Power Bus Design and Implementation This is crucial for reliable long lasting operation, especially if like me some of the wiring was tricky to install originally, is above other layers of baseboard introduced as an afterthought, so is now difficult to access. A further 10+ years on my ability to service, repair or modify has become increasingly difficult with age, what about the next 10 years or so as I head to 80+ ? This is the way I have gone about wiring my railway such that it outlives me and has a degree of redundancy built in due to how it is all connected up, different people have their own approach based on layout size, disposable income and skill set, you have to go about it in the way that fits your pocket and abilities. My preference is to solder wherever possible and the power bus is no exception. As a reminder I only use DCC for loco driving nothing else, pointwork is manual via wire in tube e.g. GEM, using conventional Peco switches and capacitor discharge units where wire in tube is not practical, or finger from the sky and Peco point motors in fiddle yards. Any accessories e.g. capacitor discharge units, turntables, cement works rotary kiln, are powered by old mains adapters as required, the exception recently being the TTR Elevator/Conveyors which are much more conveniently run as locos via heavy duty decoders. Therefore I only have one power bus, albeit split into districts, as described in the previous article, if you have DCC operated points they need to be on a different bus to avoid locos shorting out the supply against a wrongly set point and then preventing that point from being changed. Although I have split power distribution into districts, each district is wired in the same way regardless of its physical extent. Implementation I attach a pair of dropper wires from all pieces of track to the bus, even if they can only hold a small 4 wheeled shunter. I never rely on fishplates (I use Peco code 100 track throughout with their fishplates) for electrical connections as they are not reliable, even new ones have been a loose fit in my experience. I only use fishplates to mechanically align rail joins, nothing else, these spring fitted parts will weaken with age and become unreliable, DCC does not tolerate poor connections, stay alive systems are not the answer to poor connections and should be confined to locos that cannot due to design be fitted with contemporary power pickup solutions, e.g. high chassis weight, all wheel pickup, high quality conductive materials and soundly made connections. Fitting droppers to all track parts eliminates the aging spring nature due to the design of fishplates, so the fishplates then act as an absolute last resort, some built in redundancy to take over. Power loss by a dropper failure on one piece of rail may be replaced by power sourced from adjacent connected rails via fishplates at both ends of the affected part (except ends of sidings, but that is less of an issue). My layout is a fixture, it cannot be moved for exhibition use, this is important because the bare wire approach I use is AFAIK not permitted on an exhibition layout. Check this out if you consider the following approach for an exhibition layout. Specifying the Bus Details you need to consider by estimation Maximum load on a district, in amps Maximum load of all districts at one time at the power units Maximum voltage drop you can tolerate at maximum load Any overload protection devices, electronic or electro-mechanical will trigger in the event of a short circuit at the furthest point from them. From these the gauge of the bus wire can be determined and the size, current delivery capability of the power units, not to exceed 70% of their maximum rated outputs, for longevity. (1) and (2) depends on the quantity and age of the locos and their current requirements. (3) locos will still slow regardless of BEMF. BEMF will not compensate for low track voltage, BEMF will keep the loco running at the selected speed step whilst below the BEMF cut-off value, remembering that for e.g. 28 steps 28 represents the maximum DC voltage available from the decoder, it is then divided into 28 levels, of whatever the maximum available voltage is. In practice there will also be a small voltage drop across the devices that adjust the width of the DCC square wave applied to the motor (which affects speed) and to ensure it is of the polarity that matches the required direction of motor rotation. (4) If the resistance of the bus is too high then shorting the rails will not trip the overload protection devices, this could result in smokey locos or excessive heat dissipation of the power supplies. The simple coin across the rails at the furthest distance from the overload protection device will confirm whether bus resistance is low enough. Bus Wire I used bare tinned copper wire as it is easily soldered to, simple calculations were based on, from memory (because I cannot find my original calculations) :- maximum load current, e.g.4A (although breakers limit it to a maximum of 2.5A) maximum permitted voltage drop at that load, e.g.0.5V resistivity of copper wire, e.g. 1.724 x 10-8 Ω m maximum anticipated bus length, e.g. 8m From these factors the minimum cross-sectional area (csa) of the copper wire was determined, an example of the calculations is given in the part 3 as an addendum for those who might be interested. The gauge of wire