Phase-locked loop

There are several variations of PLLs. Some terms that are used are "analog phase-locked loop" (APLL), also referred to as a linear phase-locked loop" (LPLL), "digital phase-locked loop" (DPLL), "all digital phase-locked loop" (ADPLL), and "software phase-locked loop" (SPLL).[10]

Analog or linear PLL (APLL)

Phase detector is an analog multiplier. Loop filter is active or passive. Uses a voltage-controlled oscillator (VCO). APLL is said to be a type II if its loop filter has transfer function with exactly one pole at the origin (see also Egan's conjecture on the pull-in range of type II APLL).

Digital PLL (DPLL)

An analog PLL with a digital phase detector (such as XOR, edge-trigger JK, phase frequency detector). May have digital divider in the loop.

All digital PLL (ADPLL)

Phase detector, filter and oscillator are digital. Uses a numerically controlled oscillator (NCO).

Neuronal PLL (NPLL)

Phase detector is implemented by neuronal non-linearity, oscillator by rate-controlled oscillating neurons.[11]

Software PLL (SPLL)

Functional blocks are implemented by software rather than specialized hardware.

Charge-pump PLL (CP-PLL)

CP-PLL is a modification of phase-locked loops with phase-frequency detector and square waveform signals. See also Gardner's conjecture on CP-PLL.

• Type and order.
• Frequency ranges: hold-in range (tracking range), pull-in range (capture range, acquisition range), lock-in range.[12] See also Gardner's problem on the lock-in range, Egan's conjecture on the pull-in range of type II APLL.
• Loop bandwidth: Defining the speed of the control loop.
• Transient response: Like overshoot and settling time to a certain accuracy (like 50 ppm).
• Steady-state errors: Like remaining phase or timing error.
• Output spectrum purity: Like sidebands generated from a certain VCO tuning voltage ripple.
• Phase-noise: Defined by noise energy in a certain frequency band (like 10 kHz offset from carrier). Highly dependent on VCO phase-noise, PLL bandwidth, etc.
• General parameters: Such as power consumption, supply voltage range, output amplitude, etc.

Some data streams, especially high-speed serial data streams (such as the raw stream of data from the magnetic head of a disk drive), are sent without an accompanying clock. The receiver generates a clock from an approximate frequency reference, and then phase-aligns to the transitions in the data stream with a PLL. This process is referred to as clock recovery. For this scheme to work, the data stream must have a transition frequently enough to correct any drift in the PLL's oscillator. Typically, some sort of line code, such as 8b/10b encoding, is used to put a hard upper bound on the maximum time between transitions.

If a clock is sent in parallel with data, that clock can be used to sample the data. Because the clock must be received and amplified before it can drive the flip-flops which sample the data, there will be a finite, and process-, temperature-, and voltage-dependent delay between the detected clock edge and the received data window. This delay limits the frequency at which data can be sent. One way of eliminating this delay is to include a deskew PLL on the receive side, so that the clock at each data flip-flop is phase-matched to the received clock. In that type of application, a special form of a PLL called a delay-locked loop (DLL) is frequently used.[14]

Many electronic systems include processors of various sorts that operate at hundreds of megahertz to gigahertz, well above the practical frequencies of crystal oscillators. Typically, the clocks supplied to these processors come from clock generator PLLs, which multiply a lower-frequency reference clock (usually 50 or 100 MHz) up to the operating frequency of the processor. The multiplication factor can be quite large in cases where the operating frequency is multiple gigahertz and the reference crystal is just tens or hundreds of megahertz.

All electronic systems emit some unwanted radio frequency energy. Various regulatory agencies (such as the FCC in the United States) put limits on the emitted energy and any interference caused by it. The emitted noise generally appears at sharp spectral peaks (usually at the operating frequency of the device, and a few harmonics). A system designer can use a spread-spectrum PLL to reduce interference with high-Q receivers by spreading the energy over a larger portion of the spectrum. For example, by changing the operating frequency up and down by a small amount (about 1%), a device running at hundreds of megahertz can spread its interference evenly over a few megahertz of spectrum, which drastically reduces the amount of noise seen on broadcast FM radio channels, which have a bandwidth of several tens of kilohertz.

