The Throttle Problem

[pschlik]

WE DID IT PEOPLE, WSR ACTUALLY HAS DECENT DIESEL ELECTRIC PHYSICS.

And other stuff is to follow.[live.dovetailgames.com]

Updated with release of Rapid Transit-bit of a rewrite to include new details I've learned over time.
TL;DR: The way most electric motors are handled in TSW leads to an effect somewhat like the throttle has a built in cruise control which does not happen in real life, which is making driving diesel electric trains more annoying than it really should be.

Of all TSW's current, and extremely long standing bugs (this was reported over a year ago, and has been fixed in new [but not old] content already!),...the one I understand the reason behind the least is what I'm calling 'the throttle problem' but could be the 'cruise control effect' as comes up plenty as well; it's just a messed up setup with the power curves of the motors and throttle. Worst of all, the throttle and power curves were things that were (mostly) correctly modeled in TS1...but TSW somehow did it worse than TS1 did.

Anyway: the problem itself. I'm going to be referring to the SD40-2 here, but all the diesel electric locomotives do this, just at different speeds.

So, in current TSW, each throttle notch acts not only as the control for the level of power, but also also as the control of the maximum speed of the motors. The technical side of that is voltage-which I'll explain more on later. For example, notch 1 is pretty low power, and the maximum speed is set prettly low too; 23 mph. So you'll go up to 19 mph, then the amps will start dropping faster than normal as the motor nears what it thinks is its voltage limit, going down to 0 amps at 23 mph.
That's how it is like a cruise control; each notch of power has its own cruise speed in the physics. Increase the throttle, the power is increased AND the max speed of the motors is increased up to 29 mph, so you can accelerate once more. Drop the throttle to notch 1 again and the motors will be going too fast for the max speed setting, so you get 0 amps once more.

So what's the deal with voltage? Well, (DC, that is; AC ones are bit funkier) motors are nice and easy to work with; the power of a motor is equal to the voltage (speed) multipled by the amperage (torque). On a diesel electric locomotive, the throttle only controls the power (in an indirect way), and the motors use an amperage based on what speed the motor is already at to reach the constant power level. However, motors have a voltage limit before things stop working and the motor loses the ability to make the amount of torque expected. Get close to that limit, and the torque of the motor will drop, causing the power of the motor overall to drop. Exceed that voltage limit, and the power of the motor will drop to 0 (no amps, no torque).
So that sounds like what TSW does is pretty accurate to a motor hitting its voltage limit, right? Well, motors have a constant voltage limit. No matter what power you stick in, the motor will reach its limits at the same voltage, and therefore the same speed. TSW has the voltage limit connected to the throttle setting on the diesel electrics, rather than it being a constant number. And that's just not realistic.

The practical downside of this wole thing is that you get a lot less control of how much torque the train is making. You might think that having an easy 'set throttle, reach speed' system would be nice, but it frequently forces you into very strange speeds that don't correspond to speed limits while also forcing you to often apply far too much power compared to what's needed. In real life, you would probably drive an HST up to 125 mph in notch 5, then leave it in notch 3 or 4 to maintain 125 mph. In TSW, notch 5's max speed is around 140 mph, but notch 4s max speed is around 120 mph. To stay at 125 mph, you have to alternate between 1000 amps in notch 5 (too much torque) to 0 amps in noth 4 (too little torque). You either get too much power or not enough; there is no in between.


This bug needs to be fixed for all the routes, not just Rapid Transit and the ACS-64. Even though it is "playable," it's not one of those 'I can live with that, I guess' things; it is something that is pathetic to fix in one route, but then leave it alone everywhere else. This should be a universal fix, not a route by route fix.

Oh yeah, don't get me started on dynamic brakes.

[mumia]

İlk olarak Staalby tarafından gönderildi:

If you like trains. There is no excuse not to get it. Sure, it has some flaws and some bugs persists. But, trains man.. I may sometime complain, but it is always constructive. Again, trains dude..

I don't like trains with wrong physics. In fact it's worse than no trains at all.

