Reverse Superelevation ?

[gregh]

Real railways use superelevation (or ?banking? in road terms) on curves. The reason for the super is to minimize rail and wheel flange wear ? the amount of super is related to the curve radius and track speed. It also results in a little more comfortable ride for passengers.

Normally the outer rail is higher than the inner rail, but I believe that on our models we should go the other way !!

We don?t care about rail wear or passenger comfort, but it?s nice if the trains stay on the track. With my un-prototypical sharp curves and light vehicles, I often get very impressive derailments where the whole train ?falls? into the middle of the curve. This is because the train stretches right round the curve and the forces are acting towards the centre of the curve, pulling the whole thing over. It?s especially bad when the curve is on a grade.

By accident I laid some track on a 900mm radius semicircle with reverse super. ie the outer rail was a few mm LOWER then the inner. Hey presto! I can run longer trains around the curve than I could before.
Anyone else tried it or have any comments? Or does it just happen to me ?

 

[Rhinochugger]

Interesting, Greg, because I've argued that a certain amount of superlevation is necessary in the garden.

I run a minimum 8 ft diameter, and had a couple of places where the live steam Shay derailed. Superelevation cured that, and hasn't posed a problem to other stock.

Two things may be the cause. Five 1:20.3 bogie wagons (including caboose) is the maximum that I can fit into my passing loops or Six 1:22.5 bogie wagons (including caboose). However, on the 1:20.3 consist, I have the huge weight of the brass caboose at the back.

I think superelevation also suits stock with body mounted couplers.

All the 1:22.5 stock runs on metal wheels. That makes a substantial difference in the weight department, and a low c of g as well.

Just goes to prove that you've got to lay your track to suit the type of stock you run.

 

[PaulRhB]

Your idea of reverse elevation might work with low speeds as you are directing the force down rather than up, (which is why your stock stringlines at present),but it will mean the slightest rough joint will have wheels riding up and over the outside of the rail. I suspect you'll be trading derailing stock for derailing locos as the powered wheels are more likely to climb over the rail. I suspect part of your present problem is the rolling resistance of the train and you should also aim to put heavier wagons at the front of the train to prevent stringlining as they do on the prototype.

 

[spike]

I suspect part of your present problem is the rolling resistance of the train and you should also aim to put heavier wagons at the front of the train to prevent stringlining as they do on the prototype.

Most of the time I put some heavy stuff near the loco, sometimes if an empty hopper or two is up front I weigh them
down with stones.

 

[Aljosha]

This derailment touches on one of my main grieves with G-scale!

Sadly, the weight is NOT to scale.
So a real OBB 2095 weighs 32tons - the LGB one should therefore be 1.4tons !!!
Same for the wagons, they would stay alot better on the track if they only were heavier!

If our railway was at least a 'little more' to scale then
- the locos would pull more and better and uphill
- the wagons would stay on track
- alot less worry about light wagons behind loco, heavier further back!

Dreaming....

Some pics cos pics are nice:
- testing the Croc, but wagons in 'good' order...

- more testing, the circle around the pond = always good for derailments of lighter cars

 

[coyote97]

okok...it might help preventing accidents.
But it looks bad when a train whipes to the outside!

Making the rolling stock heavier leads to two things:

1.) the standing qualities of the cars are getting better.

2.) the trains get shorter automatically

Dont forget that on prototypical operation there are orders how to build up long trains when they go in steep and curved lines.
Sometimes therefore its neccessary to switch in a loco in the middle to reduce pulling forces on couplers.

To avoid derailments, just make your cars heavier and your trains shorter. Thats prototypical!
When u want to go with long trains, unsharp your curves and use lesser grades.

So, looking on it like this, the rules are quite simple.....


Greetings

Frank

 

[ntpntpntp]

So a real OBB 2095 weighs 32tons - the LGB one should therefore be 1.4tons !!!

er... no, you have to calculate weight based on volume not linear scale. If you're assuming the scale is 1:22 then the weight is 32 divided by (22x22x22) = about 0.003 tons or about 3kg.

I agree that most model rolling stock is too light and benefits from additional weight. As has been said, even fitting metal wheels can have benefit.

 

[Aljosha]

So a real OBB 2095 weighs 32tons - the LGB one should therefore be 1.4tons !!!

er... no, you have to calculate weight based on volume not linear scale. If you're assuming the scale is 1:22 then the weight is 32 divided by (22x22x22) = about 0.003 tons or about 3kg.

I agree that most model rolling stock is too light and benefits from additional weight. As has been said, even fitting metal wheels can have benefit.

Oops, this shatters my perception of this issue, how sad! ;-) Would have been nice to think of a ideal weight for the 2095 of, say12kgs! Ok, my maths is worse than bad, and this won't enter my head that quickly! But I'll try.
On the other side, it then seems that we have to make poor 2095 'overweight' (=5kgs ish) to do its job better. Especially if we make the cars heavier, too.
Anybody thought this thru at all, ever?!

