Dynamic Brakes

Discussion in 'Troubleshooting' started by pschlik, Oct 9, 2017.

  1. pschlik

    pschlik Well-Known Member

    Joined:
    Dec 8, 2016
    Messages:
    875
    Likes Received:
    1,607
    Dynamic brakes have always been a bit messy in TSW, so they deserve a thread to cover their problems all on their own. And some excuse for me to ramble on and explain complicated things.

    Blower Noise:
    Dynamic brake noise should vary with dynamic brake current. It does not in TSW (so you hear the same noise from -0 amps to -1000 amps) Thing is, it used to. And it indeed should, as for those who don't know, dynamic brake noise comes from the dynamic brake fans/blowers, and the fans are mostly driven by the current generated from the motors. The higher the dynamic braking current, the faster the fans spin, (TSW does model that chain of events in fan animations) so the louder and higher pitch the noise becomes (which TSW does not model anymore).

    Motor excitation:
    As this is a technical post, prepare to get techincal!
    The diesel engine power actually increases with dynamic brake lever position, which might seem silly, but the engine here is not being used to power the motors, it is being used to power the magnets. It’s all about activating the electromagnets in the motors so the motors actually generate electricity. Without magnets, the motors won’t generate anything, and the power for the magnets must come from the engine, so that’s why you need an engine to use dynamic brakes...(As a note, it does not take that much power to get the electromagnets doing their thing, so even full dynamics really only requires around notch 3 engine power.)
    Anyway, this is not modelled in TSW (you could even get dynamic braking with the engine off! this is impossible in real life) In fact, this is almost never modelled at all. Some highly advanced sound/enhancement packs for TS18 are the only times I'd see this in a simulator, and it is only there for showing realistic RPM numbers (you could just dub in engine noise as an MSTS dev I know of did), it has no important effect. I think SimuGraph could handle that kind of complex thing, so why not? [I get the impression that the old SimuGraph motors are magic and have no differentiation between the field winding and armature winding. The new ones clearly do.]

    Power Lag:
    Due to the way dynamic brakes work in tandem with engine output, there is some delay between setting a dynamic brake position and the dynamic brakes reaching that resistance level as the engine power must increase to increase dynamics power, and that takes time (even after the dynamics have been set up). In TSW (and usually TS17 too, for that matter) once the dynamic brakes activated, they instantly jumped up to the setting provided, which is not really accurate, and also promotes being slam on-slam off with the dynamics. You could even lose points in TS17 from the jerk created by the dynamics slamming all the way on. Only a few locos would correctly model the smooth power up and power down, and that would require fancy LUA scripting.

    And then the evil we all know...physics:
    I don't know why...but TSW decided to split off from the typical dynamic brake physics systems which were tried, tested, and accepted in at least 5 other simulators. Now, I could not make a physically based argument for why dynamic brakes work the way they do, but I can safely say the other simulators got it right, and TSW is not right. The focus of motor physics in discussion has always been power, but braking was a more ignored side of the issue, so here goes: [TLDR: TSW does dynamic brake physics wrong, look to any other simulator to get a good idea of what should happen.]

    Dynamic brake vs dynamic brake position:
    The dynamic brake current should vary about linearly related to dynamic brake position: so notch 4 would give you around twice the current as notch 2 dynamics. But, as a lot would have noticed, in TSW, dynamic brake positions from setup-3 do almost nothing to the current, but when you hit 4-6, suddenly the current shoots up, then positions 7-8 don't do quite as much to the current. This makes the dynamics very hard to control and is not realistic at all.
    I graphed dynamic brake current as it relates to dynamic brake position (in "notches") at 70 mph in the AC4400CW, and both an exponential function and even a sine curve are a more accurate predictor of dynamic brake current than the linear relationship it should be. (I think the actual implementation is sigmoidal, now that I look at it.) Which is frankly absurd when a linear relationship is the truth.
    Graph.JPG

    Dynamic brake vs speed:
    And the other thing that TSW decided to do differently was the dynamics current vs speed. It went for a strictly linear relationship: 0 mph means 0 dynamic current, 100 mph has 10x the dynamic current as 10 mph, etc. But in every other simulator, the dynamic current is actually direct for a bit, and then inverse, not totally direct: with 1. dynamics being most effective at slow speeds (25 mph or so [depends on loco]), not least effective as TSW suggests. 2. dynamics lose effectiveness when moving slower than that ideal speed, reaching 0 amps before a complete stop (speed depends on loco/traction motor type), while TSW only reaches 0 current when you are totally stopped-so that's definitely not realistic. And finally, 3. dynamics lose some effectiveness as speed increases. While in TSW, that linear relationship means current increases with speed. Up to completely insane numbers like -2000 amps, which should mean the resistors are on fire. You don't need to be an expert to know that's not right.
    20170506143734_1.jpg

    There's no point in attaching a graph for TSW, that's just a line with a perfect correlation, but here's what one of those absurdly detailed "sound"/enhancement packs spits out at full dynamics on an SD70ACe:
    As I said, I couldn't tell you why that's not linear, as linear is nice and easy and sorta makes sense...but that's not how it is in real life. (As soon as you do stuff like field weakening and dynamic braking, stuff gets strangeeee) And as a simulator, TSW should do what happens in real life, not what makes sense.
    nice curve.JPG

    Yeah I know I wrote a lot, that's what physics problems do to you. And I didn't feel like it was enough to not explain this since dynamics are not in the UK/GWE so have not been in the focus and are not commonly used by a lot of people in the sim.
     
