Update Mar 6, 2023 - added Mount Shasta fix Update Mar 1, 2023 - added DTG GP40-2 Pack fix Update Jun 26, 2022 - added DTG EMD SD9 D&RGW coupler fix Update Jun 15, 2022 - added 100 ton wagon fix for Soldier Summit coupling issues Update Jun 11, 2022 - added DTG NS GP60 Update Feb 23, 2022 - added DTG SD80MAC Update Jan 17, 2022 - added Clinchfield, Feather River Canyon, N-Line, Tehachapi Pass, Hudson Line (scroll down) Update Jan 14, 2022 - added Norfolk Southern Coal District (zero slack) As discussed in a recent thread (https://forums.dovetailgames.com/threads/es44ac-on-sherman-hill.29261/) I looked into the issue with excessive consist breakups, especially on Sherman Hill and Donner Pass. Cajon Pass is also affected. All reported consist breaks can be traced down to the same blueprint parameters. The couplers here had values of zero for slack which leads to excessive force buildup on the couplers and causes sudden consist breakings. The Kansas-Topeka coupler was also missing a spring/damper coefficient which causes the bug that a coupling task is not recognised by the game engine. This has also been fixed. The parameters I used are standard parameters for the Buckeye Type-E coupler, used before by RSC/DTG/VNHRR/HIS etc. in many of their products and proven to work well. This furthermore adds to handling of rolling stock being consistent across different routes. The parameters used for the fixed Buckeye/Type-E are the following: Strength: 150 MinDistance: 0 MaxDistance: 0.1 TargetDistance: 0.05 SpringCoefficient: 15000 Damping: 12000 MaxForce: 2.89134e+006 For those who want to install, use Railworks/Utilities.exe->Package Manager. Be sure to press the Refresh button before installing. The fix will be installed next to the default .ap files so original data will not be overwritten. Note: This fix addresses the standard rolling stock shipped with the DLCs, not 3rd party addons. As Cajon Pass uses the US Loco & Asset Pack this fix automatically applies to the other routes which use that pack. If you're familiar with TS's data structure you can install faster by opening the .rwp with 7zip and just quickly copy the files over. As the forum does not allow .rwp files to be uploaded, I have zipped them. Enjoy, and feedback is welcome.
DIFFLOCK, you remind me of my job. Used to assemble these truck differentials / axles in Gaggenau... quality assurance is important there. Nudge nudge
In real life, Buckeye/AAR couplings have 3.25 inches of play in each direction. That's why my Buckeye couplings (including those in the B&O DLC by HIS and friends) have the following values: Pivot type: Mid point Min distance: -0.08255 Max distance: 0.08255 Target distance: 0 For authentic slack action, it's also very important to have the correct spring coefficient, calculated using Hooke's Law. It comes out at 5.691622E+07. I'm pretty sure that the default value of 15000 is a cause of problems in some other US DLC. 15000 is probably alright for British hook and chain couplings but not for Buckeyes. I also set the Damping to 1E+08. Another important point is that derailments can be caused by incorrect values for the pivot points in the wagon blueprints. First of all, the distance from the front (or rear) edge of the collision box (which is collision length divided by 2) to the front (or back) pivot X should be 0.375. For example, if the collision length is 12.4, the Front Pivot X should be (12.4/2)+0.375, i.e. 6.575 (and the Back Pivot X would be -6.575). Secondly, the front (or rear) coupling pivot X should be 0.75 less than the Front (or Rear) coupling pivot X (note that the X is contained inside a position and orientation matrix). If Front Pivot X was 6.575, the Front coupling pivot X would be 5.825. Care also has to be taken with the collision boxes to make sure that they fit tightly around the actual geometry of the wagon and don't clip each other on sharp curves or when going through sudden, severe changes in track gradient.
Thanks for sharing your well researched knowledge. Is TS correctly implementing the spring/damping coefficient or does that value have to be adjusted for the game engine? 5.6E+7 seems pretty much, I've seen values as low as 750 in Canadian Mountain Passes for example. I will test your coupler for private usage, as I decided to use the well-tested parameters for upload here to fix the urgent issues. So far I haven't run into any breakups any more. And if you're in contact please send greetings to my favorite Youtuber Professor Oleander https://www.youtube.com/channel/UCtVPTOhRc65JEsTEpgGZoUQ. I take it he was in the dev team for the great UP119 DLC?
