DCC sound vs. your throttle

[slowfreight]

I'm starting a new thread to jump off the Genesis vs Proto discussion, so that I don't fully derail it.

In discussions about sound vs. throttle response in miniature, we struggle a bit because most of our control systems are a "speed knob." The speed of the train is based on how far you twist the knob.

Yes, we have momentum in our decoders now. Proto Throttle then achieved a big step forward with the hand-held control stand I always dreamt of using. But we still have limitations.

This video represents my gold standard. Note that the prime mover speed has nothing directly in connection with the train speed--or direction



What is happening is that the geep is genuinely fighting weight and momentum, and even making transition. Whether the engineer should start shoving before the backing move is complete is another discussion for the professional railroaders in the room, and some railroads lock out the throttle when the reverser is thrown to reduce wear and tear.

But the next step in DCC programming evolution will be to figure out how to simulate the weight of a string of cars appropriately, relative to the HP and number of units pulling. Our cars are much lighter, have no brakes, and don't have a feedback method to tell the decoder how to react. The momentum feature is a start, but I don't see a way around selecting some average train weights. And simulating more HP on a freight vs. less would also require selecting some options.

There's still a large contingent that need a speed knob, because momentum gets you in trouble fast and can lead to derailments in miniature. But we're a step closer to the realism of that ONR yard job, even if we can't do that gorgeous evening light in the train room.

[voyager989]

Feedback is currently based on motor load; Tsunami has had this for a while, and ESU is making their first steps into feedback (but the sound projects don't really support it yet aside from steam). Other than that, you just have systems that let you lock the throttle and use the speed knob to control PM sounds.

[cemr5396]

Even the Proto Throttle is just a glorified 'speed knob', when you notch it up your unit goes faster and when you notch it down it goes slower. It IS much more immersive than a regular DCC controller though, which I can appreciate.

That being said, my locomotives are all equipped with TSU2 sound decoders, and I've found that with a pretty healthy amount of momentum programmed in, the independent brake function programmed to actually BE a brake and not just a sound effect, and the DDE load sensing feature enabled and properly calibrated, you can get pretty close to that ONR switching video. Yes, adding a lot of momentum can have it's issues, but having a working brake helps a lot with getting the thing stopped, and if things get really hairy, there is always the Emergency Stop button on the DCC controller.

Personally I find the ESU throttle lock feature where you use the speed control as a manual notching control to be a joke, honestly. Especially for switching, it would be completely useless for simulating the throttle response seen in the ONR video.

[12bridge]

A) There are, and always will be limitations on what can be done with throttles and movements in scale. 

B) 9/10 people do not know how to properly setup and control their decoders and use them to the maximum of their abilities. 

Its as simple as that.  We have a proto throttle here.  My wife, an engineer also, spent the better part of a full day dialing that thing in.  Its almost like running a real train.  Its the closest were gonna get.  Now, obviously its not practical for everyone to have one of these, nor set it up for every engine in the fleet.  See A.  

[fishbelly]

I fall into category B. I do not take the time to really learn all the nuances of these decoders. I just get them setup and run. Now IF I could find a way to get the sound to better represent what the real world puts out, then I might be inclined to dive deeper into all the fine workings. Until then I will take the slow learning curve and hopefully move myself out of category B. Heck, I do not even know how to use the ESU setup of drive hold and independent brake. Have never done it. Going forward though I have switched to Soundtraxx Blunami. After the St. Louis RPM meet and spending a good deal of time talking and learning about Blunami, it is a much more user friendly (to me) system. Once I get my first locomotive built and setup with it, I plan to really learn how to use it. Lots of really good Youtube stuff from the Soundtraxx guy. Very helpful and visually easy to understand.

I would very much like to get close to being able to operate with sound like that video.

[abm]

Sept 28, 2023 6:55:26 GMT -8 cemr5396 said:

Personally I find the ESU throttle lock feature where you use the speed control as a manual notching control to be a joke, honestly. Especially for switching, it would be completely useless for simulating the throttle response seen in the ONR video.

