Closed Loop DCC Bus follow up question

[Anonymous]

Someone asked the question the other day about whether or not to close the loop of the DCC bus, and thus sending the signal by shortest path of resistance over the layout. For instance, the wire for rail 1 is on bus wire 1, and rail 2 is on bus wire 2, imagining the track makes a simple circle, you would connect both ends of bus wire 1 (to make a circle) and connect both ends of bus wire 2 (to make a second circle. Someone then replied something to the effect that this would potentially 'echo' the digital signal. Well... isn't that exactly what the rails do if you install them with no gaps? Don't the rails complete a full loop in the circle scenario, and thus, wouldn't you be essentially closing the same loop as the bus is powering? So, if you had a continuous loop of mainline, why wouldn't it be practical to close the loop of buswork (each end of wire 1 to wire 1, and wire 2 to wire 2) in order to do essentially the same thing that the track is doing? I think the 'path of least resistance' idea makes alot of sense, since logically, the track would be doing the exact same thing anyway.

[Blind Bruce]

I tend to agree with you entirely, IF you consider DC (direct current) only. However, DCC is AC plus a higher frequency used for encoding. I do not claim to know much about DCC but when high frequencies are involved, even a slight bend in a wire can represent an inductance that could impede signal flow. Therefore, I am taking the lead of those experienced in DCC and installing jumpers every so often and not connecting the ends of my bus wires.
BB

[Pennsy58]

I've been following this topic and do not quite understand any reason why you cant close the bus loop. I always understood it to be that it isnt required to be closed back on itself. The track completing a loop does make a compelling argument against claims that all kinds of electrical conditions are introduced by completing a bus loop.
But then, I'm not quite an electrical current expert either.

[Anonymous]

It makes no difference to me, really, which method (open ends or closed loop) to use, and I plan on using the jumpers to the track every so often to increase reliability, but when the other guy asked the question regarding whether or not to close the loop, it made me wonder about the track situation, because isn't that essentially what a continuous loop mainline in DCC would accomplish? All you're doing (if my electrical terminology is right) is making a parallel circuit by closing the bus ends. So long as you DONT connect each bus wire to the other wire (aka, wire 1 and wire 2 together) it seems to me the electrical circuit would be completed through the motor in the loco, and since the signal travels throughout the rails anyway, until it finds the loco it was sent for, it seems to me by closing the bus loop, and thus making an open circuit parallel with the rails, that the signal would in fact travel the shortest distance possible to reach the loco in a closed loop bus. Maybe I'm wrong, but the other topic got so many negative "DO NOT close the bus loop" replies, but no one really had a good (in my opinion) reason why other than "Refer ro so and so's post or website" Not any actual DCC or electrical theory behind how or why it wouldn't be a good idea. That's all I'm asking really, is electrically speaking, why is it not a good idea, or why wouldn't it be a perfectly legit idea? I did see at least one post from a guy who does exactly that, and says he's had absolutely no problems with it. Seems to me it might also reduce the need to have higher amp boosters on large home layouts that only run a few locos at a time, due to the length of the runs of wire, since I know voltage and current drops over so many feet of wire length unless you go to a large gauge wire. Seems to me, you could run an average size DCC wire, in a continuous loop paralleling the mainline, and because the wire is constantly looping back on itself would simply decrease the need to use larger current to run from one end of the layout to the other. Of course, I'm not very knowledgable about the specific electrical theory involved with AC bus work, so I also don't know if closing a loop like that would essentially make a complete (thus short) circuit, and eliminate any power going to the rail. But there again, isn't the rail itself electrified with AC current that is the carrier for the digital DCC signal? And if the rail is a continuous loop, wouldn't that essential create an AC short circuit? So it just doesn't make sense to me why closing the loop of the buswork would be a bad idea. Anybody with electrical and DCC knowledge wanna chime in?

[Anonymous]

I have some (very small) DCC knowledge, some greater electrical knowledge. Here is an opinion based on that:

DCC does encode a digital signal on the rails (and on the bus, for that matter). I think that the frequency of the signal is actually pretty low, in electrical terms, or it would not be forgiving to all of the various wiring methods that obviously work. So I'd start by saying that especially in a smaller layout there it isn't going to make a bit of difference. If the path is long enough, and the frequency is high enough, there theoretically could be trouble.

So given that, it seems that the right way is to not close the loop for the bus, and to break the loop in the track as well. Even if you don't use additional boosters Joe Fugate's FORUM CLINIC:Twelve Years Experience Using DCC http://www.trains.com/community/forum/topic.asp?TOPIC_ID=36389
shows some good reasons to break the loop, and an interesting method.

So, I'd say the answer is that strictly speaking you only want a single path from the sorce to the destination, so not completeing the loop is correct, and for the some reason, as well as others, the track should not be continuous either.

