Thank you again and will address this in the future. Think the file or Dremel approach you suggested is a great, minimally invasive 1st step. It may be a while but will return and attempt repair.
A couple of questions: Does speed or load matter? Does it derail with a nominal train load Is it an engine by itself issue Can you visualize the occurrence Preferably take a video. I would include your other loco’s The shorter two axle trucks maybe experiencing a similar movement, just not enough to derail. I find that wheels out of gage are a common situation. So common, I check & tune on every engine I work on. One wheel set could be your problem. We need to realize that tolerances really do not scale, so the smaller the engine the models the more variability is possible. On the subject of hard wired trucks. With DCC, the compression spring like Bachmann uses will cause lots of problems. It is the same issue that the wheels have. Current passing through a contact point will cause a non conducting deposit at the contact point. This is particularly true when the metals are different alloys. Hard wiring the trucks minimizes this problem. It happens with DC systems as well, just not as fast. The wires need to sized & routed to not impact the truck movement.
1-Speed does not matter. Below is a list of all my locos and rolling stock. Outside a few coupler pins that hang, none of these cars (Bluford is exception-still in case) Forward backward no issues. All purchased as new except for red CNJ #669 hopper and VO1000, since 10/20. SD35 only one with issues. 2-I'm fairly convinced that the derailment issue on 9.75 radius is track/layout related. As mentioned earlier in thread, bought new track to test on Tuesday and made 9.75 half circle, back and forth about 12x, no derailment - pretty flawless. Assuming the 9.75 is totally track-layout related, only 2 other ?s need to be answered regarding SD35. Why when going thru 1 path of double crossover (forward-clockwise inner to outer loop) severe derailment occurs and b, on occasion 1/5x, loud click heard when thru double crossover path center (forward -Counter Clockwise[normal operating direction] inner to outer loop) The SD35 does not derail though sound may stutter at low SS and generally moves on its merry way. Think Youtube video links are posted earlier in thread. Next week, the layout will be "packed up" but will be working on a 24" x 12" Kato double crossover based switching layout so need to get another Dxover anyway. Will essentially recreate layout, without spurs, on same surface as 9.75 test.
Mark, you have an S-curve when your loco/train transitions from the inner, 9.75" curve, directly into the crossover, when the crossover is thrown (i.e. directs train across from the inner track to the outer track. I use a formula to estimate the severity of a transition between opposite curvesThe formula yields the radius of the curve in that equivalent transition from curve to straight. But first, for the double crossover, they do not specify the radius of the curved rails in the double crossover. I believe they are roughly 28" radius (centerline), but I have not attempted to actually confirm that by direct measurement. With that assumption, the formula is as follows: Requiv = 1/(1/R1 + 1/R2), where R1 and R2 are the radii of the two curves in the S-curve. So, Using R1=9.75" and R2=28", Requiv=7.23" In your layout, the S-curve between the inner loop's 9.75R curve and the double crossover (thrown), is equivalently harsh to a transition between a 7.23" curve and straight track. This would explain your SD35 having issues when trying to travel counter-clockwise while transferring from inner to outer tracks via the crossover. Note this approximation is just that; it is not perfectly accurate, but it is a means to compare relative S-curve severities, and reasonably predicts the performance of the S-curves inherent in single and double crossovers: Single Crossover S curve (R19's): Requiv = 1/(1/19" + 1/19") = 9.5" (close to 9.75 minimum for most locos) Double Crossover (R28's) Requiv = 1/(1/28" + 1/28") = 14" (close to 13.75R which will easily pass most anything.) Note that sharp S-curves affect 3-axle trucks and coupled railcars/loco's more than individual railcars or loco's. Coupled cars/locos with long and or unequal overhangs (from coupler knuckle to truck pivot center) are particularly sensitive to S-curves. In Untrack, the following S-curves are generally acceptable (~equivalent to or better than a 9.75 curve adjacent to straight track): R28/R28, R28/R19, R28/R15, and R19/R19. Since you don't have room to shift the double crossover away from the curve to eliminate the S-curve (due to the yard lead switch), I would suggest either of two alternatives: Move the dbl-xover to the middle of top side tracks (in your track plan diagram), ensuring you have at least an S62 between it and any loop curves. Replace the dbl-xover with a LH sngl-xover, and include a RH sngl-xover at the left end of the upper straight doubletrack. This is not only more prototypical, but also provides a passing lane for one train to pass another without having to orbit an entire loop around the layout to do so.