[jorgen3 SE]

From physics input in TrainSim.com (5 weeks)
--------------------------------------------------------------------------------------------
The SD40-2 comes equipped with some kind of cruise control: notch1 for instance won't accelerate beyond 10mph, the amps just go to zero. Same with other notches: they all seem to have a max speed beyond that the amps go down. Don't know whether the SD40 actually comes with this feature.
Same problems in the AC4400-CW: "cruise control" Also the engines appear to be way too strong. It all feels like a train sim in arcade mode. Disappointed. At least the dynamic brake works better than in the SD40.
------------------------------------------------------------------------------------------------------
The physics are way off.
I should not be able to accelerate a mile-long laden train, powered by 2xSD40s+2xSD40s, up a 1.5% grade, when in notch 4 and cruising at 24 mph (just below the 25 mph maximum)!
The locos should be in R8 and working like buggery!
I arrived at the end of the scenario at Sandpatch Summit at 59 mph, after accelerating to a maximum of 62 mph at one point on the climb.
Ridiculous! It appears that they are using similar physics to those in TS20XX
------------------------------------------------------------------------------------------------------
TSW is going to be compared with Run 8...no two ways about it. So here are some comparisons:-
PHYSICS TSW. Can be summed up in yet another single word - unrealistic.
I expected more from this release but sadly, after getting into it for several hours, I find it disappointing. Controlling a heavy consist is almost like TS20XX but with pretty graphics; the same glaring mistakes made with the physics & dynamics of the TS series are there to haunt us.
Yet again, if this is meant to compete with Run 8 it fails miserably on that score.
----------------------------------------------------------------------------------------------------------
Pretty for sure BUT still needs a huge physics overhaul
Physics seems to be a back burner issue for the time being Oh well
Others have mentioned "cruise control".....the fact that there is no need to constantly adjust the throttle to keep from exceeding the maximum road speed. Just find the right notch and leave it there.
Then there is the completely unrealistic ability to accelerate a mile-long consist powered by four SD40s up a 1.5% grade to reach a terminal speed of 62 mph at Sandpoint Summit
------------------------------------------------------------------------------------------------------------
The problem with TSW as I last played it, is not a case of the physics being a % out on acceleration or a bit of over braking, you couldn't even move the train up the hill without it stalling. Brakes all released as it had been running reasonably well on the flat prior to that. But when people are reporting that the train splitting is not applying the brakes, let alone tripping the PCB as happens in Run8 things are way off course. If DTG wanted to put an arcade train game out with greatly simplified train performance that would have been one thing, but they billed it as a simulation.
---------------------------------------------------------------------------------------------------
That would be great if TSW stood for Tree Simulator World but until they fix the physics so we can run trains without them doing unexpected things,
------------------------------------------------------------------------------------------------
Yes, the physics are borked, which puts off the real hard core simulators
-----------------------------------------------------------------------------
Thanks for the reviews guys! I'm most interested in the physics portrayed by TSW. From what Vern has said here and others have said elsewhere, I'm not yet convinced that DTG got it right. I'm sure I'll buy in the future just to have it in the collection.

Because I'm a physics junkie, I spent my money on Run 8 v2 and was not disappointed. Run 8 physics and difficulty (level expert!) paired with DTG's beautiful scenery was kind of what I was hoping for with TSW.
--------------------------------------------------------------------------------------------------------
I think too many people are not looking at this objectively, but rather are being sucked into comparing it with their favorite, which already has been pre-determined to win.
---------------------
People want accurate, until the accurate turns into tedious things, and then they complain about the tedium and compare it to the shortcuts against tedium found in their favorites.
Yes...we want accuracy. We want simulated realism.
And when a new product is ballyhooed that is supposed to be the best thing for mankind since sliced bread, and it fails to live up to not only the bread, but also the clearly-obvious sim to which it will be compared, then yes, we get somewhat peeved.
How, therefore, are we supposed to look at it objectively?
As it stands right now, TSW has a long way to go before it can be considered a realistic or accurate simulation of railroading.
-----------------------------------------------------------------------------------------
Physics is definitely wobbly. Having re-loaded my saved game from where I'd just got the train moving and had dropped back to Notch 4, the train promptly ground to a halt again. This is on a section of line with no perceptible gradient. Once again I had to engage Notch 6 or Notch 7 to overcome the inertia and start the train moving, without any wheelslip occurring either (this is with the GP38's). After stopping at the end of Cumberland Yard to wait the oncoming train and signal to clear, same again
------------------------------------------------------------------------------
In Run8 you learn to anticipate the rollback by putting the lead engines on the independent brake and/or applying power, sometimes gradually up to Notch 4 while the brake release propagates towards the rear of the train. Having fenced your DPU properly, you can power up the mid train helpers or the rear end helpers to keep the train stalled while the brakes release, being careful not to lurch and break a knuckle because of slack action etc. etc.

