Dear AJ,

Sure, pit 1000s of horsepower against 1000s of tons, coupled thru a knuckle with a designed-in breaking strain rating of around 650,000 pounds, (http://en.m.wikipedia.org/wiki/Janney_coupler), and broken knuckles/pins/heads/shanks/draftgear are always a possibility...

NB that it's not strictly the sheer "trailing load weight" that breaks knuckles, but failure to manage the buffing force and difference-in-inertia between the locomotive and the load...
(Bunch all of the cumulative coupler slack in a mile-long train, then accelerate hard in Notch8, count the "bangs" as each coupler takes-up it's individual slack in turn, and count until one of them "let's go"...)

Those spare knuckles may not necessarily be used solely on the loco, they are just positioned such that they are handy for the conductor/brakeman when s/he has to start their walk of the train to locate exactly _which_ coupler(s) broke...

Amswering the "what happens?" Q, in _theory_
- the knuckle, pin, or shank lets go
- the train-brake airlines seperate as designed, or possibly get torn apart
- the westinghouse train brakes "go into emergency", which is visible to the engine crew as a train airline change-in-pressure, and their brakes being applied
- once both halves of the train gets stopped, the brakeman/conductor gets the task of lugging at least 1 spare knuckle (and possibly some extra brakeline hoses, or whatever else needs replacing) back along the train, as far back as needed...
(Weather notwithstanding, sun/rain/heat/snow/wind/whatever, that train's not going anywhere until the brakeman works it out...)

Suggest googling a series of essays by JD Santucci under the title "Hot times on the High Iron". He covers the down-n-dirty nature of such practical proto RR operations, inc having to diagnose, locate, and replace broken knuckles "on the road"

Happy Modelling,
Aim to Improve,
Prof Klyzlr

"How often" is kind of a relative thing. 

Knuckles fail two general ways.  First the knuckle is fatigued and cracks from repeated impacts, it wears out.  Then a "normal" strain can cause the weakened knuckle to fail.  You might hear railroaders ask the question "how much old break?"  In the fracture in the knuckle previous damage will be rusty and the final failure will the bright and shiny.  the other way is if there an excessive force in the train that exceeds the strength of the coupler.  That can be caused the engineer mishandling the slack in the train or something happening in the train causing an emergency application of the brakes, causing an excessive force. 

When a knuckle fails the train breaks into two pieces, both halves go into emergency and the train stops.  The crew then has to get a knuckle (from the engine, from the caboose or out of the last car in the train) and replaces the knuckle.  There are two types of knuckles, type E and F, that's why there are two knuckle holders, one for an E and another for an F.

Just out of curiosity, about how often will crews have to do this? I know there are tons of factors involved, but once a week/month/ year?

Steve

Prof, a quick correction:  The knuckle pin does not fail - or, more to the point, since the knuckle pin carries no draft force, the failure or loss of a knuckle pin does not affect the ability of the train to stay coupled.  I have seen many cars running with no knuckle pin at all.  The only problem is that when you uncouple one, the knuckle falls out.

Most knuckle failures are on the cars.  In my 15 years on the railroad, I have seen one locomotive knuckle break.  The cars tend to see worse abuse and less maintenance, and the locomotive knuckles are stronger than those on most cars.  In the case of that one failure I just mentioned, we had no spare locomotive knuckles.  We ended up taking the knuckle out of the front of the lead locomotive.

Also, about those brackets.  If you look at older locomotives, you will often see a box on the running board.  The ones on NS, at least old EMDs, are commonly sheet metal boxes with lots of holes in them, on the engineer's (fireman's side, for railroads that run their engines wrong way forward ) side running board just ahead of (or behind, if you're backwards) the blower duct for the long hood traction motor.  Many locomotives have tool boxes in the air compressor compartment instead.  Every locomotive is supposed to be supplied with two knuckles, a brake hose, and a wrench, hammer, and chisel.  Most also have MU hoses, glad hand gaskets, and sometimes a runaround hose for bypassing a defective train line.

Knuckles break, air hoses drag on crossings, stuff generally happens.  The train crew (mainly the conductor these days) is expected to make repairs if possible.  Having to call out the mechanical forces in some out of the way location is a time consuming and expensive proposition, so it's preferable to fix the problem yourself if possible.

Knuckles generally fail straight across the pin hole, where the metal is thinnest and stresses are concentrated.  I have also seen the draft gear retaining pin fail once, letting the entire coupler assembly simply fall out of the draft gear box.  Drawheads (the shank behind the coupler) have occasionally been known to fail, but I have heard it said that it is always a mechanical rather than train handling issue, since poor train handling should have broken the knuckle first.  I have even seen the entire draft gear assembly, box and all, ripped out of a car.  That was a mess to clean up!

There are four primary reasons that I have seen knuckles fail.  All are related in some way to the skill and actions of the engineer, though of course all can be exacerbated by prior fractures in a knuckle.

