comments about using buttons to control speed instead of knob in the other thread got me wondering, how many speed steps are needed and what should they mean, speed or horsepower.
for the moment, let's assume there are 28 steps. I assume the PWM circuit in most decoders has a resolution of 1/255. Without momentum, i assume each of the 28 speed steps proportionally maps to one of the 255 PWM settings (e.g. speed step 7 is PWM setting 64). And incrementing a speed step w/o momentum results in a jump in speed (1 step changes the PWM setting by ~9).
but with momentum, the decoder changes the PWM setting, 1/255 step at a time, each step delayed by some amount of time proportional to the amount of momentum. With momentum, speed changes smoothly regardless of the number of speed steps but there are a limited number (28) speeds that can be maintained.
If the top speed of my loco happens to be a scale 28 MPH, each of the 28 speed steps corresponds to 1 MPH which seems finer than I need.
So 28 seems more than enough for me. And for NCE controllers, there's a button that in/decrements the speed step by 4, so that there are really only 7 steps, which for me still seems adequate. The ProtoThrottle only has eight.
But should it even be called a speed step. Maybe it should be horsepower (HP). the difference being accelleration w/ momentum. HP step 28 would cause the loco to accellerate faster that HP step 8. Operation would be to go to step 28, or whatever step achieves the desired amount of accelleration and then reduce it when the desired speed is attained.
but I can also see that speed/HP does not need to be proportional for each throttle setting. It makes sense that the lower steps are finer (e.g. 1 MPH) and the higher steps are coarser (5/10 MPH). I assume speed tables allow you to control this (but are more valuable for speed matching consisted locos).
So I see no problem with a small number of steps controlled by a button and wonder if it's not more prototypical.
greg - Philadelphia & Reading / Reading
I'm not sure exactly what your asking, but I use a 128 speed step. I use Digitrax. I can use the + button on my DT400 to advance the speed, or I can use the knob. With the 128 speed step, the first "click" on the knob or the button, my locos start to creep, and I mean creep, the display on the DT400, shows 1.
From watching many real trains at crossings, and listening to my scanner as the train passes a hot box detector, the speed steps, seem to be close to the actual MPH of the train. Speed step 30 seems real close to 30 MPH.
I have an Atlas, with the dual mode decoder, that was set at the 28 speed step, and click #1, the loco started at about 5 mph. Way to fast for switching. Once I did a reset, it's where it should be.
Mike.
My You Tube
Back when they were establishing the DCC standards, a common argument was that 14 steps should be plenty, after all (most) real locos only have 8. The counter of course is that physics does not scale. Momentum features in decoders have improved, along with now having braking that actually brakes against the throttle - and operating with the ProtoThrottle looks like a lot of fun, certainly the closest thing to what I actually saw riding in the cab as far as the effect of each control on the loco's sound and motion.
At the same time, at the time I viewed the video of Dana demonstrating it, most of the comments were of the sort that it is "too fiddly" or "a huge time sink, too complex". That tells me that a lot of people aren't interested in controlling their trains in a completely realistic manner, even those who otherwise do operate and not just run trains around in circles. Or the only people that was MRVP videos are the latter. I would love to get one - it does seem like a lot of fun, but at the same time I don't know that I would want to run that way ALL the time.
With 28 steps and momentum to smooth the bigger jumps in speed per step, you kind of are - unless you really do set all locos so that they run no more than 28mph at top speed. Otherwise, you will be constantly adjusting the throttle as one step is too fast and the other is too slow (or you don't care if the scale speed is exact and just leave it on on speed or the other).
128 steps gives a far finer control. Finer than a prototypical loco, yes, but the prototype has physics to help. Yes, there is momentum, but to really work, the momentum has to adjust dynamically. And not through the user pressing a button, that immediately takes the realism out of it. Plus operating a model is always a third person experience, you are looking down, or at least sideways, depending on the layout height, at the train. Never on it, feeling that push from the cars behind, or the drag of them when starting off. Judging only by visual. That's why I think people say they "get used to" momentum - it never changes, lone loco or 30 car train, and it's all simply stright line math, at a give momentum setting, from a specific speed step, the loco will stop exactly - THERE - every time. So sure, you get used to running trains with momentum turned on in the decoders. because you memorize the stopping distance.