Typically, the reference clock enters the chip and drives a phase locked loop (PLL), which then drives the system's clock distribution. The clock distribution is usually balanced so that the clock arrives at every endpoint simultaneously. One of those endpoints is the PLL's feedback input. The function of the PLL is to compare the distributed clock to the incoming reference clock, and vary the phase and frequency of its output until the reference and feedback clocks are phase and frequency matched.

PLLs are ubiquitous—they tune clocks in systems several feet across, as well as clocks in small portions of individual chips. Sometimes the reference clock may not actually be a pure clock at all, but rather a data stream with enough transitions that the PLL is able to recover a regular clock from that stream. Sometimes the reference clock is the same frequency as the clock driven through the clock distribution, other times the distributed clock may be some rational multiple of the reference.

A PLL may be used to synchronously demodulate amplitude modulated (AM) signals. The PLL recovers the phase and frequency of the incoming AM signal's carrier. The recovered phase at the VCO differs from the carrier's by 90°, so it is shifted in phase to match, and then fed to a multiplier. The output of the multiplier contains both the sum and the difference frequency signals, and the demodulated output is obtained by low-pass filtering. Since the PLL responds only to the carrier frequencies which are very close to the VCO output, a PLL AM detector exhibits a high degree of selectivity and noise immunity which is not possible with conventional peak type AM demodulators. However, the loop may lose lock where AM signals have 100% modulation depth.[15]

One desirable property of all PLLs is that the reference and feedback clock edges be brought into very close alignment. The average difference in time between the phases of the two signals when the PLL has achieved lock is called the static phase offset (also called the steady-state phase error). The variance between these phases is called tracking jitter. Ideally, the static phase offset should be zero, and the tracking jitter should be as low as possible.

Phase noise is another type of jitter observed in PLLs, and is caused by the oscillator itself and by elements used in the oscillator's frequency control circuit. Some technologies are known to perform better than others in this regard. The best digital PLLs are constructed with emitter-coupled logic (ECL) elements, at the expense of high power consumption. To keep phase noise low in PLL circuits, it is best to avoid saturating logic families such as transistor-transistor logic (TTL) or CMOS.[16]

Another desirable property of all PLLs is that the phase and frequency of the generated clock be unaffected by rapid changes in the voltages of the power and ground supply lines, as well as the voltage of the substrate on which the PLL circuits are fabricated. This is called substrate and supply noise rejection. The higher the noise rejection, the better.

To further improve the phase noise of the output, an injection locked oscillator can be employed following the VCO in the PLL.

In digital wireless communication systems (GSM, CDMA etc.), PLLs are used to provide the local oscillator up-conversion during transmission and down-conversion during reception. In most cellular handsets this function has been largely integrated into a single integrated circuit to reduce the cost and size of the handset. However, due to the high performance required of base station terminals, the transmission and reception circuits are built with discrete components to achieve the levels of performance required. GSM local oscillator modules are typically built with a frequency synthesizer integrated circuit and discrete resonator VCOs.

A phase detector (PD) generates a voltage, which represents the phase difference between two signals. In a PLL, the two inputs of the phase detector are the reference input and the feedback from the VCO. The PD output voltage is used to control the VCO such that the phase difference between the two inputs is held constant, making it a negative feedback system. [18]

Different types of phase detectors have different performance characteristics.

For instance, the frequency mixer produces harmonics that adds complexity in applications where spectral purity of the VCO signal is important. The resulting unwanted (spurious) sidebands, also called "reference spurs" can dominate the filter requirements and reduce the capture range well below or increase the lock time beyond the requirements. In these applications the more complex digital phase detectors are used which do not have as severe a reference spur component on their output. Also, when in lock, the steady-state phase difference at the inputs using this type of phase detector is near 90 degrees.