If you're referring to the "drive hold" feature, I've found the opposite... I think it's quite useful for simulating the kind of action seen in that video. Takes some effort programming both momentum and the speed curve (MUCH flatter than what most models come with from the factory) but I found it more intuitive to set up (via the LokProgrammer) than Tsunami's DDE. But at the end of the day it's all nothing more than just some clever simulation.

For braking, I set up most engines with two brakes, and the deceleration momentum set to maximum, to provide the most "coasting" effect. Brake 1 is tied to the independent brake application sound and has a moderate braking rate; my preference is for around 4 engine-lengths of movement from about 30% speed with a continuous application. Brake 2 is tied to the coupler crash sound and has an immediate-stop braking rate; I only use this for making joints when the engine is already moving pretty slowly (though it's also useful as an e-stop at higher speeds). Loksound offers a third brake which I usually set up at about half the braking rate as the independent and tie to the automatic brake application sound, but I hardly ever use it.

[cemr5396]

Sept 28, 2023 8:28:53 GMT -8 abm said:

Sept 28, 2023 6:55:26 GMT -8 cemr5396 said:

Personally I find the ESU throttle lock feature where you use the speed control as a manual notching control to be a joke, honestly. Especially for switching, it would be completely useless for simulating the throttle response seen in the ONR video.

If you're referring to the "drive hold" feature, I've found the opposite... I think it's quite useful for simulating the kind of action seen in that video. 

to me it just seems like it would be a huge PITA to be constantly switching between using the throttle as 'not a throttle' and actually having control over the loco. Especially when switching when you are constantly speeding up, slowing down, and changing directions.

[Baikal]

Sept 28, 2023 5:57:59 GMT -8 slowfreight said:

I'm starting a new thread to jump off the Genesis vs Proto discussion, so that I don't fully derail it.

In discussions about sound vs. throttle response in miniature, we struggle a bit because most of our control systems are a "speed knob." The speed of the train is based on how far you twist the knob.

Yes, we have momentum in our decoders now. Proto Throttle then achieved a big step forward with the hand-held control stand I always dreamt of using. But we still have limitations.

This video represents my gold standard. Note that the prime mover speed has nothing directly in connection with the train speed--or direction



What is happening is that the geep is genuinely fighting weight and momentum, and even making transition. Whether the engineer should start shoving before the backing move is complete is another discussion for the professional railroaders in the room, and some railroads lock out the throttle when the reverser is thrown to reduce wear and tear.

But the next step in DCC programming evolution will be to figure out how to simulate the weight of a string of cars appropriately, relative to the HP and number of units pulling. Our cars are much lighter, have no brakes, and don't have a feedback method to tell the decoder how to react. The momentum feature is a start, but I don't see a way around selecting some average train weights. And simulating more HP on a freight vs. less would also require selecting some options.

There's still a large contingent that need a speed knob, because momentum gets you in trouble fast and can lead to derailments in miniature. But we're a step closer to the realism of that ONR yard job, even if we can't do that gorgeous evening light in the train room.

Yep, accurately modeling a combination of speed + engine sounds is difficult because of the lack of train dynamics (mass/momentum/rolling qualities, etc) feedback to the loco(s) drawbar.  And more...

I don't see modeling the sounds of what that GP was doing in my lifetime.

Yet the weight of HO scale freight cars is approximately correct.  A prototype car weighing 100 tons = about 4.8 ounces in HO.  The poor rolling qualities of model cars kills the momentum.  This will never be fixed due to the physics at the axle bearings.  A single prototype car set moving at 5 mph on perfectly flat track will may roll 10 car lengths or more.

My non-perfect solution is to more-or-less avoid sound, which will never be more than a lame approximation of what a prototype loco sounds like.  Having spent many hours on & around real locos, I can create better sound in my head than comes thru 1" speakers.  Sound in a few switchers just for the neat-o factor and guests is ok, but an entire six-unit consist, no thanks.  For me, the cost and effort required to get weak-at-best sound working, plus all the other DCC configs is better spent elsewhere.

