[Brian]

Hi all
I have been reading with interest on another forum the topic of whether of not the power bus (which also carries the DCC data signals) should/can be wired under the layout as a ring or radial circuit. Searching the internet I can't find any reference to ring circuits being used anywhere, as they all refer to radial power buses.

Will any short circuit currents be better and fast returning via a ring or a radial bus
Will the data on the power bus be best served by a radial or ring circuit?

So, is a DCC ring bus advisable or is it to be the radial or perhaps it just doesn't matter?

 

[Richard Johnson]

QUOTE (Brian @ 13 Feb 2008, 23:00) <{POST_SNAPBACK}>Hi all
I have been reading with interest on another forum the topic of whether of not the power bus (which also carries the DCC data signals) should/can be wired under the layout as a ring or radial circuit. Searching the internet I can't find any reference to ring circuits being used anywhere, as they all refer to radial power buses.

Will any short circuit currents be better and fast returning via a ring or a radial bus
Will the data on the power bus be best served by a radial or ring circuit?

So, is a DCC ring bus advisable or is it to be the radial or perhaps it just doesn't matter?

***Brian, on a small layout its not critical, but the generally recommended configuration on many fora from many wise men over many years is radial - command station at the centre and roughly balanced left/right length.

There is not benefit power wise from a ring and there are potential negatives for data as the layout size increases with a ring... so if a simple answer is wanted "radial" is best.

At this point someone usually chimes in "but the track is a ring" - the answer is no, it is not - unless it is a simple oval , otherwise its a series of short sections, as every item of pointwork is effectively creating a break in the "track ring"

Richard
DCCconcepts

 

[ebaykal]

Can you guys please explain what you mean by RADIAL?

Am abt to wire up my layout. Pretty smallish, 3mx2.5meters around the room. Planning 1.5mm sq for the bus wire and
0.25 mm sq diam. wires for the droppers.

Am I on the right track?

Baykal

 

[BVM]

QUOTE (Brian @ 13 Feb 2008, 14:00) <{POST_SNAPBACK}>Hi all
I have been reading with interest on another forum the topic of whether of not the power bus (which also carries the DCC data signals) should/can be wired under the layout as a ring or radial circuit. Searching the internet I can't find any reference to ring circuits being used anywhere, as they all refer to radial power buses.

Will any short circuit currents be better and fast returning via a ring or a radial bus
Will the data on the power bus be best served by a radial or ring circuit?

So, is a DCC ring bus advisable or is it to be the radial or perhaps it just doesn't matter?

It is important to distinguish the connections between the handsets (assuming you have more than one) which MUST be radial (that is a handset can only receive a signal from one direction) and the connections from the command station to the track which can be either but unless the whole layout is one section is most likely to be a series of spurs (that is radial).

BVM

 

[Brian]

QUOTE (ebaykal @ 13 Feb 2008, 16:40) <{POST_SNAPBACK}>Can you guys please explain what you mean by RADIAL?

Am abt to wire up my layout. Pretty smallish, 3mx2.5meters around the room. Planning 1.5mm sq for the bus wire and
0.25 mm sq diam. wires for the droppers.

Am I on the right track?

Baykal
Hi
Radial means not returning back to the place where the supply was inputted (Which is what a ring does) or in other words it starts at the supply and ends somewhere else. You an also have a radial that's 'T' feed. i.e. from supply to roughly the centre of the two main bus wires which then radiate out around the layout in roughly equal lenghts.

I have used a 2.5mm solid copper insulated wire (ex mains power point cable) as a bus on my layout (in radial format) with 16/02mm droppers coming off the bus and then connecting onto solid wire droppers from the rails above. These are made from some 1.0mm bare copper earth wire. But you can solder the 16/02mm wire directly onto the rails if wished.

 

[Chinahand]

QUOTE (Brian @ 13 Feb 2008, 22:14) <{POST_SNAPBACK}>Hi
Radial means not returning back to the place where the supply was inputted (Which is what a ring does) or in other words it starts at the supply and ends somewhere else. You an also have a radial that's 'T' feed. i.e. from supply to roughly the centre of the two main bus wires which then radiate out around the layout in roughly equal lenghts.

I have used a 2.5mm solid copper insulated wire (ex mains power point cable) as a bus on my layout (in radial format) with 16/02mm droppers coming off the bus and then connecting onto solid wire droppers from the rails above. These are made from some 1.0mm bare copper earth wire. But you can solder the 16/02mm wire directly onto the rails if wished.

Hi Guys,

I was also going to follow domestic electrical 'best practice' and install my DCC Bus as a ring main but thanks to Brian, who incidentally has many years of model railway experience and is also an ex BR signalling engineer kindly explained to me as follows why it should be wired as a Radial.