Run8 has spoiled me in realistic freight train operations, and this is what I desire the most in TSW. Coding proper physics can't be that difficult since it is all governed by Newton's law IIRC.
------------------------------------------------------------------------------------------------
.....Run8 has spoiled me in realistic freight train operations, and this is what I desire the most in TSW......
Ditto, which is why I am so disappointed with what has been cobbled together with TSW.
---------------------------------------------------------------------------------------------------------
Well, no game or computer simulation is capable of providing 100% realism; we aren't equipped with multi-million dollar simulators....so that's a moot point.
DTG will need to tweak the program to improve the physics, not do it with add-on content.
-----------------------------------------------------------------------------------------------------
I'm finding the physics puzzling, and it seems not everyone is getting the same results.
I did one of the services with a long line of tank cars. Up the hill at about run 5 or 6 and was able to keep the speed up. Hitting the crest and seeing the profile change to downhill I cut the power and figured I would need to try dynamic braking for the first time. But nope, the train started to rapidly slow down and I was soon back at run 5 or so to maintain speed...despite the fairly steep downhill. That just makes no sense.
---------------------------------------------------------------------------------
No mention anywhere of the physics issues, unless of course that is tied up with the inner tech manifesting via the graphics.
--------------------------------------------------------------------------------------
I'm just worried that after all this time, with so much wrong and the fact they allowed it out in this state, they know or even care how to fix it.
----------------------
I share this worry too: DTG lacking the desire, competence and means to sort all these bugs out and tie the game together into a true simulation. Like with TS17, only DLC that was broken under the definition of consumer protection law was repaired, the active community took care of many of the minor issues. Some things that were broken at the core have never been corrected,
It is difficult to remain optimistic and it will probably come down to this: voting with your wallet. Buy their stuff and refund it when unsatisfactory. Hopefully some 3rd parties will put pressure on DTG not to slacken.
--------------------------------------------------------------------------------
Leaving it up to third parties is not a solution. The quality control at DTG is terrible, You can't have an Sd40-2 that behaves one way, then have the exact same loco that's designed by someone else behave completely different like DTG does.

TSW proved what I had suspected, DTG had been riding the coat tails of third parties to create content, now they're in the pool trying to learn how to swim.
---------------------------------------------------------------------------
There is one very good reason to build a game where physics runs on the GPU...
GPU development exceeds Moore's Law, getting close to doubling performance every 18 months, Whereas CPU development has not shifted much over the past five years.

As mentioned previously, I have no knowledge or experience of TSW but I can tell you it is possible to have a train use detailed , complex physics when close and a much cut down simple system when distance exceeds XXX
Sort of Physics LODs

Simugraph...... I wait to be proved wrong, but I cannot see it ever producing satisfactory results, The reason (and I've had this argument before) is error multiplication
---------------------------------------------------------------------------------------

[jorgen3 SE]

From physics input in UKtranSim.com (5 weeks)
-------------------------------------------------------------------------
Physics not reflective of real life, same as early railworks
------------------------------------------------------------------------
Potential physics bug
I noticed the other day while doing the fully fueled scenario that the required force needed to traverse the grade changes if you stop the train on the slope. I needed one or two fewer notches of power (for the same grade) to keep the speed at 25 mph before I stopped the train. The train feels more "sticky" when you've stopped it recently or perhaps more realistic, depending on what is the correct behavior.
Has anyone else noticed something similar?
--------------------------------------------------------------------------------
Disturbing findings
.1 This one is probably obvious to those that have played more than a brief test: there is no wheel slip modeled or at least the rail adhesion is so high as to make wheel slip nonexistent.
It's extra sad that this is lacking, since it was already implemented in TS20xx.
I would've hoped the physics would be an upgrade compared to the old DTG train simulator (like slipping on individual car wheels) and not a downgrade.

