video; I'm sure people realize that but just thought I should clarify.)


Return to “Guest Book”

Who is online

Users browsing this forum: ClaudeBot and 0 guests