In PLL applications it is frequently required to know when the loop is out of lock. The more complex digital phase-frequency detectors usually have an output that allows a reliable indication of an out of lock condition.

An XOR gate is often used for digital PLLs as an effective yet simple phase detector. It can also be used in an analog sense with only slight modification to the circuitry.

The block commonly called the PLL loop filter (usually a low-pass filter) generally has two distinct functions.

The primary function is to determine loop dynamics, also called stability. This is how the loop responds to disturbances, such as changes in the reference frequency, changes of the feedback divider, or at startup. Common considerations are the range over which the loop can achieve lock (pull-in range, lock range or capture range), how fast the loop achieves lock (lock time, lock-up time or settling time) and damping behavior. Depending on the application, this may require one or more of the following: a simple proportion (gain or attenuation), an integral (low-pass filter) and/or derivative (high-pass filter). Loop parameters commonly examined for this are the loop's gain margin and phase margin. Common concepts in control theory including the PID controller are used to design this function.

The second common consideration is limiting the amount of reference frequency energy (ripple) appearing at the phase detector output that is then applied to the VCO control input. This frequency modulates the VCO and produces FM sidebands commonly called "reference spurs".

The design of this block can be dominated by either of these considerations, or can be a complex process juggling the interactions of the two. Typical trade-offs are increasing the bandwidth usually degrades the stability or too much damping for better stability will reduce the speed and increase settling time. Often also the phase-noise is affected.[13]

All phase-locked loops employ an oscillator element with variable frequency capability. This can be an analog VCO either driven by analog circuitry in the case of an APLL or driven digitally through the use of a digital-to-analog converter as is the case for some DPLL designs. Pure digital oscillators such as a numerically controlled oscillator are used in ADPLLs.

An example digital divider (by 4) for use in the feedback path of a multiplying PLL

PLLs may include a divider between the oscillator and the feedback input to the phase detector to produce a frequency synthesizer. A programmable divider is particularly useful in radio transmitter applications and for computer clocking, since a large number of frequencies can be produced from a single stable, accurate, quartz crystal–controlled reference oscillator (which were expensive before commercial-scale hydrothermal synthesis provided cheap synthetic quartz).

Some PLLs also include a divider between the reference clock and the reference input to the phase detector. If the divider in the feedback path divides by ${\displaystyle N}$ and the reference input divider divides by ${\displaystyle M}$, it allows the PLL to multiply the reference frequency by ${\displaystyle N/M}$. It might seem simpler to just feed the PLL a lower frequency, but in some cases the reference frequency may be constrained by other issues, and then the reference divider is useful.

Frequency multiplication can also be attained by locking the VCO output to the Nth harmonic of the reference signal. Instead of a simple phase detector, the design uses a harmonic mixer (sampling mixer). The harmonic mixer turns the reference signal into an impulse train that is rich in harmonics.[lower-alpha 1] The VCO output is coarse tuned to be close to one of those harmonics. Consequently, the desired harmonic mixer output (representing the difference between the N harmonic and the VCO output) falls within the loop filter passband.

It should also be noted that the feedback is not limited to a frequency divider. This element can be other elements such as a frequency multiplier, or a mixer. The multiplier will make the VCO output a sub-multiple (rather than a multiple) of the reference frequency. A mixer can translate the VCO frequency by a fixed offset. It may also be a combination of these. An example being a divider following a mixer; this allows the divider to operate at a much lower frequency than the VCO without a loss in loop gain.

