I'm making up a batch of DCC Terminator boards. These have the RC network needed to properly terminate a DCC bus and prevent voltage "ringing" that can damage your decoders, plus a convenient wire terminal for easy installation. Suitable for all scales Z-G Compact (less than 1" square) Convenient: No soldering. Your choice of screw or spring-lock terminals. Mounting screw hole included at no extra charge! 10-24 AWG (screw version) or 12-28 AWG (spring-lock version) wire capacity $8 each plus shipping. Small enough they should ship via First Class mail easily. Lead time should be 3-4 weeks. PM me an order if you are interested. Please specify screw or spring-lock terminals.
Uh, I'll admit to being a little rusty on my transmission line theory, but I'm not aware of a transmission line in any dcc system that would require termination. I mean, the track is a fully isolated conductor.
I'll let Allan Gartner explain. He's got pictures. http://wiringfordcc.com/track_2.htm#c2 These would be best for longer bus runs of, say, 20 feet or more. it would be "insurance" on shorter bus runs, to help make sure the DCC signal stays nice and "clean". On longer runs, you start to get "ringing" or voltage spikes on the transition edges from + to - (and back) on the signal. These spikes can get high enough voltage (in bad cases) to damage decoders. Most people can get by just fine without them, because of the shorter bus runs that are considered "best practice" anyway. I'll not hide the fact you can DIY a filter for pocket change. Allan explains how. I'm selling convenience and neatness for those folks with longer bus runs for whom it's worth $8 to avoid using a soldering iron. Another place I thought these might be handy is Free-mo style club or show layouts where you never really know until show time where the end of the bus will be, and it will probably be very long, and you want to quickly hook something up, instead of soldering or otherwise DYI'ing.
Thanks for taking the time to explain. I was under the impression that NCE could benefit from these but Digitrax does not recommend their use. I run NCE at home and Digitrax at the freemoN set ups so one day I may take you up on these gems. Having the screw or spring lock connections are nice additions. I would add Anderson connectors for ease of adding to the end of the standard freemon bus if our group feels the need.. FYI. We have a member who seems to enjoy planning things to the "T" so he always has a DCC plan to match the layout plan that was chosen by the group before the show. This pre-planning sure speeds things up.
I don't mean to cause anyone any panic or inspire any fear... if your bus wires are under 30 feet or so, this device is really not needed. Ringing bad enough to cause damage would be pretty rare, severe issue. Digitrax has a rather snarky "no" response in their knowledge base to a question about whether these are necessary, though I think they were stung a bit by their equipment being used as an example in a 2008 Scale Rails issue. But if you're one of the few who does have long bus runs (modular club layout, perhaps?) or otherwise is having trouble with noise on their bus lines, I'd hate to have built 10 only to find someone wants the 11th one.
I'm with ya TD. I was hoping my post might garner more responses from those with experiences to share. We do have long runs, but much shorter now that we distributed the power around the layout. We have yards that are divided into sub-districts and mainline both double track and single track divided based on possible demand. So the longest currently is likely 25 to 30 feet from the short circuit protection board. I would rather have them than not, but need more proof for the group to digest.
I've added a PowerPole version with the headers arranged to directly plug in to a Freemo-N power bus. They will be $10 due to the added cost of the header parts.
Here's a rough artists rendering of the board. The quarter is for size reference. The PowerPoles will be red over red, to match the Freemo-N track bus standard colors.
I'd like to order the PCB's early this week, so if you'd like one please let me know as soon as practical.
NEW LOWER PRICE! i secured a lower cost on some components. New prices: Screw or snap-lock: $6.50 PowerPole: $8.50 if you have already ordered at the old price you will get the new price.
