I am grateful that he put an accelerometer on the test train, and plotted the results: it was very cool to see effect of the dashpot snubber.
Also, the video introduced me to the term "dashpot snubber", but now when I try to learn more seems like dashpots and snubbers are distinct mechanisms [1]? I guess he demoed a snubber?
The snubber only has resistance in one direction. There's a check valve that lets air flow easily in the other direction. An ordinary dashpot without a check valve provides the same damping in both directions.
Dashpots and snubbers are not used much as emergency devices. They are for repetitive use. When you desperately need to dissipate a lot of energy, but collisions are rare, expendable devices are usually used. The article shows a pile of sand. In the early days of BART, the control system was so bad that large piles of sand were put at the end of each line for years. Truck runaway areas on mountain grades are piles of sand. The barrels seen at the gore of freeway exits are filled with sand.
Another trick is to force a collision to deform metal. When US cars had a 5 MPH bumper standard, the energy absorbing device was often a metal pin forced through a too small hole. The front ends of cars are designed to crush without damaging the passenger compartment. That's why modern car accidents usually show a crushed front end, yet even the windshield is intact.
The buffer is the final stage of an emergency stop. Something else has to handle the slowdown phase. One of the more impressive solutions is how high speed trains stop. They have a set of permanent magnets above the rails, held clear of the rails in normal operation. In an emergency, they drop to just above the rails, where they act as eddy current brakes. The energy of the moving train goes into heating the rails, which can take a lot of heating, rather than brake shoes, which can burn up.
Braking force increases with speed. No wear. Unaffected by rain, snow, leaves, or loss of power. Roller coasters also use eddy current brakes.
But because eddy current braking force increases with speed, it's nearly zero at the end. You need something else to get to a complete stop.
> Dashpots and snubbers are not used much as emergency devices. They are for repetitive use. When you desperately need to dissipate a lot of energy, but collisions are rare, expendable devices are usually used. The article shows a pile of sand. In the early days of BART, the control system was so bad that large piles of sand were put at the end of each line for years.
Piles or sand are an example of expendable devices, I think? At least barrels of sand are.
Using an expendable device at the rail ends is an indication that the control system isn't that bad. If failure to stop at the rail end was more than rare, a reusable safety mechanism would be appropriate, and that would be an indictment of the control system.
Or if the piles of sand were frequently replaced (or repiled)
I think when I hear the terms, a "dashpot" is a specific mechanism and a "snubber" is its purpose. You can use dashpots for other purposes, and you can build snubbers out of things other than dashpots.
I never realized how complicated safety devices had to be, and the dashpot snubber makes a lot of sense. However, why don't most trains the combination of the hydraulics and sliding stops as safety mechanisms? Is it because of the maintenance costs or the other external factors like the rust he discussed or the limit to the stopping space?
Were there no dead man's handles on those trains? Or do they have them and the drivers fall asleep frequently and most were caught in time but these two fell asleep at the latest possible moment?
without situational awareness, there is no good default behavior for a simple dead man handle on a hundreds ton train.
These days though i wonder why we don't have [semi-]autonomous control there. At least for emergency stop when the end of rails is 100 feet away and your speed is 20mph - any new car have such features today.
At the very least, if political opposition is a problem, they could have a very loud noise played to the conductor if they're due to stop and haven't yet.
I don't see what unions would have against automated emergency stop systems. Workplace safety is a core issue of unions, so if anything, shouldn't they demand those systems?
Obviously it’s impossible to know unless you’re sitting in on union discussions, but there is a perverse incentive at play. If one fears automation taking their job, I could see resistance to allowing any automation “foot in the door” so to speak.
If anyone is downvoting because they think this is wrong (vs just being low content and largely unhelpful) I’m interested in what alternative explanations exist.
This tech has existed for many years in motor vehicles and works reasonably well in far more complex environments than literal rails. It doesn’t seem at all like a challenging technology problem.
The things that come to mind are: failures of politicians to enact requirements, negligent or greedy rail infrastructure and/or train ownership, and unions afraid of lost jobs.
Dead mans handle devices (driver vigilance units) were introduced on UK trains in the 1960s They also certainly exist on Norwegian locomotives, I've seen them on YouTube videos.