The equations governing a phase-locked loop with an analog multiplier as the phase detector and linear filter may be derived as follows. Let the input to the phase detector be ${\displaystyle f_{1}(\theta _{1}(t))}$ and the output of the VCO is ${\displaystyle f_{2}(\theta _{2}(t))}$ with phases ${\displaystyle \theta _{1}(t)}$ and ${\displaystyle \theta _{2}(t)}$. The functions ${\displaystyle f_{1}(\theta )}$ and ${\displaystyle f_{2}(\theta )}$ describe waveforms of signals. Then the output of the phase detector ${\displaystyle \varphi (t)}$ is given by

${\displaystyle \varphi (t)=f_{1}(\theta _{1}(t))f_{2}(\theta _{2}(t))}$

The VCO frequency is usually taken as a function of the VCO input ${\displaystyle g(t)}$ as

${\displaystyle {\dot {\theta }}_{2}(t)=\omega _{2}(t)=\omega _{\text{free}}+g_{v}g(t)\,}$

where ${\displaystyle g_{v}}$ is the sensitivity of the VCO and is expressed in Hz / V; ${\displaystyle \omega _{\text{free}}}$ is a free-running frequency of VCO.

The loop filter can be described by a system of linear differential equations

${\displaystyle {\begin{array}{rcl}{\dot {x}}&=&Ax+b\varphi (t),\\g(t)&=&c^{*}x,\end{array}}\quad x(0)=x_{0},}$

where ${\displaystyle \varphi (t)}$ is an input of the filter, ${\displaystyle g(t)}$ is an output of the filter, ${\displaystyle A}$ is ${\displaystyle n}$-by-${\displaystyle n}$ matrix, ${\displaystyle x\in \mathbb {C} ^{n},\quad b\in \mathbb {R} ^{n},\quad c\in \mathbb {C} ^{n},\quad }$. ${\displaystyle x_{0}\in \mathbb {C} ^{n}}$ represents an initial state of the filter. The star symbol is a conjugate transpose.

Hence the following system describes PLL

${\displaystyle {\begin{array}{rcl}{\dot {x}}&=&Ax+bf_{1}(\theta _{1}(t))f_{2}(\theta _{2}(t)),\\{\dot {\theta }}_{2}&=&\omega _{\text{free}}+g_{v}(c^{*}x)\\\end{array}}\quad x(0)=x_{0},\quad \theta _{2}(0)=\theta _{0}.}$

where ${\displaystyle \theta _{0}}$ is an initial phase shift.

Consider the input of PLL ${\displaystyle f_{1}(\theta _{1}(t))}$ and VCO output ${\displaystyle f_{2}(\theta _{2}(t))}$ are high frequency signals. Then for any piecewise differentiable ${\displaystyle 2\pi }$-periodic functions ${\displaystyle f_{1}(\theta )}$ and ${\displaystyle f_{2}(\theta )}$ there is a function ${\displaystyle \varphi (\theta )}$ such that the output ${\displaystyle G(t)}$ of Filter

${\displaystyle {\begin{array}{rcl}{\dot {x}}&=&Ax+b\varphi (\theta _{1}(t)-\theta _{2}(t)),\\G(t)&=&c^{*}x,\end{array}}\quad x(0)=x_{0},}$

in phase domain is asymptotically equal (the difference ${\displaystyle G(t)-g(t)}$ is small with respect to the frequencies) to the output of the Filter in time domain model. [19] [20] Here function ${\displaystyle \varphi (\theta )}$ is a phase detector characteristic.

Denote by ${\displaystyle \theta _{\Delta }(t)}$ the phase difference

${\displaystyle \theta _{\Delta }=\theta _{1}(t)-\theta _{2}(t).}$

Then the following dynamical system describes PLL behavior

${\displaystyle {\begin{array}{rcl}{\dot {x}}&=&Ax+b\varphi (\theta _{\Delta }),\\{\dot {\theta }}_{\Delta }&=&\omega _{\Delta }-g_{v}(c^{*}x).\\\end{array}}\quad x(0)=x_{0},\quad \theta _{\Delta }(0)=\theta _{1}(0)-\theta _{2}(0).}$

Here ${\displaystyle \omega _{\Delta }=\omega _{1}-\omega _{\text{free}}}$; ${\displaystyle \omega _{1}}$ is the frequency of a reference oscillator (we assume that ${\displaystyle \omega _{\text{free}}}$ is constant).

