S/DS,once again

Is there any way to fix an error which happens in one tile which is totally empty, until I add one piece of track to it. That will cause tdb rebuild to (surprise!) crash at the location where the first piece is.That F#%/¤n S/DS screen burns in to my monitor soon, and the XP error log has almost 200 MSTS crashes since 8/20/03.
I wanted to extend my route to Hamburg, NY but the first piece I added today caused the crash and everything was wasted (again).
The rebuild works if I don't add anything to that certain tile, but it's on my way so it must work.

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

RE: S/DS,once again, calr, Sep-23-03, 01:04 PM, (1) RE: S/DS,once again, enocell, Sep-23-03, 01:34 PM, (2) RE: S/DS,once again, Hack, Sep-23-03, 01:45 PM, (3) RE: S/DS,once again, shawnbecher, Sep-23-03, 02:26 PM, (4) RE: S/DS,once again, enocell, Sep-23-03, 03:04 PM, (5) RE: S/DS,once again, shawnbecher, Sep-23-03, 03:16 PM, (6) RE: S/DS,once again, griffinmt, Sep-23-03, 03:57 PM, (7) RE: S/DS,once again, enocell, Sep-24-03, 03:09 PM, (8) RE: S/DS,once again, enocell, Sep-25-03, 03:15 PM, (9) RE: S/DS,once again, shawnbecher, Sep-25-03, 03:16 PM, (10) RE: S/DS,once again, griffinmt, Sep-25-03, 06:24 PM, (11)

Conferences | Topics | Previous Topic | Next Topic

A couple of things come to mind. Is the added track connected to existing track? Does the new track cross any tile boundaries--especially the corner of a tile? Try changing the starting point before doing the TDB rebuild. Sometimes just changing the processing order will affect the success of the rebuild. The rebuild starts at the stored starting point and works outward from there.Cal Rasmussen
Beaverton, OR
(My email is in my profile)Columbia Gorge Route (Phase 1 UP side 95%, BNSF side 90%)

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

It crosses the tile boundary 500m earlier. The other tracks cross boundaries too but they didn't cause problems.

Attachments

http://www.train-sim.com/dcforum/User_files/3f70840f0ca61639.jpg

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

Is this a dynamic piece that's giving you trouble? I'm wondering if you could get around this by adding in a standard piece, and then continue once you're past the tile boundary.Cheers!
Marc - 3DTrains
http://www.3dtrains.comPro Series SuperPacks

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

Or if it isn't a dynamic piece, you might try running shorter segments of track up to and over that boundary. I had the same type of problem on a route I was working on where if I added a 500m straight, I got the S/DS error, but if I removed that piece of track, everything worked fine. I think I remember seeing somewhere that long pieces of track over boundaries could cause the S/DS error, but my memory doesn't serve me all that well most of the time so I could be talking out my.....well, never mind that part (Can't wait until I actually do get old - I'll be in for a heap of trouble then).Shawn

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

LAST EDITED ON Sep-23-03 AT 03:28 PM (EDT)

I have avoided all dynamics over the tile boundary. As this piece was a long one, I'll tried with 100m straights. Somehow I suspected that it won't fix it and sadly I was right againEDIT: it crashes if there's only one unconnected piece on that tile.
EDITs2-10: changed the start tile to the one where I got several vector-in-other-vector's-@$$-errors, and then the rebuild went through. But if the route has 100 tiles, it will be long and annoying job to find the working tile. I changed it only 4 times!I suggest all route builders to write down the start tile coordinates, that makes things easier.

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

There has to be a reason for this "random" error, and if we can figure it out, we can save a lot of people future headaches. Does this section of track happen to be coming off of the heel of a switch? When I had my S/DS error, the 500M straight was coming off of the end of a small passing siding, and I had placed the 500M straight so that there would be enough room for a loco and a couple of cars to pass around a couple of other cars. I ended up moving the passing siding down the track farther, and that seemed to solve the problem. I was then able to extend the spur (off of one of the toes of a switch)out past the original problem area with no problems whatsoever. Just a thought - if this happens to be your case, you might try throwing in an invisible (0 deg.) switch and see if that solves the problem. I could be way off base though.Shawn

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

I think we have basically figgured out 'why' it happens, just not the best way to circumvent it.Given that each piece of track is identified by its starting origin (in 3 dimensions) and its angle and attitude at that point, all with about 6-7 decimal digits of accuracy, you can imagine that after CALCULATING the end locations, given various curves and crossing tile boundaries (ie: recalculation a new relative tile location), it is easy to start losing the accuracy needed to properly lineup with the next piece of track. It is my guess that as long as it is within a small degree of error, it adjusts and keeps going, but we are not always that lucky.A good example of this creeping error happens on an existing long route. If you delete a single track section somewhere, then re-install a new piece (same section), the far end most often does not seem to line up correctly. As you fix that mis-match, the next one is out of alignment and so on.Almost all of this would have been avoided if they had used double precision floating point values instead of single precision. On the other hand, it would have made a serious increase in file and memory needs, and would have slowed the simulation down as well.Martyn T. Griffin
griffinmt@mindspring.com
http://www.geocities.com/tcmrrc/

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

The start tile change fixed it (this time). Hopefully RE lets me make the rest of the route, about 25 miles. That's single track in the countryside so it should be easy...
The steelplant is still unfinished, though.

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

Just checked the route with MSTSmiles, I had no idea that it would be 293 miles long (and Buffalo Southern has only 35 miles of track!). But the amount of road sections is something: 339 miles!

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

Wow! That's a pretty long short line Shawn

Alert | IP

Printer-friendly page | Edit | Reply | Reply With Quote | Top

Keep in mind with MSTSMiles that if you have a lot of switches, you will get a somewhat inflated value. That is because there is a chunk of the switch that is treated as part of the diverging path as well as the main path. The total distance is made up of every conceivable inch of track, even if it is just a chunk laying somewhere by itself. For example, if you have Okrasa's turntable anywhere in the route, whether you make it part of a path or not, it will account for over 482 meters of track (18 x 26.81m).Likewise for any roads that are placed, whether or not any spawners are present.Martyn T. Griffin
griffinmt@mindspring.com
http://www.geocities.com/tcmrrc/

Alert | IP

[ Home | MSTS Forum | MSTS FAQ | File Library ] Copyright © 2002 by Nels Anderson. All Rights Reserved.