Trackwork, how good is good enough ?

[jbinkley60]

Thanks everyone for your input. It has been helpful. After doing numerous test runs I've decided that my maximum speed on straight portions will be around a throttle setting of 50-55, 40 on curves and 25 through turnouts. Any more than that barely looks realistic and starts to look more like I am running slot cars. Everything seems to run well now at those speeds with no dereilments or other issues. I did find a couple of spots where the outside rail on a couple of curves was slightly lower than the inside rail. This caused some sliding problems. This has been fixed. To the comment about the newer units having smaller flanges, I was in errror, they are the same size. I double checked. The Lifelike trucks seem more "springy" than the Atlas trucks. I also got my order today from Tony's with decoders for the Lifelike units so I'll be able to do DCC side-by-side testing.

Thanks again.

[Anonymous]

QUOTE: Originally posted by ehowe

One technique I've heard several people using is to use silicon caulk to glue down track to an extruded foam base. After reading this discussion, it would seem the caulk would allow slight vertical movement that could lead to derailments. Is anyone here using this technique and not suffering derailment issues? I was thinking of trying this for an N scale layout, but now you've got me concerned.

Well... I use caulk to glue down track to cork roadbed, which is caulked down to 2" foam, which is caulked down to 1/4" plywood laid over my 1x4 benchwork... No problems.

[HAZMAT9]

I'm using WS roadbed on top of 2" foamboard with adhesive caulk, not silicone. I haven't had problems with shifting or moving of roadbed and it glues down pretty darn good compared to cork. My problems derailing come from switches and out of gauge track, the foam is pretty darn stiff...I guess weight could be a factor as well.

[TomDiehl]

This sounds like an area where you've found a true parallel between the model and prototype. Trackwork quality determines how heavy of a car or locomotive can be operated on this track, as you've found out already. Curves, superelevation, and track bed quality determine not only the speed, but rigid wheelbase limits of the locomotives. 22 inch radius (I'm assuming HO) sounds a bit tight for a six axle locomotive without a lot of lateral play in the trucks. Track gauge can also be affected by driving the nails too tight and pulling down on the ties. Soldering wires can melt ties and lose the cast "spike heads" that hold the rail in gauge. Check any problem areas with the NMRA gauge.

Also, as stated above, don't assume that the locomotive (or cars) are properly gauged, check them with an NMRA standards gauge and adjust as needed. Also, check your coupler heights. Too low a trip pin on a coupler can snag on obstacles (switches, crossings, etc) and cause derailments. This paragraph is basically what I do with any new car or loco right out of the box, before it ever feels the rails.

[ehowe]

One technique I've heard several people using is to use silicon caulk to glue down track to an extruded foam base. After reading this discussion, it would seem the caulk would allow slight vertical movement that could lead to derailments. Is anyone here using this technique and not suffering derailment issues? I was thinking of trying this for an N scale layout, but now you've got me concerned.

[jecorbett]

Trackwork is one area where there good enough approach does not work. It should be top drawyer or you will have nothing but frustration. Once the track is down, it should be thoroughly tested before you proceed with the next phase. You should also stress test it by using your worst equipment at less than ideal speed. If it can stand up to that, then you best equipment running at normal speed should have no problem.

Before I retired, I programmed mainframe computers for a living. Programs need to be thoroughly tested before they get put into production. It is a natural tendency for a programmer to run tests to prove their program works so they run tests with normal conditions. A could test plan should try to prove the program doesn't work. That means doing exception testing with every conceivable factor and in combinations. Programmers who fail to do that get the calls in the middle of the night to come in and fix it and also will have to face an angry boss in the morning. Take the same approach to testing your trackwork. Try to prove it doesn't work. Try to make it break. If you can't do that, you probably have good trackwork.

[twcenterprises]

The way I see it, perfect is good enough. Seriously, though, My track is Atlas, Micro Engineering, and Walthers (turnouts) code 83, and I run my trains at mainline speeds with no derailments. I have used cork roadbed, and now use "putty tape" which is similar to AMI Roadbed, but less tacky and a bit more resilient. When laying track, I make absolutely sure everything is as perfect as I can make it. I solder joints in curves to prevent kinks, check turnouts with the NMRA gauge, carefully check for dips and humps, especially near track joints, make sure curves are as uniform as possible, and make transitions to grades as gentle as possible.

