As RioGrande mentioned, IMO, the important part is minimum VISIBLE radius.
Hidden radius simply has to be operational.
When planning, its sometimes possible to make the minimum visible radius a bit broader if nonvisible radius is cheated towards the minimum needed to operate.
- Douglas
Hidden curves have to pass the most severe tests of reliability if you want to enjoy essentially derail-free, or derail-proof, tracks where it's hard to fix things or even to reach in and drag them back out...hopefully still coupled. You'd want the probability that any one axle will derail in a given consist less than 1/1000, or close to that. As you tighten those hidden curves, and use sloppy joining techniques or smaller codes of rails, your overall probability rises, and not linearly. I have to admit that I am one of a long list of modelers who thought they could beat the odds. First time I did it with tangent track and Code 100 in a long tunnel 5' in length, very tough to reach up tight against the wall and retrieve something which I think I only had to do one time. My helix in the last build had a tricky turnout in it, the worst setup on my layout. I achieved about 1/99 out of it eventually, but that was it. Once a week I had to slide under the helix on my back and fix or retrieve a piece of rolling stock.
Take great pains to ensure reliability on hidden trackage, but double that up if it's on a curve near the minimums, especially for long strings or long cars coupled in numbers.
I have to agree with Selector. I have never been able to get my head around using a smaller radius for hidden track.
For me, close coupled passenger cars with working, touching diaphragms are a must, so my 36" in minimum is just that, hidden or visable.
Also, I have found that once you get into the 36" and above zone, it is not necessary to increase the spacing of parallel track by any noticable amount. In fact on my last several layouts all parallel tracks were spaced at 2" even on curves with clearance issues.
On the new layout, 36" will be the minimum radius on secondary trackage, and 40" minimum on the mainline.
Except for staging, I am also working hard to eliminate hidden trackage, it was one of the problems with the old layout, too much hidden track.
If my retirenent location works out, once moved to its new home, the staging tracks will be accessable from service isles behind the layout. In its current location they are reached from both above and below in different spots.
Sheldon
ATLANTIC CENTRAL I have to agree with Selector. I have never been able to get my head around using a smaller radius for hidden track.
What I was advocating was not using a smaller radius for hidden track. It was using a LARGER radius for showing track.
Yeah, it looks like I'm saying the same thing. Nope.
Taking your stated minimum of 36", I think it would be really nice to have minimum mainline radius of showing curves to be up around 100".
Easy to say, hard to do. I know.
But if you look at the track plan I showed, it CAN be done. If you have the room. It will, of course, affect the amount of track you can fit in. But that's a discussion that is ALWAYS happening with layout design, so it's not a surprise here.
Ed
7j43k ATLANTIC CENTRAL I have to agree with Selector. I have never been able to get my head around using a smaller radius for hidden track. What I was advocating was not using a smaller radius for hidden track. It was using a LARGER radius for showing track. Yeah, it looks like I'm saying the same thing. Nope. Taking your stated minimum of 36", I think it would be really nice to have minimum mainline radius of showing curves to be up around 100". Easy to say, hard to do. I know. But if you look at the track plan I showed, it CAN be done. If you have the room. It will, of course, affect the amount of track you can fit in. But that's a discussion that is ALWAYS happening with layout design, so it's not a surprise here. Ed
Ed, I was not refering specificly to your post but rather was considering the wisdom of using the bare minimum functional radius for hidden track, and then having some larger cosmetic minimum for visable track.
But even in your example I do not consider 30" radius an acceptable minimum for the equipment I run.
The layout I am currently dismantling and replacing has a 36" minimum radius. Being a double track mainline, that puts a great many curves at 38" and 40" right off the bat.
That layout included a number of "cosmetic" curves in the 50" and above range.
Not the best photo, but this curve is 54"/56" respectively:
Not your sugested 90" radius, but pretty large none the less.
The new layout will have a mainline minimum of 40", again with a double track main, making many curves 42", 44" or larger. It does have a secondary minimum of 36" radius, yards, etc. There is not much hidden track, so sharper curves there would have little space saving effect.
Most of the hidden track is long straight staging yards.
I do see where once you are above 36" as a minimum, there would be little reason to have hidden curves above that size.
I simply meant that it only needs to be reliable in any nonvisible areas. No real reason to have a 40 inch radius curve in a tunnel.
