Guitarmike posted:I like hand laying track and the lack of a gauge is a frustration. I am going to see how Gargraves track measures up with the nmra gauge. If it is in spec you could use that and a wheel set that is representative of your fleet and sort of be ok.
Oh, you're going to love this, then. I was having gauge problems with some of my equipment, so I did some measuring a while back. To my surprise, I found variation in the track. It amounted to all of 25 thousands of an inch between various brands--and to my far greater surprise, that actually made a big difference in a few cases.
Gargraves is actually 1.270 between the rails, and I have had wobbly postwar cars where the trucks would fall down between the rails on Gargraves. Ross track varies a little, in my experience, but it seems like the number they shoot for is 1.260, which works well. I had some Ross that was actually gauged at 1.250 (you know... O gauge?) or even 1.245, and certain locos had flange binding on the curves. I worked it over with pliers until I got it close to 1.260, and the problems cleared up. Fastrack and Atlas are both right on 1.250, but the Fastrack is effectively wider gauge because of the rounded rail profile. I have an MPC scale Hudson which is rated for O-31 tubular but will not run on O-36 Atlas. Yet it handles O-36 Fastrack just fine, even though both are exactly the same gauge. The difference is all in the rail profile.
If you are laying your own track, I think the number you want is probably 1.260, as long as you are not using rail with a razor-sharp corner profile.
The main takeaway (philosophically speaking) from my experience is this. The lack of actual standards in 3-rail is the product of a difference in philosophy between the scale modeler and the maker of toy trains (which is what 3-rail is derived from, of course). The scale modeler wants to reproduce the real thing accurately, so he adopts standards, and tempers his expectations for his own layout and what he will be able to run on it, realizing that there are some things he just can't do, if he is going to maintain an acceptable level of realism. But the toy train maker asks, "how much locomotive can force around this curve?" He asks that because his market is more interested in having a big locomotive than in having a realistic locomotive--"they're toys!" say they. So the toy train maker doesn't follow standards; the toy train maker cheats. He has been cheating for a hundred years now, and if there are still toy trains a hundred years hence, he will be cheating still. And as long as people expect the 3-rail scale trains to run on the same tracks as the 3-rail toy trains, there never will be any standards.
For this reason, I think any 3-railer (even a toy guy, like me) who wants to run equipment from different manufacturers, or even different eras of the same manufacturer, will just have to be reconciled to the idea of tinkering with the trains until they run on his layout.
None of that answers OP's question of course, except to say that there probably isn't an answer. If it works it's right, if it doesn't it isn't. FWIW, I found that on modern Lionel wheelsets the distance between the inner surface of flanges (i.e., back-to-back) varies, anywhere from 1.030 to 1.050, often on the same car. This never caused me any trouble, except once or twice where there was a wide variance between the two wheelsets on a single truck.
