A possible test case

[lindsayts]

There has just been released in Australia a freewhare route that OR may find usefull, the rolling stock that is. It is the "Western Australia - BHP Billiton - Mining Area C to Port Hedland" availible from "http://msts.steam4me.net/routes/wa_bhp_mac.htm".

The route is an 80 megabyte download with the rolling stock another 22 meg. The route is realtively small download inspit of being over 300 kilometers in length as route is in a _VERY_ remote area of Australia where there is little other than the mines and the railways.

The point of interest is the trains these have a gross weight of around 30,000 tons and the consist is 2 GEC Dash 9's with 104 are wagons, another 2 locos and another 104 ore wagons, total consist being 4 locos and 208 ore wagons, total length in the vicinty of 3 to 4000 metres. MSTS handles this OK although the frame rates on a fast system being around 22 fps. OR will not move the train (even empty) at all, the locos at the head move a bit to take up the coupling slack and then halt. It is likely OR is not controlling the mid train locos.
In MSTS it is handfull to control as one must be very gentle on the controls and wait to see what happens particularly with braking as if you apply these to much you will lose control of the braking as the auxilary resivours in the ore cars take AGES to recharge. It surprised me some what that MSTS has coupling slack and also correct sounds of the couplers taking up along the train, the lead tocos moving around 10 metres or so before the rear of the train starts to move.

Lindsay

[Donell]

Finally, thanks for the heads up. I have been waiting on this route for a long time now, will try it out in OR when i get the time and see if i experience the same problem. A big thank you to the people who worked on this route, your hard work is very much appreciated

Donell.

[lindsayts]

Maybe I have not made myself clear here, sorry.................

The reason I put this up as a test case is the line is a special purpose line whose only purpose is to carry the maximum load at a minumum cost safely within the limiatations of the technolgy of the time of its construction. It appears to be engineered for this and it does the job asked of it very well. The trains are very large for the motive power and braking availible so it would seem to be a good test case for the "physics",the traction, friction and braking of any sim. Neither the trains or the line is imaginary therefore any sim SHOULD be able to handle the trains reasonably well.

Lindsay

[GoGoran]

Fabulous points, Lindsay.

MSTS is really a superb simulator, regardless of its few shortcomings, and OR will take a long time to get to the same standard in many fields (though it does already have significant advantages in the field of graphics/handling of high-polygon models). I'm afraid the reason why the trains do not move in OR is that the friction is overstated in OR, but let's wait and see what the developpers say. It's a very valid point that if those trains are able to start and run in reality, they also should be able to do so in OR - if they do not move, there's something wrong with the game physics.

PS: I'm nevertheless stunned that you realize for the first time now that MSTS got coupler slack. :O

PPS:

The point of interest is the trains these have a gross weight of around 30,000 tons and the consist is 2 GEC Dash 9's with 104 are wagons, another 2 locos and another 104 ore wagons, total consist being 4 locos and 208 ore wagons

This is not entirely correct - 30,600 tons plus locos is the typical train with 204 cars, but the maximum ( 1 - 2 times a day, of a total of 16 trains a day) is 45,900 tons plus locos with 312 cars. And BHP Mining says that the typical consists should be 3 locos per 104 waggons, i.e. 6 locos per 204 waggons and 9 locos for 312 waggons.

Source: http://www.bhpbilliton.com/bbContent...Operations.pdf

[longiron]

OR will not move the train (even empty) at all, the locos at the head move a bit to take up the coupling slack and then halt. It is likely OR is not controlling the mid train locos.
Lindsay

I've downloaded the route and the trainsets. What activity are you referring to? All the ones I tried I could get the train moving and accelerate to track speed, although it takes a while. Please make sure the brakes are released on these scenarios. With 200 or so cars, it takes a while to get them to release.

[lindsayts]

Fabulous points, Lindsay.

MSTS is really a superb simulator, regardless of its few shortcomings, and OR will take a long time to get to the same standard in many fields (though it does already have significant advantages in the field of graphics/handling of high-polygon models). I'm afraid the reason why the trains do not move in OR is that the friction is overstated in OR, but let's wait and see what the developpers say. It's a very valid point that if those trains are able to start and run in reality, they also should be able to do so in OR - if they do not move, there's something wrong with the game physics.