    Last edited: Aug 22, 2018
    • Like Like x 8
    • Helpful Helpful x 3
  2. Karl456

    Karl456 Member

    Joined:
    Dec 12, 2016
    Messages:
    72
    Likes Received:
    74
    A very detailed explanation of the issue... :o
    I, too, have thought that the behavior of the dynamic brake in TSW is far away from the real-world behavior but I could never be able to say exactly what it is that makes the TSW dynamic brake feel so unrealistic. So thank you a lot for these information and lets hope that DTG will not only notice them, but also react with updates and fixes.
     
    • Like Like x 3
  3. SamYeager270

    SamYeager270 Well-Known Member

    Joined:
    Dec 8, 2016
    Messages:
    633
    Likes Received:
    667
    Interesting post pshlik. I hope you've raised a ticket and followed up by raising a bug report referencing the ticket ID. I've found in the past that this is the only way to make sure this gets noted formally by DTG. Of course this doesn't necessarily mean that anything fast will happen, if at all, but at least DTG can't then say they weren't aware of the issue.
     
  4. Karl456

    Karl456 Member

    Joined:
    Dec 12, 2016
    Messages:
    72
    Likes Received:
    74
    Really good idea. It would be such a huge waste if DTG wouldn't recognize this enormously detailed description.
     
    • Like Like x 1
  5. pschlik

    pschlik Well-Known Member

    Joined:
    Dec 8, 2016
    Messages:
    875
    Likes Received:
    1,607
    I actually did send this in to Kayako a while back, that's why I have the graph of notches vs current. And that picutre of the AC4400CW at -2019 amps is just part of the data. (I don't think you can do that any more, as the HUD ammeter on the AC44 was changed to max out at -1200.)
     
  6. SamYeager270

    SamYeager270 Well-Known Member

    Joined:
    Dec 8, 2016
    Messages:
    633
    Likes Received:
    667
    Is the ticket still open and did you also send a bug report pointing at the ticket ID?
     
  7. pschlik

    pschlik Well-Known Member

    Joined:
    Dec 8, 2016
    Messages:
    875
    Likes Received:
    1,607
    All of my tickets are closed, whether its a bug that's fixed or not.
     
  8. pschlik

    pschlik Well-Known Member

    Joined:
    Dec 8, 2016
    Messages:
    875
    Likes Received:
    1,607
    Now it is interesting to note that Rapid Transit not only addressed motor physics under power, it also addressed a lot on dynamic brake physics.

    Now, since electrics are involved and this thread is more about diesel electrics, some of the stuff here plain does not apply (notably on noise, the delay to power up, and engine RPM) but maybe whenever DTG decides to actually go back to CSX on the physics, dynamics will come out better as well.

    Anyway, the Talent 2 has addressed exciting the motors to use dynamics. Yeah, no diesel engine is involved, but even then you still need to put some power in to get power out. If you kill the circuit breaker, the dynamics stop working because A: there is nowhere to get electricity to the electromagnets from and B: there is nowhere to put electricity generated (if any could be generated.) Which is a certain step up from CSX where dynamics can work with the engine off.

    It also took care of the response between the dynamic position and dynamic effort. 25% dynamics is 25% dynamic brake effort. Not 8.5% like on CSX.

    Also, the dynamic brake effort vs speed is significantly better than in CSX. Instead of being more effective at speed, they are about the same effectiveness at speed, and die out very quickly under like 5 kph. Now, this wasn't entirely fixed, as the dynamic brake effort doesn't decrease with speed, but it is still a step in the right direction.

    So hopefully dynamic brakes will improve on diesel electrics as well, instead of doing the magic braking coming from nowhere that they make now.
     
  9. Jef-F

    Jef-F Active Member

    Joined:
    Dec 8, 2016
    Messages:
    210
    Likes Received:
    136
    I don't think we can draw any conclusions from looking at Talent performance, neither in acceleration nor braking. With AC-AC powertrain entirely driven by sophisticated power electronics it can employ much more intricate control techniques than an ancient loco with relatively straightforward control systems (although upgraded a bit). For instance maintaining constant braking effort using rheostatic, recuperative and pneumatic brakes at the same time.

    I mean, in those old EMDs you're almost directly controlling excitation with dynamics handle, so you can see nice and proportional dependencies between excitation, speed and amps. In Talent you're setting desired braking effort and electronics control everything else to maintain it. If we do not know how exactly it happens, we can't judge by eye if it works correctly or not and if it is connected to CSX locos' behavior in any way.
     