Thank you @BigJake . Love the mathematics. We can also use product over the sum R1 * R2 / (R1+R2) Appreciate all the education! From the beginning knew it was not the best thing from design point, emotions took over, and guess been very lucky outside SD35. After thinking more since the 9.75 test, best course maybe.. 1-Confirm SD35 on Separate DxOver Test with and without straight track in between. Assuming test goes OK, and I'm thinking it will, the SD35 is cleared and cancel the replacement truck(s) order. 2-Attempt to fix the derailment across 9.75 junctions using the file or Dremel technique. 3-Being a seasonal layout, going to probably leave all else same and just run the 4 axle Locos if #2 or any other forms of non-invasive remediation fail to mitigate. In the end, want more track, longer spurs (backup) regardless of direction, less scenery = new bigger layout.
That's an interesting way to look at things @BigJake and it may have been the basis for trouble with a Kato #4 that brought derailments to a few 6-Axle units. As marked by the arrow, the #4 was positioned off the diverging route of a Kato #6. As we know, Kato diverging routes are not straight, but are curved. I'm thinking that certain long trucks were slightly curving as they exited my #6, then were asked to quickly curve in the opposite direction to enter the diverging route on my #4. By notching the #4 stock rail on the diverging route (red arrow side), all was made well and I've not a had a problem since. Just a theory, but after reading your post on S-Curves, it might have had merit in my situation too.
Nobody told me there would be Math when I was planning my free lance layout @Mark Ricci Your brave testing in Reverse also, I don't dare on most areas of my layout at this time. Are you making modules to connect together at some point when you move? I think I saw where you were gonna move?
Unitrack #4 switches have long been known for operating problems due to the lack of a notch on the inside of the rail to "hide" the tip of the point rail, and prevent it from being 'picked' by a wheel flange. You can lay a 28R15 curve on top of the diverging route of a #6 switch, and it exactly coincides. Same with a 19R15 and a #4 switch. However, the through path on a #4 is NOT the same length as a 124mm straight, whereas the through route on a #6 is the same as a 186mm straight. They are not constructed the same way (in miniature) as 1:1 switches are...
Being a retired Electrical Engineer, I thought of the problem like combining resistances or inductances in parallel (or capacitances in series), where the reciprocal of the combination is the sum of the reciprocals of the constituents. I found it seems to correspond reasonably well with known/typical radius limits. It also works for severity of the transition between two curves in the same direction, if you take the absolute value of the difference of the reciprocals. In other words, sign indicates whether the curvature is in the same direction or opposite. This can help when evaluating sectional easement curves. I actually duplicated your layout with 15 degree easement sections of 11R and 12-3/8R on the main curves, and it still fits your layout space. Food for thought if you ever build the same layout again (say in a different season than winter, for Easter, Halloween or Independence Day.)
Worked in the industry a bit and taught applied algebra; circuit analysis, semiconductor devices and electronics circuits, digital electronics for almost 25 years. Former FCC Commercial Radio-Telephone License 1979 - still holding active amateur radio license 1980 .. --model railroading contributed to entering the career... The last 20 in IT. Just don't ask me to lay track... Can't believe I hosed up just about every 9.75 junction. No module approach. 1985 HO Layout 2A had turnouts to enable expansion. 1-Was initially going to add those but it was too tight to fit. 2-Had the intention from the beginning to employ the quad DS64, especially due to cost and ability to "throw" double crossover. Limiting to 3 turnouts and single control wire Dxover is perfect resulting in purchase of only 1 stationary turnout decoder for the layout. 3-While really like the seasonal look and the contrast, white is just too hard to keep clean. 4-Never intended to have layout up all year. A neighbor discarded box from new TV a couple days ago. Should slide right in with the micro legs removed. Option to ship with mover or FedEx if it does not go with me. Great theoretical application! Would not have made the leap from using sum of the reciprocals to curve radii severity. Once the layout is down, and have more time, will definitely explore the mathematical relationships further in addition to creating layout in track planning software using some of the mathematical approaches you have provided. Have 2 #4 RTs, the one attached to DxOver bought from local store in 10/20, the 2nd one, bought 6 months later online. There have been a few snafus (SD35 only) backing up in divergent path. The newer #4 works perfect but not used nearly as much. Thought Kato corrected #4 a few years ago. Local store maybe had older stock?