So, yes, theoretically a looped bus could add interference, though in most cases it will probably not be noticed. Thinking a bit more, the bus should probably be broken at the point farthest from the source, to minimize the maximum distance to the destination. In some cases more of a star bus might make sense, as long as you always know that any section of track is only powered by one leg of the star. you could even go with all "home runs", but I think that's over the top!

Jeff

[Blind Bruce]

Jeff maks a lot of sense also. Another point to ponder is wavelength and phase distortion. This can be a result of a digital signal having two unequal lengths to arrive at the destination.
BB

[Anonymous]

Well so far, the phrase "small layout" has been mentioned a couple times.... what would you consider a 'small layout' for our discussion purposes? For instance, I can understand a 4x8 or a shelf layout or small switching layout is fairly small, but for our discussion, would a continuous loop on my 16x21' dogbone around the walls with peninsula be considered 'small' since it's a home layout, not a huge 100 foot long by 50 feet wide club or show layout? You bring up some very interesting points! I'm liking this discussion more and more already! :-)

[rrinker]

What I did on my thus far 8x12 layout section with double track loops is brea the track bus into 4 sections. The connection point is in the middle fo one of the 12' sides. From there a bus for each main goes out clockwise tot he middle of the opposite 12' side. And another set of bus wires goes out counter-clockwise to the middle of the opposite 12' section. The do not connect to one another, there are 4 segments of wire, each bus run is about 20' long. The only track gaps are at crossovers between the inner and outer loops, as eventually these will be separate power districts. For now they are jumpered together at the terminal board.
For a 16x21 layout, I would certainly not run around the entire loop - even with #12 wire you want to keep your bus lengths to 50' or less to avoid voltage drop issues. If you have more than one booster, I would put them on opposite sides, and run bus wires of equal length in each direction. Naturally, between the two boosters th ebus wires do NOT connect, and the rails need to be gapped.

--Randy

[howmus]

I think the reason for not making the bus a loop with the ends connected is that it is one more chance to introduce ring into the system. Maybe you won't have any trouble with it, but you may. With pro sound systems (something I do know a lot about) the problem is in a ground loop introducing hum into the system. If the ground can be accessed by more than one route, you will almost always end up with a 60hz hum in the system. Eliminating it can be frustrating. With DCC you can create an echo and distortion in the system by looping the bus. You will also gain nothing by making the bus a loop. So....... Why bother?

[kwillfixit]

As electrical signals travel along a conductor at approximately 2/3 speed of light there wont be any measurable echoeing of the siganl at the loco. There may be other, perhaps good reasons for not closing the loop but echoeing is not one of them.

[Anonymous]

I did a little looking, and I suppose it varies based on each decoder, but it looks like the decoders have a tolerance of something like +/- 6 microseconds on the timing from the rising edge to falling edge of a '1' bit. That's on the order of a mile in distance, which is why this probably isn't a serious issue. But, as soon as you have multiple paths there will be some interference on the edges, which I would expect to cause problems sooner than that. In most cases, it isn't going to matter, but it seems that not closing the loop (on the bus or the track) is the better proctice.

[howmus]

QUOTE: Originally posted by kwillfixit

As electrical signals travel along a conductor at approximately 2/3 speed of light there wont be any measurable echoeing of the siganl at the loco. There may be other, perhaps good reasons for not closing the loop but echoeing is not one of them.

I should post after midnight after a long day...... LOL I meant to say ring not echo (more like feedback). Again jshrade may not experience any problems connecting the bus ends together. However, it may promote some problems. I have not seen anywhere that it is a good practice to close the loop and there is no reason to do so. [:D]

[rrinker]

There have been instances with very long bus runs with open ends (unterminated) havign ringback distortion issues. Thus mention of passice terminators for the DCC bus - the power/signal bus, not your cab bus! Since the DCC signal is not a constant frequency waveformYou cna have all sorts of wierd issues where certain bit combinations are a a frequency where the bus length is a natural resonant even multiple of the wavelength and other bit combinations are not.

Before we get too deep into electronic theory and confuse newcomers even more, who will likely see stuff like this and go right back to DC saying "we never had to worry about this stuff before" - this sort of issue is quite rare. DCC is quite robust, 99.9% of all users will never see a problem like this, even with 20x40 basements full of layout. This sort of discussion is more on the fringe of Nerdom, and is not something the average DCC user needs to worry about.

--Randy

[Anonymous]

I think that's the point. The signal is really pretty low frequency, really high signal to noise ratio, and probably tolerates all sorts of things. I was wondering about terminations, but didn't want to think that hard, still don't. I think that making good connnections, not using tiny wire, and breaking things up a bit should eliminate all but the strangest problems.