selected was 18 SWG (UK) which was also stout enough not to need many supports per metre when run along the horizontal wooden framework. Note: some modelers advocate the use of self adhesive copper strip, beware, although these are wide they are extremely thin and so have a very small csa which limits their use significantly. I would also be wary about the long term effectiveness of the adhesive especially if attached to timber framework. In addition electrically overloaded tape on wood is a potential fire risk. Mounting the wire This was simply done by threading it through nail in place cable clips for 3 to 5mm diameter insulated wire, spaced about 200mm apart horizontally and 20mm apart vertically, two reels were used at once, one for each of the pairs of wires, saves running out midway through a run on one of the wires. I used the convention of marking them red or black. Example below shows a section of a mainline district D2A and a marshalling yard spur taken off from D3 and in this case to both sides of the wooden upright upon which the marshalling yard baseboards are mounted. The spurs for the marshallling yard span across up to 11 roads, 8 sidings, arrival/departure roads and a bypass road, this results in a large number of droppers which are spread acoss multiple spurs. As can be seen where they went over metal framing parts or in close proximity to others in narrow spaces they were protected with red or black PVC sleeving. The ends were terminated at tag strips where flexible stranded insulated wire of the same or larger cross-section, typically 32/0.2 connects spurs to the main bus or the main bus to the wall mounted connection unit. Connecting Buses to Breaker Box This connection unit provides the interface between the fixed bus wiring on the layout frame and the wiring of the breaker box, command and booster units on a wheeled cabinet. Flying leads from the breaker boxes plug into the connection unit attached to the wall next to the framework, these are also of 32/0.2 flexible wire and utiliise 4mm plugs and sockets. Continued in part 2
Power Bus Design and Implementation - Part 2 Dropper Wires - Underside Single strand (1/0.6) PVC insulated wire, passed through the baseboard after the track was laid, red and black pairs twisted together, hooked and nipped onto the bus wire, and then soldered in place, top row first followed by the lower row. The use of non-acidic liquid flux makes for sound connections, quickly and easily. If a wire is to be removed it is simply cut close to the bus wire, as there is plenty of bare conductor available so there is no need to de-solder it and risk detrimental heating of adjacent wires. Next, a completed area which also shows the wiring to one of the few point motors, frog polarity switching is performed by switching units attached to the motor. The red and black wires attach to the appropriate power bus district, the blue attaches to the frog. Point motor control is via the purple, yellow and grey wires. Note: it is better to use fine self tappping screws to retain the wire clips to the baseboards, nails are fine in the framework but not practical into marine plywood. Dropper Wires - Top Side It is worth cleaning the rail web with a fine riffler file even if the rail is new to remove any surface oxide and certainly if it is second hand track with a coating of dirt or paint, followed by tinning the surface. The wires were then stripped, cut, tinned and bent so that about 2-3 mm pressed flat against the web, the resulting spring action meant they stayed in place for soldering. When soldered quickly with liquid flux again, no damage is made to the sleeper chairs. A modest temperature controlled soldering iron was utilised as described in my article on tools. The joint on the web is adequately disguised by the track weathering, the more adventurous and discrete way is to solder the wires to the underside of the rails before laying the track, I took the easier option as I had a very large amount of track to lay and hundreds of droppers to fit! Besides it is all invisible to me now as I thought it would become. I would advise against soldering to fishplates due to their tendency to weaken their grip on the rail over time. Removable Duck Under Sections Here I used the 4 way connectors that were originally made for connecting hard disc drives in tower and desktop PCs, they use crimp connectors and have ample current carrying capacity for a single track, just be sure they are of a decent quality such that the plugs and sockets couple firmly. They would typically be fitted with wires coloured code red/black for 5V, yellow/black for 12V, here I use red/black for the feed to the lift out section and the yellow/black pair as the loop back to the protected approach tracks. The next two images show the underside of a lift out section which supplies the double track mainline from district 2A and the port branch line from district 6, the loop back from the bridge via the yellow/black wires can be seen at the tag strips marked D2A and D6. The approach sections of track only receive power from the removable section, so, no removable section, no power to the approaches. Hinged Duck Under Section Similar to lift out sections, micro-switches on hinged sections, when the hinged section opens power is removed from both approaches. This shows the micro switches to protect the approach to the hinged section, fed