2. The friction of brake shoes against wheels does not change with speed, it's always the same. It might sound like an advanced or hardcore feature but both OpenRails and Run 8 simulate this.
I found this out by running a heavy coal train (like the one in Powering America part 2) down the east grade of the mountain with air brakes set so that only a small amount of dynamic brake was needed to modulate speed. When I wanted to stop I expected the increasing brake force of the air brakes as the speed decreased, to take over as the dynamic brakes faded. A real technique used by train engineers on mountain grades. Instead the trains deceleration leveled out and to stop the train I had to make consecutive large air pipe reductions, as if the brake force curve was flat with regards to speed.

3. Air pipe pressure change does not propagate but is instantaneous. What this means is that no matter whether your train is 100 or 15 cars long, the stopping distance will be the same as long as the total brake force of the train is the same.
Looking at the HOT Device (Head of Train), the rear brake pipe pressure drops instantly when a reduction has been made. This would suggest that there is no "travel time" for the brake pipe reduction to reach to end of the train but since it could be a bug with the HOT readout instead of a lack of mechanics I had to test it.

I took a coal train (the Newell - Brunswick service) and used the distance measure to brake from the exact same spots, from 50 mph, with a 50 car train and a 15 car train. The stopping distance was within 10s of feet. If the length of the train would have had an effect the difference should've been in hundreds of feet.
zaphr89 Getting the hang of things now Posts: 37Joined: Mon Jul 04, 2011 12:56 pm
--------------------------------------------------------------------------------------------------------
And have you informed DTG?....as interested as we all may be in your findings, we can't fix it.
--------------------------------------
Yes, of course. I've been in contact with DTG support for over a week. I'm not like some cynics I've seen (not necessarily on UKTS) who seem to almost want TSW to fail just so that they can be proven right.
The reason why I made this thread was more to share my findings so that those that feel realism is a high priority can decide whether they want to push/pressure/persuade (whatever you want to call it) DTG to make it higher on their list of priorities.

I'm afraid that DTG will take the first of the options AndiS mentioned, but hope not.
sundog: If the independent brakes were applied It's not wheel slip you encountered. It might be that wheel slip is implemented but that there are no indications of it happening, I haven't done "controlled experiment"; but I'm doubtful.
---------------------------------------------------------------------------------------------
I've got good news and bad news.

The good news is that my initial finding of the air propagation not being simulated might be incorrect. I tested a 40 car tank train and there seems to be a time difference between when the brake pipe pressure starts dropping in the cab compared to when it starts dropping on the rear. Even if I start timing after the brake pipe pressure has reached its full reduction in the front the reading on the HOT Device doesn't drop until about 1-2 seconds after that. Now the drop on the rear readout still seems to be instantaneous which I think is wrong, and even on the front the brake pipe pressure reduction seems too fast (I might be wrong here however).
Still, if the air pressure propagation is modeled the timings are much easier to fix.

The bad news is that I also tested the wheel slip on both the SD40-2 and the AC4400CW and it does not seem to be modelled, or the adhesion values are extremely high.

I applied the handbrakes on 40 cars (all of them on the SD40-2 train and 40 out of 50 on the AC44CW). With only one engine producing force I was able to stall the engine without any wheel slip occuring on the SD40-2 (1500 amps, redlining, and no slipping what-so-ever).
On the AC44CW there seems to be some anti-wheel slip system/wheel creep control because the amperage started pumping up and down when the engine stalled. There was never any slipping though.

Both of these tests were done in snow storm conditions by the way.