Something else no one has mentioned - easements.
Easements are essential on the prototype and should be considered so with our models, large radius or small.....
In fact, on the prototype, your typical curve that changes the vector of the route 45 degrees or less typically has no fixed radius at all, but is rather two long easements back to back.
There are several easy formulas for figuring this out and I have build a number of curves this way.
And, when doing this, the theoretical point of minimum radius can actually be somewhat smaller than your minimum constant curve radius, because that radius on exists for about an inch of track.........
Armstrong suggested very long flowing easements, but easily gave in to sharper curves.
Paul Mallery suggested that with large curves, 36" and above, easements could be much shorter and be just as effective.
Armstrong shows, in Track Planning for Realistic Operation, that 18" radius with easements causes less coupler offset (his "coefficient of lurch") than 22" radius without easements, and still takes up less space. Once the radius gets so large that it might be something a prototype might be able to negotiate, the need for easements goes down.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
You are correct, Randy and Sheldon, as far as I can tell. After so many inches, the greater value for an eased curve on a layout would have to be visual fidelity and appeal in order to more closely approximate the prototype. I don't think any curves greater than about 32-34" should need real easements unless we're talking highly prototypical brass or scale rolling stock. It's just that if they have the appearance of snap or set track curves, with sharp and immediate onset into curvature, it just doesn't look right, not for a main line with higher speeds on which passenger trains will run.
On the other hand, if space is at a premium, you can certainly get away with those immediate curve onsets with radii above about 24-30", depending on what you're running behind the power. For example, if you want that tunnel with curved right of way in it, and a sharp curve to save room out in the visual world, one needn't ease the curve as long as the rolling stock can handle it. But as I said earlier, the closer you creep up on those real engineered limits, the higher the probability of 'failure'. Seems to me it's a big enough bummer dealing with the failures out in the open without having to stoop, crawl, bend, contort, and reach behind scenery or into a tunnel to retrieve something intended for broader curves and finer joints.
I disagree that 18" in HO is tight even for streetcars. I belive that the tightest curve on the Boston system is 35 actual feet. I just measured, on Google Maps, the curve at Ninth Avenue and Irving Street in San Francisco, and it came out to just under 9" in HO scale. Very sharp curves can be found on "steam" railroads on industrial trackage, especially in streets. Only 40 foot freight cars and short wheelbase locomotives can use the tightest curves, and then only at speeds well under 5 MPH. A modern example of sharp cuves is the area at the east end of the Steel Bridge in Portland, OR. Passenger trains turn south from the bridge at walking speeds on a curve that scales to about 40 inches in HO scale. The grain elevator trackage just north of the bridge includes a curve of about 20" in HO.
A train traversing a 10 degree (79" HO) curve will run at slow speed, probably 10 MPH, maybe as much as 25 MPH.
Using larger radius transition at each end of a curve will help operation. For realism, avoid placing sharp curves where the viewer sees the outside of the curve, unless you are modeling street trackage.
DAN TAMSKYUsing larger radius transition at each end of a curve will help operation.
A spiral easement transition will help even more.
Layout Design GalleryLayout Design Special Interest Group
Curves are not difficult to calculate, but two elements must be known. For example the degree of curvature and the radius can be used to calculate all other elements (length of curve, tangent length, chord length, etc.). The easiest for the typical modelrailroader to understand, is radius because it is so commonly used to describe model curves.The radius of a one degree curve is 5729.578 feet. So this number divided by 2 will give the radius of a two degree curve, by 10 will give the radius of a 10 degree curve, and so on. The absolute tightest curve a Union Pacific 4-8-8-4 could negotiate was 20 degrees. Any curve tighter than that would result in either a driver climbing the rail, or in the force of the driver against the gauge surface of the rail, pushing the rails out of gauge. Either way a derailment would result. So if we divide 5729.578 feet by 20, we get a radius of 286.479 feet. Converted to HO scale this is 3.290 feet or 39.475 inches. This is considered a fairly broad curve for HO scale models, but would represent the sharpest curve your Big Boy could negotiate if you intend to be prototypically accurate. Mainline curves are much broader, and subject to speed restrictions even at less than 5 degrees which would be a 157.9 inch radius in HO scale.