Running a loco configured for the PT with a regular throttle I would think is less than ideal, You are turning a function on and off to activate the brake, constantly double tapping that key. At least with the PT there is an 'analog' brake control so you have an actual feel for how much of an applicatioon you are making.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
mbinsewiWith the 128 speed step, the first "click" on the knob or the button, my locos start to creep, and I mean creep,
if each speed step corresponds to 1 mph and there are 128 speed steps, you can set the speed to any speed in mph from 1 to 128. I can see having such resolution at slow speeds. 1 mph corresponds to ~1/4" per second in HO. But do you really need 1 mph resolutoin at high speeds? Do you need to be able to set the speed to 120 vs 121 mph?
of course, with a knob you can easily change the speed to whatever you want and easily go from 1 to 128. But as others have said, this isn't as easy with buttons.
For me a range of speeds from 1 to 28 mph is fine. I can control the speed with a resolution of 1 mph needed at low speeds. But i still don't need to be able to select 27 vs 28 mph.
i don't think 128 speed steps are needed (rule #1, the customer is always right). 1 mph resolution is needed at low speeds (e.g. 1 mph). But each speed step should be more that a change of 1 mph.
The table below indicates "progrressive" speeds for 28 steps where each step increases the speed by 17% from the previous step. Max speed is 69, and there is fine control at low speeds. Presumably CVs 4, 5 and 6 can do this.
when are 128 speed steps needed?
would button control of speed be better if progressive speeds steps are used?
That's where you use CV2-6-5 and reject decoders that do not support them. Especially having CV6 for mid - by adjusting that nearer to the start speed or nearer to the top speed, you produce a non-linear control suited for a particualr type of service - more low end fine control and bigger steps at the top end for a switcher that spends most of its time running slow, and less fine control at the low end but more at the high end for a passenger loco that spends most of its time at higher speeds and accelerating out of station stops.
Unless you need to reduce the overall top speed, this means all you need to adjust is a single CV, not 30 of them to implement a speed table.
Basically, you are emulating MRC's "taper wound rheostat" control.
i suggested those CVs above. but it doesn't look like a 3-pt speed curve results in a smooth curve (like a audio pot). You end up with small increments below the mid speed step and steeper increments above.
it seems one use for 128 speed steps is to get finer control at low speeds at the cost of having 128 speed steps which makes button control impractical
gregcwhen are 128 speed steps needed?
For me Greg, it's not about need, it's what the default setting is. I don't feel like messing around with CVs, since everything runs fine, I'll just leave it alone.
If I have three trains running at the same time while I do a little switching 128 speed steps allows me to really fine tune the three trains so they maintain their spacing. Once I get them set up with a little fine tuning that the 128 SS lets me do, they will run all day without intervention.
I also like to run a fifty car freight with diesel's and enjoy sneaking up with my BLI 2-10-4 to offer pusher service over the Rockies. I don't add the pusher to the consist, I just control it as I think it needs to be controlled, just as they did in real life. The 128 SS lets me do that with a fine touch.
It is 128 SS for me and maybe, even more, thank you very much.
Brent
"All of the world's problems are the result of the difference between how we think and how the world works."
BATMANIt is 128 SS for me and maybe, even more, thank you very much.
do you use buttons to select the speed step? if so, do you find it awkward?
Greg,
My NCE Power Cab has both gross & fine buttons, which offers you 10 & 1 speed step increments, respectively. It also has an encoder wheel. I use both the fine button and encoder wheel for accelerating and decelerating in 128 speed step mode and I don't find either one awkward to use.
Tom
https://tstage9.wixsite.com/nyc-modeling
Time...It marches on...without ever turning around to see if anyone is even keeping in step.
Hi Greg, I also have the NCE system and I use the scroll to adjust the speed unless I need to stop in a hurry and then I just tap the buttons for a quick stop. I rarely have to look at the controller anymore as I can just go by feel for most functions. I do take note of the speeds once they are set up when running more than one train. Being lazy I usually get an audible speed report from the engine itself. I do wish they would use a sexy babe voice instead of that guy with the Southern twang. Voices should be changeable through CVs.
I found the scroll speed control to be quite nice to use compared to a knob.
great. So there isn't a problem using buttons to select one of 128 speed steps because there can be buttons for different step sizes. The NCE cabbus protocol has button commands for step sizes of 1 and 5.
The 3 step speed curve ISN'T linear - that's sort of the point of using it. To move the majority of control to the part of the speed curve where you need it most based on that use of a given loco.
gregcSo there isn't a problem using buttons to select one of 128 speed steps because there can be buttons for different step sizes.