Also another thing, with as many as 25-30 hand brakes applied I was able to pull the train with a single AC44CW. That can't be right?
-----------------------------------------------------------------------------------
Railroad engineers perspective of TSW
-------------------------------------------------------------------------------------------------------
Interesting video on the sim, graphics, performance and towards the end the braking system.

https://www.youtube.com/watch?v=0rZgzqXeCck
https://www.youtube.com/watch?v=Oz0-_3oR7TQ

------------------------------------------------------------------
One wonders why DTG did not have a loco engineer on the TSW testing team. Or if they did, why they didn't listen.
-------------------------------------------------------------------
From 1hr 19min to the end he just rips the train braking and safety systems apart with the passion of someone who knows how it should work. Like he says DTG have modelled these systems and have proudly announced that a lot had been sorted since the beta but from what he says there's a lot of stuff just not happening and the stuff that is happening just isn't right. The accelerometer is obviously totally screwed, the emergency function isn't right, there are audible and visual warnings that aren't working?
--------------------------------------------------------------------------------------------------------
No need to watch another video to find out he's a '*proper*' railroad engineer, you could have just taken Markjudith at his word.
----------------------------------------------------------------------------
Lol. The problem for me is that DTG lauded this as a sim and built up the 'real as it gets' hype previous to release. I actually feel let down.
You don't need to be an engineer to know after playing for a while basic brake and throttle physics are borked. Half the time you apply the auto, the ER forgets to drop, and somehow an initial reduction ends up dropping the pipe 20psi bringing the train to a heap.

I tried the coal train, I transversed half the route before the brakes at the rear released.
The amperes drop to zero when throttle is open. Level grade, train losing speed at 35mph in notch 4 because (although the engine in audibly revving) there is nothing on the ammeter.
If they can't get that right by now lord help them.
--------------------------------------------------------------------------------
This game has got incredible potential. At the moment what we have with CSX feels like very good tech demo which didn't need too many fancy signalling systems, train management and comms with signalers. Sand Patch CSX is perfect for that. A setting to show off the improved physics and 'joy' of simply lugging 100 boxes over a gradient while enjoying a new virtual environment.
---------------------------------------------------------------------------------------------------
Yes, I do work for a railroad. My official title is a Mechanic/Engineer, so I work on them and run them depending on needs. If we are short on crews, I run trains, otherwise I'm in the shop.

As far as the simulation goes, the brakes are terrible, the throttle is wrong and the physics remind me of TANE. Basically what stevegreen has said. I've never seen an ammeter go to 0 with the throttle in notch 4, that just doesn't happen. I haven't tried the dynamics, but I've heard they aren't much better.

DTG had a working 26 brake script is TS, but they've thrown in out, it was developed by a third party. The one in TSW was built from scratch with the old code as reference. I have no doubt that they will eventually get there, but they've shot themselves in the foot with this game. It was supposed to be all that and a bag of chips, but the only thing that's obvious is they were really not prepared to make a game from scratch.
---------------------------------------------------------------------------------------------------
Hi ProfOleander, welcome. Many thanks for your video. Look forward to the next one. If you get chance would it be possible for you to leave some comments on the TSW steam forum at all? DTG are apparently reading that.

Re the auto brake, after a little more testing I noticed a pattern to the train pipe pressure drop. Apply some brake, bail and then return the auto handle to release. *Without waiting for the needle to return to 90psi* re-apply an initial reduction. The needle drops 5 pound off the current BP pressure (eg, if the BP is at 80 psi, and you put in an initial reduction the BP drops to 75) . Apply more some brake.. BP overtakes ER, and BP eventually drops to zero regardless of how much service brake is applied.

This also happens if you apply, release, then re-apply the auto without bailing off. Easily tested on a light engine where the brakes release quickly.
--------------------------------------------------------------------------------------------
DTG has seen the video, but I may wander over and post there.
Someone else that I've been talking with, that has scripted for DTG on TS, mentioned the same thing about the brake behavior. The game is really all over the place and I really wonder if they actually developed in a bubble. It really does seem like the input from the beta was thrown out the window. I know that at least half the stuff I mentioned in the video I did on the beta is still in the game.
--------------------------------------------------------------------------------------
I think the big issue here is, that DTG & Matt are claiming to have very accurate (brake) physics in the new TSW engine. They are claiming something that they can not yet actually live up to. (And eats a big chunk of performance to boot).

So if we can all live with them not having such accuracy in physics calculations, well fine I guess, perhaps. And let's move on and focus on DLC, accuracy of operations, modelling, signalling, etc. And drop the claim on accurate physics simulation.

