A few thoughts on gauges for track assembly

[Orange Cat]

For those fretting about the precision of their track building, this is the way the prototype does it:

 

 

 

Martin.

 

Can you talk us through producing a template for that in Templot and show us how to use the adjustable vertical z-axis facility?

[Horsetan]

As a non track-builder, can I just say that this thread is one of the most entertaining punch-ups I've yet read on this Forum, and I can quite understand why, after reading it, many people might choose to stay with Setrack....

[Junctionmad]

As a non track-builder, can I just say that this thread is one of the most entertaining punch-ups I've yet read on this Forum, and I can quite understand why, after reading it, many people might choose to stay with Setrack....

from a guy thats models in P4!

Can you talk us through producing a template for that in Templot and show us how to use the adjustable vertical z-axis facility?

 

yep, that track looks good for 10(0) mph , I can see the flying scotsman on it 

[bertiedog]

May I recommend the Thor Brand soft faced hammer (junior) as the tool of choice for the discerning track and point builder to use to duplicate in 4mm the heroic efforts of the Chinese to build a really bad track.

 

Apply with sufficient hammer force to the rail tops alternatively or at the same time to fettle the track into submission. You will need to check the gauge is being maintained at 16.5 or if you like even narrower gauge 16.2. Believed to work in P4 as well, but rarely seen to have been used.

please note safety glasses should be worn whilst doing the work

 

stephen

[t-b-g]

Wooden gauges: Buckingham was built with them. Sure the gauge varies by about 2mm, from 17.5 to 19.5mm but it works. I have no idea how because it shouldn't. In his writings, Peter Denny mentions that his EM was 18 and a bit millimetres because he found 18mm a bit too tight on curves with Romford axles. There are plenty of bumps and we get a few derailments mostly due to buffer locking shunting round tight curves with very wide gauge, allowing the wagons to move well out of alignment but I am gradually fettling these. In the main, it runs very nicely.

 

As for gauges, I use two flat sheets of metal, which were milled to 18.2mm but could have been cut and filed if done with care. These are about 25mm long. They give automatic gauge widening on curves. I use these to set the running rails either side of a crossing nose. Then I use a couple of bars of 2mm x 1mm metal to set check rails and wing rails and a bar of 1.5mm width to set the gap between the blade and the stock rail.

 

These are for EM where the check rail gap is 1mm and the gauge 18.2mm. Other sizes would be needed for other gauges. I have several commercially made gauges, roller ones and a triangular one but as they all vary slightly I prefer to leave them alone.

 

I have just finished building over 300 points (including some very complex formations) for a new layout and every one was done using just these gauges and every one went just right first time.

 

I have never been able to get my head around the idea that you need any more than the running rails the right distance apart, check rails the right distance from the running rails, wheels that match the standards of the track and a good alignment with no bad kinks or joints and most especially, a good pointed end to a blade that forms a smooth transition from the running rail and a good alignment from wing rails onto the crossing nose. It doesn't have to be any more complicated than that.

[bertiedog]

After asking for simplicity, lets just add that the more complex methods of accurate point construction may not even allow for the actual rolling movement of a wheelset through the point.

 

A wheelset on a curve follows a reasonably radial path from the centre of the circle, and should behave as if on straight track (dismiss the fact that one wheel is constantly slipping on curves), but such a wheelset is only found on single axle trailing or leading wheel.

 

Most wheels are constrained by the frame to travel along a curve at an angle to the true radial line, and two effects occur, the flange becomes effectively thicker and the tyre effectively narrower on curves. Also the root curve between the tyre and flange becomes less effective.(it is "shaded" by the flange).

 

On small crossing point angles, (large radius) the effect of this is too small to bother about, but it does affect the design of a tighter point, as the wheel is not running in the way a plain cross section would show.

 

The effective back to back tightens up, the flange becomes wider, the tyre smaller, and this combination causes the wheel to not be restrained by the check rail and the wheel drops because the tyre edge cannot reach both the end of the switch rail and the tip of the crossing,

 

This whole reasoning was brought up by Mr Pritchard in explaining the flangeway and check rail clearances on streamline 100 points, the effect is worst with wheels like the Dublo profile, and he had to build the point to take these.as well as more scale wheels.

