Hi all. Most of us are probably aware that real diesel-electric locomotives, at least the transition era ones, control their speed via eight notches of power. As I understand it, basically the notch corresponds to how much horsepower the prime mover is sending to the traction motors. I expect the details may be more complicated.
So my question is, why don't decoders allow you to control your model train in this manner? It seems like it should be easy to implement. Simply convert the eight throttle settings into eight corresponding wattages to send the motor, and let physics do the rest. Say notch 4 gives your engine 1.5 watts. It might be going a decent speed on the flats, then pulling cars up a grade it'll slow down and you'll have to notch up, just like the real thing. The manufacturer could have set a default notch to watts conversion, and allow the user to adjust it if wanted.
I googled this a bit to see if some decoders already support this. The closest I got was "manual notching" that seemed to only manually adjust the sound of the prime mover. The idea being, you're going up a grade, you can put the prime mover at notch 8, and dial down the speed step to what your best guess of train behavior with be. My solution seems both easier and more prototypical.
Are there decoders that can do this, or does anyone see a reason why it's not feasible? If not I'm thinking I'll make my own, though it's just another thing on a long list of projects.
Yet another one of those things that comes up all the time. The answer is pretty simple - physics does not scale.
Back at the beginning of DCC, the original spec had only 14 speed steps, 28 was proposed, and a few held out for an option of 128. The 14 step people were always going on about how real locs only have 8 notches, and 14 is almost double.
Well - physics doesn't scale. Momentum is what keeps a real train moving. A finer control than 8 notches isn't really needed (but there were some with 22 notch controllers, and Baldwin switches had an effectively infinitely variable air throttle, but those are special cases). When the engineer gets his train up to speed, and backs off the throttle, the train doesn't really slow down. Not for a long time. Unless it's going upgrade. By varying the throttle through just a limited number of sections, the engineer can keep the train moving at whatever speed is dictated by rule. I won't say it's 'easy' - it's a lot harder than it looks, just try one of the train simulator programs, and then realize even those are not 100% realistic. It's nothing like a car where you have near instantaneous control over speed.
The other half of the equation is brakes. You can certainly do the acceleration in 8 steps using momentum, and maintain the train's progress with momentum, but you also need momentum to keep the train moving, or even gain speed, with the throttle reduced. ANd to actually stop, you need brakes to actually actively slow you down.
Most of the newest DCC Sound decoders have these things. Switching between them all with buttons to activate features and functions isn't the easiest thing int he world. However, there is a solution - check out the Proto Throttle from Iowa Scaled Engineering. Watch some videos of people running locos with it. It works (and even looks) like the real thing. 8 speed notches, brakes, etc.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
Ah,I see what you mean about the momentum problem, thanks
I know of at least one very prominent model railroader (I'm sure there are others) that are currently using the 14 speed steps in their decoders / command stations. Only the first 8 steps are actually used and steps 9 through 14 are set to the same as 8.
The trick is that the acceleration / deceleration (momentum) are both set very high (if not maxed). Thereby the speed transition between steps is very smooth.
I've tried this out of curiosity, and it IS very effective. However, like operating a real train, it really takes a lot of getting used to. There are NO quick starts and stops by no means. You really have to develop a feel for the control and really plan out your starts and stops. An added bonus - and almost a necessity - is that you will really need to learn how to operate that brake feature as well.
I did get used to it and knew precisely how much time / space I needed for what I was doing, but there was no way possible I was going to hand it off to any of my regular operators and expect them to run it ! But then .... you don't give a greenhorn the "keys" to a real train and expect him to get from A to B without destroying something either !
If you're bored some time, set one up and give it a try. Give it more than a few minutes though. You might find it rather interesting.
Mark.
¡ uʍop ǝpısdn sı ǝɹnʇɐuƃıs ʎɯ 'dlǝɥ
speedybeeAs I understand it, basically the notch corresponds to how much horsepower the prime mover is sending to the traction motors. I expect the details may be more complicated.
Those pesky details —
Besides the eight notches common to many diesels there was transition to consider as well.
Once momentum got the train rolling and the amperage fed to the traction motors in series dropped off the engineer would reduce the throttle, switch the transition lever to series/parallel (later engines had automatic forward transition and manual downward transition) then to parallel shunt. So the engine RPM didn't always exactly correspond to the running speed.
The ESU "Full Throttle" feature allows you to select either the engine RPM or the motor RPM independently. Other decoders use "notch up/down" and some rely on BEMF to mimic load.
I do like the "Full Throttle" feature but it can get bothersome after a while.
Regards, Ed
Not all engines had "8".
GG-1s had 22 notches, if I remember (it's been a LONG time since I ran one of those).
Seems to me that the very early GEs like U-25s had "half-notches" along with the "full" ones. I only recall being on one U-25, ever (hostling it around in Danbury yard).
