Wire sizes: Twisting the Power Bus
Summary: A question often asked regarding twisting of the DCC Power Bus. Is it necessary? In short, no, there are other ways to achieve the same result.
To Twist or Not to Twist, That is the Question…
This topic is fiercely debated on internet forums and DCC related mailing lists, endlessly and without any real conclusions. Seems everyone has an opinion, just like they do with wire sizes…
Twisting: Is It Necessary?
Insert Heated Debate Here
There is always a large debate on the twisting of track power bus wires. Keep in mind that DCC power buses are an Unbalanced Pair: One wire (A) is held to ground while the other (B) is energized, then they flip when A carries a signal and B is held to ground. This is happening constantly.
Other debates center around transmission line theory, which fails to identify that the distances are short and the load is not fixed at the end of the circuit. In this situation the load is not stationary, but moving along the track.
By managing the inductance, the DCC signal's integrity can be preserved by reducing distortion and ringing on the power bus, with the resulting reduction in impedance allowing more power to be delivered to the vehicles on the track.
It is not necessary to twist the track power bus wires together. Track Power Bus wires should be kept close together, which is easily accomplished by twisting them together. Tying them with a cable tie is another option, and accomplishes the same purpose: Reducing Inductance. 
Controlling Inductance in the Track Bus
For more background on Inductance issues, please see the article on Inductance.
A conductor in free space will have inductance when current flows through it.
A length of wire will have impedance, caused by the resistive and reactive properties of the wire. A larger gauge wire will have less impedance, as its resistance will be lower, and self-induction (reactive) properties are also reduced. The mathematics required to demonstrate this is complex and beyond the scope of The DCCWiki.
As the image shows, a magnetic flux surrounds the wire when current flows. The direction of the current determines the direction of the flux, as demonstrated by the "Right Hand Rule", where the thumb indicates current flow, the fingers wrapped around the conductor illustrate direction of flux.
Compare two long and widely spaced bus wires (inches apart) in parallel with two identical but closely spaced bus wires (~1mm apart). The latter have much less inductance, which is better. This is demonstrated in the table below. When a pair of wires make a long run parallel to each other, the one wire induces currents in the other. Since the two forces present are in opposition, they effectively cancel out. This was discovered in the early days of telephony. In fact, Alexander Graham Bell discovered that by twisting the two wires together, interference and inductance were reduced for a stronger signal and clearer communications.
- Excessive track bus impedance can cause a multifunction decoder's PWM pulses to be superimposed onto the DCC signal. 
- Ringing: Excessive inductance increases ringing in the DCC signal. 
- The amount of inductance in the bus directly relates to the degree of DCC waveform distortion and other problems such as large voltage spikes. 
- Large voltage spikes are created during intermittent short circuits caused by derailments or other track electrical issues.
- Many boosters rely on rate of change to detect a short circuit. 
- Increased voltage drop.
- Distortion of the DCC waveform leading to runaway locomotives. 
Nature abhors a change in flux. (D. J. Griffiths)
This is an advanced topic, and it is not necessary to understand the physics involved. It is for reference.
Inductors store current in a magnetic field, when current flow increases, it resists that change. Lenz's Law states that a voltage, which will be opposite in polarity, opposes any increases in flux and current. This voltage is called Counter–Electromotive Force, or C–EMF, which is similar to Back EMF. This process is called Self Induction.
When the current is disconnected, such as when the booster output switches phase, the magnetic field collapses inducing an Electromotive Force into the inductor. When this happens the polarity of the self-induced voltage is reversed as the inductor attempts to maintain current flow by having the polarity of the induced voltage such that it adds to the source voltage, raising the total voltage in an attempt to maintain current flow in the circuit. This effect can also be seen when a relay's coil is de-energized.
Mutual Inductance is the property where the flux around one wire induces a current in another. This is the principle by which transformers work. Mutual inductance can also be additive or subtractive in nature.
By keeping the two bus wires close together, the net effect of their mutual inductance is to reduce the total inductance. As k, the coefficient of coupling increases, the resulting inductance decreases, as per the formula:
- Total Inductance = L1 + L2 – 2M
- Where M = k × √(L1 × L2)
- k is a value from 0 to 0.9. Perfect coupling is considered to be 0.9.
For example, if each bus wire were to possess an inductance of 50µHenries:
- k= 0.8 for a tightly coupled pair, LTotal = 20µH
- k = 0.4, for a loose coupling such as wires several inches apart, LTotal = 60µH
- With no coupling, (k = 0 the total inductance of the loop would be 100µH.
- The k values are just for illustration of the effect of mutual inductance and coupling effects.
This effect can be seen in the table below.
Ref: Lenk, Databook for Electronic Technicians and Engineers
At DCC Frequencies there is not much opportunity for interference with other electrical devices. While this is often given as a reason for twisting the bus, at the voltages and frequencies present the issue is not as great as many would believe.
Interference could be interpreted by multifunction decoders and could cause havoc on the system. This interference can come from outside the layout, or be caused by the bus wires themselves or nearby signal buses. Twisting the wires may reduce any interference the track bus may cause by inducing a signal in a low power signalling bus, such as your throttle network or occupancy detection system. It is also advisable to keep the various buses under your layout separated to reduce the possibility of mutual inductance. You don't want to induce signals in another bus, such as the throttle or LCC wiring. Also keep AC wiring away from your track bus and other layout buses.
This table demonstrates the differences that occur when the power bus wires are kept apart or close together.
The bus under test was made of two lengths of 12AWG wire placed in parallel, ~36 feet each for a total loop of ~72'. One end of the loop was closed to enable a reading to be taken.
- Before the inductance measurement, made with an Agilent ESR meter, each wire was measured, resulting in an Average impedance of each wire is 0.8Ω, L = 12.6μH, Resistance 0.09Ω. Measurements made at 10kHz to simulate DCC frequencies.
|Bus Wires||Z (Impedance, Ohms)||Inductance (μH) at 10kHz|
|Parallel, >1 foot spacing||1.38||22|
Referring the the illustration of the flux surrounding the wire: The right hand rule demonstates that in a DCC Power Bus, the flux in one conductor would have the opposite direction of the flux in the other conductor. Thus they would cancel out.
- TN-9, 2.2.2 Twisted bus pairs
- This issue can create problems such as runaway locomotives.
- Due to the harmonic content of the DCC signal, it has a lot more in common with alternating current than direct current.
- Excessive impedance impairs the ability of the booster to detect a short by increasing the RL time constant of the bus. This impedes the rate of change, resulting in excessive currents flowing and damaging rolling stock or track work.
- Disabling the Power Source Conversion option in CV29 will prevent this by disallowing Analog (Direct Current) operation.
- Lenk, Databook for Electronic Technicians and Engineers