[N&W] Re: Water Buffalo?

nw-mailing-list at nwhs.org nw-mailing-list at nwhs.org
Mon May 31 15:32:56 EDT 2004


Kinks and Bounces:

Just speaking as a dumb, old mechanical engineer....

Assume that the loco is a 2-8-8-2.  Assume the drivers are all 58 inches in 
diameter except for one which is 58.5 inches.  Assume this 58.5 inch driver 
is the third driver on the second motor.  Now, start the locomotive from a 
dead stop, get it up to speed where the "driver bounce" starts and run a 
high speed movie of the event with the camera on the ground.

High speed cameras use 100 ft of film or 400 feet of film rolls.  This I 
know because we use one at work for high speed filming of pneumatic 
nailers.  The camera should use the 400 footer because it takes most 
(75%)of the roll to get the camera up to constant speed. This gives about 1 
second and 3 seconds of filming time in a high speed camera before the film 
is used up.

The issue is how do you START the camera before the loco and driver of 
interest gets to the lens viewing area, get the camera up to constant speed 
and film the "bounce" as it happens.  Question:  As the driver goes past 
the lens is the counterbalance going up or down?

Clearly there is a H--- of a lot of coordination necessary to set up the 
correct sequence of events and get it filmed in the TINY "window of 
opportunity" to get the event you want to see.  Work the math and you can 
see that tripping the system needs microsecond accuracy to capture the 
event.  The loco and the drivers have to be prepositioned to fall into view 
correctly.  Any slippage ruins the set up entirely.  As an engineer , I 
hope the other guy in the office has to do this!

So, lets mount the camera to the locomotive and look at the driver of 
interest.  Does camera shake blur the scene and make the results 
unreadable?  Bet so!

Another question.  What does a rail kink look like?  It is a sharply 
defined bend with a dimension roughly 1/8" across the distortion or is it a 
gradual curvature?

Remember a 58" driver has a perimeter of Pi times D= 182.2124" and the 
58.5" diameter has a perimeter of 182.7832".  Slightly more than 1/2" 
different.

Consider that vertically BENDING the rail permanently causes two different 
events to occur.  First the rail is pushed down into a bow elastically, 
meaning when the rail will return to its original position after the load 
is removed.  Then as the "over" load pushes the rail past its elastic limit 
into the plastic deformation range the rail is permanently deformed.  But, 
some of the deflection is elastic and some plastic.  This means that once 
the load is removed, the rail comes back to some of is original position 
the "some" being the deflection in the elastic range.  This indicates that 
the bend is gradual.

The comment was made it would be simple to put a tape measure across the 
kinks and measure the distance between the kinks.  I contend that 
determining where the middle of the kink is is far, far harder to determine 
than you think.

Observe a diesel locomotive going past you on the rails.  The axle load 
"squishes' the rail down in to the tie and ballast about 1-2" inches.  The 
rail curves around the wheel (40"diameter) giving a contact patch roughly 
3-4" long.  It isn't point contact!

A permanent bend probably is a dish or bow about two feet from tangency 
point to tangency point with , say, a 1/4" to 1/2" deflection in the 
center.  Now lets lay a straight edge across this bow and pick out the 
point of maximum deflection.  Kinda tough to determine, isn't it?  Remember 
the rail head is curved.

Consider that the perimeter difference in our example is only 1/2" .  I 
suggest that the probable error in locating the center of the bow is about 
1/2" to 1".  Remember the curve is very gradual and you cannot pick out 
where it starts or ends with a high level of confidence.

So, which driver is causing the rail kink?  Don't forget this was a 2-8-8-2 
and we had to run eight drivers over the point of interest.

Determining driver "bounce" and the causes of rail kinking is a very, very 
tough engineering and measurement problem.  Be cautious about speculation 
and over simplification.

Gary Rolih, Cincinnati
_____________________________________________________________

Mark Lindsey wrote:
 >Sorry Ed, i don't buy it.  no two drivers were the same diameter.  even on
 >the same engine.

Ed replies:

Mark - on what do you base this?  Do you have some measurements from a
locomotive - or preferably from many locomotives - that verify this theory?
Or is this just another swag based on the questionable theory that no two
manufactured objects will be exactly the same, no matter how stringent the
manufacturing standards?  Everything was manufactured to certain standard
tolerances.  Driver diameters were nominal.  The J that ran 110 MPH had 68
1/2" drivers, reportedly.  But it behooved a railroad interested in having
the best possible performance from a locomotive to see that the diameters on
a single locomotive, or either  engine of an articulated, were as uniform as
possible.

 >So maybe measure 10 kinks in a row (increasing accuracy of
 >the procedure) instead of one kink to the next.

Go ahead, if you want to.  It still won't prove your point.

 >take a Y5 front end and put it on a y3.  do they have the
 >same diameters front to back??

Nominally, yes.  Practically, probably not.  But what does that have to do
with your point?

 > Hooters only has the J with arrow on #3's counterweight.
 >I saw several on the Pocohontas Glory series.

Hooters does NOT show an arrow on the #3 counterweight.  You must be looking
at the wrong thing, or the wrong driver.  I looked at the tape several times
and the one arrow on a J was on the main.

 >I have a friend at EG&G who is trying to get me info on the loco films that
 >Edgerton did with high speed cameras.  Maybe we can get a copy of "N&W
 >locos bouncing" for the collection.
 >

Ed King




More information about the NW-Mailing-List mailing list