Here we go again! I have a Walthers/Shinohara DCC friendly #6.5 HO curved turnout. Most of my locos handle it well (mogul,American, and all the diesels. But the BLI Mikado and the Bachmann Pacific both derail to the OUTSIDE when approaching the points FROM the diverging end. The approach track is 22"r Atlas. It’s hard to tell which driver is actually picking but it does look like the frog. I have checked the “check gage” with the NMRA standard as well as the flangeways and they are bang on. Both locos check out perfect with the gage as well. This turnout is well away from where I can really observe it which makes theory come into play. The Fast Tracks videos sure were helpful but the locos still derail 
Bruce,
My guess is the problem stems from the different radii in the same curve. The outside leg of the 6 1/2 turnout is a 24" radius, while the inside leg is actually about a 16.5" radius. Attaching a piece of 22" radius sectional track to either leg is going to produce a kink in the middle of the curve and is most likely the source of the derailment.
Don Z.
Bruce, I don’t know the true dimensions of the inner and outer for your turnout, but I was dismayed to find that the curved W/S #7.5’s that I ordered, six of them, have routes considerably less than advertised…so Don is almost certainly bang on.
My suggestion, as unpleasant to contemplate as it may be, would be to get a 7.5. I found the inner to be near 22", and had to break apart all the ties and webbing to get the curves, inner and outer, to reach out to the 24" minimum I set for myself.
-Crandell
If Walthers information is accurate, their #6.5 has an inside radius of 20" and an outside radius is 24." While 24" is adequate for most Mikados, 20" is probably too tight. Perhaps you should substitute a #7 with 24" and 28" radii…or limit loco size to Consolidations.
Mark
Mark,
That’s part of the problem: the information in the catalog is incorrect regarding the actual radius sizes on the curved turnouts. The outer leg is correct; the inside or tighter radius on these turnouts is actually a smaller radius than listed in the product description.
Don Z.
If that is so, I’m disappointed. Perhaps Walthers exaggerates the radius of the inner curve by averaging it with the broader “point side” curve of the turnout.
Mark
Don Z is correct - the inside radii posted in the catalog and imprinted on the underside of the turnouts are incorrect. I carefully checked the radii by drawing various curves and overlaying the turnouts and found the following:
#6 (or 6.5; some labeling confusion): 24/18
#7: 28/22
#7.5: 32/26
#8: 36/30
Something else to look for is a dip or rise in one rail, or both, of the track approaching the turnout or in the turnout itself, which can cause a lead truck or front driver to jump off the rail.
Atlas turnouts have always had a bow in the middle, and the Walthers-Shinohara may have the same problem.
Hold a good straight edge such as a 2-foot long ruler along the length of the turnout and the approach track and look for gaps under it or whether it wobbles up and down on the ends. Also run a small string bubble level sideways across the track and check for variances in rail height.