The signal / train control systems for the newer areas of BART, WMATA, Atlanta, Miami, newer areas of the CTA, and a number of other places, were all built by GRS. These use AF track circuits in most places, as does many rebuilt areas of the NEC. CalTrain is also planning on using them in preparations for electrification. In this case, the microprocessor based signalling hardware they are using, is fully capable of using both DC or AF track circuits - just by swapping out the applicable I/O board types that are plugged into it.

Moreover, in order to make PTC work in multiple track areas (or to just control trains on sidings and spurs), the use of AF track circuits or overlays of differing frequencies, will be mandatory. At the current state of the GPS art, that is the only way PTC can prove which track a given train is on before commanding it. There are other schemes proposed, but all are profoundly hair-brained, relying on circumstantial evidence and inferences, as they do.

Since DC type track circuits will not work in electrified territory; coded or not, AF or PF track circuits must be used instead, in order to prevent propulsion current from unsafely interfering with occupancy detection. PF circuits at 60hz or 100hz are still very common on electrified railways, but the AF variety is becoming more commonplace in new installations.

At BART, GRS used Insulated PF (at 100hz I believe) for relatively short (100ft - 500ft) OS tracks inside interlocking limits; but used AF tracks elsewhere. In those PF circuits, non-coded 100hz is sent in one fixed direction down the OS circuit for occupancy. A coded AF signal is also sent through the OS rails, to maintain cab signaling while traversing through. However, the direction of the coded AF is physically switched by a directional repeater relay, from one physical end of the track circuit to the other, depending upon the direction of travel.


If I recall my training days correctly, the WP did use the AFO as an overlay on a few vital pole line circuits, to transfer intermediate occupancy indications (middle blocks) to the nearest headblock hut (CP) to be put on the CTC indication line. But those particular devices weren't really designed for vital safety applications, nor for use as track circuits. I seem to recall hearing that BART was exploring their use on vital line circuits at one point, but they ultimately dropped the idea.

The AF track circuit is a different species altogether. With its carrier frequency usually anywhere from 200hz to 2.5khz, its maximum length could vary greatly, but it is always considerably shorter than a coded DC track circuit could be in the same location. It is seriously affected by track inductance and worse yet, by secondary and tertiary coupling to unintended places - sometimes compromising safety. This effect is worse the higher the frequency. The signal engineer using them must take all this into account, and provide means to safely mitigate all such possibilities.

At unusual places, such as older areas of BART, where track circuit frequencies between 5k and 10K were used, these effects are very severe, limiting block length to about 1200 feet; and causing all sorts of crosstalk problems between the track circuits. The latter are safety critical issues, requiring complex code synchronization and frequency separation schemes, in order to detect and mitigate.

But given all that, and even though the DC circuit can be nearly twice as long; the advantages of the AF track circuit in coded track circuit applications (cab signalling, PTC, etc) far outweigh the problems; especially in electrified territory, where the DC track circuit simply will not work safely, if at all.