I should also note the 5729.578 foot radius figure I use is for a 1 degree "chord definition" roadway curve not, not an "arc definition" railroad curve. The difference is slight, but worthy of mention.
Great thread!!! One thing I see missing in the discussion is the radius in relation to the grade. We all know how that smaller radii changes the reliability with longer trains...
SD Train Fan Great thread!!! One thing I see missing in the discussion is the radius in relation to the grade. We all know how that smaller radii changes the reliability with longer trains...
True.
If you do a search for "compensated grade railroad", you'll find a lot of discussion on the subject.
riogrande5761"Big" is a relative term. John Armstrongs Track Plannining for Realistic Operation book lists 18" as sharp, 24" as conventional and 30" as broad. For some, 30" would be considered a big radius, if they are very cramped for space. But times have changed and standards with it. If I were to revise John Armstrongs terminology for today, I'd list 18" as very sharp, 24" as sharp, 30" as conventional and 42"+ as broad.
I believe Armstrong was not creating the terminology, simply reporting what the terms used regarding curves were. I suspect you can find sources before that have used the same terms; not sure who first stated them.
I wonder if Walthers saying their HO passenger cars work on 24" radius curves is related to 24"R being a "conventional curve"? Kato HO Unitrack starts at 24"R then has three sizes bigger and three smaller curves, all 2-3/8" apart.
IIRC John Allen's first layout had some 14"R curves, and the final G&D had curves as tight as 26"R.
The B&O had at least one mainline 14 degree curve on their Pittsburgh & Western subdivision west of Pittsburgh, Pa. Trains of all kinds including Amtrak and intermodal with 89' cars used this route. The speed was 30 mph.
The flanges definitely contacted the rail. Metal filings could be seen on the ties throughout the curve and the rail head was worn to match the wheel/flange shape.
I don't know if they were used in regular service, but I have seen a picture of a B&O EM-1 2-8-8-4 running a fan trip on this route.
Mark Vinski
I doubt that there are many modelers of true high-speed rail; the numbers just make it impossible. High-speed lines in France and China with train speeds of around 200 mph have minimum radii of 5,000 - 7,000 meters (3 - 4.5 miles). In HO, this would translate to around 200 feet (or 2,400"). Interestingly, high-speed trains are much better equipped to deal with vertical grades, because of there very high power-to-mass ratio. The German high-speed line from Frankfurt to Cologne has 4% inclines, and there is no noticable drop-off in speed.
Traction and interurban modelers definitely have it easier in the "realistic sharp curves" departmemt, as some level of compensation for our relative lack of variety in ready-to-roll equipment! My 15" curves are very reliable (with easements and very careful work at the tracklaying stage) but sometimes spotting cars on my tightest curve, 12" radius on an industrial spur, gets a little hairy. Recently I bought an estate collection of a fellow who was into heavy steam, and I tested some of the engines on my layout; while the 2-10-4 and articulated loco didn't clear my curves, the other steam engines did fine, including a heavy Mogul, even at relatively high speeds.
However, as scale modelers, in addition to the reduction in proportional size for the scale we operate, we often use "selective compression" to represent a larger area than the confines of garage, basement, spare room or bookshelf. Industrial buildings are shrunk from their original proportions, towns moved closer together, mainline distances and yard capacities are frequently reduced beyond their scale size, but they still look good to the viewer, because the emphasis is on the trains instead of the landscape. We use theatrical tricks to bring the eye where we want it, and away from places that break the illusion. We even use "fast clocks" to increase the illusion of greater distances traveled. In the same way, we use sharper-than-realistic curves for practical reasons, but there are many ways to disguise them and minimize their aesthetic impact. One of the most popular ways to make a sharp curve look broad is to view it from the inside; there's a popular railroad planning idea that trains look better from inside a curve rather than outside, and monumental efforts are made to avoid "blobs" in planning both because they're seldom found on the prototype and they draw attention to the trains' bad side. The other method, used for blob-hiding, is to conceal the trains in a tunnel, but others have already mentioned that hidden track, due to its inaccessibility, has an even greater need for reliability than exposed track.
So, often, the question isn't so much "Are my layout's curves realistic?" as much as "Do my layout's curves look realistic?" These can be two substantially different questions, depending on the perspective of the viewer--so be mindful of where your trains will be seen, and from what direction, as much as prototypical practice and dimensions.