[abm]

For the switching I do on my little layout, most of the time I just use the throttle as a throttle, the brake(s) as brakes, and let the combination of the momentum settings, speed curve, and braking rates do their thing. I use drive hold to simulate very low speeds with very high throttle settings, such as dragging a long/heavy cut of cars out of one track, and then shoving them into another. The thing I had to come to grips with (you probably already know this) was that on a model railroad, the speed and time scales too. What seems fairly slow to us, 1:1 scale standing over 1:87 scale, would actually be really fast if we were shrunk down to 1:87 scale as well. Imagine that ONR video posted above... at ground level, they're jockeying that thing back and forth pretty quick, but viewed from above, I bet it would seem remarkably slow, and the distances they cover rather short. The video is just shy of 2 minutes long and all he did in it was pull down about a dozen cars, and shove back about half that many for the kick... it really wasn't much movement in almost 2 minutes. So what we end up doing (or are tempted to do) is switch must faster (elapsed time) than we would in real life. And hey, it's a model railroad, do whatever makes you happy. Personally, I switch pretty slow, because that's how I remember it being back when I did it for real 20+ years ago (yikes... I can't believe it's been that long). Point being, at the speed I work, turning the drive hold on and off isn't that big of a pain.

One thing I haven't been able to figure out is how to effectively kick cars in HO scale... now THAT would be fun. As it is now my brakemen are stuck riding every shove all the way to the joint... sucks to be them!

[Baikal]

Sept 28, 2023 9:18:59 GMT -8 abm said:

For the switching I do on my little layout, most of the time I just use the throttle as a throttle, the brake(s) as brakes, and let the combination of the momentum settings, speed curve, and braking rates do their thing. I use drive hold to simulate very low speeds with very high throttle settings, such as dragging a long/heavy cut of cars out of one track, and then shoving them into another. The thing I had to come to grips with (you probably already know this) was that on a model railroad, the speed and time scales too. What seems fairly slow to us, 1:1 scale standing over 1:87 scale, would actually be really fast if we were shrunk down to 1:87 scale as well. Imagine that ONR video posted above... at ground level, they're jockeying that thing back and forth pretty quick, but viewed from above, I bet it would seem remarkably slow, and the distances they cover rather short. The video is just shy of 2 minutes long and all he did in it was pull down about a dozen cars, and shove back about half that many for the kick... it really wasn't much movement in almost 2 minutes. So what we end up doing (or are tempted to do) is switch must faster (elapsed time) than we would in real life. And hey, it's a model railroad, do whatever makes you happy. Personally, I switch pretty slow, because that's how I remember it being back when I did it for real 20+ years ago (yikes... I can't believe it's been that long). Point being, at the speed I work, turning the drive hold on and off isn't that big of a pain.

One thing I haven't been able to figure out is how to effectively kick cars in HO scale... now THAT would be fun. As it is now my brakemen are stuck riding every shove all the way to the joint... sucks to be them!

I don't understand what you are saying re: scale speed & time.  You are over-thinking this(?).

Speed is defined as distance divided by time.  Scale has nothing to do with it.  The scaling of speed is accounted for in the scaling of distance.  In HO, that distance is approx 1:87  [ or 1:(254/35)*12 exactly ].  Time stays the same, no scaling.

For example, a prototype car moving at 5 miles/hr takes about 6.8 seconds to move one 50 ft car length.  And an HO scale car moving at 5 scale miles/hr takes about 6.8 seconds to move one HO scale 50 ft car length.  So moving a train 100 car-lengths at X miles/hr takes the same time on the prototype as it does in any scale.

Not mysterious or complicted.  It's just math.

