Summary: Digital Command Control may seem very complex, but armed with the proper knowledge, you'll soon understand what makes model railroads operate with DCC.
Overview of NMRA Digital Command Control – Power
You don't need to fully understand the technical details of how DCC works to use DCC on your railroad. If you don't understand everything, don't worry.
NMRA Digital Command Control, unlike other analog and digital Command Control systems, puts a 100% Digital Signal onto the rails, delivering both power and data in the same signal. The digital information is encoded in the time domain by pulse width, not amplitude. It is not superimposed on a DC or AC waveform, nor does it use a high frequency carrier, unlike past and current command control systems. Since the rails are alternately energized or held to zero, there are no issues with polarity.
Since the signal is completely digital, the NMRA Digital Command Control  waveform is a square wave. The nature of a square wave results in more demanding requirements for wiring to avoid voltage losses and signal distortion compared to analog control methods.
Too Long Didn't Read (TLDR)
There are too many myths surrounding the DCC signal. Most myths result from applying analog ideas to a digital concept. Some come from a literal interpretation of the NMRA DCC Standards. Others have been promoted over the years due to misunderstandings or dislike for the DCC technology. Another source of confusion is the application of incorrect terminology.
- The booster supplies a binary ON (HIGH) / OFF (LOW) signal which is applied to the power bus, which is carried to the track, then the train.
- The pulse widths are asymmetrical; the amplitude is symmetrical.
- The data value is determined by the period from one HIGH state to the next: 106µs (58 + 58) has a value of One, 200µS is a Zero.
- The data is encoded in the time domain. Rise times and the amplitude of the signal carry no information.
- There is full power on the rails at all times while the booster output is turned on.
- Voltage (amplitude) does not determine locomotive speed.
- There is a signal phase present on the rails, where one rail is energized (HIGH/ON) while the other is LOW/OFF. One rail will always be in the opposite (inverted) logical state compared to the other.
- The phase of the rails does NOT control the direction of the locomotive.
- There is no polarity, there are only two possible states: HIGH or LOW. There are no negative voltages present.
- Current flows from HIGH to LOW.
- The rails are identified as A and B, whose phase can be handled by auto reversers when A meets B in a reverse loop, without need to flip toggle switches.
- The DCC Signal is in the audio range. There are no high frequency carriers in the Megahertz region.
- Each output of the booster can be a Source or a Sink, their status is controlled by the data signal from the command station.
The DCC Signal
The booster is a digital amplifier. It receives a logic level waveform from the command station. Logic level signals have two states: On or Off. The data is contained within the period of the on/off sequence. This logic level voltage is amplified to the necessary track voltage determined by the scale of your model trains.
Power is a measure of energy. Energy is required by the multifunction decoder to operate the motor and lighting functions, as well as any other features it has. This power is delivered to the track by the booster.
The amount of power a DCC system can deliver is determined by the capacity of the both the booster and the power supply. An undersized power supply will limit the amount of power available. For N scale, the amount of power needed is low, from 36 to 60 Watts. For HO it can range from 50 to 80W. With larger current boosters, HO can reach 125W, and with larger scales, 240W.
Analog control systems rely on multiple small power supplies delivering low power to the track, such as 12 to 15W. The potential for damage when a short circuit occurs is limited, unlike a DCC booster which can deliver 5 or more amps into a short. For proper operation DCC requires more robust wiring (the power bus) than analog direct current, both for delivery of energy to the decoder, and quick reaction to a short circuit condition.
The voltage output of the booster is dependent on your scale. N and Z Scales require much less voltage than O or Large Scales. Where N Scale may require 12V, a large-scale locomotive may require up to 24V for proper operation.
|Common DCC Track Voltages by Scale|
|10 − 12||12 − 14||14 − 16||22 − 24 |
The amount of current required is determined by two parameters: The maximum amount of current the booster itself can deliver before overheating, and the current the power supply can deliver. Ideally the power supply should at least match the booster's current handling capacity. A power supply with a lower current output will work. 
Of those two parameters, the booster limits the maximum current can be delivered to the track. The booster will also have the capacity to disconnect its output from the power bus should a short circuit occur, to protect itself from damage from excessive heat.
Signals on the Power Bus
Oscilloscopes display signals as vectors, as they have both amplitude and direction.
The oscilloscope traces on the left show (in Purple) the sum of Rail A and Rail B. Rail A is the trace in Yellow, Rail B is in Blue.
The yellow and blue traces show the booster's output transistors switching states as they follow the data signal from the command station. The numbers 1 and 2 at the left represent the 0 VDC reference point. Both output channels connect to a common point while the probes connect to the Rail outputs.