Judging by the experiments I've done in TS, I'm fairly confident that it does indeed implement the coefficients correctly. So it does pay to put in the proper values. The spring coefficient is a measure of the stiffness of the spring. If F is the force (in Newtons) needed to displace the spring by x metres from its resting position, then the spring coefficient is F/x. I do try to keep in touch with Professor Oleander. He helped with beta testing on the UP119 and gave me a lot of valuable input from his experience as a fireman on steam locomotives.
Again thank you for that very valuable information. Like I said I had to make a compromise between ultra-realism and what is widely used and free of reported issues. But interesting what TS is really capable of. Thanks to Kuju who coded the base engine which is still in use, slightly modified albeit. If you look at MSTS train .eng files you see those values already in there. And folks like you or Searchlight Simulations, Bossman Games, AP pushing it to the limit and showing what is possible. Kudos!
I want to use Smokeboxes' buckeye as a reference/standard coupler on my game, and then upload a pack using his parameters. I am still a bit unclear as for the negative min values, is there adjustment on the cars' and engines' .bin files needed to properly work with the coupler? And if yes, would shifting the values to min 0, max 0.1651, target 0.08255 work for unmodified stock?
The cars' and engines' blueprints would need adjusting to make sure that the distance between the pivot X and coupling pivot X is 0.75, as well as to ensure that the pivot X is 0.375 away from the edge of the collision box. It's better to set the coupler target (where the coupling is meant to come to rest) to 0 and have the min and max as 3.25 inches in each direction, because otherwise they wouldn't come to rest at the same position as the car's front and back pivot. Ideally the point where two buckeyes are coupled and in equilibrium (not stretched or compressed in their sprung sockets) should be exactly the same as the pivot X, and that's what you achieve by setting the coupling target to 0.
As adjusting all cars is not what I want to do for a global fix, and it kinda "feels right" now I'll stick with my parameters at the moment. HIS used them in CSX Hanover and they just feel fine. Will continue testing though, thanks Smokebox.
Back in 2000 or so, Codemaster released an offroad racing game called 1NSANE (they did use a number 1 rather than a letter i) why had multiplayer options. I was hooked and thus I wanted a username to reflect genre - so DIFFLOCK. It also became my handle on some Land Rover forums I used to frequent when I owned a Discovery 2 TD5 (ironic that the Disco only used traction control, but to good effect I might add!). It's also part of my Xbox gamertag
Update Jan 17, 2022 - Added more coupler fixes As there is only five uploads per post allowed, here's more fixes for the same issues. I remember having a consist split in Feather River's Keddie Consignment scenario, not knowing why back then.
Hello torfmeister and DTG Jamie I am vouching for this fix and recommend updating the default couplers in the mentioned packs. I have had regular issues in the past across much of the content, frequency varying. Since installing the override I have driven 16 freight hours over Donner Pass without a single train separation (1-3 would be due). Maybe less slack is enough - just to reiterate, the problem isn't normal operation (weight, grades, handling), but some arcane issue causing parts of the consist moving different distances, causing either a separation or a collision. There might be residual failures based on theory I'm not scribbling now It happens in real life anyway.
Added DTG Norfolk Southern GP60 (all zeroes...). This fixes the issue with the front coupler not working (TargetDistance and loco collision box didn't match) What a beautiful loco this is... Functionality, Cab, Model and Texturing are on par with TSW. One of DTG's best! Nice not having to use the Gen. Field switch to turn on the wipers...
I have updated the GP60 fix. It has some more issues with the collision box. The Cold&Dark version had strange coupling matrices, I've worked on that too. A user on the steam forum noticed it would not couple a SH tanker to the front. I verified this issue. The collision box needed an offset of -0.2 to be centered between the coupling pivots. Because I had to move the center of the collision box, the cabview is affected - you are now sitting 20 cms back of the default position. A cab view update which fixes this is included. One additional note, I found those sunny/cloudy autumn scenarios on the N-Line to be best looking if you adjust AmbientLight to -3, Contrast to +2 and Sunlight Intensity to +4 in the Ingame Options, because of its bright TimeOfDay colour defs.
