> Cylindrical wheels reduce hunting of trucks. Hunting was a big problem with the first
> BART cars. You could really feel it at speed. That's why BART stopped using tapered tread
> wheels. This was the result of extensive research by BART.

Not sure I agree that there even was a hunting problem, or anything wrong with the ride. I did ride them during early testing as a recorder, until I was transferred to wayside work. Didn't notice it at all. But then, maybe I was so used to swaying rail cars that I just didn't notice.

It is true enough, cylindrical wheels do seem to reduce hunting, but only if gauge tolerances are also tightened sufficient that both flange radii simultaneously contact the rails. But this causes very rapid wear, which soon enough negates any such benefit. And as the wear continues, the car not only hunts badly, but will sometimes slap side to side violently; because without the tapered tread, there is then no control at all until a now worn flange finally HITS a rail.

Solving the so called hunting problem this ways merely trades the cost of fixing it the right way now, paid for by the design consultant; with a very high maintenance requirement, paid for later by BART (the public actually). The fact is that it could have been completely solved back then by proper mechanical damping of all oscillations; but it was not in Bechtel's best interest to do that.

There is no way in hell that BART could have conducted any extensive research effort on any subject whatever. That research was done by Bechtel, which was biased as mentioned above. Given that BART's first design and construction contracts were "Turnkey projects, BART had no employees capable of such research at the time. Those were hired much later, after it had become apparent they needed them.

Moreover, as the British railways discovered through extensive research, some minor amount of oscillation is needed, to keep the rail and wheels clean enough for safety critical train detection to work reliably. As discussed below, cylindrical wheels did not allow enough of that - to BART's great detriment.



> Even the BNSF has been prone to track circuit shunting problems

I believe all railroads have had track circuit shunting issues at one time or another. I remember my Grandfather having to lay down (weld) a bead of relatively soft nickel alloy on the rail tops in some places. It was soft enough that the trains quickly squished it flat. The SN was having problems with rusty rails approaching grade crossings on some light traffic lines. Since iron oxide is an effective insulator, the light trains (44 ton diesels) would not shunt very well. But the oxides of this nickel alloy were relatively conductive. Problem fixed, at least to a tolerable level. But because of its softness, it wore off and had to be redone every 3 or so years.

I recall that BART tried this at the Dalay City Station once. But they apparently welded on a bead of stainless steel instead of a softer alloy. Trouble was that the bead was too hard and caused cracks in the rail head - meaning that all the rails they did this too had to be replaced - prematurely.

But as for shunting problems in general, I believe BART has had way more than its share. Wheel contour contributed greatly to this. Because of the cylindrical tread and tight gauge tolerances at the flanges, the cars would not "wipe" the rail heads very well. BART's wheels polished the gauge corner and the far side of the rail head, but left a think bead of rust on top of the rail the rail head at the gauge side. When enough wheels did manage to hit the rust spot, the track circuit would malfunction and loose the train. As the wheels and rails wore down to a closer mechanical match, the rust disappeared and detection improved, but there were still issues.

But by far the worst shunting problem with the wheels, was the dis-similar metal used in wheel hubs. Carbon steel tires were hydraulically pressed onto aluminum-magnesium hubs, which were likewise pressed onto carbon steel axles - sometimes using a semi-conductive grease to facilitate the process. That means there are four sometimes contaminated galvanic junctions on each wheel set; whereas, with solid steel wheels on solid steel axles, there would be little of such problems.

To shunt the track, the track circuit current needs to traverse all four galvanic junctions, as well as the metal to metal contact at the rails. Unless track circuit voltage (or some other current) is high enough to break down those galvanic junctions, BART wheel sets thus present a net impedance across the rails two to three times higher than that of normal railroad wheels. Thus leading to chronic marginal detection on any standard 60 milliohm track circuit, which is what was specified by Bechtel, and supplied by Westinghouse (barely making the spec - but not entirely). I have never found any Bechtel spec for the shunting ability of the cars - Bechtel probably entirely overlooked that.

It still works well enough at BART, so long as propulsion current (that other current) is flowing, breaking down those galvanic junctions. But not so well when the third rail is off; or if any particular car is not drawing third rail current, there is a marginal chance that it will not be detected.

So again, BART's Bechtel re-invention of the wheel has proven a great dis-service to it, and to the public which daily has to put up with and pay for BART's woes. All told, an across-the-board unmitigated disaster - actually!

OPB