Wednesday, September 24, 2014

War Stories: Episode 29 - "Sadder but Wiser"

In the musical, Music Man, there is a song, "Sadder but Wiser Girl for Me".   Howard Hill's interest is for the girl that has learned life's lessons the hard way.

Lessons learned the hard way tend to stick.

Conrail and Amtrak had learned a hard lesson on January 4, 1987.  (see here:

Things started changing right away.

Conrail reinstalled train stop on their cab signal system and replaced the annoying and easily tampered with whistle with a hard to defeat electronic beeper.  The risk of a penalty brake application causing a derailment seemed almost irrelevant compared to what had just occurred.

Conrail also looked  to CSX for some space to reroute some trains off the NEC between Washington and Philadelphia.

Amtrak immediately slapped a 30 mph speed limit on all Conrail freight trains.  Shortly after, they changed it to almost no Conrail train operation except between 11 pm and 5 am (or something like that), but that would be allowed 50 mph.

The Federal Railroad Adminsitration (FRA) and Congress started legislation and rulemaking that led to a requirement for speed control on all trains on the NEC.

Conrail pushed back on this one.  Using the old "suppression style" speed control was going to cause problems with the train sizes currently being operated.  Lots of dangerous slack action and potential for derailment.  Exactly what you DON'T want on the NEC!

Working with the FRA, Conrail developed a specification for a speed control system that would satisfy the FRA and safe train handling.  They called it Locomotive Speed Limiter (LSL).  It required the engineer to slow his train down so he'd approach the next signal at the right speed for that indication according to a generic braking curve programmed into the device.  It was the best attempt that could be made to turn the existing cab signal system into a predictive speed control system.

The specification was sent to both "old line" cab signal manufactures, Union Switch and Signal of Pittsburgh PA and General Railway Signal of Watertown NY. It was also sent to an upstart signal manufacture, Harmon Industries of Grain Valley, Missouri.  Conrail's Communication and Signal Dept. had a good relationship with this vendor and suggested we give them a shot.

Harmon's bid came in far below the other two.  Harmon it was - a bit to the chargin of some of the US&S "true believers" - PRR, PC and Conrail had used US&S exclusively for locomotive cab signaling longer than anyone could remember.

Things progress quickly. Six prototypes were constructed and installed on SD40-2s for testing by the end of October, 1987.  We were going to run them between Oak Island and Alexandria (Conrail trains OIPY and PYOI) which used the NEC from Philadelphia and Bowie MD.

Each test train operating with LSL had to have one Conrail and one Harmon rider.  We had a very early laptop computer that we connected to the device to log activity.   It was small and compact for it's day, but weighed a ton.  You had to boot it up with a floppy disc containing DOS in one drive and it would log data to the floppy in the other drive.  There was no hard drive in it.

 I drew the first southbound OIPY2 departing Oak Island the evening of November 2, 1987.  We gave the locomotive engineer a brief explanation of how the system worked, put the system in "no cab" mode, since there was no cab signalling until Philadelphia.  The train followed the old Lehigh Valley line to Bound Brook where we swung onto the ex-Reading and headed for Phila.  We wiggled through Philadelphia, up to Zoo where we climbed up onto the ex-PRR Highline - the freight bypass around 30th Street Station.  It was here that we first hit cab signal territory.

The device automatically changed to "cab" mode, as it was programmed to do and would now enforce speeds associated with the signal system.  We entered the NEC at Arsenal tower and proceeded south.  Each time the signal would change to a more restrictive aspect, the LSL device would dutifully start counting down the seconds until a penalty brake application would occur based on the current speed and position relative to the braking curve, just as it was programmed to do.

That same night, there was a northbound test train departing from Pot Yard in Alexandria.  We should have passed it near Philadelphia, but now we were south of Wilmington and there was no sign of it.  We arrived at Baltimore to work the yard there, and there was the northbound trying to get out of town.  It seems their LSL had been nothing but trouble, crashing and freezing up many times.