At least with the NCE ProCab throttles. With the fine speed step button on the Power Cab, you can press it individually to increase or decrease the speed step one increment at a time...or, hold the button and it increases or decreases at a rate of two speed steps per second. The encoder wheel also allows for "ballastic" changes in speed - slow or fast, which is determined by how quickly you move encoder wheel. Because of that I rarely use the emergency stop button to stop my locomotives quickly.
The placement of the four speed buttons and central encoder wheel are conveniently located on the throttle (Area 2):
So, they are easily accessible - whether you are left-handed or right-handed. And, like Brent - I know exactly where they all are and don't even need to look at them when changing speeds on the fly with my thumb.
tstage, et al - i now see that button control of 128 speed steps is fine. Having four buttons for speed control makes sense
rrinker The 3 step speed curve ISN'T linear - that's sort of the point of using it.
of course it's not.
linear would be each speed step it the same (i.e. 2-pt curve). but it's not clear that it isn't linear from min to mid, and from mid to max, as the chart i posted from the DCC Guy, Larry Puckett, illustrated.
I don't see where the NMRA configuration variable spec describes how to generate the values for a 3-pt speed curve. I don't assume all manufacturers do it the same way.
the plot compares 3-pt curve using linear pieces as Puckett describes and a smooth cubic curve.
i appreciate all the comments. They help me better understand how DCC works, how to use it and what to questions to ask.
I know of a few people who use 14 steps with momentum set to nearly max on both acceleration and deceleration. It makes for surprisingly realistic operation. It is still six steps more than a real engine has.
In a real engine, you don't set the throttle and forget it. You are consistently moving between notches to maintain the desired speed. Notch 4 is too slow and notch five is too fast. With the huge momentum, you just keep moving between the two notches with your model. Movement in a real train is hardly instantaneous.
That being said, judging from what I've read from most people, everyone wants instant and accurate control over what their engine does, even though it's not how a real train functions. It requires a lot of practice to run a train set up like this - but it IS more like how a real train functions. I had the pleasure of running a local for a shift and it was a huge eye-opener as to how thing really worked .... NOTHING happens quickly. It totally changed how I run my trains - both in switching and mainline running.
I've challenged people once before to give this a try, and not one person would. I don't expect anyone will this time around either. Probably the mentality that 14 steps are severely outdated, so why would I want to do that ?
Mark.
¡ uʍop ǝpısdn sı ǝɹnʇɐuƃıs ʎɯ 'dlǝɥ
Part of the issue here is that two vastly different things are being conflated as if they are the same.
The digital representation represents actual speed, and other considerations such as acceleration or momentum have to be handled in metadata as a combination of time and step. Throttle settings, particularly the 8-notch system imposed by the workings of a Woodward governor, represent power input, which only gets to road speed by way of the effects of acceleration, momentum, etc. -- on a model, this is a lot of algorithmic activity and involves at least a passable Davis-style approximation of train resistance to work.
The 'correct' way to implement this on a model is to have a display that shows the simulated throttle 'notch' for the consist, and the road speed (changing up and down): this gives prototypical action and by extension a real-world version of "dynamic braking" (it's of course nothing of the sort on the actual model) when operating with one of the 'prototypically correct' handheld or desk devices. The 'button' control now goes up or down by notches, with the implementation of 'constant speed' then being evoked through the equivalent of a macro assigned to a button or control that works the notch up and down with appropriate time constant between to keep the speed in the desired range. You then have a couple of parameters for setting "speed" with the buttons: similar to a thermostat, you have high and low setpoints (which are essentially offsets to a speed-table CV) and the desired speed as represented as a value in the table. Now there needs to be a realtime display of the speed value from the table, and not the step from 1 to 128 or 28 or 14 or whatever, and there probably needs to be some way to adjust the upper and lower action setpoints in conjunction with the desired level of momentum, etc., but this shouldn't involve much if any custom programming.
Note that most of the functions on a modern system like LEADER can be very simply added to this, including the ability to simulate GPS-based 'lookahead' for more effective power management. I would be tempted to have a variometer-like rate-of-speed-change indication to simplify reasonably-steady-speed operation without "constant fiddling" -- many modelers have gotten the idea that setting the speed directly is the way that real trains work, which is flat wrong, but still want to be able to set one knob to make the train move at constant speed -- this could be automated between limits.
i'm discovering that there are different types of modelers that DCC manufactureres need to satisfy:
Mark R.In a real engine, you don't set the throttle and forget it. You are consistently moving between notches to maintain the desired speed. Notch 4 is too slow and notch five is too fast. With the huge momentum, you just keep moving between the two notches with your model.
i'm glad Mark added this and Overmod commented futher
the problem with this is the need for an accurate measure of speed on the throttle which is not possible with existing DCC because it's one way communication. Maybe with WIFI/Bluetooth throttles.