If you have a DCC ring main and you send a signal/instruction to a loco which is on a section of track which is connected say 1/3rd of the way round the DCC ring main the signal will actually travel in both directions around the ring main and will arrive firstly along the shortest route between the controller and the loco.

However, milliseconds later, the signal will arrive again along the longer route between controller and loco.

The loco will therefore receive the same signal twice which can, apparently, cause its little brain to become confused and may result in unpredictable consequences. Hence you should use a DCC Radial Bus to ensure that the signal/instruction only arrives once at the loco.

Brian, I hope I have done justice to the explanation you gave me.

Regards,

Expat.

 

[Gordon H]

QUOTE (Expat @ 14 Feb 2008, 12:11) <{POST_SNAPBACK}>If you have a DCC ring main and you send a signal/instruction to a loco which is on a section of track which is connected say 1/3rd of the way round the DCC ring main the signal will actually travel in both directions around the ring main and will arrive firstly along the shortest route between the controller and the loco.

However, milliseconds later, the signal will arrive again along the longer route between controller and loco.
Not milliseconds. Nanoseconds at the most - one million times less.
To get a difference of one half of a DCC '1' bit (58 microseconds), the path difference would need to be about 17.4km!!

QUOTE The loco will therefore receive the same signal twice which can, apparently, cause its little brain to become confused and may result in unpredictable consequences. Hence you should use a DCC Radial Bus to ensure that the signal/instruction only arrives once at the loco.
Once more...
The frequencies involved in DCC are way too low, and layouts sizes are way too small, to cause problems of this kind.
Try it both ways - see if you can detect a difference.

 

[Brian]

Hi Expat
I think following some heated debates elsewhere, the ring really isn’t necessary so long as the radial bus wire is of the correct minimum size wire gauge and infact I can find no reference anywhere by anyone in the US (Where DCC has a much larger following) wiring a bus as a ring!
I'm no computer buff!!! and my understanding of data transfer is exactly as you have copied from one of my earlier comms with you.

Richard Johnson has I feel summed it up rather nicely too in the above, with the comment "There is no benefit power wise from a ring and there are potential negatives for data as the layout size increases with a ring... so if a simple answer is wanted "radial" is best."

While my web site also states..
"There is great debate as to whether or not the data bus should be in a ring"

I think the reference I made to milliseconds later should have been micro/nano seconds later but it was only a simplistic generalisation.

So for me it's the radial bus

Regards
Brian

 

[Richard Johnson]

I think both the attempts to quantify it and their qualifications are well meaning guesses or "illustrative", and these will always fail as they presume something in isolation.. which related problems never are..

Trying to express the cause and effect in simple terms will always fail. The point is that the bus on a model railway is always imperfect and will always have more or less of other influences on it.... Inductance, resistance, accidental shortcomings of various kinds.... careless connections, bad wiring layout etc etc.... all accidental and really largely unnoticed and unseen.

DCC is rubust & reasonably tolerant but often, irrespective of layout size, circumstances can combine to make things work less than perfectly.

All we can do is take aspects of best practice and apply them where we can - trying to create a real world set of recommendations that will give the "average but careless with wiring" modeller a best possible chance of a low fault or frustration rate:

Therefore we who advise consistently try to recommend certain ways of creating the wiring configuration such as separated or twisted bus, certain sizes of cable and certain other practices. Each may indeed seem to be slightly overkill if taken in isolation, but all taken together and applied as recommended, will result in an almost 100% success rate.

So - forget rationalising one part of it: It will always fail as a debate. We are trying to give pragmatic real world advice that will always work for the average railway modeller who considers most things connected to wires as akin to witchcraft. A tee or radial bus is part of this.

In the real world too: layout size isn't always needed for ring Vs raadial to have a real world difference

I have modified busses in layouts large and small to resolve problems. In at least two real world cases, both smallish layouts by world standards, removing the ring and then adding termination/suppression to each of the bus ends resolved several problems that had bedevviled the layouts and frustrated the owners for months.

Technically why - it des NOT matter - it did the job... removing the ring worked. Adding the 150 ohm resistor+ 10microfarad apacitors as resistors worked too. Which was most responsible for the improvement, I have NO idea - I don't take a digital storage scope with me to look before I act.

The point is, BOTH are issues which are agued over ad nauseum for the sake of it by those who just like to argue.... but the truth is, pennies in parts and ten minutes did the trick...less time than I've had to spend on the last wo replies. Why waste time in questioning a pragmatic reality when there is no benefit to the argument, and no real cost in time or money to the adoption of the "conventional wisdom"

Sometimes, ignoring the debate and "just doing it" is the right answer .