 

Also the other forgotten problem with wheels in the argument above is that the effect varies with wheel diameter, ever not noticed that diesels sail through complex point work, whereas a steam loco will rough ride, despite having perfectly good wheels set to correct back to back??

 

It also helps explain why 16.2 works..........helps keep the "effective" flange width in the correct path through the point.

 

These effects are slight, but measurable, and no, I have not sat down and worked the geometry out to the nth fraction of a degree or part of a thou... but it is there and usually never even considered in making points.....except that the NMRA standards did include such a study and may help explain the better running than 00 on the same track.

 

The whole issue vanishes on P4, the points are too large to show the problem, and all the wheels match each other bar diameter.

Neither does it bother ordinary 00 much, the tolerances of the commercial stuff cover it up, but it does affect better made scale track built to tighter standards, and may explain some jittery running on well made points with good wheels, you get an accumulation of tiny errors and end up with wheel drop.

 

Perhaps an answer is that each particular radius or length should have the flangeway and check rail set for that radius, one size may not suit all points.

[bertiedog]

Only a very rough sketch, but it shows the effective change in cross section, and tyre width between a wheel running on straight track and a curve.

 

 

[nswgr1855]

Terry,

 

I think we have it in a nutshell. You are, by training an metrologist, ie in the business of measurement precision. I am, by training, an engineer, and as such interested in what will make things work. A consequence of that is that I am interested in the limits within which our wheels and track will work with each other, rather than the precise tolerancing of a set of dimensions that are within those limits. Part of that is knowing how far you can depart from the ideal before, in our case, the perfomance of the rolling stock starts to suffer noticeably.

 

Jim

Jim, you have it the wrong way around. I have trained and worked as a fitter machinist, mechanical engineer, then metrologist.

 

Cheers,

Terry Flynn

[Pannier Tank]

The first measurement tool you need is a good stainless steel rule, 6 inch is handy, and it should be marked in inches and metric. Get a magnifying glass to inspect the scales accurately. The main use of the ruler is a general checker for overall dimensions, not to actually gauge the track that you are laying.

 

My preference is to use a Rustless Chrome Faced Rule, much easier to read than a shiny Stainless Steel Rule.

[nswgr1855]

As a track builder , I find myself in complete agreement with Martin. The two relevant gauges are flangeway and check gauge , obviously one must keep the track gauge within limits , but the fact remains variations in it have little effect

 

In the crossing V changing the track gauge impacts on either the flange way distance or range of wheel back to back and flange width, a mathematical fact.

 

Whether wooden gauges etc are acceptable , remains within the realms of a curiosity as I've never met anyone using them. Making a metal gauge ( or buying one ) is not complex.

 

Given my track intentionally varies from 16.5 nominal to 16.2 nominal I suggest the gauge is not " that " important

 

 

True only for  track beyond the crossung V. My H0 plain track varies between 16.2mm to 16.8mm.

Note , I'd suggest P4 is easier to build because of the smaller number of " standards " and the fact that things like crossovers are easier to build.

 

 

P4 is harder to get right as the tolerances for the crossing V are smaller than coarser standards.

Ps : could I suggest that to avoid pissing contests , that telling me that professional railway engineers have any relevance to model railways , is rather like a brickie advising on building a model station. !!! ( a toolmaker , metrologist and a railway engineer walked into a bar .......)

 

That's the great thing about standards , there's so many to choose from

 

 

 

The problem is many are not compatable with RTR 00 and some  standards have mistakes in them

The other issue being missed here , is that most builders use gauges as jigs rather then gauges.

 

 

Both terms are interchangeable

Furthermore AMRA, DOGA. Etc , Don't produce standards , they produce specifications , standards are standards because they are widely accepted and adopted. Specifications are something you'd like people to follow. Peco is a standard and is not a specification, RP-25 100 is both as its widely adopted to be judged a standard , DOGA is not a standard.

 

 

I disagree with your definition of a standard and specification. PECO do not publish a standard, they make track that is compatable with various standards and wheels.

