Why do diesel-electric locomotives still use "notch" load control?

And you use the GPS on your cell phone when you’re pulling into a parking space? Just to make it challenging, you turn off your headlights at night.

There are times when a Mk I eyeball, a lightbulb, and experience are better than tech.

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.

Give it a think. Please.

Yes, I’m a road engineer at a major Class 1 U.S. railroad. I’m also a long time flight simmer and watch a lot of the pilot vlogs on youtube and have also noticed many similarities.

My company actually used the sterile cockpit concept, and told us it was a safety measure they adopted directly from the airlines when the first rolled it out. We call our version sterile cab.

Maybe I will start a thread.

In aviation, a cockpit is “sterile” during critical phases of flight. What are the critical phases of train operation?

For us its things like after passing a signal that requires us to stop the train at the next signal, going into a work zone where work is being done on the tracks, temporary speed restrictions, things like that.

In Japan Point and Call has been adopted where the driver points to the critical item – wayside signal, speedometer, etc. – and calls aloud its aspect. Wiki says it has been adopted by North American transit operators, notably NYC and Totonto.

https://en.wikipedia.org/wiki/Pointing_and_calling
Pointing and Calling Japanese Safety Standard at Railway Companies & Toyota (HD) - YouTube

Equally important, backward compatibility with the existing staff of locomotive engineers.

There’s roughly twice as many of them as locomotives, and they all were trained on the eight-notch system, and have many years of experience working with it. What benefit would there be in changing it?

The benefit would be infinitely variable throttle control - although, as I’m learning from this thread, that benefit is not big enough to justify the backward compatibility issues that would ensue. But as far as locomotive crews are concerned, it’s hard to imagine that much (any?) training would be required, as the throttle lever that I envision would exhibit the same behavior as on a boat, scooter, or snowmobile.

Except that in alot of situations, but particularly when starting a heavy train from a standing start, you have to be really careful with how much tractive effort (T.E.) you apply. The limits are on the cars in the train not the engines. With an infinitely variable throttle you could easily exceed these T.E. limits if you moved the throttle just a bit too far.

A notch however, will give you far better control of the in train forces. For example, say I have a 15000 ton train that is 10000 feet long, which is really nothing for U.S. railroads today. If I go to notch 1, I know that all my engines will load around 13000 pounds of T.E. I select N1 and slowly stretch the slack in the train out. Some times that won’t be enough to get the train moving, so I’ll go to N2. Now I know I’ll be around 26000 pounds of T.E. I’ll remind you that number is for each locomotive. A train like this would take 3 on my territory so 78k pounds of T.E. total, in that notch.

With an infinitely variable throttle, I could easily spike that up to 100k pounds total while I’m looking for the right setting, but if I do that I rip the train apart. Then all the bosses come out and I get in trouble.

I could maybe see the infinitely variable throttle on passenger trains, but it would just introduce to many variables on freight trains.

And, as far as I can tell, Machine_Elf has yet to adequately explain the advantages infinitely variable T E would have over eight discrete steps. “Not enough oomph? I’ll try twice as much,” as opposed to, “Not enough oomph? I’ll try 1.3452 times more.”

My perspective is that of a car/motorcycle operator, where infinitely variable load control is pretty much indispensable. For trains, having solid-state infinitely variable load control would enable cruise control, for example, as well as the potential for remote control. From my perspective as someone who has never operated a locomotive, these seemed to me like they might be useful features to have.

From my perspective as someone who has worked with control systems, and never been involved in the rail industry, the rail industry was a leader in designing systems to prevent cruise control and to prevent remote control.

People had horses which, when well trained, could navigate and operate without direct supervision. Then they had trains, which, when not constantly supervised, ran into people, crashed, ran off the rails – at a speed and with consequences that were unprecedented.

FWIW, I was looking at high-speed, Australian, passenger, rail cars, which run in sets of 6. They’ve got notch control. So I guess it’s not just for very long freight trains.

In any control system you’re dealing with several layers of mathematical derivatives. You pick one in the chain to connect directly to the operator’s UI and let the ones both up- and down-stream in the math be the consequences.

Sometimes it makes sense to have speed be the thing controlled and other times it makes sense to control for torque or some other upstream variable.

Per the expert upthread, managing tractive effort is the variable to control (at least under acceleration and hill climbing, and presumably also under braking and descending). The alternative is broken trains. So you build your control system to make that the primary variable, and speed just happens.

In my world things are different in detail, but similar in effect. During climb we manage the engines to thermal and RPM limits. The amount of thrust that makes varies based on altitude and other environmental conditions. And how new/old those engines are. It is whatever it is, and we’re not trying to control that variable. Then we control speed by altering pitch. What pitch gets us what speed varies not only on engine output but on other environmental factors, our weight, how dirty the machine is, etc.

Going back to trains …

I used to ride this light rail system to work each day. It appeared to me the operator controlled acceleration, speed, and braking with a pair of pedals closely analogous to what we have in conventional cars. No notches there. Whether for reasons of operator training pulling from a pool of bus drivers or for the difference that each car is self-propelled so drawbar limitations are totally NA I don’t know. But bottom line is a continuously variable control system is used someplace sometimes on modern light rail systems.

My parents had a truck that I used to drive occasionally. It had an accelerator rather than a speed control.

In the car I normally drive, selecting a gear and pushing the pedal down determines how fast the car goes. On the flat, I can drive with my foot more or less continuously in one pedal position. Push down for faster, lift off for slower.

In the truck, the pedal controlled how much acceleration I got. Push down to accelerate faster, lift off to accelerate slower. No matter what speed I wanted, the accelerator had to be approximately in the ‘neutral’ position. Any more than that, and the truck accelerated. Any less than that, and truck slowed down. When slightly accelerating, the truck didn’t simply accelerate to a new slightly higher speed: it had far too much power for that. A small difference from the ‘neutral’ position would cause it to slowly accelerate, and slowly accelerate, and slowly accelerate, and damn, that thing was hooting.

Simply having the same pedals and continuous control didn’t make it the same to drive.

Please do. One thing I’ve been puzzling over is, fifty years ago in the US about 75% of locomotives were EMD and 25% GE with a smattering of other manufacturers like Alco or Fairbanks Morse.

Nowadays the small guys are gone – not surprising because they were on their last legs even then – but the Class 1s (at least) are now running about 90% GE and EMDs are down to 10%.

I’m guessing it’s because EMD was slow in adopting AC and/or more sophisticated control systems, but don’t really know what caused the shift.

My understanding is that it started in the late '80’s with EMD’s 50/60 series offerings of the time requiring more down time for maintenance then GE offerings at the time.

Here’s the old thread from 2013 for some the topics covered. New stuff would be great. Ask a train engineer

The airline industry went through a similar turnover in powerplant manufacturers. The leading light of the early 1960s stumbled badly on reliability in the late 70s/ early 1980s and is just now beginning to recover market share after a 30+ year spell in the lonely cold profit-free wilderness.

Transportation companies can abide a lot of mistakes from their suppliers. But not unreliability. That’s the death knell for a supplier.

Over at an australian train forum, its explained why there are notches for the first 5 notch positions, at least for the NSW electric passenger trains… its actually a whole set of solenoids, not just 3 power levels solenoids (like 4, 2 , 1 solenoids,) It sort of is, but I gather its not the only control. eg where it says that the resistance is “high”, they also have resistance control right beside the notch control ?

I think if a notch wasn’t working ,they would skip it, and make the train jerk along as it gained speed, as it was too much force…traction was too high … floored it.

But its also good to lock the control in place, so its not like the foot pedal on a car ?