DCC Bus Terminators

Thanks for the encouragement, Bob.

Never said I was giving up. Just no longer think this particular "solution" has (much of) a problem to solve. I still plan to do some of my own testing to come up with a cogent explanation of what is going on. Call it "Mythbusters - DCC Edition". It will be a while, though. We'll see whether it's busted or confirmed.

In the meantime, I've got some much more interesting projects in the pipe, that I hope to have out in the not too distant future.

 

Well, if he took all those pictures with no load on the system, then they are all, to be blunt, stupid and useless. As now the circuit is being completed by the Scope and those transients could be related to the scope itself even. The fact that placing any load on the bus removes the issue is an indication that there essentially is no issue.

They should have only done tests with a decoder on the track and the scope in parallel to it.

As to why the track matters, well it's all moot now, but if we were to believe that there's a need for termination, then, feeders and track will screw up the transmission properties even further.

Typically, bus lines are relatively large gauge and feeders quite small. Then the track itself, is not only a different dimension, but a different physical shape and a different metal.

From a transmission line point of view, this is a nightmare piled onto a travesty wrapped up in a catastrophe.

A little transmission line theory tells us that this string of conductors is a horrible transmission line with electrons bouncing who knows where. The fact that it works at all, is because the system is designed to use a crappy transmission line and as such is designed to not need to even consider termination. Because how in the heck would you even do that? You wouldn't just terminate the bus, you'd need to terminate each stub section of track on the entire layout...if you were to believe that such a thing were needed at all.

 

Can you move it?

 

Keep obscuring the true science.

The true frequency content of the DCC waveform is the convolution integral of the varying PCM modulation and the packet repetition frequency which is again convolved with the mutual coupling function of the parallel tracks above. That gives the final time domain answer. Next we perform a Fourier transform of the time domain function and see what really happens. There is substantial power in frequencies way above the nominal 8 KHz DCC base signal.

This is how science is done. That is not to say that the terminator will not work, it will work in some cases just not all. The best analogy I can come up with involves America's Pastime, Baseball. Just think of all the games that have been pitched, that is the general solution. Now look at all the perfect games and no-hitters, they are anecdotal because they do not happen every day. The ratio is something like 1 perfect/no hitter game for every 27,000+ games pitched.

All I am trying to get across is that it isn't going to work for everyone. Its numbers may be better than major league pitchers, but depending on complexity of the track etc, it may make things worse.

If you use high quality cable for your bus wire. And similarly high quality wire for your feeders will actually help a great deal. High copper purity oxygen free wire will do more to help than anything else. The best size wire is actually 16 gauge and 20 gauge feeders spaced every feet. In this arrangement, the feeder wires act as a low pass filter cleaning the ringing without terminations. This can all be proven mathematically, and empirically.

 

After a lot of web research, this is my understanding of the need for terminators.

“A terminator is not necessary on a DCC layout. If such devices are necessary on a DCC layout, then it is likely that the layout has other electrical problems which should be investigated prior to fitting a terminator.”

However, the transients associate with shorts and the short-detection method of the power station could be a reason to install terminators. These shorts are somewhat like miniature lightning strikes on you railroad. The size of these strikes are directly related to the amperage output of the “power station”, the energy storage capacity of everything connected to the power-station output, and the lack of resistance (or other protection) between the energy storage capacity and the susceptible equipment. In other words deviating for normal good practices like more power-station amps, less resistance in the DCC path, and ... are not nessessarily a good thing.

As with lightning, it is a matter of individual choose about now much protection you want and what you are willing to pay for it.

My best investment in DCC was for a remote AC power switch on my power-station. I wear the remote around my neck so a can quickly kill track power for those inevitable shorts.
Bob

 

I'm also a little concerned about modular setups... which happens to be where I seem to find the most interest for these things.

At a modular show, you're likely to find very long bus runs of varying types/quality of wire, closely spaced DCC wire pairs, and lots of inter-module connectors creating discontinuities in the bus wire. Electrically, as far as wires go, it would seem a very "dirty" situation, and with few opportunities to clean it up.

I just don't know enough about all the ways things can get messy in a setup like that to either prove or debunk the need for them under such circumstances.

 

Well here is a thought for your endeavor. If you change it to be a variable RLC filter you could then make it somewhat tuneable. You could make two ways one as the end terminator, and the other as a dongle between modules. This way, with the help of an oscilloscope, you would be able to tune the termination over a much wider audience and over many more possible solutions.

The RC function works on inductive reactance where the RC equivalent reactance reduces the inductive reactance. if the end of the line looks capacitive, then adding the capacitive termination would make things worse because it will increase the capacitive reactance.

To reduce a capacitive reactance requires an inductive reactance. When they are equal then there is no reflection that leads to the spikes or the degradation of the DCC signal. When the sum of the reactance is zero, you have what is called a matched load. That is the best you can get.

So setting up an RLC type termination could be adjusted either way. This would make the product most adaptable. It would not solve all issues, but it would solves a great deal more than just the RC solution.

 

TwinDad,
You could use a not-so-valuable low-quality DCC locomotive to run some tests on any track sections that you are concern about. In addition to the usual tests, put the locomotive on the track as near to the DCC track connection(s) as you can, turn the light on, and create a few shorts near it. Of course you should probably clear this test with whoever is in charge, and remove all the other decoders. These short tests should at least provide some confidence in the DCC setup.
Bob