But modelers often run too fast, esp doing yard work.  There's a lot of waiting between moves and even though moves may be made at 10 mph we often neglect the long acceleration & decelleration periods.  Plus waiting for crew members to walk, apply hand brakes, etc.  I was a switchman & yardmaster, and it wasn't unusual for a crew to take 20 minutes just to get a light loco to the opposite end of the yard.  At the last yard I worked, we'd typically set-out 2 to 4 bad-order cars from departing 80-car trains.  This would take 45 min to 2 hours, sometimes more if there were any issues.  How many modelers spend an hour setting out two cars?  Just pumping up the air could take an hour IF there were no problems, like finding & fixing leaks.

[riogrande]

I wonder if it is that important for more realism, what about the Proto Throttle used with custom decoder settings?

[sd80mac]

In my experience, the decoders that change the engine throttle position based on how much load the decoder "thinks" its pulling are not very good. I had a trio of Tsunami 2 equipped GP9s set up this way and they were constantly notching up and down, far more than you ever would in real life (I can back this up, I've been a licensed engineer for over a decade at this point) to the point where it became incredibly annoying. I have some BLI steam locomotives as well as some steam with TCS wowsound that do the same thing. It's so annoying I've disabled it on everything. I much prefer ESU's Drivehold feature, where I can just set the train's speed to a prototypcial setting and manipulate the engine sounds as I see fit.

[kpack]

For what it's worth, here's how Railpro tries to achieve what the ONR video shows.  You can adjust momentum in acceleration and deceleration as normal, but there are two additional functions to improve on the experience.  One is "Load" and the other is "Brake"  Essentially they are function buttons that help to mimic the movement of a prototype train.  The Load setting can be adjusted on the fly and simulates the mass behind the locomotives.  Crank it all the way up to 100% and then spin the throttle up....the prime mover takes off but the locomotive will take some time to start moving.  The top speed is affected as well as the stopping distance.  Sound is also affected.  Under heavy load the prime mover is much louder, then quiets down to a whisper when the throttle is backed off.  It's helpful for switching ops because you can change the load settings quickly and from there it's just using the throttle knob to control the speed of the train.  Minimal input.

The brake is just what it says.  It slows the train down and is heavily dependent on the load setting.  It's not like a prototype train brake, but it does the trick.  

These aren't perfect solutions but they seem to help.  I use the load settings with automatic notching when I'm switching.  When I'm running mainline trains I use manual notching to have greater control over the prime mover sound.  It annoys me when model trains are stuck on a single notch sound at a given speed.....it gets repetitive and doesn't feel or sound like the prototype.

Here's a video I did a few years back that shows some of what I'm talking about:



-Kevin

[Baikal]

Sept 28, 2023 10:51:56 GMT -8 kpack said:

For what it's worth, here's how Railpro tries to achieve what the ONR video shows.  You can adjust momentum in acceleration and deceleration as normal, but there are two additional functions to improve on the experience.  One is "Load" and the other is "Brake"  Essentially they are function buttons that help to mimic the movement of a prototype train.  The Load setting can be adjusted on the fly and simulates the mass behind the locomotives.  Crank it all the way up to 100% and then spin the throttle up....the prime mover takes off but the locomotive will take some time to start moving.  The top speed is affected as well as the stopping distance.  Sound is also affected.  Under heavy load the prime mover is much louder, then quiets down to a whisper when the throttle is backed off.  It's helpful for switching ops because you can change the load settings quickly and from there it's just using the throttle knob to control the speed of the train.  Minimal input.

The brake is just what it says.  It slows the train down and is heavily dependent on the load setting.  It's not like a prototype train brake, but it does the trick.  

These aren't perfect solutions but they seem to help.  I use the load settings with automatic notching when I'm switching.  When I'm running mainline trains I use manual notching to have greater control over the prime mover sound.  It annoys me when model trains are stuck on a single notch sound at a given speed.....it gets repetitive and doesn't feel or sound like the prototype.

Here's a video I did a few years back that shows some of what I'm talking about:



-Kevin

Interesting concept, but too complicated if there's any changes in the load, like during switching.  You'd have to keep going back and changing settings.  I'd rather keep my eyes on the loco & cars.