 

[ntpntpntp]

On the other side, it then seems that we have to make poor 2095 'overweight' (=5kgs ish) to do its job better. Especially if we make the cars heavier, too.
Anybody thought this thru at all, ever?!

Hi Markus,
well some guys in the smaller scales certainly get hung-up about "scale weight" etc. Must admit it's not something I worry about, but I know what you mean - why do some of our toy locos need additional weight just to pull a few wagons that you'd expect the real thing to handle? I guess there may be several issues here: one that I suspect has a lot to do with it is the relative increase in drag and lower momentum of our stock in comparison to the real thing, ie. it just doesn't roll in the same way. Plus probably some physics involving available traction to loco wheels as size and weight decreases?

But whatever, it seems to be a fact of life that we have to artificially fiddle with weight of models to get closer to how we think they should behave to emulate the real thing.

 

[gregh]

There are many good comments here from everyone, so I?ll try to address some of them without quoting you all.

Rhinochugger is certainly correct in saying ?just goes to prove that you've got to lay your track to suit the type of stock you run.?

Anyhow, today I bit the bullet and put 1.5mm thick styrene pieces under my sleepers on the inside of the 1m radius semicircle. Got out the same train that derailed yesterday and today no problem! May be due to quantum theory or something else, but it seems to be an improvement. Time will tell.

I have all plastic wheels. Steel wheels or a lower centre of gravity are certainly useful. Where I have a wagon with high c of g, I glue steel washers to the back of my plastic wheels and put some lead on the bogie.

Not sure that putting reverse super will cause the leading wheels to derail more often, but it?s a thought to consider.

I?m not sure any prototype trains were limited in length by sharp curves. Just the mass was limited by grades and to lesser extent, curves.

I like steep grades and am stuck with some 1 metre rad curves. I am happy to double head or push-up, or shorten trains to get them up the hills if one loco can?t make it, BUT what I?m aiming for is this:

if a given loco can haul a certain load up a grade, then I want to get it around my curves too.

I?m quite astounded by the weights people speak of. 5 kg locos and 1.5 kg wagons are way beyond what I have. My heaviest locos are 2.5 kg and bogie wagons 500-600grams.

The derailed train in the pic in my original post, had seven bogies and seven 4-wheelers, weighing 7kg and had a drag force of about 300g on the level. On the 1:25 grade this drag force goes up to 650g. My loco has no problem hauling this up the grade. Just for interest, this is how I measure drag force using a 5kg fishing scale.

 

[CoggesRailway]

Hey this a really interesting thread.

I two have some of these issues. I have found I CAN run long trains with weight in the loco and leading third of the train AND running at a nice crawl. To do this I have a weighted LGB Alco and it's tractive effort is amazing. I have never been able to stop it- if it does "stringline" it will just pull the derailed train like a plough through the veg beds till I run to the controller :-(

However my main point is this- we can't expect our trains to behave prototypically when our track simply isn't. Many of the shots of railways on here (inc mine) have bends that would not even be present on a tiny industrial railway let alone any kind of mainline. So my solution to this it to going to get a bit more like Brunel and rebuild letting nothing comprimise my gradients and curves. This is going to involve cutting across a lawn and knocking through a wall- but I am going to do it anyway. If you look at the railways on youtube from the states- they run protypical consists (or closer than we do) why? cos they have prototypical bends and gradients. From the videos looks like main131's set up is ike this too. I suppose it is a lot about how blessed with space you are, and how willing you are to "force a route" through SWMBO's flowers etc.

But love the thread- this is my main disapointment since starting. Really my railway whilst a long run must really be viewed as a little industrial railway- and as such- 0-4-0s with a couple of wagons are very happy- big lashups less so.

 

[JRinTawa]

As the H&MGR only runs short trains I can't speak from practical experience on how long trains perform but I do prefer a bit of superelevation on a railway if your can. I think to keep rolling stock bearings/axle boxes well lubed and my personal preference, using metal wheels, helps greatly in reducing rolling resistence, and hence tipping over.

 

[PaulRhB]

Not sure that putting reverse super will cause the leading wheels to derail more often, but it?s a thought to consider.

It raises the chance of derailing on models and the prototype especially at higher speeds. The flanges are there as the last resort safety when the wheel tread profile fails to keep the wheels in the centre of the track. If the outer rail is lower the Centre of Gravity is now towards the outside of the track rather than in the centre increasing the centrifugal forces and putting more pressure on the flanges and inner face of the outside rail. A powered wheel can climb off the track if the flange digs into the rough inner face of the rail. It's known as override and there were two derailments caused around London about two years ago.
It works for you in the garden so thats fine but if you suddenly get locos derailing after the flanges have roughed up the inner face for a while you'll know why

I?m not sure any prototype trains were limited in length by sharp curves. Just the mass was limited by grades and to lesser extent, curves.