The digital data is represented by the time (period) the rail is energized. There are no positive and negative voltages on the rails, the rail is either On or Off. The multifunction decoder only sees its input signal go high (On) and low (Off). A total period of 58µS High + 58µs Low (116µS in total) represents a binary value of 1, a total period of 200µs represents a 0 value.
This is not an AC signal, as the rail will have a voltage or no voltage present at any point in time. What is happening is the current is changing direction as it moves from one rail to another. Measuring this with an oscilloscope will display a peak-to-peak signal, which many will claim supports their assertion that the signal is an AC waveform. Since this measurement is made with the track outputs floating, the trace indicates which connection was more positive than the other at a point in time. 
Mathematically, the signal would be expressed as X +jY + X − jY, where X is time and Y is the amplitude or voltage.
This method of transmitting data and power in the same signal results in a very robust signalling technique with a high signal to noise ratio, while reducing the space charge around the rails and the electrostatic attraction of contaminants. A multifunction decoder can receive data regardless of the orientation of the locomotive, as either rail has data available.
The ideal Digital Command Control waveform is a square wave. Yet, the influence of the power bus and track, as well as the design of the booster present a number of challenges, such as distortion of the waveform.
Distortion can include ringing and switching (zero crossing) issues. The NMRA Standards include the requirement that both booster and multifunction decoder minimize any issues which may arise, such as the waveform quality at the booster output and requiring the decoder to accept a predetermined amount of signal degradation.
Matching the impedance of the track and power bus to the booster is impossible given the multitude of factors present in layout wiring, so the engineers designing DCC boosters have to try their best to ensure a good quality signal within the tolerances specified. They optimize the time the output is off during the transition from one state to another, and how quickly the on/off sequence should occur during the bit transitions. At the same time, they are trying to minimize excess heat production in the output stage.
Increasing the dead time between switching from one pair of transistors to the other comes with the possibility of increased distortion. Decreasing the period between switching can introduce ringing (overshoot) in the waveform. Ringing can be controlled using a termination on the bus. 
Unlike analog (DC) where polarity of the track controls the direction of the train, direction is controlled by the multifunction decoder in the locomotive. The decoder receives speed and direction instructions from the throttle via the command station and acts accordingly. Physically re-orienting the locomotive has no effect.
Digital Command Control has no concept of polarity. Binary signals do not have the concept of negative. A negative voltage would be considered an Undefined Value.
The track voltage on an oscilloscope's display indicates the presence of a plus/minus signal. It "sees" a negative voltage because the reference point is floating.
The track voltage has phase, one rail is always the inverse of the other. Phase is an issue with reverse loops and crossovers. A short circuit occurs during a phase mismatch, as current can now directly return to the power source. As mentioned above, the booster output has two pairs of transistors, or switches, each pair consists of a switch which connects the load to the power source, the other allows the current to return to the power source, completing the circuit. The state of these pairs is constantly changing in step with the data signal from the command station.
There are numerous claims that the DCC track signal has positive and negative values, yet this would only be possible if:
- The power supply, with two wires, has an output voltage of at least twice that of the apparent peak to peak track voltage, for example 30VDC at 5A
- The power supply has three wires supplying both positive and negative voltages with a common return/reference point (15VDC / 0 / –15VDC)
- True, the booster could have a voltage doubler incorporated into its design. That adds cost and complexity with no real advantage.
- Simple On/Off (binary) signal, where one rail's state always the inverse of the other.
- It is not an analog waveform, where frequency, phase or amplitude have meaning.
Many argue polarity is the reason reverse loops introduce problems.
The reverse loop is not the problem, the issue lies with the turnout at the entry/exit point . As the train exits the loop, the stock and point rails will have the wrong phase relationship with respect to the loop, creating short. A mechanical or electronic switch is used to invert the phase of the loop  to maintain the correct relationships upon entry or exit from the loop.
NMRA Definition of the Positive RailThe NMRA defines the Positive Rail as the right hand rail when the locomotive is facing forward.  The actual direction of travel is determined by the multifunction decoder. The purpose of this convention is to maintain compatibility when a multifunction decoder is operated in Analog Mode. 
Reading the DCC Data by a Decoder
The input circuit for a multifunction decoder consists of an opto-isolator, with a "freewheeling" diode to protect it from a reverse bias condition. This circuit allows the vehicle to face either direction while providing a usable data signal to the multifunction decoder.
Digital Command Control Signals
- Main article: Digital Packet
Advanced DCC Power Topics
The Advanced DCC Power Topics are provided for information purposes. It is not necessary to understand these topics to enjoy Digital Command Control
NMRA Digital Command Control Standards
The DCC protocol is the subject of two standards published by the NMRA: S-9.1 specifies the electrical standard, and S-9.2 specifies the communications standard. Several Recommended Practices documents are also available.
