Phase Offset

Phase Offset (Phase Error)

Phase offset/phase error is the time difference between the reference input clock and the feedback input to the phase detector of a PLL. The two types of phase error, static and dynamic phase errors, are defined below.

Types

1. Static Phase Offset

Static phase offset (t(∅)) is the time difference between the averaged input reference clock and the averaged feedback input signal when the PLL is in locked mode. The word average implies that a comparison is made between the input of the PLL and its feedback over several thousand periods, and the resulting time differences are averaged. This method excludes jitter components, hence the name static phase offset.

2. Dynamic Phase Offset

Dynamic phase offset (td(.)) is the phase difference between input clock and output clock due to the PLLfs inability to instantaneously update the output clock when the period of the input clock changes (due to input-clock SSC). This is also referred to as tracking skew. The dynamic phase offset includes jitter (specification not yet finalized in JEDEC).

Propagation Delay

Propagation Delay

Propagation delay (tpd) is the time between the specified reference points on the input and output voltage waveforms with the output changing from one defined level (high-to-low) to the other (low-to-high) defined level (tpd= tPHL or tPLH).

Types

1. Propagation Delay Time, High-to-Low Level Output

Propagation delay time, high-to-low level output (tPHL) is the time between the specified reference points on the input and output voltage waveforms with the output changing from the defined high level to the defined low level.

2. Propagation Delay Time, Low-to-High Level Output

Propagation delay time, low-to-high level output (tPLH) is the time between the specified reference points on the input and output voltage waveforms with the output changing from the defined low level to the defined high level.

Skew & Its Types

Skew

Skew is the time delta between the actual and expected arrival time of a clock signal. Skew can be either extrinsic or intrinsic. The latter is internal to the driver (generator circuitry) and defined as the difference in propagation delays between the device outputs. On the other hand, extrinsic skew is the time difference due to unbalanced trace lengths and/or output loading.

Types of Skew

1. Output Skew

Output skew (tsk(o)) is also referred to as pin-to-pin skew, output skew is the difference between propagation delays of any two outputs of the same device at identical transitions (i.e., compares tpd(LH) versus tpd(LH) or tpd(HL) versus tpd(HL) for any two outputs). For example, if the propagation delay of the fastest output (tpd(LHn)) is 2 ns and that of the slowest output (tpdLH1) is 2.165 ns, then the output skew is: tsk(o) = tpd(LHn) − tpd(LH1) =−165ps JEDEC defines output skew as: the skew between specified outputs of a single device with all driving inputs connected together and the outputs switching in the same direction while driving identical specified loads.

2. Part-to-Part Skew

Part-to-part skew (tsk(pp)) is also known as package skew and device-to-device skew. Part-to-part skew is similar to output skew, except that it applies to two or more identical devices. Part-to-part skew is defined as the magnitude of the difference in propagation delays between any specified outputs of two separate devices operating at identical conditions. The devices must have the same input signal, supply voltage, ambient temperature, package, load, environment, etc.

3. Pulse Skew

Pulse skew (tsk(p)) is the magnitude of the time difference between the high-to-low (tPHL) and the low-to high (tPLH) propagation delays when a single switching input causes one or more outputs to switch, tsk(p) = tPHL − tPLH . Pulse skew is sometimes referred to as pulse width distortion or duty cycle skew

4. Process Skew

Process skew (tsk(pr)) is the difference in propagation delay times between corresponding outputs on any two like devices when both devices operate under identical conditions. Process skew quantifies skew due to process variation in the manufacturing process (skew caused by lot-to-lot variation). It excludes variations in supply voltage, operating temperature, output loading, input edge rate, input frequency, etc. Conceptually, process skew is output skew over several devices.

Process skew is generally specified and production tested under fixed conditions (e.g., VCC = 3.3 V, TA= 25°C, CL = 25 pF, all inputs switching simultaneously).

5. Bank Skew

Bank skew (tsk(b)) is the output skew between outputs (at same bank), of a single device with a single driving input terminal. The main difference between bank skew and output skew is that the latter is the worst-case delta between outputs in any output bank.

6. Inverting Skew

Inverting skew (tsk(inv)) is the skew between specified outputs of a single logic device with all driving inputs connected together and the outputs switching in opposite directions while driving identical specified loads.

7. Multiple-Frequency Skew

Multiple-frequency skew (tsk(ù)) is the skew between the controlled-edge position of two different output frequencies of a PLL or counting device that has more than one output frequency, when both signals are rising or both signals are falling.

