Re: B&M Train Steering
Sorry I am late to the party...it's been a busy day ( cat >/dev/null 
if you must know). So let me take us all through a bit of Coaster Steering 101.
First, let me plug my ancient essay on "How Roller Coaster Cars Work (or don't)". There is a lot in there, but it should be noted that when that was written, B&M was new and the first Millennium Flyer train had not been built yet.
Let me start with the original questions, then move on to the subsequent questions. Steve Guilmette asked:
...do the trains without tie rods steer (on the outside rail) exclusivly by the spring load on the rear guide wheel?
The B&M wheel carrier assembly has six wheels on three carriers. I'm pretty sure that all of the wheels are mounted on pivoting assemblies that allow the position of the wheel relative to the track to dynamically adjust; a flexible bumper between the wheels maintains the tension. But for our purposes, this is really unimportant. The spring loading reduces minor bounce but it really isn't important for steering the train. Take away the springs and you'll get a less smooth ride, but tracking should be about the same.
The B&M wheel carrier assembly is mounted on a bearing which can yaw about the Z axis of the bearing...that is, it can steer. The trick is that the point of rotation is at the center point of the road wheel carrier. This means that when the track turns to the outside, the front guide wheel will be pushed outward by the rail. When that happens, the wheel carrier will yaw, causing the rear guide wheel to be pushed *inward*. That's well and good, but what if the track curves to the inside? The track won't make contact with the lead guide wheel, in fact it will pull *away from* the lead guide wheel. That's where the tie rods come in, right? To keep the wheel carrier from pulling away from the track and allowing the train to derail. But look more closely. The distance between the center of the front guide wheel and the rear guide wheel is far enough that when the track pulls away from the lead guide wheel, as the train is pulled laterally (with or without tie rods) the track will actually push out on the rear guide wheel on the inside. So in a nutshell, when the train hits a curve, the axle is steered by two guide wheels: the front guide wheel on the inboard side, and the back guide wheel on the outboard side. Since the wheel carrier is a rigid assembly, the whole thing rotates to follow the rail.
This also explains why the road wheels usually start with a flat surface even though they are running on a curved surface. When the wheel carrier rotates, there is going to be some error, where the road wheel does not run exactly parallel to the rail. But the distance between the front and rear road wheels on each wheel carrier is short enough and the curve is broad enough compared with that distance that the error is relatively slight. You will see a greater error on the Arrow multi-element train axle, which has the guide wheels further apart than on the B&M train...but I am getting ahead of myself here.
If the answer to the above is yes, why do any B&M trains have tie rods, and what makes the inverteds not need them?
[url][url]
Honestly, I am not sure how B&M decided whether to use tie rods or not. It may be instructive to note that the inverted coaster, which has no tie rods, was their first design. I don't have an answer to this one, though it may be related to the specific geometry of the guide wheel assemblies and the maximum or minimum allowable curve radius on a particular ride design. Yes, that's a complicated way of trying to duck out from beneath the green slime.
The old-style PTC junior cars, which have steerable flanged wheels, do not have guide wheels and have only one flanged road wheel on each corner of the train. These trains have tie rods, because the wheel flange is only steered from the outboard side, unlike the modern design that is steered on both sides.
Could a true trailered train function properly with fixed wheels (one road, one guide, and one upstop)?
The answer to this one is an unqualified "yes". The trick is that the hitch point has to be positioned correctly, which is to say, right in the middle of the axle. If you do it that way, the back of the car can sit with its axle completely fixed. The whole axle can yaw, staying centered and parallel to the rail. The hitch point, positioned in the middle of the axle, will rotate as the axle shifts, but most important it will not translate...that is, regardless of the angle of the axle, the hitch will remain in the same position. That means the next car can move with some independence, following the curve as it arrives. The Morgan wood coaster cars work in exactly this way, except that many of those don't have up-stop wheels because they were built for Traver-style track that doesn't have enough clearance for up-stop wheels. This is exactly the configuration used on Colossus, Dragon, Giant Dipper, Giant Dipper, Grizzly (CGA), and the Jack Rabbit (Seabreeze), among others. Those Morgan trains are pretty much crap, but they track pretty well, and they don't tear the track up much. That is also the configuration of the original Prior & Church trains, and it is basically the configuration of the GCI Millennium Flyer. The GCI train actually has a floating wheel carrier that dampens some of the vibration, but from a high-level perspective, it makes about as much difference as the spring assemblies on the B&M train: it makes a difference, but it doesn't fundamentally change the design.
FUN asked:
...do wooden coaster cars have wheel assemblies with independent movement? I always thought they were rigidly fixed the chassis.
It depends on the wooden coaster car you are talking about. The older PTC trains, and the trailered trains I just described, offer no movement in the wheel assemblies. The GCI trains have some motion, but it isn't significant to the way the trains track. The exception with the trailiered trains is that the lead axle on the Morgan, GCI, and Prior & Church cars can roll on its longitudinal axis. The wheels are in fixed positions on the axle, but the axle can *roll*. It does not steer, but by rolling it allows the train to enter a bank without lifting any axles. In fact all of the hitch points can roll, so all of the axles can roll relative to each other. The more recent PTC trains, the ones Jeffrey Seifert is talking about, have a rear axle that can roll about three degrees in either direction. So the front wheels are attached directly to the chassis, and the rear wheels are attached to a steel frame mounted on a pivot shaft under the back seat. For Gwazi, the back seat of the car was raised about an inch to give the axle a little more room to swing, and that same modification has become standard on the Gravity Group coasters. The extra inch gives the train about +/- four degrees of roll, but that's really quite a lot...that means that within the six feet or so from the front axle to the rear axle the track can roll four degrees. Four degrees in six feet means that in 540 feet you could do a complete barrel roll. 
More usefully, it means that when you see a 90-degree banked curve, you know that unless the train is lifting wheels, there is a minimum of 135 feet of running length between the flat track and the maximum bank angle on the curve, a little more than three full train lengths.
PTC puts the roll pivot on the back axle instead of the front axle because that makes the most sense. As the train starts to roll, the front wheels will follow the track, and the car body will roll to match, while the car's rear axle swings a little to stay behind the roll. It's counter-intuitive, but once you realize that the train is following the track, not the other way around, it makes sense. But it isn't a requirement; Morgan, GCI and Premier Rides allow the front axle to roll.
Also, it's worth noting that except for the GCI and Gravity Group trains, most wood coaster trains are basically copies of the PTC articulated design.
I think I have some photos somewhere that illustrate some of this...unfortunately some of them are photos that I promised I wouldn't spread all over the 'Net, so I'll have to make sure the context is suitably obfuscated, or get permission to share...
I think that covers pretty much everything covered so far. What questions are still unanswered?
--Dave Althoff, Jr.
Last edited by RideMan, October 28, 2009, 11:51P
/X\ _ *** Respect rides. They do not respect you. ***
/XXX\ /X\ /X\_ _ /X\__ _ _ _____
/XXXXX\ /XXX\ /XXXX\_ /X\ /XXXXX\ /X\ /X\ /XXXXX
_/XXXXXXX\__/XXXXX\/XXXXXXXX\_/XXX\_/XXXXXXX\__/XXX\_/XXX\_/\_/XXXXXX