> without situational awareness, there is no good default behavior for a simple dead man handle on a hundreds ton train.
Depending on the train system, engines to idle, medium level braking is probably a reasonable response if there is no operator detected.
> These days though i wonder why we don't have [semi-]autonomous control there.
This is likely the right question. Closed systems have all sorts of mechanisms to prevent overruns or at least reduce speed. Typically something sticking up from the track that will engage emergency brakes on all the cars when the train passes a certain point; and they can be controlled so that the post goes down if the train is has clearance and is traveling at a safe speed.
I suspect NJ transit does not own all the rails their trains run on. A mandate for positive train control on rail lines keeps getting pushed back, but would likely have reduced the speed of this overrun.
> These days though i wonder why we don't have [semi-]autonomous control there.
The US does. That's what Positive Train Control does.[1]
For a long time, there was reluctance to apply automatic train stops to freight trains, because long freight trains don't do emergency stops well. Just slamming on the brakes on a heavy freight can cause derailments on some steep curves, and may cause wheels to develop a flat side. What PTC does is to display a graph of speed vs. distance ahead, and the driver must stay below the graph. The system will apply brakes if the current speed exceeds the speed limit line, but only as much brake as is required to get below the speed limit. Classic emergency stops forced a full stop, and then someone had to walk the length of the train and examine all the cars before a restart. With PTC, the speed graph will drop to zero at the end of a line, which prevents overruns. It took years, but PTC is now deployed on all US track.
There were generations of previous systems. Here's the New York Subway's classic control system, from General Railway Signal.[2] This is safe, but drastic - it stops trains by raising a metal arm at each signal, which, if raised, hits a valve lever which directly applies air brakes. There are timers, and if a train is going too fast, the brakes will be tripped by the trackside systems. At the ends of tracks, "too fast" is very slow.
I am grateful that he put an accelerometer on the test train, and plotted the results: it was very cool to see effect of the dashpot snubber.
Also, the video introduced me to the term "dashpot snubber", but now when I try to learn more seems like dashpots and snubbers are distinct mechanisms [1]? I guess he demoed a snubber?
[1] http://gotmotion.com/actuators-dashpot-snubber-airpot/
The snubber only has resistance in one direction. There's a check valve that lets air flow easily in the other direction. An ordinary dashpot without a check valve provides the same damping in both directions.
Dashpots and snubbers are not used much as emergency devices. They are for repetitive use. When you desperately need to dissipate a lot of energy, but collisions are rare, expendable devices are usually used. The article shows a pile of sand. In the early days of BART, the control system was so bad that large piles of sand were put at the end of each line for years. Truck runaway areas on mountain grades are piles of sand. The barrels seen at the gore of freeway exits are filled with sand.
Another trick is to force a collision to deform metal. When US cars had a 5 MPH bumper standard, the energy absorbing device was often a metal pin forced through a too small hole. The front ends of cars are designed to crush without damaging the passenger compartment. That's why modern car accidents usually show a crushed front end, yet even the windshield is intact.
The buffer is the final stage of an emergency stop. Something else has to handle the slowdown phase. One of the more impressive solutions is how high speed trains stop. They have a set of permanent magnets above the rails, held clear of the rails in normal operation. In an emergency, they drop to just above the rails, where they act as eddy current brakes. The energy of the moving train goes into heating the rails, which can take a lot of heating, rather than brake shoes, which can burn up. Braking force increases with speed. No wear. Unaffected by rain, snow, leaves, or loss of power. Roller coasters also use eddy current brakes.
But because eddy current braking force increases with speed, it's nearly zero at the end. You need something else to get to a complete stop.
> Dashpots and snubbers are not used much as emergency devices. They are for repetitive use. When you desperately need to dissipate a lot of energy, but collisions are rare, expendable devices are usually used. The article shows a pile of sand. In the early days of BART, the control system was so bad that large piles of sand were put at the end of each line for years.
Piles or sand are an example of expendable devices, I think? At least barrels of sand are.
Using an expendable device at the rail ends is an indication that the control system isn't that bad. If failure to stop at the rail end was more than rare, a reusable safety mechanism would be appropriate, and that would be an indictment of the control system.