    Last edited: Dec 21, 2017
  10. Anthony Pecoraro

    Anthony Pecoraro Well-Known Member

    Joined:
    Dec 11, 2016
    Messages:
    3,711
    Likes Received:
    4,323
    Dynamic braking is bad in TS20xx too.
     
  11. jENKS

    jENKS New Member

    Joined:
    Mar 21, 2018
    Messages:
    20
    Likes Received:
    32
    This needs to be addressed as well. Holy cow, what a mess this sim is. It's so far from reality, it might as well literally be using space shuttle physics.
     
  12. Anthony Pecoraro

    Anthony Pecoraro Well-Known Member

    Joined:
    Dec 11, 2016
    Messages:
    3,711
    Likes Received:
    4,323
    Your making it worse than it is. Northeast Corridor New York is their best TSW DLC yet.
     
  13. pschlik

    pschlik Well-Known Member

    Joined:
    Dec 8, 2016
    Messages:
    875
    Likes Received:
    1,607
    Well the ACS-64 does have decent dynamics, actually implementing a correct power curve for once, but that is just one locomotive out of 9.
     
    • Like Like x 1
  14. nic el loco

    nic el loco Guest

    Hi Pschlik and everybody, first post here.
    Thank you so much for this explanation, I'm new to train simulators (only played a few time the first Microsoft TS) and since TSW lacks in tutorials, sometimes it's hard to understand if I'm doing something wrong (and that's very often the case) or if the physics in game is not properly simulated. I struggled a bit to understand how to manage breaking while going downhill from Sand Patch to Cumberland in a realistic and efficient way, and while looking for a tutorial or some advice I found this that I think is a good and cool guide on how to approach the problem (since I'm not an expert at all it can actually be rubbish, it just seems to be legit to me...). But when I tried to apply what was explained in the text in CSX HH, the result was not as smooth, since I tried to remain around 3 or 3.5 with the dynamic brake. I had to completely release and reapply the auto a couple of times (which I think isn't possible in reality, at least with this efficiency) because I went too deep, then understand that my train responded suddenly more to DB when I applied it around 5.

    Overall I like this game a lot, it's just a shame that there are not many in-depth tutorials for somebody like me who doesn't come from TS20XX and doesn't have more than a vague idea on how to drive a train. I really hope that CSX HH will not be left behind since it will not be part of the July release...
     
    Last edited by a moderator: Jun 6, 2018
    • Like Like x 1
  15. Anthony Pecoraro

    Anthony Pecoraro Well-Known Member

    Joined:
    Dec 11, 2016
    Messages:
    3,711
    Likes Received:
    4,323
    New Diesel-Electric physics are coming!
     
    • Like Like x 1
  16. pschlik

    pschlik Well-Known Member

    Joined:
    Dec 8, 2016
    Messages:
    875
    Likes Received:
    1,607
    Well, bumping this topic again. (oooh rebel me) As you may have observed in prior studio update articles, the 'secret locomotive' with dynamic brake equipment and all was the 143. As was said in the June update, the crazy wiring and switches would be useful for CSX stuff, and since then, we've got a chance to work with the 143. And frankly, the dynamic brakes are really disappointing, and as this is related to the CSX locomotives, I'm worried about this same code going in there.

    Despite all this fancy rewiring, the dynamic brakes on the 143 still behave as if more speed = more brake power. Which is far from what it should be for both a 143 and a generic dynamic brake curve. And is probably the most fundamentally wrong thing about the dynamic brakes. What's even more bothersome is that the 185.2 has an electric brake effort curve that is basically perfect for that kind of locomotive....yet it has none of these extra bits and bobs for the electric braking. So somehow the locomotive without the fancy stuff is the realistic one, and the one with the fancy stuff is the unrealistic one.

    That's really darn confusing and really worrying. Power physics may be great and all...but dynamic brakes were a big problem too, and its not solved yet. At least not as far as I'm convinced.

    EDIT: here is a graphic that's not totally accurate to represent whats going on:
    generic dynamic brake.png
    So, the red line is your real life dynamic brake curve. Increases linearly up to a certain speed, where it then decreases in a logarithmic style.
    The dotted line indicates the dynamic brake effort limit. This depends on a locomotive by locomotive basis, but usually German stuff limits it pretty low, so that at any speed you can count on getting 150 kN of effort, no more, no less.

    Then you got the 143's stuff with that blue line. Does not take an expert to understand that this relationship doesn't match with any real life principles at all.
     
    Last edited: Aug 23, 2018
    • Like Like x 3
    • Helpful Helpful x 1
  17. Anthony Pecoraro

    Anthony Pecoraro Well-Known Member

    Joined:
    Dec 11, 2016
    Messages:
    3,711
    Likes Received:
    4,323
    I seriously hope this gets fixed properly.
     

Share This Page