I'm uncertain, but there may be something to it. Some years ago I bought a Kato #4 to study Unitrack reliability. It exhibited the common tendency to cause derailments. The #6s were flawless. I recently ordered four #4s and three have been excellent with all of my locomotives, with one troublemaker that can perhaps be pardoned because of where it was placed.
The mis-alignment of parallel straight track and both ovals junctions' as shown below. The other 3 line up perfectly. This has been a concern from the beginning.. Checked layout against plan track segment by track segment 3x and all fit good and aligned as shown in SCARM above. The #4 length difference was not noticed until swapping DxOver and temporarily replacing the #4 with 124mm straight during troubleshooting last April. The cause of this mis-alignment and impact on the rest of the ovals, contributing to SD35???
Mark, Not sure what's going on there... my first thought is an incorrect radius curved piece somewhere, but you checked that 3x already. I like to use doubletrack straight sections of Unitrack where practical, just to keep everything aligned (track sections' endpoints abreast of each other). A framing square or a square edge of a piece of plywood, poster board, etc. can help check/transfer alignments across to the other side of small ovals.
Thank you for the info.. Interesting... Yesterday's Test-- 1/2 dozen times, at speed step 1/128 ran SD35 back and forth thru the #4 in the spur section alternating between thrown and closed paths. The SD35 performed never pushing the point away from the rail. Maybe #4 and/or less acute curve into it.
LOL Have to admit, yes for the first few laps, but not anymore... That #4 is good. The fingers were clinched, 3x in a row made it thru the troublesome #4. Stopped there, didn't tempt fate. Wanted to try 4 locos on track. Tough on small layout. More locos than spurs, passenger service train parked overnight in service loop..
Mark, it sounds like your problem is behind you, but if it rears its head again, by centering your crossover in the middle of the straight section it's on, you might incent your SD-35 to track better through it. To do that, you'd need to lift out your yard tracks and move them to the other side. The red arrow shows where the turnout would need to be. Just a thought.
Thanks all for the input and feedback.. Will probably hold off doing anything with layout since moving soon and only the SD35 A)derails at 9.75 junctions b)-50% time opens the point on the first #4. Still leaves 3 remaining locos and all the rolling stock that run AOK. May attempt several repairs.....Using @Hardcoaler #4 technique -- no track needs to be removed and derailment most likely due to the manufacturing of the #4. The 9.75 junctions, if can be repaired using file technique described by @BigJake would be next. Concern over the possibility that attempted repair of any kind can result with other locos and/or cars having track problems. Moreover, inner loop feeders are located at the 9.75 arc middle junction and both sides of the 9.75 are close to the vertical rock further complicating removal. Its hard to believe that every 9.75 junction is bad but guess its possible. Thinking that the arcs of the entire 9.75 segments are off. The mis-aligned 2 straight sections shown in prior post might be the reason. The immediate need is to determine whether any SD35 issues remain. Have run the SD35 more in last 2 days than in the 12 preceding months. The last Atlas service on SD35 has made a tremendous difference not just in the ability to creep thru double crossover, but in overall performance as well, especially moving in the spurs. That leaves 2 other things to clear on the SD35.... Both of which are related to double crossover. As a novice, they seem inconsistent. 1-The occasional click going inner to outer forward counterclockwise.. Setting the throttle at a fixed speed with the double crossover thrown letting loco go around and around many times thru both loops. Why click doesn't occur at every pass?? 2-Severe derailment going forward clockwise inner-outer. The derailment occurs with the #4 closed, presumably acting like a straight track section preceding the DxOver leg. If anything, would think the derailment would occur when forward counter clockwise, entering either leg of the DxOver from a 9.75 curve?? Neither one is a significant problem in terms of operations --no derailment when click is heard and forward clockwise is rarely used. It would be good to know. Maybe a combination of 6 axle loco, track design and or track install, and still something to do with SD35. Thinking the new DxOver test will shed some light.