I do however agree with bertiedog, gauges only need to be precise/accurate " enough " , even in p4. Many builders do not achive " thou " accuracies all through a typical turnout even in p4. Some time ago , when I did a fair bit of copper clad , I built a few p4 points , I didn't find it much different to building 00 ones. The process was the same, the rail was code 75 and the gauges were the same type albeit wider. It did seem less tolerant of poor installation more then build accuracy.

 

The standard you are working to determines the accuracy and manufacturing tolerances your gauges need to be in order to build accurate turnouts.

 

Cheers,

 

Terry Flynn.

[nswgr1855]

Mention is made that the gauge standard of accuracy is not that important and that is true on track, but it is vital to the spacing of the flangeways in points, where they must be in relation to the rails, and you have to assume the gauge is correct, or your heading for troubles.

In model railways terms a Gauge and a Jig are interchangeable, lets not make it complex.

Also the use of wood was suggested as I have used steel pins in wood many times, but usually make a jig in brass in about 5 minutes, checking with a vernier caliper.

On speed of working to get a sound point, I have just made four for 009 gauge and they took two evenings to complete on PC sleepers, The running through them is smooth and drop free, and the action light as there is a phosphor bronze strip hinge in the blade, looks solid at set road, only bending the hinge on the dead switch blade. The testing showed up problems with two Peco items for the L&B stock, they needed the back to back re-set

.

Last night I did a custom curved point for the Brewery layout, at 9 inch inner radius, 16 outer, in code 75 flat bottom on PC  sleeper in about 2 hours to test and fitting. Because of the tight curves the Hornby Peckett and 4 wheels stock can take, the crossing is done with fully curved parts. I used no gauges at all, just the vernier, and again the operation is smooth, the loco barely notices the point, with stock wheels just making a faint click as they go over the crossing V. It varies with different makers wheels as wide tyres run better over a frog.

The PC construction was allowed as the points are to be buried in cobbles, so no chairs needed.

 

Also I did a couple of more conventional large radius points in C&L and Exacto chairs, with plastic sleepers, these took about 3 hours each, again no jigs, except to fit into the flange way to glue in the half chairs on the checkrails. No issues except the cost of the sleepers, which in future I am replacing with Plasticard in black, epoxy glued to a sheet of thin ply and then sawn up in my bench top bandsaw. It provides a better choice of lengths. The grain is done with a wire brush.

 

Running is perfect, nothing unusual, as long as the wheels are compliant to something near the RP-25 spec. Very narrow tyres drop a fraction, all modern RTR sails through the points. These longer ones have no hinge fitted, solid except for insulation, with switch blades electrically bonded to the stock rail, to be able to reduce the clearances as much as possible when the rails move on switching.

 

The chaired version is not to P4 standards as to the correct chairs exactly in each position, greatly simplified to just chairs and sliding chairs, plus half chairs cut from whole ones. The check rails are supported by p/b strips from the running rails, covered by the half chairs.

 

Stephen

Hello Stephen,

 

If you had set your digital calipers to 16.26mm for the track gauge and 1.01mm for the flangeways your narrow wheels would not have a noticeable wheel drop for the 00 wheels.

 

Cheers,

Terry Flynn.

[Colin parks]

Hi, Bertiedog,

 

I am getting confused by some of the statements you are making, especially in post #81. Why would one wheel always be slipping on a curve?  I thought the coning of the tyre enabled one wheel to be running faster (i.e. the outer one) than the other.  Where you say 'switch rail', surely you mean closure rail/wing rail?

 

Could we see some examples of track you have constructed using these wooden gauges?

 

Colin 

[nswgr1855]

It is indeed swings and roundabouts, it works both ways, but I think for points that involved a curved route through them, the most commonly used in models to save space, the datum is the track course, and the crossing is fitted exactly where it has to go, +/- nothing, aided by precise gauging and nothing else.

 

No No No again. No one can make track exactly to +/- nothing. Precise gauges are made within a tollerance which is bigger than nothing.

 

 

 

Cheers,

 

Terry Flynn.