We finished our trip without a hitch, but that northbound was a sign of things to come.  We rested at a hotel in Alexandria and departed the next night on train PYEN for Enola PA.  Our LSL froze a couple times and needed to be reset.

That was the end of the beginning.  Harmon would take the data gathered and make a myriad of hardware and software changes - sometimes wholesale.  We'd ride the trains and gather more data - and return with even more failure modes.  By then, Conrail had diverted most traffic off the corridor.  Nearly all the rides were to/from Baltimore and Enola.

Some recollection of rides:

Early on, I caught a juice train from Pot Yard.  It was a solid 60 car train of refrigerator cars filled with frozen orange juice.  It ran as a unit train from Orlando, Florida to Newark NJ.  This day, it ran up the corridor during daylight, so was limited to 30 mph.  Nothing like going 30 mph on 125 mph track, bouncing off the low rail in curves with 6" superelevation!  The engineer told me he often could feel the juice sloshing around in the reefers.  I didn't have the heart to tell him it was frozen OJ and it was likely the end of car cushioning he was feeling. (It still runs:

While the train to Oak Island was still running up the NEC, we'd get off right on the High Line in Philadelphia and walk down the stairs to Market Street, where we could catch a transit ride to work or back home.  We were issued a switch key for this as the stairs had a locked gate.  I still have the key...

The usual drill from Baltimore to Enola was to take Amtrak to Baltimore, call for a jitney to the yard and ride to Enola.  The train would pull down to the yard limit and wait on the Amtrak curfew before toodling up the NEC to Perryville, then up the Port Road to the Enola Branch at Columbia, to Enola.  We usually arrived about dawn, got a jitney to Harrisburg train station and then went home to rest.

One night, we crossed over to the middle track of three going through Aberdeen, Maryland.  They ususally left us on the most eastern track up to Perryville.  As we're cruising through Aberdeen, a headlight from an Amtrak train appears around a curve to the north.  Minutes later, a second headlight appears around the same curve.  Nervous glances in the cab.  This was not what typically happened.  The first train closes the gap rapidly and is definitely on the track to the west of us, but the other one?  Could all three tracks be signalled for southbound operation?   Yes, and it became apparent a few seconds later -  blam - whoosh - whoosh - whoosh - whoosh - whoosh - whoosh - whoosh - zinggggg!

For the Enola to Baltimore trip, we would take Amtrak to Harrisburg in late afternoon and get to Enola about 5:00 PM.  The crew was usually on duty about that time and we'd depart at 7:00, cruise on down the Port Road and get to Perryville about 9:00 PM - only to sit until the curfew at 11:00 PM.  We'd get the signal, swing onto the bridge over the Susquehanna and be to Baltimore by 1:00 AM or so.  From there, we could catch a late night Amtrak train about 4:00 AM and be back in Phila by 6:00.  Then home, then bed.

One night, we had an older engineer who was really quite skilled and nice to ride with to boot.  The locomotive not only had the new cab signal beeper replacing the whistle, but had and alerter as well, which would also beep.  He started to nod off at one point during the trip, and a minute later, the alertor started sounding.  He woke up with a start  and was slapping the alerter reset and stomping on the cab signal pedal - each many times with great gusto until the beeping stopped.  I guess these things worked!

Another trip, I rode with an engineer who: 1) had just "bumped back" into freight service after a couple years with Amtrak and 2) had spent all day in divorce court.  That he hadn't run a freight train in a couple years meant he's never run with LSL.  That he was in divorce court all day, he struggled to stay alert.  This made it doubly tough to remember the railroad and how to handle a freight train over it.  At one point, we were pulling over a crest into the next swale when he slapped the throttle shut.  The his eyes got big!  He sat straight up and started notching back out rapidly, a notch at a time.  He was about to notch 6 when the slack ran in.  WHAM!  He later told me he used to just let the train run up to 60 mph or more through the swale and the coast back down to track speed.   No harm. No foul.  Not anymore...