8. PLL Tracking Skew

PLL tracking skew is the phase difference between the input clock and output clock due to the PLL’s inability to instantaneously update the output clock when the period of the input clock changes. Tracking skew normally applies to a PLL with SSC induced input clock [4]. Therefore, tracking skew is the phase offset of a PLL resulting from a time-varying reference input. If the total measured phase offset due to tracking skew is lumped with phase jitter, including input jitter, then it is referred to as the accumulated tracking skew. Note that tracking skew can either lead or lag the reference clock input.

9. Input Skew

Input skew (tsk(i)) is the difference between any two propagation delay times that originates at different inputs and terminates at a single output. Input skew describes the ability of a device to manipulate (stretch, shrink, or chop) a clock signal. This is typically accomplished with a multi input gate wherein one of the inputs acts as a controlling signal to pass the clock through. Input skew describes the ability of the gate to shape the pulse to the same duration regardless of the input used as the controlling input.

10. Limit Skew

Limit skew (tsk(l)) is the difference between the greater of the maximum specified values of tPLH and tPHL, and the lesser of the minimum specified values of tPLH and tPHL. Limit skew is not observed directly on a device; rather it is calculated from the data sheet limits of tPLH and tPHL. Limit skew quantifies how much variation in propagation delay times are induced by operation over the entire ranges of VCC, TA, output load, process variation and any other specified operating conditions

11. Board Skew

Board skew (tsk(pcb)) is introduced into the timing system by unequal trace lengths and unequal loading. It is independent of skew generated by the clock driver

ACR-F

• ACR-F, formally known as Equal Level Far End Crosstalk (ELFEXT)
• In June 2008, Fluke Networks adopted the new nomenclature ACR-F as defined in ANSI/TIA-568-B.2-10 and the forthcoming ANSI/TIA-568-C Series as well as various European and International Standards.
• ACR-F is an acronym for Attenuation Crosstalk Ratio Far-end.
• ACR-F is a calculated result, rather than a measurement.
• It is derived by subtracting the attenuation of the disturbing pair from the Far End Crosstalk (FEXT) this pair induces in an adjacent pair.
• This normalizes the results for length. Consider the FEXT and attenuation measured on two links constructed of the same materials with the same workmanship, but different lengths. 

50 m link example:

FEXT = 45 dB and Attenuation = 11 dB

ACR-F = 45 – 11 = 34 dB

Another way to understand ACR-F is to think of far-end Attenuation Crosstalk Ratio (ACR) as the same thing.

Results Interpretation
Compare the results of measurements made from both ends of the link to the appropriate ISO or TIA limits. There are 12 ACR-F measurements made from each end, for a total of 24. This is because the attenuation can vary slightly depending upon which pair is energized. So as an example, the field tester will energize Pair 1 and listen on Pair 2 at the far end. Then it will energize Pair 2 and listen on Pair 1 at the far end.

ACR-F that is too high is indicative of either excessive attenuation, higher than expected FEXT, or both.

Troubleshooting Recommendations
Experience has shown us that ACR-F issues are normally caused by a cable issue/fault.

What is Crosstalk?

What is crosstalk?

Crosstalk is the unwanted coupling of energy between two or more adjacent lines. This unwanted coupling can change the required signal.

Crosstalk size estimations are commonly divided into two separation calculations

1. Energy that is coupled from the actual signal line, the aggressor, onto a quiet passive victim line so that the transferred energy “travels back” to the start of the victim line. This is known as the backward or near-end crosstalk.
2. Energy that is coupled from the active signal line, the aggressor, onto a quiet passive victim line so that the transferred energy “travels forward” to the end of the victim line. This known as forward or far-end crosstalk.

If the victim line is not terminated at both ends in its characteristic impedance the induced spurious signals can reflect at the ends of the line and travel in the opposite direction down the line. Thus a reflected near-end crosstalk can end up appearing at the far end — and vice versa.

Scope

The crosstalk figures discussed in this note are the peak values as calculated by popular and frequently quoted formulae for maximum ease of comparison. These equations only hold when coupling is weak and the following assumptions apply:

1. The voltage and current on the aggressor line are unaffected by coupling to the victim line
2. The terminations of either the aggressor or victim lines have no effect
3. Propagation on either line is that of an uncoupled line
4. The average attenuation along each line is the same
5. The characteristic impedance of each line is that of an uncoupled line, Z₀

Near-end crosstalk

• The magnitude of the near-end crosstalk is dependent upon the mutual capacitance and inductance between the two interacting lines and it will increase to maximum amplitude as the coupling length increases.
• The maximum near-end crosstalk is a dimensionless ratio of voltages between the victim and aggressor lines.