[jim.snowdon]

I do detect an obsession with precision that is, in practice, unwarranted. There is , obviously, Terry's way, and the somewhat more practical methods that are practiced by the rest of us, and which are getting good results.

 

Jim

[nswgr1855]

After asking for simplicity, lets just add that the more complex methods of accurate point construction may not even allow for the actual rolling movement of a wheelset through the point.

 

Most standards have allowed sufficent clearances for this not to be a problem.

 

A wheelset on a curve follows a reasonably radial path from the centre of the circle, and should behave as if on straight track (dismiss the fact that one wheel is constantly slipping on curves), but such a wheelset is only found on single axle trailing or leading wheel.

 

Most wheels are constrained by the frame to travel along a curve at an angle to the true radial line, and two effects occur, the flange becomes effectively thicker and the tyre effectively narrower on curves. Also the root curve between the tyre and flange becomes less effective.(it is "shaded" by the flange).

 

On small crossing point angles, (large radius) the effect of this is too small to bother about, but it does affect the design of a tighter point, as the wheel is not running in the way a plain cross section would show.

 

The effective back to back tightens up, the flange becomes wider, the tyre smaller, and this combination causes the wheel to not be restrained by the check rail and the wheel drops because the tyre edge cannot reach both the end of the switch rail and the tip of the crossing,

 

You only get noticeable wheel drop if the wheel width is less than 2 times the flange way width for typical model railway crossings.

 

This whole reasoning was brought up by Mr Pritchard in explaining the flangeway and check rail clearances on streamline 100 points, the effect is worst with wheels like the Dublo profile, and he had to build the point to take these.as well as more scale wheels.

 

Also the other forgotten problem with wheels in the argument above is that the effect varies with wheel diameter, ever not noticed that diesels sail through complex point work, whereas a steam loco will rough ride, despite having perfectly good wheels set to correct back to back??

 

More to do with the geometry differences between steam and diesel models and flangeway / wheel width ratio.

 

It also helps explain why 16.2 works..........helps keep the "effective" flange width in the correct path through the point.

 

And the fact 00 wheels are more than 2 times the flangeway width.

 

These effects are slight, but measurable, and no, I have not sat down and worked the geometry out to the nth fraction of a degree or part of a thou... but it is there and usually never even considered in making points.....except that the NMRA standards did include such a study and may help explain the better running than 00 on the same track.

 

The whole issue vanishes on P4, the points are too large to show the problem, and all the wheels match each other bar diameter.

Neither does it bother ordinary 00 much, the tolerances of the commercial stuff cover it up, but it does affect better made scale track built to tighter standards, and may explain some jittery running on well made points with good wheels, you get an accumulation of tiny errors and end up with wheel drop.

 

That's why you do the caculations when designing a wheel and track standard, and make the flangeway equal or less than 2 times the minimum wheel width. Also the accumulation of tiny errors are designed out if you do the sums correctly and use practical tollerances.

Perhaps an answer is that each particular radius or length should have the flangeway and check rail set for that radius, one size may not suit all points.

 

The simple answer is to use is the AMRA minimum radius standard http://amra.asn.au/2010/05/22/standards/. Then finescale standards work for all cases.

 

Cheers,

Terry Flynn.

[nswgr1855]

I do detect an obsession with precision that is, in practice, unwarranted. There is , obviously, Terry's way, and the somewhat more practical methods that are practiced by the rest of us, and which are getting good results.

 

Jim

Clearly you have ignored the details of my posts. My way to build my H0 turnouts is the long proven conventional way of using roller gauges and track soldered to PCB sleepers. I use RTR flex track whick has a broader tollerance on my general trackwork than most posting hear use as US H0 RTR flex track  gauge can be as wide as 16.8mm. The precission in track building gauges is warranted if you want smooth reliable running. These days I am able to push or pull trains made up of rakes of more than 30 RTR 4 wheel wagons at scale running speeds through complex trackwork.

 

Cheers,

Terry Flynn.

[gordon s]

It also helps explain why 16.2 works..........

 

Oh gawd, now you've done it....

 

Gordon S, sitting happily building 00-SF pointwork using a check rail gauge, 1mm shim and a 16.2mm track gauge.  Also enjoying every word and standing back in admiration of the combined knowledge of those contributing to the topic.

