"In theory", a loop of the bus could cause trouble. In practice, I don't think I have heard of any cases where it actually did. Personally, I would not complete the loop, but I wouldn't undo it if it was already that way. If there were any isolated power districts, for short management, the track would not complete the loop either.
The one issue with this table is the domain of validity. All American Wiring tables were developed using 60Hz AC power. DC and DCC are not anything like 60Hz AC so results can vary wildly. There are these things called frequency dependent reactances that are different at different frequencies. It may actually have less loss at 2000Hz than at 60Hz. Like it or not, the DCC buses are in fact transmission lines and the rules are different. Just a word of caution to not blindly follow these "rules" In fact, since DCC is higher in frequency content, you can actually use smaller wire and not lose and power over that lost in the larger wire. It also is a function of the quality of the wire too. High purity copper 16 gauge wire will have substantially less loss than the same length of low purity copper in 12 gauge wire. Blindly following rules like this is akin to thinking the sun revolves around the earth, and that the earth, in fact, is flat. This was the common thinking during the dark ages. If you check your local library for the Radio Engineers Handbook or the Antenna Engineers Handbook, you will find a very different story about the gauge of wire based on the frequency of operation. Oh, and you will find that the twisting is not quite right, that a simple conjugate load at the end and parallel lines is what is known as a balanced transmission line. There is no ringing and maximum power transport. The twisting of the wires removes that balance and can actually increase the ringing. It is really a happy coincidence that some twisting has helped. In actuality it did help somewhat because it creates a reflection point that induces a small amount of the power to flow in opposition to the main pulse, effectively reducing the tops of the pulses a little bit. This is seen a lot in microwave and other high frequency circuits as the theory of multiple reflections. It is not a desired way to move power across the wires. The use of an arbitrary RC circuit at the end may help some what, but if you have the correct load circuit on the end tthat forms the conjugate match, you will have no ringing at all. But in truth the correct circuit would be an R-L-C. The really crazy part is that if you use the same R-C circuit at the ends and the load is not that same at both ends, ie not symmetrical, then you actually can increase the ringing. All that said, going with quality wire over the cheap stuff will always be better an going bigger on the wire can also help. For the record, I have powered a 200 foot set of Ntrak modules with one controller and no boosters. I used 16 guage, properly balanced and terminated wire on all of the modules. The voltage at each end of the layout was 11.25 volts. With the oscilloscope, there was no ringing and only 1% distortion of the pulse train. As a microwave and antenna engineer, I can make the calculations for the end of bus loads to provide this level of operation. This is my area of freakdom. My point is that with a little more knowledge about what is really going on, better decisions can be made and you can have cleaner and more reliable operations.
I'm not an expert on DCC, but don't you think twisting the bus wires together introduce Charecteristic Impedence (resistance) in the bus? Particularly at the 8KhZ DCC frequency
i have a question on twisting the bus wires togather. How does this work when your useing blocks? lets say you have 6 blocks. do you twist all 7 wires togather? or is this not required when useing block wireing for detection?
Guys The NCE Power Cab manual (p13) recommends 18 gauge for runs up to 25'. I am designing a 4x10ish layout. An early response in this post called for a 4' bus down the middle, with 5' lines going out from the middle. So thats about 44' of wire, then a multitude of feeders, so total length might be a lot. At most any point might be 9 - 11' from the power supply. Am I considering the longest distance from the power supply or the total wiring, when considering gauge? Thanks
Think the relevant length measurement is the shortest path to the track from the posts on the booster, not the total length of all bus wires. Look at your track plan and decide where you will do the power drops of light gauge wires. Your bus need only serve those drops. My solution for 4x10, 4x12, and 4x14 layouts used one 16 ga. bus. Longest distance from posts to drop was certainly less than 5'. All drops were about 8" and 22 ga. There was only one drop for each isolated section of track. I spliced 16 ga. branches into the main bus to get the 16 ga. near the drops. When I changed layout sizes I just ignored unused branches. Used solderless connectors except for the 22 ga. drops at the track. And this eas a little overkill, because I now am doing a 2x8 layout for which all layout wiring is in the surface of the foam base and is 22 ga. In this case I have run very short 16 ga. from booster posts to three connection points for the track wiring. Again only 1 drop per isolated segment. The Peco c55 track carries the electricity just fine w/o soldering track joints. Simple, easy, and works great. No idea how large a layout would be that would require a more elaborate solution. Pic shows the bus that served all sizes of layout. I twisted the wires for ease of handling, not for any electric or electronic issues. You can see & touch the wires from track to posts to troubleshoot any problems.
Term "common rail" flies a big red flag. Do you mean it in the traditional sense where one rail/wire carriers the polarity throughout the entire layout? In a later post I see you said "a rail & b rail buses" which is okay. Just checking.
DCC ! I like to peek in here now and then to see if I can learn anything. I think I'm finally getting it. :wideeyes: BUT! I still don't understand this booster stuff. How do you know if you need one?
To determine if you need another booster, you need to add up all of the potential consumers of power. Locomotives Sound decoders Lighted passenger cars Cars with lights, Fred's that are powered by the track Losses in the wiring you use Losses in the track Losses in the connections if not well soldered Accessories like Occupancy detectors, stationary decoders and signaling controllers Add all of that up, and if it is more than 75% or the power out of the current controller/booster, then you need to add another one. The idea is to have a bit of room in the event of other unforeseen factors, plus it is never a good idea to run your power supply at full power, causes overheating and the is a slow death to electronics. Another thing that is sometimes overlooked is the grades and the curves. As an example, if memory serves me correctly, John Armstrong showed that a 180 degree curve with a radius of 16 inches is equivalent to a 2% grade. So if these are also on the layout, you may want to drop the safety margin from 75% to 60% to account for the extra power the locomotives are going to draw. It isn't really that complicated, it is just a lot of accurate book keeping on the power budgets. Here is a sample from an article I am writing. The idea is to figure out how much power the layout will use, and that will be a significant part of the decision making process. Let's work an example, numbers are assumed for this example: Layout Sources: 1. Command Station 8 Amps, 12 Volts, 96 Watts Layout Users 1. Locomotives, 10 count, .5 Amps each, 60 Watts 2. Circuit Breaker, 2 count, .1 Amp each, 2.4 Watts 3. Stationary Decoders, 3 count. .3 Amps each, 10.8 Watts 4. Wire Loss, .3 Watts So in this model we have a 96 Watt Supply and we are consuming approximately 73.5 Watts or just over 76% of the supply power. As a general rule, you never want to allocate nearly all of the power supplied.
I have an HO layout and I have used solid. I found it easy to bend and when I solder it it did not splay out as stranded tends to do.