OvermodThe 'correct' way to implement this on a model is to have a display that shows the simulated throttle 'notch' for the consist, and the road speed (changing up and down)
correct for who? (could be an option if possible)
I have to ask why prototypical locos are controlled by a processor that modulates the notch to maintain a desired speed given GPS location? (not complicated)
gregcBut should it even be called a speed step.
no one has commented on this. when starting out, do engineers always apply full power resulting in max accelleration or might they use a lower notch resulting in less accelartion but less stress (heat) on motors, less wheel slip?
I think the momentum setting would be better described as Mass and the speed step as horsepower step which would result in different acceleration for different horsepower and mass settings. (what about brake)?
Mark R. I know of a few people who use 14 steps with momentum set to nearly max on both acceleration and deceleration. It makes for surprisingly realistic operation. It is still six steps more than a real engine has. In a real engine, you don't set the throttle and forget it. You are consistently moving between notches to maintain the desired speed. Notch 4 is too slow and notch five is too fast. With the huge momentum, you just keep moving between the two notches with your model. Movement in a real train is hardly instantaneous. That being said, judging from what I've read from most people, everyone wants instant and accurate control over what their engine does, even though it's not how a real train functions. It requires a lot of practice to run a train set up like this - but it IS more like how a real train functions. I had the pleasure of running a local for a shift and it was a huge eye-opener as to how thing really worked .... NOTHING happens quickly. It totally changed how I run my trains - both in switching and mainline running. I've challenged people once before to give this a try, and not one person would. I don't expect anyone will this time around either. Probably the mentality that 14 steps are severely outdated, so why would I want to do that ? Mark.
Yes, the 14 step mode is obsolete, and it is an optional feature.
The 28 step mode is mandatory, and if the command station sends 128 step commands, the decoder, if compatible with that (optional) mode, will automatically switch to it.
Many expect the locomotive to move on step 1. IIRC, Rapido found that out when they programmed their decoders not to do that, just increase the RPMs, like the real thing. Some people didn't like that level of realism.
gregc tstage, et al - i now see that button control of 128 speed steps is fine. Having four buttons for speed control makes sense rrinker The 3 step speed curve ISN'T linear - that's sort of the point of using it. of course it's not. linear would be each speed step it the same (i.e. 2-pt curve). but it's not clear that it isn't linear from min to mid, and from mid to max, as the chart i posted from the DCC Guy, Larry Puckett, illustrated. I don't see where the NMRA configuration variable spec describes how to generate the values for a 3-pt speed curve. I don't assume all manufacturers do it the same way. the plot compares 3-pt curve using linear pieces as Puckett describes and a smooth cubic curve. i appreciate all the comments. They help me better understand how DCC works, how to use it and what to questions to ask.
I've probbaly forgootten more math than I know, but I'm not sure I would refer to the 3 step speed curve as a cubic. Every use and represenation I've see has been as two lines, each with constant slope, but different from each other (if CV6 is set to other than exactly the middle value, or left at default for most decoders). The flat lines on that chart, not the curves. If the actual behavior is the curves lines, that's news to me, and the behavior of locoos I have adjusted seems to agree moore with the straight lines than the curves, but I do not have a speed trap of any sort to actually measure the speed at each step to plot it out.
i'm not suggesting that a 3 point curve is a cubic. i am suggesting that it wouldn't be difficult to generate a smooth cubic curve.
Larry Puckett agrees with you that it describes two linear lines. The NMRA recomendation is not specific. It states "Vmid is used to generate a performance curve ...".
betamax .... Many expect the locomotive to move on step 1. IIRC, Rapido found that out when they programmed their decoders not to do that, just increase the RPMs, like the real thing. Some people didn't like that level of realism.
.... Many expect the locomotive to move on step 1. IIRC, Rapido found that out when they programmed their decoders not to do that, just increase the RPMs, like the real thing. Some people didn't like that level of realism.
You have that backwards.In a real engine, notch 1 just engages the circuit to the traction motors - there is no rpm increase. RPM doesn't actually increase until notch 2.
People were saying it was wrong that way (obviously never ran a real locomotive) and wanted it to rev up before moving off. A real engine WILL rev up before moving off, provided the engineer puts the throttle in a higher notch position to do so.