The June 22, 2009 Train Wreck

On June 22, 2009, near the Fort Totten Metro Station, 9 people (the train operator and 8 passengers) were killed and scores more were injured unnecessarily in a horrific wreck on Metro’s Red Line. It could have been avoided and should never have happened.

Incidentally, because Metro carries Workers’ Comp (WC) insurance, the family of the train operator, Janice McMillan, cannot sue Metro. Everyone else can sue but Ms. McMillan’s family will only receive a pittance in relation to what the passengers and/or their dependents will receive. A passenger who had her arm broken in the wreck (for example) will almost certainly receive a much larger settlement than Ms. McMillan’s family. This is grossly unfair and needs to change. Employees (anywhere) should not be forced to give up their right to sue as a condition of employment.

Most of the details about the worst accident in Metro’s history have been reported in the Media. There’s no point in me repeating them here.

However, there is something that to the best of my knowledge was not mentioned in the MSM and that is the test or PMI we performed to check the track circuits. The “Track Circuit Data Sheets” that we had used since I began working for Metro in 1983 had two (2) columns for the verification procedure (where we pretend to be a train). We ATC technicians had been specifically instructed to only use one of the two columns. We only performed one of the two procedures because that’s how the PMI was written and it’s how we were taught to do it. Had we been performing the second half of the verification all of those people would almost certainly be alive and uninjured – celebrating the holidays with their family and friends as I write this.

That’s really all you need to know. For those who are interested in some brief explanatory details, they follow.

If you’ve got coffee, drink it now!

Seriously, if you’re at all interested in how subway trains can move automatically, this should be somewhat interesting. If not, well, read at your own risk.

This photo should be helpful in visualizing what I am referring to below.  The black box in the foreground is an ‘impedance bond’.  The small cable carries the train detection and speed command signals from the TCR.   The yellow sign is a ‘chain marker’.  A ‘chain’ is 100 feet.  This is 71,700 feet from Metro Center.  ‘A1’ is the A line, track#1.  This is just north of the Twinbrook Station.  The tracks on the other side of the fence belong to CSX.


The railroad tracks are divided into sections, aka “track circuits”. Like city blocks except track circuits vary in length from less than 100 feet to maybe 1,300 feet or more.

In order to detect the location of the trains, a signal is fed through the rails from one end of the track circuit to the other. It is an audio frequency signal which you can actually hear in some places. The rails act as wires and conduct the signal. Where the signal enters the rails is called the transmitter (Tx) end and the other end is called the receiver (Rx) end.

Normally, with no train in the circuit, the signal goes out to the tracks through a cable to the Tx end, travels down the rails to the receive end and back to the TCR. As long as the signal makes it back to the room the track circuit is assumed to be vacant (no train). Any approaching train should get unrestricted speed commands (the equivalent of the posted speed limit on a public road).

When a train enters the circuit, its steel wheels and steel axles form a short circuit, connecting one rail to the other which prevents the signal from making it to the receiver end. When the signal does not return to the room the track circuit is assumed to be occupied by a train.

It is done this way for safety. Anything that causes the signal to not complete its journey out to the rails and back (a cracked rail, a blown fuse, a broken rail clamp, etc) will be interpreted as being a train in the circuit. It is supposed to be a “fail-safe” system, and it is – unless Metro “engineers” start mixing new equipment from one mfr with old equipment from another mfr. More on that later.

A ‘track circuit verification’ is vital but pretty basic. We would use a ‘shunt strap’ (like a single jumper cable) and short across the rails – to simulate a train – and confirm that our equipment indicated that the circuit was occupied. We called this ‘dropping’ the circuit (causing it to show occupancy).

Above I said we only did half of the procedure. The data sheet had columns for both the Tx and Rx ends. I recall asking my supervisor back in 1983 or ’84 why we never did the verification at the Rx end. He said it wasn’t necessary and clearly wasn’t interested in talking about it any further.

In fairness, what he said made sense intuitively – if the signal could be shorted out at the Tx end where it is the strongest, verifying the Rx end would be an unnecessary waste of time. To my knowledge no one really questioned this over the years – not even the most intelligent and conscientious technicians and supervisors (some of which are _real_ engineers).

We continued verifying the transmitter end only, until after the wreck in June of ’09.

After the wreck, what we found was that contrary to what our training and common sense would suggest, it was actually more difficult to ‘drop’ most track circuits at the Rx end. What that meant from a safety perspective is it would be possible for a train to be in a track circuit and not be detected – for all practical purposes it would disappear!

