DCC Bus wiring - to twist or not?

[Bill H]

Group;
Finally my basement is nearing completion into a finished more habitable space and in anticipation I am finishing up the first N scale module in my workshop that will move into the basement as part of a permanent layout. I am running Digitrax DCC and using BDL168s for detection and SE8cs for signalling.  ME Code 40 track is laid and I am beginning to start wiring. I made an organizational commitment to keep components like the BDLs and SE8cs un-centralized and as close as possible to their primary uses to minimize detection and signalling wire runs.  Given the comparatively increased resistance in code 40 vs larger code rails, I am using drops very frequently, about every two feet, and all rails have drops. Rather than having all the drops hitting the bus wires directly, I was planning on using distribution boards spaced across the layout with the drops connecting to their closest distribution boards and the bus wires interconnecting the distribution boards.

I am at a point of installing the DCC bus wires and finding contradictory recommendations regarding twisting or spacing bus wires and secondarily twisting or spacing bus wires that are part of a detection section.

Thoughts based upon your experiences?

Kind regards,
Bill

[C855B]

I twist. I have a lot of years in 2-way radio installations and it's a necessity to keep the RF gremlins out of things like power supply wiring. I treat DCC buses the same way. Twisted pairs in a bundled situation also prevent crosstalk between pairs. That's why Ethernet and telephone cabling are bundles of twisted pairs.

Twisted pairs are also a lot neater. I create my own pairs by clamping one end to a table and chucking the other end(s) into a drill.

[Bill H]

Mike;
Thanks, that what I was taught years ago in networking from punch down blocks to token ring, Arcnet, Ethernet and onward. I have seen a number of well functioning layouts that I operate on that do it both ways. Digitrax notes it does not matter, but I think NCE suggests it.

I have also read not to twist near detection devices, even though the BDL168s cores are on the board itself and not remote. Last, out of an abundance of caution / concern I wonder if the use of distribution blocks might cause some harmonics even though I used a ton of them in the ancient past wiring up networks?

Kind regards,
Bill

[John]

Mike;
Thanks, that what I was taught years ago in networking from punch down blocks to token ring, Arcnet, Ethernet and onward. I have seen a number of well functioning layouts that I operate on that do it both ways. Digitrax notes it does not matter, but I think NCR suggests it.

I have also read not to twist near detection devices, even though the BDL168s cores are on the board itself and not remote. Last, out of an abundance of caution / concern I wonder if the use of distribution blocks might cause some harmonics even though I used a ton of them in the ancient past wiring up networks?

Kind regards,
Bill

Larry Pucket has a series of good videos on DCC wiring - pros and cons on youtube . also check out his video on snubbers ..

[peteski]

Since I cannot unsee it, I will again chime in on the naming convention.   "Buss" means "kiss", or also a brand name of electrical fuses.  The electrical wire arrangement is called "bus" (one es, just like the vehicles which carry passengers).

"Buss" fuses are actually "Bussmann" fuses, but as many other names it was shortened to "Buss"
https://bussfuses.net/

https://www.merriam-webster.com/dictionary/buss

https://www.merriam-webster.com/dictionary/bus 
Definition #3 points to:  https://www.merriam-webster.com/dictionary/bus%20bar  which colloquially gets shortened to just "bus".

[Chris333]

He's just bussing your balls. 

[Bill H]

Peteski:
Clearly did not mean to constipate the thread, particularly as that is how I was taught by IBM how to spell the term when I was working on very early 360s in the sixties. Obviously I have Been Misled. Corrective spelling applied. Meanwhile, in anticipation of the usual thread drift - any thoughts on issues that might be caused by using distribution blocks, other that some drops being longer than usual?

Kind regards,
Bill

[peteski]

LOL!  Friend of mine also worked for the same computer company, and he has a nice N scale  layout. He is very fastidious, but the DCC bus on his layout is not twisted.




The DCC bus wires are running along other wires (for Tortoise switch machiens) in the wire bundle.  The visible suitcase connectes are on the DCC bus wires).

I'm using his layout as a example because I don't have a layout, but I spent good chunk of my hobby time on  construction of his layout, so I'm pretty familiar with it.

While I'm not a radio-frequency engineer, IMO the possible issues with DCC signal degradation have been greatly exaggerated, unless you have very long buses (like on Mike's layout), or have a bus with multiple branches.

DCC is a high power, relatively low-frequency (as far as radio emissions go) signal.  Yes, it is a square-wave signal which is rich in higher frequency harmonics, but since it is high power, I believe that odds of external noise affecting the DCC signal are very , very small.  If anything it would be the DCC bus itself (with those high frequency harmonics) emitting electromagnetic (radio) waves which might interfere with some external devices.

We also don't know any details about your layout and its future DCC bus.  Will it be a a simple straight bus, or will it have multiple branches?  Where will it be fed by a DCC booster? While N scale locos require fraction of the amperage of larger scales, having many feeders to the track is IMO good idea.

[John]

Peteski:
Meanwhile, in anticipation of the usual thread drift - any thoughts on issues that might be caused by using distribution blocks, other that some drops being longer than usual?

I don't think it will be an issue .. DCC is pretty forgiving if you take a little care and wire it up neat (don't look under my layout  ) ..  I didn't wire for block detection initially, and retrofitted BDLs and other detectors .. which resulted in a less than neat wiring job .. but so far no problem ..  good luck

[davefoxx]

Five years ago, when I started my layout, I could not find a definitive answer regarding twisting bus wires.  So, I decided to just spread them apart (they’re certainly no closer than the parallel rails of HO track), which does make it much easier to tie the feeders in.  I’ve had no problem with the DCC signal.



Hope this helps,
DFF

[Ian MacMillan]

My DCC bus has never been twisted and I've never seen the need to. It's pretty forgiving and I'm not concerned about packet loss, interference, etc.

[wvgca]

i have a 15 foot by 16 foot layout, and the bus wires are seperated by about four inches, no twists .. been reliable for seven years now ...

[nickelplate759]

Seems to me that if twisting the bus were necessary, we would also need to twist the rails 

[MK]

Seems to me that if twisting the bus were necessary, we would also need to twist the rails 

Don't do that!  Easier to twist the trucks!   

[peteski]

How about what DCC manufacturers suggest?  Does any of the DCC manufacturers literature (or even NMRA DCC specification) recommend twisting the DCC power bus?  I don't recall seeing that anywhere, but I have also not immersed myself in that literature.   I only remember some manufacturers recommending (and even selling) snubbers to deal with DCC bus signal degradation. Yet, there is lots  of "helpful information" from the "experts" sprinkled all over the Internet.