Early diesel-electric trains employed a load control (“throttle”) lever that featured eight discrete forward settings, called “notches” (in addition to another position for idle, and still another position for dynamic braking). Why eight settings? Because the system employed three solenoids that could combine in eight different ways to change the settings of both the engine’s electromechanical governor (to adjust fuel quantity) and the main generator (to change the RPM/voltage/current relationships). There were a few weirdo manufacturers over the years who used four solenoids to achieve 16-notch control, but the rest of the industry settled on 3 solenoids/8 notches, which was found to provide adequate control resolution for something that accelerated/decelerated so slowly.
This eight-notch throttle lever scheme persists today, which surprises me. I would assume the engine on a modern locomotive uses electronic injection control, and also electronic/solid-state motor control, which would enable infinitely variable load control. So why have “notches” in the throttle control at all on a late-model locomotive? Why not a continuously-variable lever with a display that reads “% load”?
The same reason your landline telephone uses at its core, technology from the 1880s: Backwards compatibility. On the North American continent a train can have one, two, or several locomotives at the head end depending on the tonnage of the train and the grades it will run over. Long ago the American Association of Railroads standardized the way to run multiple units (MU) using a 27-pin connector for – among other things – the throttle setting, and several pneumatic connectors for the brake settings.
Eight throttle settings suffice for what the locomotive needs to do an nobody’s seen fit to change a system that’s been in place for quite some time. All modern locomotives from both manufacturers left (GE and EMD) are compatible with each other and anything made since the mid-1950s or so.
It’s done so that if you have multiple engines, you can easily set them all to the same “notch” and thus more easily balance the load. It also simplifies the wiring and signalling between engines so that one person can control multiple engines.
In the USA development of about all train technology came to a halt in the late 70’s. It was good enough for hauling loads of freight and passenger trains became a secondary industry kept afloat by the government. Many development and manufacturing facilities just shut down or maintained a skeleton crew for supporting service because it wasn’t profitable anymore. Meanwhile the rest of the world passed us by and our train system is the joke of the world that has gone high speed rail. We are a first world country with a third world rail system.
Excepting the Channel Tunnel* nowhere does freight move on high-speed rail – that’s passenger service only. One fundamental difference in railroads between Europe and the US is that in the former, the railroads were built to primarily move people with freight as an afterthought. Here it is the other way around. In the Northeast corridor, where population densities and densities are similar to Europe or Honshu we do have Acela with an average speed of 85-mph between Boston and DC, not up to TGV or InterCity speeds but without major renovations of the right-f-way, that’s the best we’ve got.
With the pandemic drop in traffic, all passenger service west of Chicago is down to tri-weekly. There just isn’t any way for an upgrade to high-speed service to pay for itself.
After WW2 air travel in America was cheaper and faster than trains. You could drive to a (fairly) local airfield, and catch a plane just as we catch trains in Europe.
Of course, 9/11 put a stop to that, but train travel will never be appealing where such vast distances are involved.
Freight, especially bulk freight, is a very different animal: it is not usually in a great hurry and it doesn’t mind being shaken around.
In the UK, canals were the answer to poor roads and slow horse-drawn transport that made shipping goods and raw materials risky and expensive. Railways put the canals out of business and it is only due to the tourist industry and volunteers that so many still survive. Motorways almost destroyed the rail network.
This happened in a country that is smaller than some US States, so it is hardly surprising that transport here developed differently.
We found a way to connect the cell phone network to the landline network. It seems like it would be trivial to have a locomotive with a variable output translate it to one of eight notches when it sends the signal to a notched unit. Or, the other way around, if the notched unit is in control, the variable slave unit would know to interpret each notch as a specific throttle position.
Now, I’ve never heard of any of this, but it doesn’t seem like an overly difficult task so I assume I’m missing something.
I anit broke don’t fix it is what you are missing. It adds nothing to the bottom line and adds cost. Its an old industry that just keeps chugging along while making a consistent profit as it is and has been. No one wants to retrofit the fleet without a measurable increase in profit margin.
And just to give an idea of the magnitude of any such refit of the fleet, the seven Class I railroads in the U.S. have a total of over 25,000 locomotives. All currently pretty much compatible with one another.
So the other engines would be going close to the same speed as the primary locomotive?
It’s as bad as it sounds: as you advanced the throttle on the lead engine, the other two would throttle back to compensate, then when you hit the magic threshold, they’d come in with a rush – and if that caused the driver to throttle back, they’d throttle back too. So the only way to drive it would be to learn where the notional notches are, and move the ‘infinitely variable’ throttle distinctly between notch positions.
OK, so the notch system persists to assure backward compatibility with the existing fleet.
How is the notch system implemented on late-model locomotives? Are they still literally using a trio of solenoids to change alternator settings and engine fueling quantity? Or is the throttle lever just telling a computer “notch X”, and the computer takes care of everything (including sending the corresponding 3-bit load signal to the rest of the locomotives?
Especially since the throttle does not directly control how much fuel goes to the prime mover in the first place but rather how much tractive effort you want at the rail. Circuitry in the locomotive(s) adjusts the fuel flow to the prime mover and field windings in the generator (the load on the prime mover) so that the motors driving the wheels have the right oomph. The fact that this oomph is in eight discrete steps is immaterial.
I have seen consists where the circuitry in one locomotive is defective and it is revving the prime mover up and down – and presumably changing the field strength – as it hunts for the desired tractive effort.
Locomotives used in switching service have a light right under the cab that shines onto the ground. This is because at switching speeds – 3 to 5 mph – the speedometer is useless and you have to see how fast you’re going. How fast the prime mover is turning over is useless because there’s no direct correlation between that and your speed and you can’t see that in the dark.
People have the impression driving a locomotive is like driving a really big truck. It is not. You will notice, for example, I have said nothing about transitioning.
Drifting a little off topic, my understanding is that road locomotives (as opposed to switchers) had ground lights to be able to see motion at night. When you’re starting a train, you get a few seconds for the train to actually start to move. Too long a delay risks overheating the motors. With a couple thousand horsepower worth of diesel engine pounding away just behind you, it may be hard to sense motion at night. Thus, ground lights. Or so it was explained to me by a friend, an engineer back in the '70s and '80s. I’ve also read (sorry, no cite) that modern locomotives have Doppler radar focused on the ground for the same purpose, detecting motion.
Is there a problem with notched controllers that’s solved by using variable speed controllers and doesn’t add more problems? Maybe eight notches is enough to cover the different load cases out there, so why change if it’ll break compatibility and not add any meaningful benefit?
This was kind of true in the past with DC traction motors. AC tractions motors can be fully loaded at 0 MPH without overheating. We use the ground lights on road power to make sure we are moving and not spinning the wheels. That will melt the rail.
Just to make it clear, all the diesel engine does is generate electricity. The actual motors that move the locomotive are electric. They could just as easily be powered by overhead wire.
@split_p_j: If you’re a railroad engineer or conductor, especially of long haul road locos rather than just yard switchers or city light rail transit machines, AND you’re interested, I think it’d be really cool if you created a thread in IMHO or MPSIMS along the lines of “Ask the locomotive engineer”. It’s been awhile since we’ve had one of those “ask me …” threads from an esoteric career, and I for one would really enjoy learning more about that world as it appears to you. I’m a pro pilot and there are vast similarities in the problems we both solve but vast differences in how we do that.