Unless "100% load" means a single empty car, the loco is still accelerating much too fast.  100% load should mean something like "barely able to start a train" and it would be moving at 1 or 2 mph a minute after starting to move.

[cemr5396]

Sept 28, 2023 10:47:25 GMT -8 sd80mac said:

In my experience, the decoders that change the engine throttle position based on how much load the decoder "thinks" its pulling are not very good. I had a trio of Tsunami 2 equipped GP9s set up this way and they were constantly notching up and down, far more than you ever would in real life (I can back this up, I've been a licensed engineer for over a decade at this point) to the point where it became incredibly annoying. 

That's an easy problem to fix, just lower the sensitivity.

[timvanmersbergen]

Doesn't really relate to sound but I've found that people that have worked on the prototype switch differently than those that have not. We are more in tune with what the crew is doing and where they are at.

If you want to simulate this, get a couple cheap figures or even chess pawns and place them on the layout to simulate where a brakeman would be. Pick them up when they are riding. Try to minimze how much they walk. One switches differently when keeping this in mind. In the freight yard, it was always preferable to move the locomotive rather than moving yourself.

[rockisland652]

I have all my power programmed to respond as an actual locomotive would, with the inertia, momentum and braking that exists in the real world. This annoys some operators who are used to typical power that starts and stops far too quickly using only the throttle position. A welcome addition would be an easy way to adjust the train weight so that the inertia and momentum could reflect that.

I also have the DDE programmed so that the load on the power matters. The annoying 'hunting' between notches in a consist was mitigated by reducing the DDE sensitivity. No big deal. The ProtoThrottle provides a decent interface that at least superficially mimics the control inputs from a real control stand. Not sure that model railroaders need to know what the brake pipe/brake cylinder/main rez pressure is on their HO scale power, but I'm sure that overkill will be available to purists someday.

I have yet to find a 'Drive Hold' button in any real locomotive that I have operated.

[kpack]

Sept 28, 2023 11:16:37 GMT -8 Baikal said:

Interesting concept, but too complicated if there's any changes in the load, like during switching.  You'd have to keep going back and changing settings.  I'd rather keep my eyes on the loco & cars.

Unless "100% load" means a single empty car, the loco is still accelerating much too fast.  100% load should mean something like "barely able to start a train" and it would be moving at 1 or 2 mph a minute after starting to move.

Think of it as an easier way to adjust the momentum. It takes 2 seconds to do. When I'm switching I'm not going and changing the load setting on every pull. I either have it at 50 or 100, depending on what I'm trying to simulate. Basically give it enough load to get the price mover spooling up and give the train a sense of mass. When the load is set, I use the throttle like normal and don't bother looking at it.

If you wanted to make the take off much slower, you can increase the base acceleration settings. But I don't bother with that. Works well enough for what I'm trying to do.

[sd80mac]

Sept 28, 2023 11:19:47 GMT -8 cemr5396 said:

Sept 28, 2023 10:47:25 GMT -8 sd80mac said:

In my experience, the decoders that change the engine throttle position based on how much load the decoder "thinks" its pulling are not very good. I had a trio of Tsunami 2 equipped GP9s set up this way and they were constantly notching up and down, far more than you ever would in real life (I can back this up, I've been a licensed engineer for over a decade at this point) to the point where it became incredibly annoying. 

That's an easy problem to fix, just lower the sensitivity.

Still hunted far too much for my liking. But I also hate Soundtraxx so I'm not too worried about making them run nicer. They'll get Lok'd eventually.

[rockisland652]

Sept 28, 2023 12:02:08 GMT -8 sd80mac said:

Sept 28, 2023 11:19:47 GMT -8 cemr5396 said:

That's an easy problem to fix, just lower the sensitivity.

Still hunted far too much for my liking. But I also hate Soundtraxx so I'm not too worried about making them run nicer. They'll get Lok'd eventually.

Well, at least you're unbiased in your assessment.

I can't stand the weak sound you get from a Lok, so I guess we're even.