That’s exactly what happened the day of the accident. An inbound train was holding outside Fort Totten, waiting for the train ahead of it to clear the platform. It happened to be sitting close to the Rx end of the circuit and was not detected. Therefore, the train approaching from behind on the same track received full, unrestricted speed commands. The operator of that train, Ms. McMillan, saw the stationary train in front of her and hit the emergency brake (or “mushroom”, named for its mushroom shape) but it was far too late. Her train slammed into the rear of the stationary train causing catastrophic damage to the first two cars. She was presumably killed instantly.

Early warning:

Prior to the accident, ATC technicians had been instructed to install new ‘impedance bonds’ (the equipment that couples/transfers the train detection and speed command signals to and from the rails) on the tracks. The bonds they were told to install were made by a different mfr than the rest of the equipment in the TCR. More than one of the techs questioned the wisdom of this but they were blown off, as per usual. They noticed that some track circuits weren’t adjusting properly and reported this but were told to “crank up the power” which is not a good idea from a safety standpoint. If the signal is too strong the circuit might not ever drop (show occupancy).

A particular circuit (the one the struck train was in) began picking and dropping – alternately indicating occupied then vacant – days before the accident. This was confirmed by the system log (sys log) that was leaked and published in a local paper.  A ‘work order’ was opened for a ‘bobbing track circuit’ (a circuit that is alternately indicating occupied then vacant – ‘train/no train/train/no train’). It was that same circuit. A bobbing circuit would be obvious on the huge display at Central Control (OCC) downtown at the Jackson Graham Building (JGB). In fact, that’s almost certainly why the work order was opened! After the accident, it came out that there was a letter issued by the mfr of the bonds clearly warning against mixing them with equipment from other mfrs. At the NTSB hearing the “engineer” claimed that he “never saw the letter” (although others in the industry had). WMATA claimed that they “never knew” about the track circuit problems before the accident when there was a clear paper trail showing they absolutely did know. Was anyone fired or demoted due to the accident and subsequent blatant lies?  Was anyone prosecuted for manslaughter?  Hell no – it’s Metro! “Mistakes were made. Heads will roll. We’re on top of this…”. OK, back to sleep now…

And in a nutshell, THAT’s what’s wrong with Metro. There is absolutely ZERO accountability. Whoever was down at OCC when my two coworkers were killed this past January is guilty of voluntary manslaughter but will they face any consequences? No way. Can my coworkers’ widows sue them? Sorry, WC means that not only can they not sue WMATA, they can’t sue any Metro employee acting in his/her ‘official capacity’. “Sorry about your dead husband/wife; sister/brother; mother/father. Go away now, you’re bringing me down. I have to get back to killing people…”.

The “engineer” responsible for instructing us to install the mismatched equipment? Not only is he not an engineer, he doesn’t even have a high school diploma! He has a GED, that’s it. He and two other uncertified “engineers” are the subjects of a recent newspaper article which I linked to in a previous post. He retired after the accident but now works as a “safety consultant” for Metro!!

This is so outrageous it would almost be funny if people weren’t injured and killed.

Hopefully this accident will forever prove to be the nadir of the history of Metro.

This entry was posted in Inadequate Training, Management Follies, Safety Incidents, The Metro Hall of Shame and tagged , , , . Bookmark the permalink.

2 Responses to The June 22, 2009 Train Wreck

  1. Ted K. says:

    Cranking up the power is a no-no in other fields :
    1) As a data comm. tech. I’m well aware that the telephone company has a hard max on copper circuits;
    2) The FCC has published limits for equipment like citizen’s band radios; and
    3) Aircraft engines are de-rated for safety and service life reasons.

    By the way, that supervisor’s instruction to boost the power could be considered an FCC violation if it caused ANY interference in nearby radio gear. It’s against the law to jam somebody else’s radio channel (check the harmonics !).

  2. Metro doesn’t concern itself with petty details like FCC regulations — they don’t apply to the 51st State. 😉

    This accident should never have happened. Even though I wasn’t involved I feel somewhat responsible, because the primary cause was ATC equipment. There are always ‘what if’s’ after an accident, but I can’t help thinking, what if we had pressed harder about why we were only verifying track circuits at the transmitter and not the receiver?

    The only thing worse than an accident in which people are hurt and/or killed is an avoidable accident.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s