from district 6 bus at the Nena docks, both rails are switched independently, more safety, in case one micro switch fails to open. Testing This was done frequently in stages, say 10 dropper pairs at a time using a simple 12VDC power supply and a multi-meter, wire too many at one time and a wire or two is crossed it will be extremely tedious to locate. Testing with DC keeps testing easy, then once an area is in good order, including running a DC loco through pointwork, connect the DCC power units and final test. Rushing straight into applying DCC to a significant new build is just asking for trouble. Summary That concludes how DCC power is distributed around my railway, kept simple for ease of maintenance rugged to reduce the need for maintenance safe from me driving trains into an abyss Addendum: Example bus wire gauge calculations follows in Part 3 Next, dedicated lighting. Jim Index of Articles with Links
Power Bus Design and Implementation - Part 3 Addendum Example Bus Wire Gauge Calculations For those who are really keen, here is an example of calculating the diameter of the wire required for the power bus, I cannot find my original calculations but they would have been like this. Setting the Scene. The output voltage of the Lenz command and booster (aka amplifiers) units are adjustable in steps which is useful in our case as well as for being appropriate for different scales of locos. I will never be running my locos train-set style flat out so somewhat less than maximum voltage at the motor is acceptable, e.g. even 75% for realistic running speeds, which of course is a matter of personal choice. I have had no problems with any locos not reaching their maximum scale speeds as per their prototype based on my power bus calculations and power supply output voltage since implementation. In passing, there will also be a small voltage drop across the decoder, from track power input to loco output depending on the technology of the parts it contains and the efficiency of its design. These calculations consider the bus length as being for each of the forward and return paths, 8m each, not the total round trip of 16m. This also ignores a slightly higher resistance path in parallel being present, that is the track rails which in practice would lower the total path resistances. The power supplies nominally see the attached circuit as three resistances in series (ignoring the parallel track path), i.e. forward bus resistance to the load e.g. loco(s), the load and the return bus resistance. In practice it is further complicated by how far the loco is from its power supply so here I just calculate worst case, the furthest distance. As an illustration, if the power supply output is e.g. 13V, and 0.5V is lost in each of the forward and return paths at 4A that still leaves 12V across a motor drawing 4A at the furthest point. In practice the breakers are set at a maximum of 2.5A, and if that was exceeded by a double header something would be amiss even with my vintage locos, so I am being quite conservative in my calculations. As the amount of wiring for my railway was going to take a lot of time to complete it was better to over engineer than under followed by increasing capacity, so the calculations are based on the longest paths, many of the rest are probably about half that length. Common materials and techniques throughout makes purchasing of materials more cost effective too. If all the layout track was fed from one bus then a much heavier bus or multiple ones would still be required, besides I have stated earlier in the article on districts, feeding 4 amps+ into a loco and a short occurs will also burn pickup wipers which are unlikely to be easily replaced in any contemporary models. So, on to the calculations. Taking the simple approach in three steps Step 1) To find the tolerable bus resistance using Ohm's law of V = I R, V = permissible voltage drop, e.g. 0.5V I = maximum current in amps, e.g. 4A R = resistance in ohms, to be determined Re-arranged in terms of R, i.e. R = V/I, to determine the tolerable resistance e.g. for 0.5V drop with a 4A load, R = 0.5/4 = 0.125 Ω Step 2) To find the minimum cross-sectional area (csa) of the wire Pouillet's law R = (ρ x L )/A, (Ref 25.1) R = resistance in ohms, 0.125 Ω ρ = resistivity of copper in ohm m, 1.724 x 10-8 Ω m L = length, approx 8m (8m forward, 8m return as reasoned above) A = wire csa to be determined, m² Re-arranged to determine the area (csa) Area = ρ x L / R e.g. csa = (1.724 x 10-8 x 8 )/ 0.125 = 1.103 x 10 -6 m² Which is 1.103mm² So use this to find the wire gauge number in a conversion table or if you need the wire diameter apply step 3. Step 3) To find the diameter of the wire Use area of a circle knowing its diameter, A = (π x d²)/4 A = csa in mm², 1.103mm² π = mathematical constant pi defining the relationship between a circles diameter and circumference, 3.142 to 3 decimal places or 22/7 is often good enough for us. d = diameter in mm, to be determined Re-arranged to determine the diameter d = √ (4 x A / π) = √( 4 x 1.103 / π) = 1.18 mm Result Looking this diameter up in a conversion table, it is near as matters to 18 SWG at 1.219 mm diameter. Conversion tables (diameter or csa to wire gauge) to suit your country are readily available on line. Jim Index of Articles with Links References 25.1 Pouillet's law https://en.wikipedia.org/wiki/Electrical_resistivity_and_conductivity