A long standing issue has now been fixed. Many players reported issues with the [GP9] 5. Roper Yard Switcher scenario on Soldier Summit - the 100 ton cars would not couple up. Another scenario in the EMD SD9 pack is also affected. I've had a look at the wagon's blueprint and used Mike "Smokebox" Rennie's method of adjusting the pivot points. I've adjusted the coupling pivots to be -0.375 away from the collision box edge, and the front/rear pivots (=receiving points) to be +0.375 from the collision box edge. By default the distance between the front pivot and the collision box was only 0.2365 metres (~9 inches). That's the reason the collision sound instead of the coupling sound was played... The default coupler blueprint already uses the parameters I use, so no fix for the buckeyes required. Bug seems to be gone. Happy shunting!
DTG's EMD SD9 used on Soldier Summit. Just saw a split consist screenshot on Steam forums. Here's the fix.
DTG GP40-2 Here's one that fixes derailing issues due to the coupler pivot being too close to the collision box of the loco. Also enables loco to loco coupling which is impossible in default state. https://forums.dovetailgames.com/threads/dtg-gp40-2pack01-coupling-problems.65556/ The collision box had to be reduced in size to keep the couplers at their (correct) position as opposed to the RWA fix which moves them out.
I don't want to go into these discussions about "stretchy" couplers... and I think this forum is more active. I find the RWA forums confusing. I'm considering putting up my own website, wixsite or something.
Mount Shasta also has almost no slack, making it impossible to pull off a heavy train from a grade. Here's the slack back DTG Jamie Can this thread be pinned? A lot of issues are adressed here and proven successful. (Edit: Ignored again by our great Community Managers.)
^ This! You get my vote for Train Simulator Classic MVP, and a single location to find all your mods would benefit the TSC community immensely! Dovetail should put you on their payroll I've downloaded all of your coupler fixes from this thread. Thank you so very, very much!
Just some notes to add here. As I started replacing the couplers to get rid of the consists split successfully, I'm now going over locomotives as I play to relocate the couplers and fix their pivots and collision boxes incase they're really off. There's some really "stretchy" couplers out there - but this is not due to the distances specified in the couplers' blueprints - but due to wrong distances between the CouplingPivots and the PivotX's in the engine blueprints. A fine example is Arizona Divide's Dash8-40BW. Couplers stick out a bit too far when uncoupled and seem to be stretched excessively when coupled - this is due to the PivotX being too far away from the CouplingPivot. I've experimented with that one - if you want to edit yourself here's the correct values (CP, PX, CL: ±9, ±9.75, 18.75) Default: Fixed:
Torf - The same provider (JL) also did Raton Pass which has Dash8-40BW and they have similar numbers (not exact - puzzling). Edit: The fix works on the Raton units as well and if they are intermixed.
Thx John, I'll go through each locomotive. Instead of posting all the modified engine.bins, I'll post an Excel sheet with accurate values for each unit. Will do JL stuff first.
Yes the Pepsi Can and the Dash8-40BW go with these values. For Ratons 40CW use: FrontCouplingPivot 10.3 RearCouplingPivot -10.1 (yep it's off-centered, but that's not a big deal. Calculate the collision box based upon the smaller value - moving its center is also an option but that affects cabview, so I leave it off-centered) FrontPivotX 11.05 BackPivotX -10.85 CollisionLength 20.95 [ 10.1 * 2 + 0.75 ] While Shasta came with zero slack couplers, JL used correct blueprints on Raton for the coupler itself (like mine).
Yes, as the Buckeye is of type "eMidPoint" - receiving pivot matrix is needed for 3-link couplers and is ignored for the buckeye.
Might be worth talking to Steve or Jamie to sticky some of these threads and maybe move them to their own subforum. I have saved them by watching them. I keep that list clean. Some others (the All Aboard specifically) could be resolved by a content patch, but this one is quite generic, useful even if supplied as a content patch.