1. The most common problem is when locomotives start to slip while climbing a grade. and pulling hard.  When they regain traction, they can lunge forward suddenly.  Any knuckle close to the point of failure already can easily be broken by the sudden jerk.  The solution is to use no more power than is absolutely needed, and to carefully manage slipping.  I will frequently climb a grade with one hand on the throttle, one on the brake, instantly dropping a notch if I feel the wheels start to slip of the engines buck in any way.  I will also sometimes use a little engine brake to help control slipping.
2. The second most common problem is when coming out of a bottom with a poorly built train, especially if there are loaded tank cars toward the rear and empty cars on the head end.  What happens is that the loads settle down in the bottom and give a sudden tug as they are forced to start up the hill again.  It is even more of a problem because, when using dynamic braking, the slack is bunched up going down hill, and pulled out again going up hill.  Thus, at the bottom of the hill, the slack is transitioning from in to out, and the sudden change can be enough to break a knuckle.  The solution is to predict (based on length) when this will happen and reduce the throttle to compensate.  This was also done in the days of manned cabooses, to reduce the shock on the cab, and is known (at least locally) as cab notching.
3. Another case of poor train handling is when starting a train, not allowing the slack to adjust properly.  The worst case scenario is having stopped with dynamic brakes and having the train draped over a hill, with the slack in.  When the brakes are released, the rear portion begins rolling backwards down the hill, while the head end also runs down hill.  Pulling at the same time can induce enough of a shock to break a knuckle at the top of the hill.  The same thing can happen whenever the train is stopped with the slack bunched, especially if the engineer pulls too hard, too fast.  This is one of the reasons I always try to stop with the slack out whenever possible, and ALWAYS when I'm draped over a top.
4. Finally, it is possible to pull a knuckle apart with sheer brute force.  Four large six axle engines have enough draft force to break a brand new knuckle, which is the reason why we are restricted to three on freight trains on NS.  Loaded coal trains are allowed four, since coal hoppers have higher strength knuckles than regular cars.  If there is any prior damage to a knuckle, it is possible to break it with fewer engines.  I have had this happen a few times, and the only solution is to inspect and replace the knuckles prior to total failure.

Just out of curiosity, about how often will crews have to do this? I know there are tons of factors involved, but once a week/month/ year?

That depends a lot on the engineer.  When I was first starting, I would get a knuckle every few months.  Now (knock on wood!) it's been years since I've been responsible for one.  I did have one break about a year ago, but it was nothing I could have prevented, but rather an old break on the first heavy grade leaving the yard.

A given crew might get a few a year or less, but overall it's something that happens on our district about once or twice a month.  I have no idea how frequent it is system-wide.

Thanks Ken for my answer and the general overview too.  Learn something new everyday

Steve

Interesting stuff!!

John C

Engines and cars use the same knuckles, but there are two types, E and F, and they are not interchangeable.

Figure 1 knuckle per 50,000-100,000 train miles for a modern class 1 railroad with good maintenance and well trained engineers averaged over a large system.  So if your model railroad represents 100 miles of railroad, one train every 75 operated might get a knuckle.  If you operate 20 trains a monthly op session representing a day in the life of the railroad then you might get one knuckle every 4th op session or 3 per year.

Realizing that knuckles are not evenly distributed, they are mostly concentrated on trains going up a grade.  So some routes might have one a year and others one a week.  If you have a flat territory on your layout, one knuckle during the life of your layout is about right.  If you model a line climbing the Rockies, maybe one every other session or two.

When a train is unable to recover from an emergency brake application, the conductor (or brakeman if there is one) will walk the train to look for a separation in the brake line. If he finds a broken knuckle, he'll have the engineer drop the correct knuckle (E or F) on the ground and mark it with a burning fusee. The conductor will close the angle cock, hop on the train and ride up to the spare knuckle th engineer has left behind. The conductor picks up the spare knuckle, sets on the rear car and rides back to the rest of the train and replaces the knuckle. No need to lug that knuckle around, let the train do the work.

Greg Amer

The Industrial Lead

  This kind of stuff is AWESOME!  It would be nifty to simulate this on a layout that has grades, either with a random dice roll or defect detector.   Similar to the other thread about derailed rolling stock and getting it re-railed.

  Even adding the "drop a knuckle, pull forward, pick it up and back up to drop it off" maneuver.

  Other topics to think about are special loads (over length width, or weight).  Derailments with Hazmat cars in the train. Setout procedures for shifting loads, defective equipment.  Maybe rockfall/landslide detection?

I'm always happy to help and share my experiences (what railroad man isn't ready to talk about work?), but I don't know that it would make a good column.  The trouble is that there's only so much to share, unless you were to make it a sort of running diary of life on the railroad.

Here's one for you guys.. If you have a tonnage train descending a heavy grade, it's likely that you'll need air brakes to control the speed.  However, there is always a risk that one or more of the cars in the train will have a sticky brake valve, and as soon as the engineer applies the brakes the train will go into emergency.  Given that tonnage trains are usually long, that can mean a very long walk for the conductor to inspect the entire train.  Just recently, the rules were relaxed to allow a train to proceed without inspection in certain circumstances, but it's still a delay.  Also, after that happens, the track for the length of the train, from the location of the rear when the emergency application occurred to the head when the train stopped, will have a 1/2 authorized speed restriction until it can be inspected for damage and/or the cause of the problem.

A problem like that is most likely in one of two places - the first descending grade leaving a yard, or the heaviest grade on the line.  Once an engineer knows he has bad air, he'll do everything possible (including going a lot slower over the tops) to prevent having to use it again.