 

.....and not a bump or click to be heard or felt....

 

[jim.snowdon]

Clearly you have ignored the details of my posts. My way to build my H0 turnouts is the long proven conventional way of using roller gauges and track soldered to PCB sleepers. I use RTR flex track whick has a broader tollerance on my general trackwork than most posting hear use as US H0 RTR flex track  gauge can be as wide as 16.8mm. The precission in track building gauges is warranted if you want smooth reliable running. These days I am able to push or pull trains made up of rakes of more than 30 RTR 4 wheel wagons at scale running speeds through complex trackwork.

 

Cheers,

Terry Flynn.

No, Terry, I haven't. I have long been building track in both EM and 0 gauge, the former using rivet & ply construction, gauged with a steel rule, a wooden slip gauge for the crossings (as you can't solder wood) and a check gauge for the check rails. For the 0 gauge, it has been all plastic chairs on plastic sleepers, set to 31.5mm gauge with etched fold-up gauges for the track and check gauges. And like yours, I was able to push and pull trains through my trackwork without lumps, bumps or derailments. It does help to understand which dimensions matter and which don't.

 

Jim

[martin_wynne]

I am getting confused by some of the statements you are making, especially in post #81. Why would one wheel always be slipping on a curve?  I thought the coning of the tyre enabled one wheel to be running faster (i.e. the outer one) than the other.

 

Hi Colin,

 

Even on the prototype, coning/steering works only on very gentle high-speed curves. On sharp curves and most model curves, one wheel must slip to follow the difference in lengths between the inner and outer rail. Likewise on such curves the steering is provided by the flange running against the outer rail.

 

regards,

 

Martin.

[t-b-g]

My 1mm check rail setting gauges (I have 2 for making up crossing noses) are metal. They are chemically blackened, so I can solder all around them with no problem, as long as I use my favourite paste flux and not phosphoric acid, which eats the blackening off.

 

The cone shape of the tread on wheels does help with the slip on curves as in effect, the diameter of the outer wheel can increase slightly as the wheel moves to the outside rail. This increases the circumference so the outer wheel is slightly bigger than the inner wheel. I have yet to see or hear of any evidence that the slipping makes any practical difference in the scales and sizes that we are working in and in my view it is a red herring and a complication that we can ignore. We can certainly do nothing about it other than run wheels that are not fixed on the axles.

[bertiedog]

No point(Sic) in showing the points or the track, as a wooden gauge and a brass one give EXACTLY the same results, the gauge is correct or it is not correct. The usefulness of a pin gauge is that the nails are easy to file to the correct width for the flangeway.

Any wooden gauge would be checked before use!

 

On wheels going around a curve, if the tyres are flat, then the inner one journeys a shorter distance than the outer, one must slip. In real life on the railway the curves are gentle and elastic effects in the rail compensate and slip does not occur that much, but can set off screeching noise from the tyres,

 

In  theory coned wheels and inclined track top overcomes the problem, if the track is correct and the wheel unworn and to perfect standard profile..

 

With models though you are in the hands of the gods as to what taper is there or whether the top of the rail is inclined (as in DCC).

 

The self steering and the diameter difference on the wheel are swamped by other factors, like the slack in bearings allowing the axle to run out of true etc., etc.,

 

All these little points add up to explaining why some makers just get "bad" points, constant jittery running, stalls, and unneeded roll etc.,

 

Mind you, if you want such track and the look of an old industrial track..............

[bertiedog]

Out of curiosity a few days ago I got a couple of nails and knocked them into a piece of hardwood, I think maple, and set them by filing to 1 inch exactly, by comparison to a BS slip gauge, and measured with a certified 2 inch micrometer.

The gauge has got wet, been dried and put outside to freeze and warmed up again.....it is about .0005 bigger..........worry about that ?!? a casual check with a vernier would not show anything.

 

You do not need such accuracy, you need consistency and method to get the best from points and it is so easy when you know how and why problems arise, Oh......... and actually admit to yourself you have problems, which is another matter, as so many times I have stood looking at fine layouts, then see a loco and stock shake and jitter its way through pointwork.