Consider sinusoidal signals

${\displaystyle f_{1}(\theta _{1}(t))=A_{1}\sin(\theta _{1}(t)),\quad f_{2}(\theta _{2}(t))=A_{2}\cos(\theta _{2}(t))}$

and a simple one-pole RC circuit as a filter. The time-domain model takes the form

${\displaystyle {\begin{aligned}{\dot {x}}&=-{\frac {1}{RC}}x+{\frac {1}{RC}}A_{1}A_{2}\sin(\theta _{1}(t))\cos(\theta _{2}(t)),\\[6pt]{\dot {\theta }}_{2}&=\omega _{\text{free}}+g_{v}(c^{*}x)\end{aligned}}}$

PD characteristics for this signals is equal[21] to

${\displaystyle \varphi (\theta _{1}-\theta _{2})={\frac {A_{1}A_{2}}{2}}\sin(\theta _{1}-\theta _{2})}$

Hence the phase domain model takes the form

${\displaystyle {\begin{aligned}{\dot {x}}&=-{\frac {1}{RC}}x+{\frac {1}{RC}}{\frac {A_{1}A_{2}}{2}}\sin(\theta _{\Delta }),\\[6pt]{\dot {\theta }}_{\Delta }&=\omega _{\Delta }-g_{v}(c^{*}x).\end{aligned}}}$

This system of equations is equivalent to the equation of mathematical pendulum

${\displaystyle {\begin{aligned}x&={\frac {{\dot {\theta }}_{2}-\omega _{2}}{g_{v}c^{*}}}={\frac {\omega _{1}-{\dot {\theta }}_{\Delta }-\omega _{2}}{g_{v}c^{*}}},\\[6pt]{\dot {x}}&={\frac {{\ddot {\theta }}_{2}}{g_{v}c^{*}}},\\[6pt]\theta _{1}&=\omega _{1}t+\Psi ,\\[6pt]\theta _{\Delta }&=\theta _{1}-\theta _{2},\\[6pt]{\dot {\theta }}_{\Delta }&={\dot {\theta }}_{1}-{\dot {\theta }}_{2}=\omega _{1}-{\dot {\theta }}_{2},\\[6pt]&{\frac {1}{g_{v}c^{*}}}{\ddot {\theta }}_{\Delta }-{\frac {1}{g_{v}c^{*}RC}}{\dot {\theta }}_{\Delta }-{\frac {A_{1}A_{2}}{2RC}}\sin \theta _{\Delta }={\frac {\omega _{2}-\omega _{1}}{g_{v}c^{*}RC}}.\end{aligned}}}$

Phase locked loops can also be analyzed as control systems by applying the Laplace transform. The loop response can be written as

${\displaystyle {\frac {\theta _{o}}{\theta _{i}}}={\frac {K_{p}K_{v}F(s)}{s+K_{p}K_{v}F(s)}}}$

Where

• ${\displaystyle \theta _{o}}$ is the output phase in radians
• ${\displaystyle \theta _{i}}$ is the input phase in radians
• ${\displaystyle K_{p}}$ is the phase detector gain in volts per radian
• ${\displaystyle K_{v}}$ is the VCO gain in radians per volt-second
• ${\displaystyle F(s)}$ is the loop filter transfer function (dimensionless)

The loop characteristics can be controlled by inserting different types of loop filters. The simplest filter is a one-pole RC circuit. The loop transfer function in this case is

${\displaystyle F(s)={\frac {1}{1+sRC}}}$

The loop response becomes:

${\displaystyle {\frac {\theta _{o}}{\theta _{i}}}={\frac {\frac {K_{p}K_{v}}{RC}}{s^{2}+{\frac {s}{RC}}+{\frac {K_{p}K_{v}}{RC}}}}}$

This is the form of a classic harmonic oscillator. The denominator can be related to that of a second order system:

${\displaystyle s^{2}+2s\zeta \omega _{n}+\omega _{n}^{2}}$

where ${\displaystyle \zeta }$ is the damping factor and ${\displaystyle \omega _{n}}$ is the natural frequency of the loop.