The Port Road was intereting.  Right along the river.  Lots of curves. Tunnels, too.  The speed was 40 mph, but there were lots of 30 adn 35 mph curves.  Most engineer just ran a steady 30 mph and didn't bother trying to make time just to have to sit and wait at Perryville.   However, one engineer was a cut above the rest.  He ran the train right on the limit, speeding up and slowing down exactly in accordance with the timetable....and he did it effortlessly!

All along the way on these trips, I heard stories of what was where - or what used to be where along the line.  Most of these stories involved women who were visible from the tracks in various states of dress...

Finally, by the end of 1988 or early 1989, we were done testing and started installing production LSL units on a chunk of SD40-2s and B23-7s.  The B23-7s were for the Hudson Line, which had become included in covered territory after a Turboliner went nose first into a Conrail freight at Tarrytown.

Not good...
Harmon took what they had learned about cab signaling and developed their own system - Harmon Ultracab.   They took that plus their signal expertise and developed a turn-key transit train control system which they sold to St Louis's light rail line.  Later yet, they expanded the LSL system to become the ITCS system now used to govern the operation of 110 mph passenger trains in Michigan.  Conrails started buying Ultracabs with their first order of C40-8Ws.  They were less than half the price of the normal US&S system.  Today, Harmon/GE's Ultracab II is the standard system on NS.

Ultracab on Conrail's first C40-8W, CR 6050 during testing at Erie

This is quite a bit of good.

After the innocence ended, we were ultimately wiser.

There must be fifty ways to feed your diner

With apologies to Paul Simon...

"Don't give him a flapjack"

"Get a steak from the Outback"

"Or bring on a bag snack"

"But don't make it free."

"Put on a machine, Gene"

"And fill it with cuisine"

"No attendants that are mean"

"Chow down with the scene"

I just read this blog at Trains Magazine.

Eating in the diner on a train can be a great pleasure. ( )  But, Amtrak loses gobs of money providing it.  They are so bad at it, that Congressman Mica decided to have a hearing about it a while back.  It made good political theater, but it was one more log on the fire for those who would burn Amtrak to the ground.

Can anything be done about it?  Sure.  The problem is that Amtrak is still running their dining car service like it was 1950.  They stock food in a commisary, load it on the diner, cook the food on board, and have a wait staff take orders and serve it.  When the diner staff is not working, they are sleeping on the train.  It's been this way since trains first started crossing the great open spaces between the cities.  Back then, restaurants were few.  People cooked most meals for themselves.  The railroad dining cars needed a "closed loop", vertically integrated system to deliver meals to passengers. It's a very expensive way to do it.

Since 1950, a lot has changed.   Cooking meals at home has wane.  People eat out a lot.  Much if it at fast food joints, but quite a bit at chain restaurants.  The country has sprawled.  The interstate highway network is complete and every 30-60 miles, there's an "interstate" town by an exit complete with a variety of restaurants.  Prepared food used to be a rarity.  Now it is everywhere.

There is a whole industry to support and deliver prepared food.  Everything from large national and regional restaurant wholesale suppliers to restaurant "groups" that own several chains.  For example, did you know that Bloomin Brands owns Outback, Bonefish Grill, Carrabba's, Flemings and Roys?  Darden owns Longhorn, Olive Garden, Season's 52 and Capital Grill, among others.

These people are experts at making money serving prepared food.  If they can do it, so can Amtrak.  There are several ways to go about it:

  1. Concessionaire:  Rent them the space and get out of the way. Menu, prices, etc. all up to them.  They keep the revenue and bear the costs.  This way the only cost to Amtrak is to provide the diner in the train.  They want to house the staff on the train?  Fine.  They can pay for the space.  They want to cycle the staff on and off the train around meal times?  Also fine.
  2. Contract provider:  Set the requirements for the service and bid it out. Write performance measures into the spec with bonuses good performance.  Might wind up with a negative winning bid, that is, Amtrak might have to provide a small subsidy. 
  3. Benchmark:  See what the industry leader business and logistics processes are and copy them. 
In all cases, things would change.  It might mean the food might be cooked at restaurants along the route and delivered to the train, ready to reheat.  It might mean that favorite, branded food would be available. It might mean that food is still cooked on board, but is served like "just another location" in a restaurant chain.   It might mean that staff rotate off to hotels instead of taking up valuable revenue sleeper space.