Or if the piles of sand were frequently replaced (or repiled)
I think when I hear the terms, a "dashpot" is a specific mechanism and a "snubber" is its purpose. You can use dashpots for other purposes, and you can build snubbers out of things other than dashpots.
I never realized how complicated safety devices had to be, and the dashpot snubber makes a lot of sense. However, why don't most trains the combination of the hydraulics and sliding stops as safety mechanisms? Is it because of the maintenance costs or the other external factors like the rust he discussed or the limit to the stopping space?
Were there no dead man's handles on those trains? Or do they have them and the drivers fall asleep frequently and most were caught in time but these two fell asleep at the latest possible moment?
without situational awareness, there is no good default behavior for a simple dead man handle on a hundreds ton train.
These days though i wonder why we don't have [semi-]autonomous control there. At least for emergency stop when the end of rails is 100 feet away and your speed is 20mph - any new car have such features today.
At the very least, if political opposition is a problem, they could have a very loud noise played to the conductor if they're due to stop and haven't yet.
Unions.
I don't see what unions would have against automated emergency stop systems. Workplace safety is a core issue of unions, so if anything, shouldn't they demand those systems?
Obviously it’s impossible to know unless you’re sitting in on union discussions, but there is a perverse incentive at play. If one fears automation taking their job, I could see resistance to allowing any automation “foot in the door” so to speak.
If anyone is downvoting because they think this is wrong (vs just being low content and largely unhelpful) I’m interested in what alternative explanations exist.
This tech has existed for many years in motor vehicles and works reasonably well in far more complex environments than literal rails. It doesn’t seem at all like a challenging technology problem.
The things that come to mind are: failures of politicians to enact requirements, negligent or greedy rail infrastructure and/or train ownership, and unions afraid of lost jobs.
Dead mans handle devices (driver vigilance units) were introduced on UK trains in the 1960s They also certainly exist on Norwegian locomotives, I've seen them on YouTube videos.
They bring the train to a stop.
> without situational awareness, there is no good default behavior for a simple dead man handle on a hundreds ton train.
Depending on the train system, engines to idle, medium level braking is probably a reasonable response if there is no operator detected.
> These days though i wonder why we don't have [semi-]autonomous control there.
This is likely the right question. Closed systems have all sorts of mechanisms to prevent overruns or at least reduce speed. Typically something sticking up from the track that will engage emergency brakes on all the cars when the train passes a certain point; and they can be controlled so that the post goes down if the train is has clearance and is traveling at a safe speed.
I suspect NJ transit does not own all the rails their trains run on. A mandate for positive train control on rail lines keeps getting pushed back, but would likely have reduced the speed of this overrun.
> These days though i wonder why we don't have [semi-]autonomous control there.
The US does. That's what Positive Train Control does.[1]
For a long time, there was reluctance to apply automatic train stops to freight trains, because long freight trains don't do emergency stops well. Just slamming on the brakes on a heavy freight can cause derailments on some steep curves, and may cause wheels to develop a flat side. What PTC does is to display a graph of speed vs. distance ahead, and the driver must stay below the graph. The system will apply brakes if the current speed exceeds the speed limit line, but only as much brake as is required to get below the speed limit. Classic emergency stops forced a full stop, and then someone had to walk the length of the train and examine all the cars before a restart. With PTC, the speed graph will drop to zero at the end of a line, which prevents overruns. It took years, but PTC is now deployed on all US track.
There were generations of previous systems. Here's the New York Subway's classic control system, from General Railway Signal.[2] This is safe, but drastic - it stops trains by raising a metal arm at each signal, which, if raised, hits a valve lever which directly applies air brakes. There are timers, and if a train is going too fast, the brakes will be tripped by the trackside systems. At the ends of tracks, "too fast" is very slow.
[1] https://railroads.dot.gov/research-development/program-areas...
[2] https://www.nycsubway.org/wiki/Subway_Signals:_A_Complete_Gu...
I have also seen a separate deadman device installed on some trains, it beeps, you push a button or the train stops.
I find it very strange that the example trains crashed at 21mph....was the deadman brake already slowing them down?
Because if the driver was decelerating deliberately I find it difficult to beehive that is when they fell asleep
I've never driven a train, but i have driven a car.