But they do want to have that accurate simulation. So they can be held accountable for the claims they make and don't deliver.
-----------------------------------------------------------------------------------------------
It's too early to write TSW off. It has a powerful physics engine under the hood and if DTG can't tune it properly then I'm sure those same experts will step in.
------------------------------------------------------------------------------------------------------

[Staalby]

jorgen3 SE.. The hero we need.

[josef55]

jorgen3 SE.... excellent..

[pschlik]

At least there is a lot going on with this...DTG better take notice before people lose interest.

[bigbadvuk]

I doubt they can ignore it.. People will loose just interest and 40$, but DTG has to loose much more.

But they cant fix it in 1 patch for sure... It will take a month or two.

[jorgen3 SE]

From physics input in Railworks America (5 weeks)
-------------------------------------------------------------

Same Old Throttle Flaw?

Unread postby SD40Australia » Sat Mar 18, 2017 5:35 am
One small (well not so small actually) detail is the throttle.

On the SD-40-2 I noticed that once 25 MPH is reached in Notch 3 it no longer registers amps. So in Notch 4 you get a surge of power, but if dropping back one notch it goes to zero. It shouldn't. I am wondering if the same algorithim was used to calculate power and throttle since the TS2017 or Railworks.

A notch is not 'speed centric' it is 'power centric'. In other words as speed increases for ANY notch on a freight locomotive the amps should decrease, but not to zero and then having to go to the next notch and you suddenly get 400-700 amps.

US Railroad Engineers should be able to verify this.

The engine is still producing power and generating AMPS and the TSW is congiured wrong if the Amps show 0 in Notch 3 at 25mph. Going to Notch 4 should only increase the amps output by 10-20%. So if Notch 4 produces 400-500 Amps then Notch 3 should produce around 350 Amps approximately.

Take a ride on a real train.
--------------------------------------------------------------------------------
Dude no you can't. Starting movement even on dry rail with a Geep or even SD40-2 or advanced AC traction locomotives. You're going to be grinding and hitting the wheel slip protection until you're going about 5-6 MPH if you're trying to lay down more power than N5.

That ammeter will be bouncing back from 11-1200 amps until the thing can finally bite in, and it only gets worse with wet or greasy rail.
TSW doesn't show any semblance of wheel slip even in snowy wet conditions, or auto sand, or power cutting.
-------------------------------------------------------------------------
why do I suspect that the brits developing this, are using youtube videos to develop physics on US locos? Would be curious to know if they even consulted anyone involved with OpenRails on how they did their physics. But I doubt they did, because they seemed hell bent on keeping all development in house, even deciding not to use searchlight simulations crew to do the sounds. But I see on FB, that the searchlight guys are already on the ball in a few areas with TSW....all on their own accord mind you, it doesn't appear to be at DTG request. I bet anything, this release occured, because they knew they ran into a brick wall of problems, and instead of giving the non DTG people a chance to get a piece of they pie, they are hoping 3rd party people will fix their stuff for free for them. Much like how MSTS was fixed by third party free help both before and after abandoned by MS.
------------------------------------------------------------------------------------------
At least the Dash 9 was able to start its train without having to melt the traction motors in Notch 7 as opposed to my earlier experience with the GP38's, but this of course was followed by the excessive acceleration. Absolutely no sensation of "heavy haul" at all, accelerated like a Class 158.
----------------------------------------------------------------------------------------------------

Why would DTG talk with the Open Rails People or Run8 people? I could just see Paul Jackson asking OR if he can see their physics.
-----------------------------------------
OR is open source. Anyone can download the coding and look at it and make or propose modifications, or do mods for their own cause and not share it. It would have been easy for DTG to take a look at their coding, and even read the forums that have lengthy posts on how they were getting the physics improved.
-------------------------------------
No but physics are physics. And that does not change. The only thing that changes is coding depending on what one uses as a base code. But the physics that need to be coded, those do not change. Between TS, and now TSW, it is very apparent they are taking an MSTS approach to arrive at some form of nutty physics coding that they hope will simulate the feeling. Don't know how these programmers after what, 9 years, can't even begin to get close with the physics. The only people close to getting it right are the 3rd party devs.
--------------------------------------------------------------------------------
Well here's an oddity, which maybe shows the physics are all over the place. Running one of the AC4400 hauled intermodals in Services, as noted earlier initially took off like a bat out of Hell. However, once I hit the climb up to Mance (around 1.5%), after slowing for the 25 MPH section speed continued to drop eventually under 10 MPH, then under 5 MPH all in Notch 8 then wheelslip then I started to roll back. Sanding didn't help. This is also at odds with other reports I've seen of having to throttle back on the hill. The only thing I changed during the run was to increase the tree quality from High to Ultra which, ironically, boosted my FPS by around 3-4 FPS. However I can't see how this would have borked the physics calculations.