Then the next type of train does not have problems, probably a diesel, I have seen many owners checking track with gauges during exhibition days, seeming to clear it from being wrong, but if you go to a position where you look down the track more precisely, you can see the kinks in the crossings and also tell if the track is flat by the reflections from the rail surface.

 

On mentioning 16.2 and it working , it does, no dispute from me, my only point is that we already have a major problem with gauge, just do not add to the misery by .3mm extra.

 

Stephen.

[t-b-g]

No point(Sic) in showing the points or the track, as a wooden gauge and a brass one give EXACTLY the same results, the gauge is correct or it is not correct. The usefulness of a pin gauge is that the nails are easy to file to the correct width for the flangeway.

Any wooden gauge would be checked before use!

 

On wheels going around a curve, if the tyres are flat, then the inner one journeys a shorter distance than the outer, one must slip. In real life on the railway the curves are gentle and elastic effects in the rail compensate and slip does not occur that much, but can set off screeching noise from the tyres,

 

In  theory coned wheels and inclined track top overcomes the problem, if the track is correct and the wheel unworn and to perfect standard profile..

 

With models though you are in the hands of the gods as to what taper is there or whether the top of the rail is inclined (as in DCC).

 

The self steering and the diameter difference on the wheel are swamped by other factors, like the slack in bearings allowing the axle to run out of true etc., etc.,

 

All these little points add up to explaining why some makers just get "bad" points, constant jittery running, stalls, and unneeded roll etc.,

 

Mind you, if you want such track and the look of an old industrial track..............

 

I wasn't suggesting that wooden gauges don't work, just that if they are being used to avoid problems with solder, that there is an alternative.

 

In my experience 99% of problems on model railways are caused by poor alignment at baseboard joints, rail joints, or through crossing noses, or blades that don't form a good running surface with the stock rail. Gauge matters a bit but isn't critical within sensible limits.

 

Checkrails only cause a problem for me when they are too close to the running rail and wheels set at a good back to back dimension catch on them.

 

I test all my pointwork by running a long wheelbase rigid 4 wheeler wagon through before fitting checkrails. If your crossing nose and blades are good, it goes through with no click or bump.

 

Having said that, the smallest point I am using is a B6, so that may not work on shorter points with tighter curves.

 

Buckingham has all sorts of wheels, from Triang 1950s plastic glued to metal axles through to modern Gibsons. In theory, building track that they will all work well on should be very difficult. In practise, if you make the check rail gaps big enough for the biggest flanges to go through and your alignments are good, it all works.

[bertiedog]

Off the direct subject, but every wondered why a particular loco pulls more than a similar one, it barely happens with mass produced 00 locos, but often plagues live steam where the wheels are turned at home.

 

One Tich may pull four people another struggles with one, one 060 pulls like blazes, another slips at any touch of the throttle.

It is because the wheels have been machined different diameters, and often only by a few thou difference, but as the loco rolls along each wheels travels a different distance, and a surprisingly low figure of difference can set up problems, as little as 2 to 3 thou can induce slipping. The largest wheel usually grips, and the others do little to aid the running of the loco.

 

The effect would not really show in 00, but it does demonstrate that if you do not know about it a loco can give problems. The same applies to track, you may use a gauge, but need to know why the dimensions are so important.

[t-b-g]

One thing I forgot to mention is that the curve through a point should be even and as long as possible. By that, I mean that I have seen some points that are nominally built as, say, 4ft radius where there are two long straights and a very tight, short curve between them. This can lead to problems with buffer locking.

 

A good template can help here and having things like Templot available can be a huge help. I used to judge the curve by eye but one time I checked one with a "Tracksetta" and it wasn't quite as good a curve as I thought it was. Nowadays I place the nearest available Tracksetta gauge along the curve and even if it is not exactly the same radius, it shows up and kinks or changes to the curve that shouldn't be there. It is possible to make your own templates to different radii but I haven't found it necessary. You could draw a curve of the correct radius in Templot, print it out and stick it to a bit of plasticard or thick card and cut it out.