All this extra work paid off in smooth flowing and trouble free track. Any dips or humps may not cause a problem with shorter cars, but as I found out, longer cars such as 86 foot boxcars and auto racks will end up with a high-low coupler situation. Six axle engines will often derail on such "joint humps" as the center axle is "high centered", and the end axles may actually leave the rails. At this point, they are free to go where they wish, and on a curve, they will usually go over the rails. I do believe the LL units have some give (equalized trucks, maybe?) in the 3 axle trucks, so they are less likely to show this problem, and 2 axle units don't mind the humps so much.

As far as speed and derailments, I don't have this problem. I have used 22" curves and run Athearns through them at full throttle without derailments. I don't usually run my freight trains at more than about half throttle, but occasionally I will open 'em up briefly. Passenger trains may run at full speed on the main, less on the branchline. So, the way I see it, yes, you can run trains fast enough to derail on a curve, but mine won't even go that fast. The only way mine will derail on a curve is if the centrifugal force lays them over. Of course, once the inner wheels leave the track, the power is cut and the engine slows down. So, in order for this to happen, some way of keeping power to the motor must be there long enough (1 second or so), to maintain the speed to overturn.

If your trains are derailing on curves, then you would need to check the track, and the engines. Make sure the wheels are properly gauged, centered on the axle, and centered in the truck. Make sure all 3 axles (on 3 axle trucks) ride level with one another, not one high/one low. If placed on a piece of glass, all 6 wheels should touch the surface. Check to see if the trucks are swiveling freely, both side to side (like going around a curve), and end to end (like going over a dip or hump). I don't think the flange depth has anything to do with it. Real trains use (proportionally) smaller flanges, and they don't have any problems at high speed.

This all may sound like a lot of work, and for a large fleet it can be, but most of it is a one time thing, once an engine is all tuned up and properly adjusted, it should not need any more tinkering until it needs its regular service (clean and lube). The track should be relatively maintenance free once all the kinks (pun intended) have been worked out.

[MisterBeasley]

I guess my answer is that it's never good enough. I can run most of my trains around at near-Nascar speeds, but every now and then there's a derailment. If I can really find the cause, I'll stop running the trains until I fix it. These days, though, it's mostly a problem with my 40-year-old rolling stock, which I'm slowly bringing up to modern standards.

[cmrproducts]

Jeff

I have a few BLI, Proto and Atlas engines all with sound and my layout is Digitrax DCC controlled. I have ZERO derailments, unless someone runs a turnout!

Now I really do not want to get into a p---ing match here with others on the forums about what causes derailments but I am using OSB (Orientated Strand Board) for a sub base, Homasote on top of that and then HO & N scale cork for the actual roadbed under the track. My track is tacked down with Atlas track nails until I get the ballasting done.

None of my 2700feet of track moves up & down or sideways. All of the joints are straight and I used my eye and sited down along each joint to make sure when I soldered each joint it had no kinks.

I have no problems running steam although my layout is all diesel. Anytime the track is able to move there is going to be a problem with derailments. And anytime I am ask to take a look at a layout and ask about derailments it always seem that the owner has used some material that allows the track to move, even slightly (foam, the stickey tape and the insta roadbed). And the rails have kinks, some very minor, but they are there and they are having problems.

I try and explain this to them but they feel that it is too much work to do it right and they live with the problem. One layout owner even knew where the train would derail ever time and was ready with his hand at the spot to catch the train and set it back on the rails.

Also every car and engine on my layout gets the wheel gauge checked with a NMRA gauge, Kadees and the weight set properly before it ever hits the rails. This is why I can state that I have ZERO derailments. And if one does have a problem I have the operator that finds a problem car make out a bad order card for it and I look the car over after the OPs session and I fix the car. If the same car keeps coming back with repeated problems and I can not seem to fix it, it is soon at the flea market as it will not be on the layout to cause problems. And with over 800 cars this can become a problem if it is not done right, right from the beginning.

Speaking from years of experience!