For the one-pole RC filter,

${\displaystyle \omega _{n}={\sqrt {\frac {K_{p}K_{v}}{RC}}}}$

${\displaystyle \zeta ={\frac {1}{2{\sqrt {K_{p}K_{v}RC}}}}}$

The loop natural frequency is a measure of the response time of the loop, and the damping factor is a measure of the overshoot and ringing. Ideally, the natural frequency should be high and the damping factor should be near 0.707 (critical damping). With a single pole filter, it is not possible to control the loop frequency and damping factor independently. For the case of critical damping,

${\displaystyle RC={\frac {1}{2K_{p}K_{v}}}}$

${\displaystyle \omega _{c}=K_{p}K_{v}{\sqrt {2}}}$

A slightly more effective filter, the lag-lead filter includes one pole and one zero. This can be realized with two resistors and one capacitor. The transfer function for this filter is

${\displaystyle F(s)={\frac {1+sCR_{2}}{1+sC(R_{1}+R_{2})}}}$

This filter has two time constants

${\displaystyle \tau _{1}=C(R_{1}+R_{2})}$

${\displaystyle \tau _{2}=CR_{2}}$

Substituting above yields the following natural frequency and damping factor

${\displaystyle \omega _{n}={\sqrt {\frac {K_{p}K_{v}}{\tau _{1}}}}}$

${\displaystyle \zeta ={\frac {1}{2\omega _{n}\tau _{1}}}+{\frac {\omega _{n}\tau _{2}}{2}}}$

The loop filter components can be calculated independently for a given natural frequency and damping factor

${\displaystyle \tau _{1}={\frac {K_{p}K_{v}}{\omega _{n}^{2}}}}$

${\displaystyle \tau _{2}={\frac {2\zeta }{\omega _{n}}}-{\frac {1}{K_{p}K_{v}}}}$

Real world loop filter design can be much more complex e.g. using higher order filters to reduce various types or source of phase noise. (See the D Banerjee ref below)

Digital phase locked loops can be implemented in hardware, using integrated circuits such as a CMOS 4046. However, with microcontrollers becoming faster, it may make sense to implement a phase locked loop in software for applications that do not require locking onto signals in the MHz range or faster, such as precisely controlling motor speeds. Software implementation has several advantages including easy customization of the feedback loop including changing the multiplication or division ratio between the signal being tracked and the output oscillator. Furthermore, a software implementation is useful to understand and experiment with. As an example of a phase-locked loop implemented using a phase frequency detector is presented in MATLAB, as this type of phase detector is robust and easy to implement.

% This example is written in MATLAB

% Initialize variables
vcofreq = zeros(1, numiterations);
ervec = zeros(1, numiterations);
% Keep track of last states of reference, signal, and error signal
qsig = 0; qref = 0; lref = 0; lsig = 0; lersig = 0;
phs = 0;
freq = 0;

% Loop filter constants (proportional and derivative)
% Currently powers of two to facilitate multiplication by shifts
prop = 1 / 128;
deriv = 64;