Richard

 

[Gordon H]

QUOTE (Richard Johnson @ 14 Feb 2008, 14:01) <{POST_SNAPBACK}>Technically why - it des NOT matter - it did the job... removing the ring worked. Adding the 150 ohm resistor+ 10microfarad apacitors as resistors worked too. Which was most responsible for the improvement, I have NO idea - I don't take a digital storage scope with me to look before I act.
But it does matter.
Apparently solving a problem without knowing how or why the cure works does not really constitute a solution, certainly not a repeatable one. Without understanding its mechanism, you cannot know for sure that the problem won't re-occur when presented with a slightly different, but equally valid set of conditions.
A DSO is precisely the sort of thing that SHOULD be taken along when investigating this kind of matter, so that the effects can be understood and quantified for future reference and explanation. Otherwise, we go round the whole thing time and again when others inevitably ask the same question.

 

[Richard Johnson]

QUOTE (Gordon H @ 15 Feb 2008, 22:15) <{POST_SNAPBACK}>But it does matter.
Apparently solving a problem without knowing how or why the cure works does not really constitute a solution, certainly not a repeatable one. Without understanding its mechanism, you cannot know for sure that the problem won't re-occur when presented with a slightly different, but equally valid set of conditions.
A DSO is precisely the sort of thing that SHOULD be taken along when investigating this kind of matter, so that the effects can be understood and quantified for future reference and explanation. Otherwise, we go round the whole thing time and again when others inevitably ask the same question.

*** Gordon, you may enjoy the technical discussions / have a far deeper set of knowledge than most on this sort of subject and I will give you a qualified answer.

"Yes of course it matters that in the background, each recommendation is based on fact - but it is NOT helpful to over-technify issues on a general list where modellers want simple best practice guidelines and fixes".

It is of course possible to quantify anything in isolation - however every layout is a non-repeatable set of "wire knitting", with more or less of the small errors or issues that create problems - so any qualification has little or no merit.

If I was to take an analytical approach to WHICH of the problems was mostly guilty on a layout before I touched it then I'd waste hours when I could already have most of them fixed. Each contributes more or less depending on the layout, each is a real issue - that is all that matters.

To see the scope of such things / if you wish for an example please look at the latest NMRA "Clear Block" article by Didrik Voss, the technical chair of the NMRA - he simulated DCC bus configurations and prolems and analysed them VERY well. ie - he created a simulated typical modellers bus configuration, then measured the results - following this with various approaches and filters to fix it. His analysis was professional and really well presented. You'll need to be an NMRA member to get the article though...

THIS is what you are saying we must do on list??? I respectfully disagree - it is several pages of very technical data that would put modellers who just want to run trains off in droves when the "useful answer" is:
(1) don't make a ring main
and
(2) add a resistor+ capacitor across the ends of the bus to suppress voltage spikes and assist maintenance of good CC signal.

So...

I maintain that the backgound of such issues is actually totally irrelevant as long as they are founded in fact and they work - just as a modeller does NOT need to know how his loco's motor works or how his DCC system knows which loco is which. I can say with absolute confidence that all the things recommended ARE based on fact and DO contribute to better layout performance - so any layout created with the guidelines expounded earlier in the thread WILL work well and reliably..

So - in fact, my general answer is "Nope - it doesn't really matter - and most definately it is of no interest to the modeller whose layout is fixed and not to the one who follows advice and avoids problems.

Surely... If we can give a simple set of instructions that when combined will sort the problem, then that is all that matters.

Kind regards

Richard
DCCconcepts

 

[Brian Considine]

QUOTE (Richard Johnson @ 14 Feb 2008, 14:01) <{POST_SNAPBACK}>Sometimes, ignoring the debate and "just doing it" is the right answer .

Richard

Exactly - you don't have to know why a car (or computer, or telephone or..........) works to use it & use it effectively & reliably !

 

[Doug]

Richard, you say:
QUOTE add a resistor+ capacitor across the ends of the bus to suppress voltage spikes and assist maintenance of good CC signal.
I know that you have said this before. Can you elaborate please.

What resistance and capacitor values should be used and should these values change if the voltage goes up or down or the size of the DCC power bus changes?

 

[Richard Johnson]

Hello Doug

Use a 1 watt 150 ohm resistor in series with a 0.1mfd ceramic or "monolithic" capacitor. Neither is polarity conscious. wattage of the resistor isn't critical - anything 1 watt upwards will be fine if 1 watt not available locally. Cost of these parts will be less than 50p

wiring is bus wire 1 > resistor > capacitor > bus wire two.