PS: I'm nevertheless stunned that you realize for the first time now that MSTS got coupler slack. :O

PPS:



This is not entirely correct - 30,600 tons plus locos is the typical train with 204 cars, but the maximum ( 1 - 2 times a day, of a total of 16 trains a day) is 45,900 tons plus locos with 312 cars. And BHP Mining says that the typical consists should be 3 locos per 104 waggons, i.e. 6 locos per 204 waggons and 9 locos for 312 waggons.

Source: http://www.bhpbilliton.com/bbContent...Operations.pdf

I mostly drive 2 car DMU's in the sim these are not long enough to show coupling slack (also as both are powered it does not show up) hence I did not realise it had same.

I have been impressed with MSTS's simulation for some time. Although it is fairly simple it is capable of excellent results. The major shortcoming of it is that steam loco simulation cannot be varied in effeciency on the move as happens in real steamers.

Lindsay

[lindsayts]

I've downloaded the route and the trainsets. What activity are you referring to? All the ones I tried I could get the train moving and accelerate to track speed, although it takes a while. Please make sure the brakes are released on these scenarios. With 200 or so cars, it takes a while to get them to release.

Thanks for the reply. I did try a couple of activities must not have given enough time for the brakes to release, I will give it more time. As I said though my main aim in bringing this to attention is I believe it is one of the more extreme test cases.

Also for anyone who has downloaded this route there is an error in one of the path files "Nelson Point to MAC.pat" The corrected version can be found at http://mstsforumsforaustralia.freefo...route-t24.html

Lindsay

[lindsayts]

Now I have run the activity "Yandi to the coast" now a number of times now on four of these (all of them) when I pass the first AI train OR freezes and some seconds latter Windows (I think) opens a window saying something like OR runactivity has run into a problem do wish to report this to Microsoft. Now on one of these occasions I had saved the activity only a few minutes before the lock up. I resumed the activity and drove for around another 20 to 30 minutes then exited via the "escape" key, OR crashed immediatly. It did however print an error message something to the effect the sound system of OR was having memory problems most likely caused by corrupt memory.
It sort of seems possible OR somehow does not like train consists this large?????????????

Wih anykind of luck attached is the log of the crashed session.

Lindsay

[callagst]

I gave the route a whirl in MSTS and in ORTS using the BHP rolling stock and my own unreleased QNS&L ore car/locomotive rolling stock. In MSTS the default activity consist of 4+208 ran (albeit at about 15 fps inside the cab and slowed to a crawl outside-6-10 fps). In ORTS it started, went into negative fps, and locked up solid. My own stuff ran fine with a short consist of 2+90 (22-26fps in MSTS in the cab, 12-16 fps outside 3/4 view and heads out view). This is with sliders maxed in MSTS. In ORTS, my stuff ran fine. I was testing my loaded ore car physics from Yandi 1 balloon northbound and walked up the grade at 20-22 mph, which is about right for 8400 hp and 12000 tonnes. The other ascending grades were similar. Descending, I stayed within the BHP speed limits using an 8 psi brake pipe reduction with varying DB force levels. In real life you do not go down 1.5% descents more than 6 kilometres long at 65 kph, 55 is more reasonable, otherwise you overheat the wheels and get into thermal cracking problems with the wheels medium to long term. I am looking into the sound issue (mentioned in the docs that came in the route), and taking a look at the vegetation and static model LOD levels, as the areas with lighter vegetation have significantly higher frame rates. I like this route. It is pure high volume heavy haul and great for physics testing (I like MLTs KHP2 for the same reason-severity of testing). More later.

Steve Callaghan

[lindsayts]

In MSTS I get around 16 to 20 FPS outside the cab with the std BHP consist this being on an ASUS P6X58D Premium with a i7 950 cpu and and Radeon 5870 graphics card. Most of the low frame rate appears to be due to the large amount off rolling stock visible and the detail level in the ore cars. I was aware the need to decend the grades at a relatively slow pace to protect the braking system. This also has the benefit in the "Yandi to the coast" activity that one does not have to stop for the signal in the loop after the junction to (I think) Newman (the loop at the bottom of the first grade) as the first down train has already past this loop and this saves around 12 minutes of time.
As I have said I put this up because it is such a good test both for physics and "limit testing" of the sim. I may also say the route is very interesting and it is of interest what railways have acheived up there in such an inhostpitable place.

I had a note from the routes author stating he did not complete Port Hedland as he did not know enough about the town and the port itself, also that other projects mainly I think rolling stock were calling. Unfortunately Port hedland is about as far as one can get in Australia from the main centre of the countrys population and still be in Aus (be something like 4000 kilometers one way by road).

Lindsay