As I said, something very strange going on under the hood of this game.
------------------------------------------------------------------------------------------
Previously in TS17 cars loaded in a scenario didn't always put on weight, making visibly loaded trains very easy to handle.

It takes considerable time to release the brakes on longer trains in TSW. Check the brake pipe pressure at the EOTD (Shift F3 will take you to the end of the consist and there is a LED readout) or in the box on the control stand. With the brake handle in release position, put the throttle in Notch 1 to pump the air somewhat faster. Unfortunately EOTD's appear and disappear as by magic, haven't quite figured out how and when ...

Sandpatch grades are sufficiently steep that dynamic brakes alone aren't sufficient to hold back a fully laden train. Dynamic brakes work best in the 20-25 MpH speed range and train handling rules require a certain number of dynamic brake axles per ton of train, with a maximum axles number to prevent drawbar failures. So it is necessary to put a minimum service reduction on the train just before cresting the mountain to reduce any excess speed above 20-25 MpH. And use dynamic brake modulation to gently ease the train down the grade at no more than the equalising speed. (This varies per locomotive type, AC traction fares definitely better than straight DC here.)
When maximum DB power isn't sufficient, brake pipe pressure is lowered in small increments to find the equilibrium. When full service reduction is reached, the train must be put into emergeny and secured with a sufficient number of handbrakes once stopped. Then, more DB power must be added, preferably at the rear, or retainers must be set on the cars before the train can proceed safely down the mountain.

Since the minimum service application cannot be lifted before the train has come down the grade for the real and present danger of multiple applications depleting the air reservoirs, one sometimes has to pull the train down the less steeper parts of the descent

So long and heavy freight trains descent the mountain at much lower speeds than the track limit.

I'll try to find a copy of CSX's Air Brake and Train Handling rules, or any employee timetable/rulebook. CSX's recent publications are very hard to find online though.
----------------------------------------------------------------------------------------------
Completely Incorrect throttle/amperage modelling

Postby sd40Driver » Tue Apr 18, 2017 9:39 am

Hello all,

First time poster here, long time reader. A little about me. I am a second generation railroader, been working for a large class 1 railroad as a locomotive engineer now for 7 years. I purchased this simulator because it touted its realism, and I wanted to test that. Largely, it fails an nearly every area. The old Microsoft Train Sim was even more accurate in many areas. Because I have real world experience, I thought I'd share my observations.

Aside from the incorrect braking modelling (wrong in so many ways it's hard to explain in a short forum post), the throttle modelling is what absolutely unequivocally kills the immersion for me. First, the locomotives load (traction motors see current) WAY too fast, and start moving (overcome inertia) way too quickly. Completely wrong. Next, the amperage falls off to zero in between notches. For example, each notch (say, notch 1) has a preset "zero point" based on speed. For notch 1, once you reach 10 miles per hour, the amps go to zero in the game.

This is completely in violation of the basic theory of how diesel electric locomotives work. Amperage can only reach zero two ways - if the unit is transitioning to dynamic braking, or, if the locomotive is off. Even at 60 miles per hour, if you place the throttle in notch 1, you'll still get a tiny positive amp reading to the traction motors. You probably wont be able to see it on the gauge. And it wont be enough to hold that speed, and you'll slow down, but you should EASILY be able to reach 30 miles per hour or so with light power in notch 1. This is the case for each notch as modelled in the game. In my view, this absolutely ruins the immersion in the game. It gets even worse at higher speeds.

