Tuesday, April 30, 2013

How to Patch Your Signaling System (in 3D)

In this week's installment I will travel to Georgia to highlight an interesting development on the CSX Abbeville Subdivision that I discovered while visiting a friend last year and how it ties in with a pet peeve of mine that wound up influencing signaling in the UK. The Abbeville Sub was a former Seaboard Air Line route between Atlanta and the port city of Wilmington, North Carolina. The line is your typical single track CTC with passing sidings and sees only freight traffic including double stack trains.

I have mentioned before that the current practice in North America regarding signaling system alterations is to typically replace the existing signaling even if the existing signaling if relatively new with solid state components. My suspicion is that it is easier to test new signals in situ than to try to patch in new vital logic to the older, in service equipment. However every so often there are exceptions to this rule, usually where a single signal is replaced due to structural deficiency or damage. Less frequently single signals will be replaced as part of a change in the physical plant. (And least frequently will the old hardware be modified.) This post is an example of CSX performing a simple signaling upgrade and only replacing the signals that were directly affected instead of everything in sight.

The location in question is between milepost 506.8 and 513.4 near Athens GA. The upgrade in question was the installation of a new interlocking at milepost 509.6 to serve a large grain customer so that crews no longer had to operate a hand throw switch. Previously there had been an automatic signal at 510.1, an interlocking at the south end of the Fowler Junction controlled siding and another interlocking at the north end of the siding. The distances involved were 2.3 miles between the auto and SE Fowler and 9,800 feet between SE Fowler and NE Fowler. All signaling was of Seaboard origin and dated from the 1970's or early 1980s. On CSX a controlled siding is one with interlocked end points, but no train detection so all diverging moves into the siding receive a Restricting aspect. Signal progression for a Restricting signal on CSX is just Approach to Restricting so the 511 automatic could make do with a single 3-lamp head in each direction. However the new interlocking at milepost 509.6 would split the 2.3 mile long block into two blocks of 1.5 and 0.8 miles respectively. While a 1.5 mile block wasn't a big deal, the 0.8 mile block would be too short for a train to safely stop and it was decided to install 4 block signaling on either side of the short signal block. This would require the signals at the 513 auto, the 511 auto and NE Fowler Jct to display an Advance Approach aspect, which they were not set up to do.

Now had this been working under NORAC, Union Pacific or Canadian signal rules the change would have been trivial because Advance Approach (preceed prepared to stop at second signal) is represented by a flashing yellow (*Y*). However CSX's Seaboard System rules use Y/Y for Advance Approach. This means a hardware modification to install a second head or, completely new signals and that is what the signal vendor encouraged CSX to do. So here we see the new 5110 automatic signal that I happened to catch displaying Advance Approach. Note the old (P) board at the base of the signal which was used in the Seaboard system to mark Permissive automatic signals as opposed to the more common numberplate.


I was pretty lucky to catch the 511 auto displaying Advance Approach as the signal itself is approach lit providing only a brief window of time when a train is in the block where one can catch anything lit up. Aside from the actual signal hardware everything else about the TCB signaling in the area had been left unchanged including the relay cabinet and the original concrete signal base that the new mast appeared to be glued on top of.



This is a good opportunity to explain why I dislike the Y/Y advance approach indication.  The obvious problem which is demonstrated by this complex patch job is that you have to add an additional lamp or signal head when shortening block lengths instead of just setting the yellow to flash.However from a signaling theory point of view Y/Y Advance Approach makes the signal aspects much less consistent. Basic block type signals can no longer all be displayed on one head (or with one lamp). Moreover on CSX Y/Y for Advance Approach displaces what would normally be Approach Slow into a three lamp aspect, Y/R/G. This then serves to dilute the impact of having only absolute signals three lamps as well as disrupting the pattern of the "Approach Speed" series of signals listed below.

Y/R = Approach Stop
Y/L = Approach Restricting
Y/Y = Approach Slow
Y.G = Approach Medium
Y/*G* = Approach Limited

By stealing Y/Y for advance approach the pattern is disrupted. Y/Y was originally the standard way to display Advance Approach and was actually copied by the Big 4 British railways when that system adopted color light signaling in the 1930's.  The reason for this was because at the time flashing relays were both expensive and unreliable, but as technology marched on flashing signals became common and when Conrail was forming what would become NORAC in the 1970's the decision was made to dump Y/Y Advance Approach and replace it with flashing yellows.  The real headscratcher is that even if CSX's embrace of Y/Y advance approach precluded wholesale conversion to the two headed style of Approach Speed signals, the Seaboard rules do not currently make use of *Y* for another indication.  Much like Conrail did CSX has been free to phase it in ever since it embarked on its massive re-signaling spree starting around 1999.

Anyway, in addition to the 511 Auto the 513 Auto also needed replacement to fit Advance Approach due to the now "short" 1.5 mile block between the 511 auto and the new interlocking.  Again had *Y* been available all the original signaling hardware would have been able to be left in place.


Now because the North End of FOWLER JCT is an interlocking only one of the three Seaboard era mast signals were replaced to reduce fiddling with and re-certifying too many new components. So for at least the time being at NE Fowler Jct we have classic "large target" US&S color light signals for northbound trains.


And one rather out of place Darth Vader mast for southbound trains once again placed on the same foundation as its predecessor.


Looking closely at the new signal we can see that the top head has G/Y/R, as is expected, and the bottom head has Y/L/R. The lunar is for Restricting aspects into the un-signaled siding and the yellow is for the Advance Approach.

If you're a clever cookie you may have noticed something interesting about the pair of classic signals for northbound train each lack the ability to display a Restricting aspect. This is another big irritation of many southern and western railways as they, at least in the past, didn't feel it was worth it to outfit absolute signals with some sort of "call-on" light. Therefore if the block beyond is occupied the trains can no longer proceed on a Restricting aspect, but instead must either wait for the block to clear or contact the dispatcher for permission past a Stop signal. This practice evolved in the south where traffic densities were much lower and thus there would be little demand for such permissive practices. Another point of note is that the siding exit signal lost its third head in the year or so before I managed to visit. I am unsure even why it had a third head let alone why it was then decided to remove it.

Of course I should mention what's going on at the new interlocking called, in what might just be a product placement, PILGRIMS PRIDE, after the grain processing center that the sidings serve.  In the eastbound direction the new mast signal mirrors the one at NE Fowler Junction with both a yellow lamp for Advance Approach and a Lunar for Restricting indications into the siding.  The Advance Approach is due to the short block to SE Fowler Jct, although if the old block between NE Fowler and SE Fowler was long enough for 3-block signaling I am not sure why is suddenly needs 4.


Heading eastbound from the new interlocking the block between there and the 511 auto has been deemed long enough for 3-block operation, even thought in the eastbound direction it required an upgrade to 4-block.  Again, go figure.



Ironically, at least as I have been able to gather using overhead intelligence, SE FOWLER JUNCTION, interlocking seems to have been left completely unaltered.  So as you can see CSX can't seem to act consistently either in the rulebook or during the implementation phase.

No comments:

Post a Comment