There are probably more ways to do to.  These are only the ones that have popped into my head.

All of these would be better than the status quo.   And, they might help Amtrak's image, to boot!

Paul Simon might even be able to think of a few more.

Wednesday, September 17, 2014

War Stories: Episode 28 - "The End of Innocence"

Trains were always fun.  Watching them go by as a kid, with my dad.  Model trains in the basement.  Train rides as part of vacations and trips.  This was all fun.

Then I got a job with the railroad.  It was serious work, but still fun.  I got to see trains, climb on them and ride them.

Then, January 4, 1987, it all changed. I wasn't even at work or  near a train.  I was home watching and NFL playoff game.

The game was interrupted for breaking news.  There had been a train collision at Chase Maryland.  People were dead.  The helicopter pictures showed a horrendous scene.  The northbound Colonial travelling over 100 mph had struck three stationary Conrail locomotives just north of Gunpow interlocking on Amtrak's Northeast Corridor.  There were fourteen passengers and two Amtrak employees dead and scores injured.  It overwhelmed the local residents and EMS.  It would have been much worse had the lead coach been occupied.  It was completely crushed.

More detail here:

This was not fun.  This was depressing.

Details started coming out.  The Conrail engineer had been smoking marijuana.  The cab signal system on his locomotive had a muffled whistle and missing bulb - which he should have caught when he tested the system prior to departure.  He had blown by an approach signal, past a cab signal cut point and through a stop signal, trailing through the interlocking switches and into the path of the Amtrak train.

At work on Monday, the mood was somber, but the wheels of change had started to turn.  This couldn't be allowed to happen again.

Some background, first.

At the turn of the 20th Century, railroads were not a safe industry.  Death and injury to employees and passengers was not uncommon.  Technological improvements to safety were developed and implemented as time went by, often with some pushing from government in reaction to particularly gruesome train wrecks.  Steel passenger cars replaced wood ones after a particularly deadly, firey passenger train wreck, for example.

Advance signal systems became available and railroads slowly installed them.  Train stop systems that could apply the airbrakes if a train went past a stop signal appeared.  Later, more sophisticated systems that could show the current signal indication inside of the locomotive's cab were developed.  These systems were known "cab signals", generally. An added feature was called "train stop".  If train went by a more restrictive signal, a whistle would sound, and if the engineer didn't acknowledge it by pressing a pedal, the brakes would apply.

The Pennsylvania Railroad (PRR) installed quite a bit of this inductive cab signalling with train stop on their railroad, including the entire railroad from New York City to Washington DC, the part we now call Amtrak's North East Corridor (NEC).   Later, they added an additional feature called "speed control" to it.  It not only required the engineer to acknowledge the signal change, it required him to apply the brake and stay under the speed for that particular signal indication.

Nice explanation and demonstration at 3:40.

The PRR eventually installed it on every passenger train and electric locomotive.  This meant that nearly every train on the NEC was covered with speed control in the 1960s.  This condition lasted until roughly 1980.

Then, three things happened.

One is that Conrail stopped using electric locomotives (more here

Second, trains got longer and a penalty brake application from the train stop system, under certain conditions, might cause a derailment.  So, Conrail took the train stop function from the cab signal system.

Third, speeds and number of trains on the NEC were increased.  First, Metroliners were allowed 110-120 mph.  Later, in the early 1980s, tracks were improved and new electric locomotives and coaches were acquired and speeds were raised up to 125 mph on long stretches for nearly all Amtrak passenger trains.  Amtrak increased the number of trains as ridership responded to the faster trip times.

Add in a rather laissez-faire attitude to drug and alcohol use at that time and you had a deadly mix. (More here

The result was, in the 1980s, Conrail diesel hauled trains were moving up and down the NEC without the benefit of train stop or speed control, mixing with an ever increasing number of faster passenger trains. During the dark years of the "railroad crisis" and Penn Central bankrupcy, the world had apparently forgotten what the PRR had known.

And a price was paid.