The AEM-7's had a single "notch", just off idle. Beyond the first notch the throttle was stepless to maximum.
The HHP-8s and Acelas have completely notchless throttle levers.
Lastspikemike Prototype diesel electric use the electric power part of the drivetrain as a torque converter in an automatic car transmission would do. In fact, all those electrical gubbins are "just" the transmission, electric power only need just the traction motors. Diesel electric also carry a portable electrical generating power plant, just as a hybrid car does, sort of. There are Diesel railcars that have been built with mechanical gearboxes, like a bus on rails, but no mechanical gearbox can efficiently handle the massive torque delivered by a prototype locomotive. Torque delivered to the rails maybe as low as zero rpm (zero speed) there and as high as maximum rpm at the Diesel engine. Those steps are not speed steps but torque steps (equivalent to "power" at the Diesel engine output but not at the rails, the locomotive has to move for there to be power there.) Apart from the physics issues, which are not insurmountable, you can't model this with a direct geared drive. What a sound decoder can and does now do is model the sounds of the Diesel engine running at different rpm (and loudness!!) while the locomotive does not change speed, as per prototype. Acceleration and free running or deceleration effects can also be modelled by the sound board, the motor cannot. Model locomotives do not have variable ratio transmission.
Prototype diesel electric use the electric power part of the drivetrain as a torque converter in an automatic car transmission would do. In fact, all those electrical gubbins are "just" the transmission, electric power only need just the traction motors. Diesel electric also carry a portable electrical generating power plant, just as a hybrid car does, sort of. There are Diesel railcars that have been built with mechanical gearboxes, like a bus on rails, but no mechanical gearbox can efficiently handle the massive torque delivered by a prototype locomotive.
Torque delivered to the rails maybe as low as zero rpm (zero speed) there and as high as maximum rpm at the Diesel engine. Those steps are not speed steps but torque steps (equivalent to "power" at the Diesel engine output but not at the rails, the locomotive has to move for there to be power there.)
Apart from the physics issues, which are not insurmountable, you can't model this with a direct geared drive.
What a sound decoder can and does now do is model the sounds of the Diesel engine running at different rpm (and loudness!!) while the locomotive does not change speed, as per prototype. Acceleration and free running or deceleration effects can also be modelled by the sound board, the motor cannot.
Model locomotives do not have variable ratio transmission.
Maybe just me .... I've read this three times and still don't understand what you're driving at.
Not just you.
Disclaimer: This post may contain humor, sarcasm, and/or flatulence.
Michael Mornard
Bringing the North Woods to South Dakota!
decoders are given a speed value and provide a proportional voltage to the motor. momentum determines the time between in/decreasing the speed one step at a time. but for constant HP, acceleration varies with speed.
speedybeeSimply convert the eight throttle settings into eight corresponding wattages to send the motor, and let physics do the rest.
the relative power of an electric motor in a model is vastly greater than a prototypical locomotive, such that it can accelerate a model train much quicker
but even the small processor in a throttle can calculate the drawbar force and resulting acceleration and speed given the tonnage (# cars) of the train (~50 ton/car) knowing some simple frictional forces. the throttle would give the speed to the decoder.
imagine the difference in performance between pulling the tonnage of 10+ fully loaded cars and then just pulling one car to switch in an industry
since the knob or lever on the throttle doesn't dictate speed, a separate braking control is necessary to slow/stop the train. most modelers are unfamiliar with westinghouse air brakes which can't be partially reduced like car brakes. so both increasing and decreasing the speed of a train would be different from what most modelers are use to.
the ProtoThrottle has an unused tonnage setting that suggests they were thinking of doing this. my understanding is many ProtoThrottle enthusiasts are fond of the need to use brakes.
437
greg - Philadelphia & Reading / Reading
LastspikemikeApart from the physics issues, which are not insurmountable, you can't model this with a direct geared drive.
The complication is in simulating the power output of an engine with, say, an 8-notch Woodward governor commanding crankshaft output speeds, directed through a generator/alternator with variable excitation, with DC traction power being periodically switched series/parallel or field weakening being applied to motors. This is a complex thing but can certainly be simulated; the question is not 'if' but 'how much extra will it cost' and 'will model railroaders want it'. We had a recent thread on this in which the preponderance of 'vocal' commenters did not like the idea much, either in general or in how one manufacturer, ESU, appears to be directing their product evolution.
AC locomotives of course control their transmission much differently, and their motor output characteristics are very different from DC designs. But there, too, the motors are permanently geared to the axles, so emulating the torque commanded to the motor over time could duplicate the 'real' locomotive's performance well enough on a model.
LastspikemikeThe only prototypical reason for diesel electric drives is the advantages of the portability of the electrical power plant. Otherwise it makes no sense. You would never need that in a model.