for it = 1:numiterations
    % Simulate a local oscillator using a 16-bit counter
    phs = mod(phs + floor(freq / 2 ^ 16), 2 ^ 16);
    ref = phs < 32768;
    % Get the next digital value (0 or 1) of the signal to track
    sig = tracksig(it);
    % Implement the phase-frequency detector
    rst = ~ (qsig & qref); % Reset the "flip-flop" of the phase-frequency
    % detector when both signal and reference are high
    qsig = (qsig | (sig & ~ lsig)) & rst; % Trigger signal flip-flop and leading edge of signal
    qref = (qref | (ref & ~ lref)) & rst; % Trigger reference flip-flop on leading edge of reference
    lref = ref; lsig = sig; % Store these values for next iteration (for edge detection)
    ersig = qref - qsig; % Compute the error signal (whether frequency should increase or decrease)
    % Error signal is given by one or the other flip flop signal
    % Implement a pole-zero filter by proportional and derivative input to frequency
    filtered_ersig = ersig + (ersig - lersig) * deriv;
    % Keep error signal for proportional output
    lersig = ersig;
    % Integrate VCO frequency using the error signal
    freq = freq - 2 ^ 16 * filtered_ersig * prop;
    % Frequency is tracked as a fixed-point binary fraction
    % Store the current VCO frequency
    vcofreq(1, it) = freq / 2 ^ 16;
    % Store the error signal to show whether signal or reference is higher frequency
    ervec(1, it) = ersig;
end

In this example, an array tracksig is assumed to contain a reference signal to be tracked. The oscillator is implemented by a counter, with the most significant bit of the counter indicating the on/off status of the oscillator. This code simulates the two D-type flip-flops that comprise a phase-frequency comparator. When either the reference or signal has a positive edge, the corresponding flip-flop switches high. Once both reference and signal is high, both flip-flops are reset. Which flip-flop is high determines at that instant whether the reference or signal leads the other. The error signal is the difference between these two flip-flop values. The pole-zero filter is implemented by adding the error signal and its derivative to the filtered error signal. This in turn is integrated to find the oscillator frequency.

In practice, one would likely insert other operations into the feedback of this phase-locked loop. For example, if the phase locked loop were to implement a frequency multiplier, the oscillator signal could be divided in frequency before it is compared to the reference signal.

As an analogy of a PLL, consider a race between two cars. One represents the input frequency, the other the PLL's output voltage-controlled oscillator (VCO) frequency. Each lap corresponds to a complete cycle. The number of laps per hour (a speed) corresponds to the frequency. The separation of the cars (a distance) corresponds to the phase difference between the two oscillating signals.

During most of the race, each car is on its own and free to pass the other and lap the other. This is analogous to the PLL in an unlocked state.

However, if there is an accident, a yellow caution flag is raised. This means neither of the race cars is permitted to overtake and pass the other car. The two race cars represent the input and output frequency of the PLL in a locked state. Each driver will measure the phase difference (a fraction of the distance around the lap) between themselves and the other race car. If the hind driver is too far away, they will increase their speed to close the gap. If they are too close to the other car, the driver will slow down. The result is that both race cars will circle the track in lockstep with a fixed phase difference (or constant distance) between them. Since neither car is allowed to lap the other, the cars make the same number of laps in a given time period. Therefore the frequency of the two signals is the same.

Phase can be proportional to time,[lower-alpha 2] so a phase difference can be a time difference. Clocks are, with varying degrees of accuracy, phase-locked (time-locked) to a leader clock.

Left on its own, each clock will mark time at slightly different rates. A wall clock, for example, might be fast by a few seconds per hour compared to the reference clock at NIST. Over time, that time difference would become substantial.

To keep the wall clock in sync with the reference clock, each week the owner compares the time on their wall clock to a more accurate clock (a phase comparison), and resets their clock. Left alone, the wall clock will continue to diverge from the reference clock at the same few seconds per hour rate.

Some clocks have a timing adjustment (a fast-slow control). When the owner compared their wall clock's time to the reference time, they noticed that their clock was too fast. Consequently, the owner could turn the timing adjust a small amount to make the clock run a little slower (frequency). If things work out right, their clock will be more accurate than before. Over a series of weekly adjustments, the wall clock's notion of a second would agree with the reference time (locked both in frequency and phase within the wall clock's stability).

An early electromechanical version of a phase-locked loop was used in 1921 in the Shortt-Synchronome clock.