Loco Stall Repro Test Route and Session

I believe I have been able to reproduce the locomotive stall bug that myself and others have seen.

I have uploaded a .zip file of a simple route and session using all built-in assets.

https://1drv.ms/u/s!AkMIc8G3YArLgq8ZI1egBPhTA77IOw?e=qf9IFA

The task is simple. Using DCC mode, spot each of the 5 tank cars on a separate yard track and return the loco to the starting position. I use the on screen icon to uncouple so do that please. In 24 of 25 times running this session the loco will stall at a random point after uncoupling from the tank car and moving a short distance away. Turning the DCC dial will increase the amount of smoke and noise from the loco but it will not move. If you use the 3 dots to change to Cab Control then you find that all the settings are set correctly and the train starts to move. Stop the loco and then return to DCC mode and you have control again. The bug may repeat again if you choose to continue but it is random at which point it shows up.

It does not appear to matter what order you choose to spot the tank cars or on what track you choose. The bug always appears after uncoupling and the loco moving away even if the loco is still coupled to the other tank cars.

I wanted to keep this as bare bones as possible so I did not include signals or trackmarks but I suspect that this might be the same bug that effects driver commands.

If anyone would like to test my findings please feel free. I created this in the latest version of Trainz Plus build 117669

William

Sorry William just lost my whole post as I got logged out while writing it. About 1 out of 10 times that happens I can recover it but no joy in Mudville tonight. I'm not a happy camper here on the Trainz forum.

I just ran your session and my results are a bit different. I only stalled 3 times in spotting the 5 tank cars.

I don't think this is a random event. As I have seen it on my own MR of the EBT that I'm building in TRS19 and on a MR by Phil Skene. It happens when the loco stops on a curved section of track with a small radius of curvature. In DCC mode the train resistances in the past haven't affected the pulling force of the loco (as in counteracting the tractive force supplied on the rail). It's affected the rate at which the train reaches a new speed set by the control dial but not the speed that's eventually reached.

I'll add that the center point on the train car is the important point. If it's located on the curve the car is affected. In my experience in testing Trainz physics, the track conditions at the center of the train car affect the calculated resistances. The curves thru your turnouts show a 30m radius in Surveyor. On my EBT (HOn3) model railroad design the min curve is 24 inches or 53 m in my Trainz version. When the loco was stopped on these curves, I could not restart it in DCC mode.

With TRS19 (at least as early as version 100240) it seems the curve resistance does affect the pulling power in DCC mode. It seems it does not take much momentum to keep a train moving thru the curve but if it stops on the curve it can't restart. So not real physics here as we might expect but some hybrid bug. If it is using a calculated value of the resistance on a curve it is an excessive value as a loco with a realistic power output based on its espec will start on the same curves in CAB mode.

In DCC mode I can run a train up an almost vertical slope in total defiance of real physics but the same train can't start moving on a small radius curve that it can start on in CAB mode.

This is a bug that I thought was fixed in TRS19 SP5. I guess I will have to go back and re check SP5. This bug could kill Model Railroading in Trainz among other things.

Bob Pearson

Interesting, I'll try a smaller loco like a SW1500. I did notice that the switches are way to sharp for my taste. That is the pre-fab RH Yard Ladder set to 100 radius. I could not find a way to make the yard longer without making it wider. If you got 3 stalls in one try that is higher than what I was getting. I would often get just one stall per try but I was varying the distance the car was spotted down the yard track so sometimes the loco had a longer run at the switch. I only had one clean run out of 25 tries.

Thanks for giving it a go.

William

wreeder said:

If you got 3 stalls in one try that is higher than what I was getting...
William

Click to expand...

Sorry I wasn't clear but I only had 3 stalls total in spotting the 5 cars. Each spotting movement included a backup up, stop, uncouple car, pull forward and final stop beyond and clear of the turnout to the next ladder track so it could be switched. So in 5 of those movements I had 3 stalls total. The only times I stalled was when I stopped the loco on a turnout's curves track. Not surprising as the loco represents 74% of the initial consist weight. But in DCC mode it should not happed period.

Your curves are tight but if you look at actual turnout construction several real turnout sizes fall close to that.

AREA standards for #4, #5, #6, #7, #8 frog turnouts have closure rail radius of 33.7, 53.1, 80.9, 110.4, 148.6 meters respectively.

I haven't taken the time to see what the min radius in Trainz DCC mode is currently in order to avoid this feature as I see it as a bug that has to be fixed. Your test route is TB 5.1 but I can make a quick re-check on my MR route with the 53 m radius min curves again in TR19 SP5 before I take a closer look at this.

Bob Pearson

You are totally correct with the engine always stalling on the curve of the turnouts. I would say that in 25 times of running the session I had a total of about 35 stalls total out of 125 times of spotting the car, uncoupling and backing the locomotive away. I try to keep the speed of the loco constant at around 5 mph. So I agree this appears to be a bug.

I re-did the whole setup this time using 80m radius on the prefab yard template and got a totally different yard arrangement with beautifully smooth turnouts. I'm still learning to use the prefab templates so I don't know why the radically different results. I probably did something different without knowing it. Anyway, using the built-in SW1500 loco I experienced no stalls on the much wider radius of the turnouts. So I will now re-run the tests with the original layout and the SW1500 loco.

William

Thank you, @wreeder and @RPearson for looking into this stalling on the curve. It is a most annoying bug. I am trying to get a 2ft/600mm industrial diesel around an 18m radius curve, but it keeps stalling under DCC control, as well as Cab control. Such small-radius curves, however, are typical of narrow-gauge industrial layouts, so I would like to keep them.

I understand that there is currently no config or script modification to work around this bug. In that case I'll keep my fingers crossed for it to be resolved in a later incarnation of Trainz.

For the record, I filed a bug report ticket about this, and within a day got a reply from Zec at N3V Games (thank you!):

N3V Games (Zec) said:

(...) it appears that the locomotive may not have a suitable amount of traction to handle the resistance within such a sharp curve.

In this case you may either need to try using a different locomotive, or try using a different enginespec asset on this locomotive (...)

Click to expand...

If the cause is in the enginespec, that may explain why the effect occurs both under DCC and Cab Control. I'm assuming that both forms of control use the entire enginespec. I had no problem replacing it with the specs from another loco, which solved the stalling problem.

RobKi said:

For the record, I filed a bug report ticket about this, and within a day got a reply from Zec at N3V Games (thank you!):



If the cause is in the enginespec, that may explain why the effect occurs both under DCC and Cab Control. I'm assuming that both forms of control use the entire enginespec. I had no problem replacing it with the specs from another loco, which solved the stalling problem.

Click to expand...

This makes sense. I remember encountering this issue way back when people were "fixing" content as part of the original content repair program. Someone "fixed" a bunch of Alco RS-3 locomotives making them totally useless because these switchers would get stuck on curves when switching all the time.

Increase the curve radius to at least 125degres.

John