First, in building a model of a prototype, it is certainly not inappropriate to model its prototype behavior. And that is what the very complex PWM synthesis of the behavior over time of the 'rest' of the diesel-electric powertrain accomplishes ... for those who appreciate the fidelity.
Yes, the waveform that is sent to the motors will be complicated, and yes, many parts of it will be deemed unneeded or even undesired by many model railroaders. Likewise the single electrical signal out of a high-end hi-fi system contains far more information about its prototype 'sound' than, say, a MIDI electric piano reproducing the musical score would. The result at the 'output device' is not constrained by the relative mechanical simplicity of the power transducer used in either case ... and to the extent the result is unpropirtional 'as a result' we can apply the equivalent of Dolby correction or wave shaping.
Second, there are in fact well-known examples of motor-electric drive in larger scales, although none with the pretense of miniaturizing a 'scale' 645 or traction alternator....
LastspikemikeFor a model railroad model it makes no sense to model the prototype diesel electric drivetrain. It can be modelled much more easily by mere representation.
The discussion is about the other 'half' of the "representation" which is allowing model controls to be made and manipulated in 'prototypical' ways that produce the 'prototypical' simulated behavior in the model.
As noted, not all modelers will find this attractive, or interesting, or necessary for anyone to do ... all that is in previous comments. On the other hand if you were building a training simulator for real railroaders, this would provide it... or any more 'forgiving' adaptation for early training or just 'more prototypical' fun.
Well, there are the Kato P42's with axle mounted coreless motors in HO, simulating traction motors.
But no internal cumbustion prime mover turning a generator inside.
Hornby did live steam in HO, but it's not exactly practical when you get down to that size. The most common diesels in larger ride-on scales seem to be gas-hydraulic, mainly because appropriate hydraulic pumps and motors are pretty easy to source. Ones that size with electric motors tend to be battery powered, no engine of any sort.
rrinkerWell, there are the Kato P42's with axle mounted coreless motors in HO, simulating traction motors.
Plus the Rapido ex-New Haven FL9s where the prime mover is shut down and the traction motors are powered off the DC third rail. About as close to model railroad simulation you can get.
IMG_6820_fix by Edmund, on Flickr
Cheers, Ed
That's just simulated in the sound decoder though, those locos have a typical electric motor and gears driveline. I don't think the 3rd rail pickups on the Rapido models even actually work.
LastspikemikeMy opinions on this subject are influenced by my perception of what this hobby is about.
what's being modeled?
if locomotive, is it mechanism or behavior?
LastspikemikeSo, yes, I bought a used Mantua 2-6-6-2 logger that is deliberately made to be out of scale and really doesn't run that well, never did even from new apparently. I had at that time no power nor track to run it on. If you understand why I would do such a crazy thing then each of your questions will be answered.
Do you use it for display or operation? Do you build a module or layout around it ... or fit it in to something else? Do you rebuild it for better scale, or superdetail it, or weather it with care and skill? Do you rebuild the drivetrain for smoothness, or remotor for 'moar power'?
And all the other things about structures to go with it, tracklaying connected with it, follow-on purchases or products ... and, above all, the joys (and other emotions) that go along with it. (We already agree it's been stipulated to be an 'obsession')
The potential fun in this specific context is whether you want to model how a steam engine 'drives' or not. If you think eight-notch plus physics is counterintuitive (as is three-notch DC MU electric control to an even worse extent, btw) there is a voyage of discovery just raising anchor: throttle, reverse, drifting methods, use of wire drawing for slip control... all of which can be 'emulated' in a device gear-driven by a variable-speed electric motor but only with great effort and care...
LastspikemikeWe speak at cross purposes.
Architectural models serve the same purpose as design models or stylists' clays: they are a part of a good design process, not an end result to show off craftsmanship or 'wow' clients. Architecture crits often contained the dread words "there was no model" (meaning that designs that looked fine in the side and end elevations or in the axons just Did Not Compute in as-built 3D). The same applies in spades to many automobile designs and details... and yes, this includes steam and other locomotive streamlining. But very seldom do we find this in actual model railroading...
The word "model" has several denotations and even more connotations.
I confess I like to hear your opinion on the invaluability of railways. There's a poster with the handle 'ttrraaffiicc' who occasionally posts over on the Trains Magazine forum; I'm sorely tempted to sic you on him. Read some of his posts and you'll see why...
I do not think an 8 step throttle would work in the model world.
Just as an example: A GP-9 will act much differently in notch 2 with one empty hopper car than it will in notch 2 with six loaded boxcars and a caboose.
There is just no reasonable way to simulate all the real world effect of hundreds of tons in miniature.
-Kevin
Living the dream.