- DCC Tutorial (Power) - The tutorial version of this document.
- Power supply - Details, requirements, and recommendations for power supplies.
- Root Mean Square - Root Mean Square or RMS, additional information (lots of math).
- DCC Ammeter: What you need to build an ammeter that can measure DCC current. Plus other electronics projects for DCC.
- H-Bridges – the Basics
Square Waves and Fournier Series
All about how you can create square waves from a sine wave and its harmonics.
- This measurement is floating as there is no reference to a common circuit ground. Images such as this reinforce the myth that DCC is an AC Waveform.
- These are usually integrated into a single IC package, rather than using four discrete devices and their supporting circuitry.
- NMOS FETS turn on when a positive voltage ((VIN + VGS) > VGS) is present on its gate. PMOS FETS (the ones where the arrow points toward the speaker) switch on when the gate voltage is not present.
- VGS is the voltage between gate and source.
- Inverters, or NOT gates, invert the input. When the input is Low, the output is high.
- The NMRA Standard mandates that there be no output in the absence of a signal from the command station.
- These voltages are based on which scale the booster is set for, not on which scale you're actually running. The booster has no way of knowing what gauge track it's connected to! The user must select the appropriate voltage for their application.
- Source: Digital Command Control
- The full NMRA specification of 24 V is recommended if you run anything but slower narrow gauge steam locomotives. Large scale diesels from Aristo-Craft for example cannot achieve prototype mainline running speeds on 22 volts, they need the full 24 volts to overcome the low gearing. Also speed matching requirements for consisting means you need some "headroom" in voltage.
- It may also interfere with the booster's short circuit protection.
- These are single ended measurements, taken from the source to a common reference point.
- The 0V reference point is the chassis of the booster.
- The term Ground is incorrect, as ground should only refer to the point where the electrical system is connected to earth ground. The correct definition is floating, as there is no fixed reference to a point on the booster.
- Read the entire document and any notes. The standard is deliberately not specific, as it only describes what the result should look like, not the exact construction. Also note it does not define polarity, as that is a quality wherein the direction the locomotive defines the "positive" rail.
- Differential Signal: One signal is in the High State while the other is Low, allowing Common Mode Rejection of any noise which may be present.
- The correct terminology in the NMRA Standard should be "floating". The track voltage is “floating”, meaning it has no reference to a common point. The measurement is taken as the voltage difference between the two wires. As the measurement is floating, this gives rise to the myth that the DCC signal has both positive and negative components.
- The data is represented by the time between the leading edges of the pulses.
- Alternating current is continuously changing in amplitude and direction in time of both the voltage and current.
- The output of the booster is floating, as there is no fixed reference point such as a chassis ground for zero volts. In this case, the "ground" clip is connected to one of the rails, making this trace possible.
- TN-9:2.2.1 Ringing is a distortion of the DCC signal where there is a spike at the leading edge of each pulse. The duration of each pulse determines if the byte is a 0 or a 1. This short duration spike can be several times the amplitude of the nominal peak voltage of the DCC signal. In some cases, it can make reading the signal difficult and, in the extreme, can damage the decoder.
- TN-9:2.2.1 There are several factors that affect the amount of ringing present. Length of the bus run, load on the bus, how the wiring is installed and the power station. Each power station is different. Some produce more ringing than others. Often this is a design compromise. If the output is driven too hard ringing is present. If driven too softly the slew rate3 becomes too large (too long) making it difficult to read the DCC signal.
- TN-9:2.2.3 Bus terminations: The DCC bus may be fitted with a resistor capacitor (RC) filter using a 150Ω resistor of adequate wattage in series with a 0.1µf 50V capacitor across the bus. The purpose of such filters is to reduce ringing and to shunt any voltage spikes created when there is a short circuit created by a derailment or equipment running into a turnout set against it. Results will vary for each situation depending on the length of the bus, the load on the bus and the power station.
- TN-9 2.2.3: Determining if there is ringing on the bus and where to place RC filters to reduce the ringing is an advanced technique requiring an oscilloscope and somewhat advanced electronic expertise. Such RC filters will draw a small amount of current and should not be placed down line from any current sensing occupancy detector.
- There is no common reference point held to a zero potential. Another way to look at this is how the current is flowing: It will flow from A to B, then B to A. The scope trace will display this.
- The entire turnout must have the point and stock rails isolated from the loop.
- Phase relationships are maintained by reversing the connections to the track
- S 9.2: Forward in this case is in the direction of the front of the locomotive, as observed from the engineer's position within the locomotive.
- See Alternate Power Source
- S-9.1 Read the entire document and any notes. The standard is deliberately not specific, as it only describes what the result should look like, not the exact construction. Also note it does not define polarity, as that is a quality wherein the direction the locomotive defines the positive rail.