Well since this is my area of expertise and someone mentioned Transmission Lines, what is need to reduce the ringing is what is called a conjugate match. This is calculated from the by combining all the resistive elements and reactive elements, and then calculating their values at the length of the transmission line and then applying a load that has the resistance at that point and the opposite reactance at that point. That is the only way to stop the ringing using any sort of termination. And the real resistance isn't necessarily what you may think it is. These are all calculated based on the length of the line at the frequency of interest. Since DCC is such a low frequency that the wavelength is just a little over 123 thousand feet. Since all layouts are much, much shorter than even 1/4 wavelength, the calculations get even a bit more difficult. That is because you now have to consider the mutual effect that each component will have on each other. This also gets blown out of simple theory if there are any active devices on the lines as well like stationary decoders. And it is possible that the wrong termination could make things work. There is no true all purpose termination, and if it does help it is strictly anecdotal, which means it does not work in all cases. Your best case is to use the best quality wire and components for the frequency of the operation, which means a lot of what you learned in DC is out the window. It is patently untrue that 12 gauge wire is less lossy at DCC frequencies because it holds a much higher reactance than smaller wire. This is holdover thinking from DC and not really applicable. Add to that the the waveform isn't even a sine wave, so the solution of an RC circuit is a standard solution for a sine wave signal, but not a solution for a digital pulse train. The reason that a DCC run is limited to 300 feet, is because that is when the harmonics have enough reactive energy to cancel out the signal almost entirely. No termination can be made to stop this. Proper science and proper materials is the best way to have a clean DCC signal path. Sorry to throw water on this.... David Eaton, Professional Electromagnetics, Antennas and Microwave Engineer
After doing some more studying, I don't think it's transmission line effects either. The distances and frequencies involved don't add up. If I used the phrase earlier, it was as a layman's term, not the technical term. Probably shouldn't have done that. I believe the circuit is merely helping to damp the overshoot on the signal. I'm making arrangements to get an oscilloscope for some real world testing, but my SPICE models generally align with the scope shots on Gartner's website. They show an improvement for lightly loaded circuits, less so for heavily loaded ones. I'm anxious to see how the real world testing goes, as the more I look into this, the more I'm intrigued by the "controversy" over the circuit. With one major DCC vendor recommending it in their user manual and the other one practically laughing over it, what's not to like? And if selling a few of these triggers someone (me?) to do the analysis to prove or debunk it, either way it's good. My main "personal goal" in offering them is more to test out my development and production toolset and flow than to make any profit. Certainly not at anyone's expense.
So, I used the term transmission line, because you called it a terminator. Which is something you do to a transmission line. In particular a high frequency transmission line. But DCC is low bandwidth baseband PCM and the bus lines are the best part of the signal path. The track itself would, if this mattered, undo any advantages. Now reading further, it sounds like what you're talking about here is a filter, but I don't get what is causing the need. The base units and boosters should already have appropriate filters.
If these terminators make a specific problem you are having go away, then I would say use them. If you are trying to fix potential problems, then you just are adding what I would call unnecessary complexity. Everything that you put in or on the DCC signal path is a potential problem. The waveform characteristics of anything but a simple and static layout are highly complex and constantly changing. The amount of testing that it would take to generally validate the value of these terminators is unrealistic. This is why it is good to also have a valid theoretical explanation. Bob
I think he's wrong about the theory. It isn't transmission line effects. I think it's just slew-rate-limiting the square wave. I'm not sure in what way the track itself affects this... Can you elaborate?
I'm pulling this product. Well, sort of. The questions raised here have spurred me to look more deeply into the science behind this, and I now believe the information I based my rationale on is well intentioned but ... flawed. Looking specifically at the information presented on Gartner's site, I now believe this circuit merely presents a load to an otherwise open bus, and really has no effect on an otherwise loaded bus (anything that actually has a decoder on it). Note in particular photos #7 and #8 show essentially identical effects on the bus from the snubber circuit and a locomotive decoder. My own simulation testing bears this out, and I expect the real-world tests I have planned will as well. Since no one actually uses an unloaded DCC bus, the signals present on an unloaded bus are of little interest to anyone but the booster manufacturers and the FCC. However, a damping circuit like this is still recommended in the NCE user's manual, and there may be some circumstances where its use is warranted that I just don't understand yet. Especially with very large, modular layouts with very long buses dotted with many intra-module connectors and feeder tap points. I have not yet modeled -- nor did Gartner -- a "messy" track bus such as this, so I can't say one way or another what that might look like. So, I'll be offering refunds to any current orders who may wish to change their minds, but if anyone actually wants one, I'll still sell it to them. I just won't be touting it as a cure-all for what ails your track bus. Sorry for any inconvenience I may have caused. Mods, feel free to delete the thread, though it may actually be of some educational value to anyone who reads it in the future. I'd edit the OP, but it appears that I can't anymore.
TwinDad, don’t give up trying to improve the performance of DCC. If I had just a nickel for ever not-so-good idea I pursued, I be very rich. Bob