Prototype trains are governed by the gauge limits for a line and that means physcal dimensions, (what most people call the loading gauge), weight and the length. An out of gauge load can be too heavy or too long to safely negotiate a line.
In the US they use helper locos for heavy trains but they are distributed differently depending on the terrain. Drawbar loads are affected by loco power, (sudden snatching if you have too much), train length and weight.
On the flatlands you'll often find them all up the front as it's quicker than splitting the train in two or three to add helpers.
In the mountains with stiff grades and sharper curves you'll find the helpers split into the train or pushing at the rear as it eases the drawbar load reducing the chance of broken couplings and stringlining. It all depends on the weight and length of the train.

 

[Gizzy]

On the prototype, the tire profile on the wheel is conical. I saw this on a recent visit to Norwich Crown Point Depot on a faulty wheel set.

The wheel sets are turned on a lathe so that the diameter on the outside of the wheel is less than that on the inside, hence you get a slight slope on the running surface of the wheel, whereas on our models, this surface is flat. I was informed that there is very little wear on the tires in normal use.

You sometimes get flat spots if the train wheel slips, due to loss of traction, and this is more prevalent on high speed stock with disc brakes, such as the 100 mph Mk III Intercity coaches, but not on vehicles with brake blocks like the 75 mph class 153/156 DMUs, as the block tends to rub these out.

The idea of a conical tire is so that the wheel flanges do not contact the rail. (When they do, you get the classic squeal that you hear on tight curves or on points in stations. And obviously the associated wear on the flanges and rails too.)

Because of the tire profile, super-elevation can be used so that the tires still ride in the centre on curves at speed. However, there is a limit to the amount of elevation, dependant on the speed of the trains (both passenger and freight running at different speeds) and also passenger comfort. For trains to go faster on a 'standard' railway, you have to use tilting trains such as the APT or the Pendolino as used on the WCML....

 

[coyote97]

steep grade and heavy train operation with double heading doesent mean just to have long trains.

lets take an example:

The downscaled weight of 1:20,3 models is by factor 1/8365 from the original weight.
For a Connie with a tractive force of about 21.000 pounds that gives (out of the baldwin catalogue) an prototypical weight of about 49 tons (just loco) and with this a model weigth of 5,85 kg (just loco).

With this weight (what my connie has) it hauls relatively exact the cars and loading that would be prototypical, and -for surre- without rubber rings. Noone operating "prototypical" needs rubberings!

Taking an incline of about 4% (what my line has), the connie will haul 155 tons. That is quite right for a loaded 3 (!) cars coal-hopper train with a caboose. And enough for a 7 cars hopper train with a caboose double headed.

i have curves with 1200mm radius and inclines of 4 %. My locos and cars weigh nearly prototypical downscaled (flatcar 850 gramms, Hopper 2,6 kg). Some cars like the AMS Passenger coaches need roller bearings for.
I can do everything on my layout. Doubleheading, pushing, pulling, helper in the middle of the train.

I have all body-mounted couplers (nothing else works with this conditions), even on my 3 little bachmann flatcars that are NOT from the AMS or Spectrum-line.
But these 3 cars need much load, because beeing the first cars in a heavy train, they dont roll, they "fly", uplifted between the couplers.
Railroads need some weight to work.

All that can stop my operations are snow, feet of not downscaled people, little stones, little branches, not-downscaled cats.
But not the rolling stock.


Greetings

Frank

 

[stevedenver]

well not too high brow an observation coming

weight and depp flanges help a bit-
reveres curves are usually a problem

i have run really long n scale trains-50-75 cars on heavy grades and relatively tight curves

and 
over twenty years ago, when i was younger and easily amused....., and into LGB on a budget 

out to the patio from the kitchen on a modest grade but with 1100 curves 

to accompany my Indian theme dinners in the very hot dry Colorado July days

a really heavy lgb martini trains-2 ice gondolas , box cars sans roofs with several beer bottles, 3 low long gons with squat but heavy glassware,  four full tankers-scotch, vodka, gin, and rum, nuts and olives, mixers and lemons-in the orange hatch covered gons..not exactly elegant or effective but unique for most guests 

we're talking quite heavy and high center of gravity powered by two 2015 and my green stainz-no problems so long as steady and slow each engine kept things more or less from getting too much lateral strain

with both n scale and g scale

i have found that using a pusher and mid line engine of exact same speed as the head engine is really an effective way to avoid string lining as well as excessive draft and buff (ie slack in the coupling when up or down or on top of a grade) 

it may not be prototypical given the number of cars -but it is sort of prototypical in that one is reacting to the actual needs of effective train operation

it was fun-but always a bit dicey-more so on the descent

the guests handling of cars was much more a derailment issue however

and it is disappointing to learn that the lgb tankers arent always water tight-my BASF desingated for Gin had a slow leak-alas a puddle of Tanqueray on evening 