BOB H - Clarion, PA

[Stevert]

QUOTE: Originally posted by jbinkley60


I have fixed the derailments at what I consider reasonable speeds but I am wondering what others consider reasonable or how they know it is good enough ?

As the other folks have already pointed out, good trackwork is like a good foundation for your house: Nothing else will be right if that foundation isn't right.

I hate to say this, but you might have to rip it up, get rid of that WS foam trackbed, and use cork (or whatever). I used the WS stuff on a 6 foot spur on my layout, and while it was okay there, I'm certainly glad I didn't use it on the entire layout.

As far as reasonable speeds, your loco choice leads me to believe you're doing HO. In that scale, five "real" MPH, which is the speed at which an average person walks, translates into 435 HO scale miles per hour!

In other words, if your train can keep up with you as you walk along the edge of the layout, it's going way too fast Even half as fast as you walk is still over 200 scale MPH.

But here's how you can figure out scale speeds. Based on the formula:

SMPH = 0.6818 x Scale Factor x Distance / Time

Where SMPH = scale miles per hour, Scale Factor = 48 for O-scale, 87 for HO, 160 for N, etc. and Distance is expressed in feet and Time is expressed in seconds;

Then in HO scale, if your train goes one foot in one second, it's moving at about 60 HO MPH. For anything other than serious mainline running, about half that would be better: Moving one foot in TWO seconds is about 30 HO MPH.

HTH,
Steve

[jbinkley60]

The main difference is that on the newer sound unis the flanges are smaller, the trucks are a little stiffer and such but the point of my question is to understand how far folks go in testing. I have fixed the derailments at what I consider reasonable speeds but I am wondering what others consider reasonable or how they know it is good enough ? Trains will derail at a certain speed, regardless of how good the track is. I am just looking for some advice/opinions here. The super elevation seems to have helped the most on stopping derails on downhill runs with curves. I run code 83 track with a minimum radius of 22" and #6 turnouts. Again I know at some speed anything will derail, even if I had 48" radius turns and #10 turonouts, it would just be a higher speed than 22" radius track.

[selector]

Your trains just sit if you have no reliable track on which to run them. Where do you run your car if you have bad roads? On the bad roads.

I am not a diesel buff, so I am unable to respond more helpfully than to suggest you find out if the trucks on these locos are longer (6-axle) and if they are encountering gaps, misaligned rails in either elevation or lateral position, breaks in electrical connectivity, and so on.

Also, gauge your wheels for distance apart, and for their spacing relative to the ends of the axles. You want this: <=()=======()=> and not this: <==()======()>

Is your track in gauge everywhere?

[jwmurrayjr]

I think that your trackwork may be the most important part of building a satisfying layout. Even if it means tearing out some a redoing it.

I've never used foam roadbed (except where the track is lain directly on the blue foam) so I don't know about that.

Do not proceed until you are satisfied with your trackwork. Then you know that your trains will run.

[jbinkley60]

I've spent the past week fixing minor trackwork issues that have caused the new Atlas 8-40CWs and a BLI SD40-2 sound unit to either cause intermittently or derail. I built this portion of my layout 2 years ago and have not had any of these problems with the other 30 or so DC locomotives. My main test units in the past have been a Lifelife SD60 and an Atlas 8-40B but these new units have smaller flanges and less give in the trucks for the axles. I bought the new units and switched to DCC recently. Not all of the old locomotives have decoders yet.

I've used the Woodland Scenics Trackbed and the foam seems to allow more give and be a source of some of the problems, especially since I glued the track down. I've since driven a lot of spikes to fix vertical movement which causes stalling on the DCC sound units but I have to be careful not to drive the track into the trackbed foam and make it uneven. I've also had to file a few joints to even up the railhead heights. I've superelevated a couple of curves where the leading trucks would slide over the railheads if it was going too fast. The point of all of this is how good is good enough ?

What I mean is how do you know when you've gotten the trackwork to where it will be reliable and not have any problems ? I've been raising the speeds of my tests but there is always a speed you can run that will cause a derailment. Right now I am testing at a maximum speed of 30-35 on my DT400 Digitrax throttle through turnouts and curves. I think that is more than sifficient. I am wondering what others do ?