SeeYou190I do not think an 8 step throttle would work in the model world.
why not if you have a tonnage input such as the ProtoThrottle has (unused) and use the control to determine the drawbar force (HP) and not simply as a speed setting. acceleration is proportional to the mass (tonnage) / drawbar force.
the loco would accelerate much differently w/o any car vs 5 cars
i'm curious if each notch is a proportional amount of horsepower: 12, 24, ... 88, 100%? how much horsepower do you want in notch 1 if you're just moving the loco?
my understanding is that prototype engineers are switching between two notches to maintain a relatively constant speed. wouldn't modelers interested in prototypical operation like that? select a notch to accelerate the train up to speed, possibly reduce it to zero and coast, then use the brake to slow and finally stop.
How can you scale down a real locomotive to perform like it does on the prototype to any of the scales used in model railroading? You can't. So, why are we trying so hard to do it?
Rich
Alton Junction
richhotrainHow can you scale down a real locomotive to perform like it does on the prototype to any of the scales used in model railroading?
don't understand why you think controlling the speed like a real loco is that hard? it's just math
of course you don't want a model pulling 100 cars to take as long as a real locomotive, but you can cheat a little so that it accelerates faster.
i think there are at least 3 things to understand for more realistic control
joe fugate has described more prototypical operation of a loco/train w/ sound as "micro operations"
the ProtoThrottle is selling to the market interested in more micro operation
gregc richhotrain How can you scale down a real locomotive to perform like it does on the prototype to any of the scales used in model railroading? don't understand why you think controlling the speed like a real loco is that hard? it's just math of course you don't want a model pulling 100 cars to take as long as a real locomotive, but you can cheat a little so that it accelerates faster.
richhotrain How can you scale down a real locomotive to perform like it does on the prototype to any of the scales used in model railroading?
gregc joe fugate has described more prototypical operation of a loco/train w/ sound as "micro operations" the ProtoThrottle is selling to the market interested in more micro operation
richhotrainbut you can't replicate.
why not? but i doubt modelers would want that, taking 10 mins to get up to speed. so you scale the mass computation
i've worked on code to determine the speed of a train based on the tonnage, HP, bearing friction and grade. not perfect, close enough. does a train stall/slip going up a grade? does it start rolling backwards?
it appears there are modelers interested in more accuarate behavior of a train. trains that more realistically accelerate (not constant), are limited in speed due to HP/tonnage/grade, slow down or accelerate on grades and require use of brakes to maintain (down a grade), slow and stop a train.
and they want sound that matches the situation. brake squeals, lots of noise accelerating a lot of tonnage and less noise when pulling no cars.
Hence the Proto Throttle. Even if it were half the price, I can't see most people replacing all their unrealistic knob throttles with them. But as seen in numerous videos of them in use, for the people that do enjoy it, the Proto Throttle gives them that realistic feel.
Now they just have to make one modeled after an Alco control stand - running an RS3 with an EMD control stand just feels wrong.
(no, no pictures of those abominations of Alcos converted to EMD prime movers)
rrinker Hence the Proto Throttle. Even if it were half the price, I can't see most people replacing all their unrealistic knob throttles with them. But as seen in numerous videos of them in use, for the people that do enjoy it, the Proto Throttle gives them that realistic feel.
I have said before... if I had to start all over I would have a G scale ISL shelf layout with one GP-9 and a proto throttle.
That would feel like railroading.
LastspikemikeThat 8 step throttle control on the prototype locomotive is not as simple as the fuel injection controller on a truck Diesel engine, although now even automotive diesels use quite complex software to deliver the desired torque characteristics.
Please, please, please...
You have made a lot of bad comparrisons lately about DCC, and the DCC people have tried to correct you.
PLEASE leave on-highway diesel engine controls and injection systems out of your under-informed statements. I know you are liking to think you are an expert on everything, but now you are dipping your toe into the pool of my true field of expertise. I do not want to be trolled or pulled into a conversation where I try to explain to you how you are wrong like others have been.
Thank you.
LastspikemikeNow I don't know how modern diesel electric locomotives control torque output...
I do know just from standing beside moving locomotives that the 8 step throttle (which isn't technically a throttle just btw, diesels are not throttled) does not correlate to engine rpm nor locomotive speed.
The electric transmission in these locomotives is infinitely flexible which is why it is used.
Hydraulic transmissions have been used, pump and motors correlating exactiy to their electric versions but not as durable, flexible or ultimately efficient as the electric versions.
All the electrics are just an infinitely-variable-ratio transmission, by definition a stepless transmission like a CVT now in automotive use
None of this is particularly arcane,and the 'why and wherefore' is not difficult to acquire, but you have to learn it 'as it is' and not pretend it's the way you want it to be from assumptions.
Of course, ultimately we reach the point where we need to ask "Do I want a control system that's so realistic that only a trained locomotive engineer can use it?"There is such a thing as too much realism.