 

[TerrySoham]

I've been thinking about scaling weight. I think that you would relate weight to volume. Thus the volume of the model OBB 2095 would be 22.5 x 22.5 x 22.5 (L x W x H) smaller than the real thing. 32 metric tonnes is 32000 kg. Thus to be to scale, the weight of the model will be 32000/11390 = 2.8 kg. This is quite a heavy model!!! This assumes, of course, that the model is actually a scaled down replica of the real thing - which it isn't.

Terry




This derailment touches on one of my main grieves with G-scale!

Sadly, the weight is NOT to scale.
So a real OBB 2095 weighs 32tons - the LGB one should therefore be 1.4tons !!!
Same for the wagons, they would stay alot better on the track if they only were heavier!

If our railway was at least a 'little more' to scale then
- the locos would pull more and better and uphill
- the wagons would stay on track
- alot less worry about light wagons behind loco, heavier further back!

Dreaming....

Some pics cos pics are nice:
- testing the Croc, but wagons in 'good' order...

- more testing, the circle around the pond = always good for derailments of lighter cars

[/quote]

 

[muns]

I've been thinking about scaling weight. I think that you would relate weight to volume. Thus the volume of the model OBB 2095 would be 22.5 x 22.5 x 22.5 (L x W x H) smaller than the real thing. 32 metric tonnes is 32000 kg. Thus to be to scale, the weight of the model will be 32000/11390 = 2.8 kg. This is quite a heavy model!!! This assumes, of course, that the model is actually a scaled down replica of the real thing - which it isn't.

Terry




This derailment touches on one of my main grieves with G-scale!

Sadly, the weight is NOT to scale.
So a real OBB 2095 weighs 32tons - the LGB one should therefore be 1.4tons !!!
Same for the wagons, they would stay alot better on the track if they only were heavier!

If our railway was at least a 'little more' to scale then
- the locos would pull more and better and uphill
- the wagons would stay on track
- alot less worry about light wagons behind loco, heavier further back!

Dreaming....

Some pics cos pics are nice:
- testing the Croc, but wagons in 'good' order...

- more testing, the circle around the pond = always good for derailments of lighter cars

[/quote]

Hmmm, 22.5 x 22.5 x 22.5 - tis a cube then?????

 

[Woderwick]

Prototypical rail is canted inwards to match the cones on the wheels, as was my C& L O Guage track to match the cones on my O scale wheels.

I have nt ever seen that on G stuff.

Also on most real stuff the wieght is a lot lower down, mostly under the sole bar. a lot of the modle stuff has unreal mass up high and coupled with tight curves the results are unsurprising.

 

[coyote97]

Hi muns,

for all of you that havent thought about scaling down weight (or dont know about...we cant know everything..thats what forums ar for...):

22,5 x 22,5 x 22,5 would give a cube, and it stands for what is on meaning behind it, but maybe, the 22,5 cube is not suitable for explaing it.

take a cube of 10 mm x 10 mm x 10 mm. In there is a volume of 10ml and it weights 1 gramm filled with water.

if u now DOUBLE the length of the cube (scale 2: 1), u get 20mm x 20mm x 20mm.
If u imagine that in the origin 10 x 10 x 10 mm cube, the "cube10" will fit in the "cube20" 8 times
doing the same with scale 3:1 (cubelength 30mm) the "cube10" fits in the "cube30" 27 times!

so, scaling a body with factor 2 makes 8 times more volume (and therefore weigth)
scaling a body by factor 3 makes 27 times more volume.

8 = 2 x 2 x 2
27 = 3 x 3 x 3

it works the same with scaling down.

so, scaling down weigth from prototypical data means:

with a scale from 1:Y

the factor (and formula) to calculate on the prototypical weight is:

model weight = 1 / (Y x Y x Y) x prot.weight or model weight = prot.weight / (Y x Y x Y)



So, my 1:20,3 scale Railroad has a downscale divisor of 8365 (20,3 x 20,3 x 20,3). That seems much, but it fits well and is true!

All my roling stock or load can be "overcalulated" with that. And i found out that even on load, it fits well. Give the thing some time and do revival of the maths lessons:

the tankcars, the hoppers....."loading" them will nearly match the prototypical weight. I found, that both AMS and Spectrum rolling stock are quite near to the prototypical unloaded weigth.
Fitted with cadee couplers, that rolling stock allows all prototypical operations while pulled with a weightscaled loco without rubberrings.
i found out that my loco is beginning to slip before my cars are pulled out somewhere.

Greetings

Frank