No need to change values at all withing the normal 11~18volt range of DCC bus voltages.

Richard
DCCconcepts

 

[Gordon H]

QUOTE (dbclass50 @ 18 Feb 2008, 12:06) <{POST_SNAPBACK}>Exactly - you don't have to know why a car (or computer, or telephone or..........) works to use it & use it effectively & reliably !
But you will need to know something of how a car works if you going to be wiring it up yourself.
That is effectively what buyers of proprietary DCC equipment are doing, quoting this example.

Admittedly, as an engineer, I will always want to know the 'whys' and 'hows' of such matters rather than just being presented with an apparent fix with nothing to back it up.

What most people seem intent on overlooking is that creation and running of railways, both real and model, is an engineering task whether they like to admit it or not. Electrical, mechanical and civil engineering are all involved to some extent, and until you can go to a shop and buy a completely finished RTR layout with rolling stock there will always be a need to have some understanding of the principles involved.

OK, transmission line effects on DCC waveforms are a bit esoteric for a list such as this, but the question did arise, and solutions were offered - which in principle ought not to be necessary based on the engineering figures available. In that circumstance, where the theory doesn't necessarily agree with practical experience - surely it is acceptable to ask for an explanation?

 

[TonyDaly]

Hello Everybody.
I have just read this technical discussion & I have to say if I went as deep into the DCC subject as some here I would run for the hills as they say.
To DCC I am a newbie. I am just back to Railway Modelling after many years out of the hobby. If I didn't know better & took notice of all the deep discussion here then I wouldn't take up DCC in a million years.
I don't need to know the in's & out's of why something works, I just need it to work. When my layout is ready for DCC wiring I will take a practical route to doing it. I will use a Radial form of Power bus & I will twist the bus wires & terminate them as described. Not because I understand how it all works but because it is said by many with more knowledge than I to be the correct way. If I can't get it right I will ask & hopefully someone will give me a practical answer to my problems. I have a general knowledge of DCC but I am no "Expert"
By all means continue the deep discussion but you will have left me & many more "Railway" modellers far behind.
Don't frighten people off by being too technical. Sometimes a simple picture is better than a volume of knowledge.

Regards.
Tony Daly.
Southern Ireland.

 

[poliss]

If anyone new to the hobby read this debate, then they'd probably go and take up something simpler, like nuclear physics.

 

[Brian Considine]

I am also an engineer (electrical) & there was a time when I wanted to know how everything worked. Now, I tend to ask myself the question "do I really need to know how this actually works ?" - often the answer is no, not really, I have other things to do. My time is not infinate.

 

[Hugh Williams]

QUOTE (poliss @ 18 Feb 2008, 17:23) <{POST_SNAPBACK}>If anyone new to the hobby read this debate, then they'd probably go and take up something simpler, like nuclear physics.
Hi Poliss,

Can't help but observing that, despite the learning scale that has to be surmounted, DCC makes things immeasurably easier and more fun. It also allows you to get on with running locos/trains faster in the end.

Cables and Buses:-

For others in the debate, the radial bus (or if you want to think Data Networks, think of a 'Star Network') is just the bits and easy to put into effect. It starts with one pair of cables coming out from the relevant connection on the DCC Controller which then connected upto a choc box and then run off a daisy chain. From each of the pairs in the daisy chain you then run off a separate bus. The daisy chain replicates the number of running lines, say four if your layout has four parallel track circuits. These radial pairs/buses are separated on the track by insulators between facing pairs of points. If there are extra lines within this type of circuit, say loops/sidings, then use a further leg. To make matters easier for identification purposes, use different colours of cable for the droppers between the track circuit and the bus. I hope this makes sense, as it is easy to do. I hope I haven't gone on too much, or confused the issue. This construction of the bus is working fine on a 20' x 12' loft layout, with additional power districts (another subject) attached to it.

Additionally, a separate bus is used for the control of accessory decoders, used for point control. This is important, and, should influence the type of controller you buy.

Finally, follow as much advice as can be obtained, and, in particular, the terminators to the bus. If I may make a suggestion check out the DCCUK forum, a useful area for learning and batting ideas around, not to mention read around the DCC Section in this forum.

Best of luck to the newcomers in DCC, it is worth the work and the layout of cash for the loco decoders and ancillary equipment. It is also possible to run standard point control systems (AC/DC) separately from the DCC point control on the layout, so the cash layout for these systems isn't wasted!

Cheers,

Hugh

 

[poliss]

The language needs to be simplified. To most people, a bus is something you catch, radials are the tyres on the bus. Choc boxes are what you give your girlfriend on Valentines day. Daisy chains are things girls make at school.