On Monday, December 28, 2015 at 12:30:37 PM UTC-5, Stephen Sprunk wrote:   
      
   > On 27-Dec-15 20:32, hancock4 wrote:   
   > > I believe much of the Long Island Rail Road and ex PRR SEPTA lines   
   > > still use the amber position light signals, though upgrades may have   
   > > eliminated some signals.  FWIW, relatively new pictures on nycsubway   
   > > of the LIRR trunk line show signal bridges with PRR style signals.   
      
   > Most RRs upgrade signals only when they fail, and that won't happen very   
   > often for position-light signals; they use long-life, hard-duty bulbs   
   > for obvious reasons.   
      
   I believe sometimes there is a economic motivation even if the   
   existing signal system is working as expected.  I believe old   
   systems with relay logic were still labor intensive and modern   
   systems are more reliable and automated.  Further, a replacement   
   system may increase track capacity by allowing higher speeds   
   and/or more trains.   
      
   But also, on some old railroads the signal systems might be 75 y/o   
   and no longer reliable.  The IND in NYC is supposedly a challenge   
   since it has many miles of original 1930s wiring and relays.   
   (I believe some of the money intended for the Second Ave   
   Subway in the 1950s was spent instead of redoing the then   
   50 y/o IRT signals to modernize them and increase capacity;   
   which was probably a worthwhile expenditure).   
      
      
   > Also, unlike the transition from incandescents to LEDs, this is a change   
   > that really needs to be done for an entire gantry (or set of masts) at   
   > once, meaning there's an incentive to just keep replacing failed bulbs   
   > individually rather than potentially convert 12+ entire heads.   
      
   Signal lamps do burn out.  Even on approach-lit signals there are   
   outages, which can be a nuisance.  For instance, a burnt-out green   
   on a wayside means every train has to creep through.  Ex-PRR   
   position light signals often had one lamp of the row out, but since   
   there were three lamps the signal still functioned.   
      
   FWIW, it appears that the signal heads in Pennsylvania Station and   
   its approaches are the 100 y/o units.  I have no idea about their   
   innards or control logic.  Supervisory control is modern computer-   
   based.   
      
      
      
   > OTOH, what all is needed to do a conversion?  Remove the center bulb and   
   > block the hole, change two outer lenses from amber to red, two from   
   > amber to green, two from amber to lunar white, and then what?  That   
   > seems simple (and cheap) enough to just all the rest at once.   
      
      
   I believe there are modfications to relay controls, perhaps the   
   wiring, not just the signal head.  For example, at one NEC   
   location, I noticed that the hardware holding up a signal was   
   badly corroded.  It appeared when they went to color they   
   replaced all the hardware and put it an all new signal head.   
      
      
      
   FWIW, SEPTA has bid specs for a CBTC system. (500 pages, 6 meg,   
   very detailed, but worth a look).   
   http://www.septa.org/business/bid/100k/detail/MSHL-I&TC-%20FINAL   
   20OPEN%20SPECS-_ISSUED%20FOR%20BID-1.pdf   
      
   --- SoupGate/W32 v1.03   
    * Origin: LiveWire BBS -=*=- UseNet FTN Gateway (1:2320/1)