But it does. The distance the following train needs to keep depends on its speed and its distance from the lead train. A block signal system uses only some of the data on the position of the two trains in question, and no data on speed.
Not true. Timers that start when a train enters one block can control the aspect of the next block (in addition to occupancy ahead). The net result is to ensure that a train is going below a maximum speed.
- In a block system, the position of the lead train is only approximated. There is no difference whether the last few feet of the train are in a particular block, or the whole train.
There is also uncertainty with regard to the position of a train in a CBTC system. Position is usually determined by dead reckoning with some form of correction at fixed points. The sensors count the revolutions of the drive wheel. The nominal wheel diameter is 34 inches. Suppose the actual diameter differs by 1/8". This is an error of 0.36% or 190 feet for every 10 miles travelled. The second source of positional uncertainty for a CBTC system is the maximum tolerated communications latency. Most systems tolerate a 1 second delay on each transmission. Consider a train were travelling at a maximum speed of 75 ft/sec (50 mph). The train will transmit its "state" to the central computer in 1 second and receive a command in additional second. Therefore, the system must assume a maximum delay of 2 seconds or 150 feet travelled before an emergency brake kicks in for lost communications. Thus, the CBTC system must assume a suitable cushion which is equivalent to the quantization error for a block system.
- In a block system, the speed of the following train is irrelevant to the amount of space it must leave. Yes, block signals can demand that a train slow down. But they don't say: it's OK for you to go 45 mph here, but if you do you need to leave more space than if you go 20 mph.
Consider the real world. If there were no stations, then the spacing between trains going 75 ft/sec at 40 tph would be 6750 feet. Most block lengths are in the 1000-1200 foot range between stations. Trains get close to one another only as they are approaching stations. The NYCT signals have taken advantage of this fact by decreasing the block length at station approaches. This reduces the positional uncertanty where it is important. At some stations, e.g. 125th and 59th on the Lex, they have dual signals within the station. These signals will go from red to a timer and then to a yellow aspect depending how far the train has left the station. Yes, the following train can enter the station at 40 mph or be told to slow down to 20 mph, if there is not sufficient stopping distance beyond the station.
The ability to place complexity where it is needed is one of a block system's economic advantages. A CBTC system must be to handle the worst case, communications-wise, regardless of context.
In practice, a current NYCT block system will not allow as close spacing at full speed as a PROPERLY DESIGNED AND TESTED CBTC system.
Actually, the current NYCT block system is pretty good. As I mentioned above, operating between stations at full speed is not a critical problem. The problem is at the stations. The NYCT system is uses block lengths in the 100-200 range for station approaches.
This is not to say that NYCT's CBTC will be properly designed and tested. They have the potential to screw it up.
They have not defined what a successful CBTC installation would be in terms of train speed and service levels. Therefore they have not defined what a screw up would be considered. :-)