[mdq]

insert quote here
B) 9/10 people do not know how to properly setup and control their decoders and use them to the maximum of their abilities.

guilty!

[mdq]

insert quote hereEven the Proto Throttle is just a glorified 'speed knob', when you notch it up your unit goes faster and when you notch it down it goes slower. It IS much more immersive than a regular DCC controller though, which I can appreciate.

I think that is the whole point

If it wasn't for the $$$ I might just get one

[mdq]

Sept 28, 2023 8:28:53 GMT -8 abm said:

Sept 28, 2023 6:55:26 GMT -8 cemr5396 said:

Personally I find the ESU throttle lock feature where you use the speed control as a manual notching control to be a joke, honestly. Especially for switching, it would be completely useless for simulating the throttle response seen in the ONR video.

If you're referring to the "drive hold" feature, I've found the opposite... I think it's quite useful for simulating the kind of action seen in that video. Takes some effort programming both momentum and the speed curve (MUCH flatter than what most models come with from the factory) but I found it more intuitive to set up (via the LokProgrammer) than Tsunami's DDE. But at the end of the day it's all nothing more than just some clever simulation.

For braking, I set up most engines with two brakes, and the deceleration momentum set to maximum, to provide the most "coasting" effect. Brake 1 is tied to the independent brake application sound and has a moderate braking rate; my preference is for around 4 engine-lengths of movement from about 30% speed with a continuous application. Brake 2 is tied to the coupler crash sound and has an immediate-stop braking rate; I only use this for making joints when the engine is already moving pretty slowly (though it's also useful as an e-stop at higher speeds). Loksound offers a third brake which I usually set up at about half the braking rate as the independent and tie to the automatic brake application sound, but I hardly ever use it.

That is the thing. momentum. I find it complex to reduce (sometimes to zero) because it can be annoying (ESU) and sometimes a little is just right. 

[mdq]

Sept 28, 2023 11:23:37 GMT -8 timvanmersbergen said:

Doesn't really relate to sound but I've found that people that have worked on the prototype switch differently than those that have not. We are more in tune with what the crew is doing and where they are at.

If you want to simulate this, get a couple cheap figures or even chess pawns and place them on the layout to simulate where a brakeman would be. Pick them up when they are riding. Try to minimze how much they walk. One switches differently when keeping this in mind. In the freight yard, it was always preferable to move the locomotive rather than moving yourself.

That is a good point. I'll try that.

[rockisland652]

Sept 28, 2023 11:23:37 GMT -8 timvanmersbergen said:

Doesn't really relate to sound but I've found that people that have worked on the prototype switch differently than those that have not. We are more in tune with what the crew is doing and where they are at.

If you want to simulate this, get a couple cheap figures or even chess pawns and place them on the layout to simulate where a brakeman would be. Pick them up when they are riding. Try to minimze how much they walk. One switches differently when keeping this in mind. In the freight yard, it was always preferable to move the locomotive rather than moving yourself.

A good switchman never walks!

[michaele]

Sept 28, 2023 10:47:25 GMT -8 sd80mac said:

In my experience, the decoders that change the engine throttle position based on how much load the decoder "thinks" its pulling are not very good. I had a trio of Tsunami 2 equipped GP9s set up this way and they were constantly notching up and down, far more than you ever would in real life (I can back this up, I've been a licensed engineer for over a decade at this point) to the point where it became incredibly annoying. I have some BLI steam locomotives as well as some steam with TCS wowsound that do the same thing. It's so annoying I've disabled it on everything. I much prefer ESU's Drivehold feature, where I can just set the train's speed to a prototypcial setting and manipulate the engine sounds as I see fit.

I have an older Roco ÖBB 2143 locomotive using a Zimo decoder that can be set up to do that. While it works, it fluctuates the engine sounds much too often when encountering even only a slight change in grade. I've also found it accelerates the engine sound much too quickly too for the amount of grade encountered.

I turned it off. It becomes an annoyance after a short time.