On mild curvature track, with no grades, it is common to run an average train with the power in notch one or two at track speed of 50 or 60 mph. But in the game, you would have to be in notch 5 or 6. In notch 5, you have zero amps, but go to notch 6 and it shoots way up to 400 or so. So in the game, you're constantly bouncing between decelerating and yanking hard. In real life, would certainly break couplers, and possibly put cars on the ground. Totally wrong. This isnt how any real life locomotive works. In real life, the amperage falls off in an exponential type curve. For example, last night I had a single SD40-2 on level track and in notch 1, at 10 miles per hour, I had about 160 amps. At 15 miles per hour, it was about 120 amps.

There are so many other flaws, and not minor nit-picks. Major issues, Like the fact the ammeters dont even register negative when you go into dynamics. Or the fact the accelerometer is useless because it bounces around so much. Emergency brake applications dont trip off PCS. Brakes dont charge up right and the brake pipe values are wrong during application. When you shoot the brakes from the FRED, it doesnt cause the locos to register an emergency application. None of the locomotives sound like the actual locomotives at all. The EMD's have whisps of the right sounds, but the GE is completely wrong. On and on.

Just a few thoughts from a guy who does it for a living.
--------------------------------------------------------------------------------------------
If open rails can get relatively decent physics, why can't DTG? DTG has never given a poopoo about proper physics. They are brits, trying to model American diesel electric physics, using manuals the found on ebay. Between TS20XX and TSW, they are banking on the common end user not knowing the difference in how an american loco behaves. Who cares if it can go from 60mph to 0mhp almost as fast as a 2017 zl1 camaro. The common user doesn't know the first thing about amperage and how it's supposed to behave and look on the gauge. Green is good. Go with it. I also clearly remember when they were touting the amazing new simugraph, that someone at DTG commented somewhere that they had a realworld, sd40-2 operator, try the TSW sd40-2, and that this "experienced" sd40-2 operator, called simugraph a success, and that the sd40-2 in TSW was as close as you could get to the real thing. That all turned out to be a load of over the pond manure for sure.
------------------------------------------------------------------------------------------------------

[jorgen3 SE]

Thank you for all the good response I received.
Has been perhaps too harsh in criticism. But I would like them to succeed with this.
If they get to the physics then it will be a dream train simulator.
Have some more info I can add tlll next weekend.

[Captain1]

Great work Jorgen,

all we now need, besides the required fixes, is for an official recognition from DTG that this is a problem and that they do intend to fix it.

DTG have been pretty quiet about this issue to date.

It's about time we had some clarification!

[SteakOplomb]

Yep, I would love to hear what's comin'up our way in regards to these issues.

It looks like the locos behave in a similar way as the diesel-hydraulics in TS, although, it is far worse in TS since the diesel-hydraulics won't even retain their acceleration momentum. They will instantly stop accelerating when the engine and transmision RPMs match, instead of having a smooth drop to zero torque being transferred to the wheels, and have the acceleration smoothely decreasing as the momentum drops down.

Anyways, it almost feels like DTG did not apply the right "traction profile" for the diesel electric locos in TSW, and slapped the one meant for diesel-hydraulics instead.

[pschlik]

Actually, that is a really good point! Reminds me of a quote from the Simugraph dev diary where someone mentioned modeling a diesel hydraulic locomotive. That confused me since we are only talking about diesel electrics here, so I pushed that off as a hypothetical situation not referring to CSX Heavy Haul. But....Maybe that was referring to CSX Heavy Haul and whoever did the physics work was dumb enough to think a diesel electric worked the same way as a diesel hydraulic.

[Armorin]

It's hard to think that DTG lacked the knowledge to get this wrong. Some of them have been working on train simulators for a decade.

[Randorian]

Good discussion here. I hope that DTG takes note and gets the core simulation aspects right for all future DLC to share the same baseline as intended.

[Armorin]

İlk olarak jorgen3 SE tarafından gönderildi:

They must know that the physics is almost completely wrong in TSW

What does that mean. Are they completely satisfied with the physics of the simulator and do not want to change anything.

Train simulator then it must behave right close to reality if it does not, it is a train game.
Right now, TSW is a Train game not a simulator.

I'm afraid this selection of your quote is the answer. No update fix on the horizon. They are already working on the next big thing by the sounds of it.