Far-end crosstalk

• The far-end crosstalk can be considered as an effect caused by the difference in velocity between the odd and even modes of propagation and thus a difference in edge arrival times at the end of the line.
• Crosstalk that is propagated in a disturbed channel in the same direction as the propagation of a signal in the disturbing channel.
• The terminals of the disturbed channel, at which the far-end crosstalk is present, and the energized terminals of the disturbing channel, are usually remote from each other
• The crosstalk measured at the far end of the cable from where the transmission was sent.
• The far-end crosstalk coupling coefficient (FEXT) is a unitless ratio of the maximum voltage perturbation caused on the victim line.
• Far-end crosstalk increases with a sharper risetime, a longer coupling length and a higher K_f factor. In an ideal homogeneous stripline situation there will be no far-end crosstalk.

Summary

The electromagnetic fields between two closely coupled lines interact with each other and will affect the behavior of the signals on both lines. The formulas given here will enable the maximum peak effect to be predicted.

Near End Crosstalk

• A measurement of the ability of cabling to reject crosstalk.
• Crosstalk is an undesirable condition in which the signals traveling through adjacent pairs of wire in twisted-pair cabling interfere with each other.
• Near-end crosstalk (NEXT) measures the ability of a cable to reject crosstalk between pairs of wire at the near end of the circuit.
• The pair causing the interference is called the “disturbing pair,” while the pair experiencing the interference is the “disturbed pair.”
• Channel NEXT is the NEXT value measured between one wire pair and another in the same cable; it is measured at both ends of the wire.
• The NEXT value for a given cable type is typically expressed in decibels per 1000 feet and varies with the frequency of transmission.
• The higher the NEXT value, the greater the cable’s ability to reject crosstalk at its local connection. For example, the specifications for category 5 cabling include the minimum NEXT values shown in the following table.
• The NEXT value generally decreases with increasing frequency, indicating increasing interference due to crosstalk at higher frequencies.

The NEXT types are as follows:

1.      Pair-to-Pair NEXT:

NEXT between adjacent pairs of wire in a twisted-pair cable. A typical four-pair (eight-wire) unshielded twisted-pair (UTP) cable has six possible values for pair-to-pair NEXT, which are then averaged. This simple measurement is not adequate, however, because every pair of wire generates crosstalk with every other pair in the cable.

2.      Power Sum NEXT (PS NEXT):

A more rigorous way of rating a cable’s crosstalk that measures the total amount of crosstalk between one wire pair and all its neighboring pairs in the same cable. PS NEXT is particularly important for cabling used in high-speed networks such as Gigabit Ethernet and Asynchronous Transfer Mode (ATM) networks.

3.      Far-End Crosstalk (FEXT):

A measurement of how the far end of one wire pair affects the near end of another pair.

Desired Data Rate and the Cable’s Minimum NEXT Value

Frequency	Minimum NEXT Value
4 MHz	53 dB/1000 feet
10 MHz	47 dB/1000 feet
20 MHz	42 dB/1000 feet
1000 MHz	32 dB/1000 feet

 

Crosstalk

In electronics, crosstalk (XT) is any phenomenon by which a signal transmitted on one circuit or channel of a transmission system creates an undesired effect in another circuit or channel. Crosstalk is usually caused by undesired capacitive, inductive, or conductive coupling from one circuit, part of a circuit, or channel, to another.

Crosstalk in cabling

In structured cabling, crosstalk can refer to electromagnetic interference from one unshielded twisted pair to another twisted pair, normally running in parallel.

Near End Crosstalk

• NEXT
• Interference between two pairs in a cable measured at the same end of the cable as the transmitter.

Far end crosstalk

• FEXT
• Interference between two pairs of a cable measured at the other end of the cable from the transmitter.

Alien crosstalk

• AXT
• Interference caused by other cables routed close to the cable of interest.

Other examples

In telecommunication or telephony, crosstalk is often distinguishable as pieces of speech or signaling tones leaking from other people’s connections. If the connection is analog, twisted pair cabling can often be used to reduce the effects of crosstalk. Alternatively, the signals can be converted to digital form, which is much less susceptible to crosstalk.

In wireless communication, crosstalk is often denoted co-channel interference, and is related to adjacent-channel interference.

In integrated circuit design, crosstalk normally refers to a signal affecting another nearby signal. Usually the coupling is capacitive, and to the nearest neighbor, but other forms of coupling and effects on signal further away are sometimes important, especially in analog designs. There are a wide variety of possible fixes, with increased spacing, wire re-ordering, and shielding being the most common.

In a music recording setting, the term “crosstalk” can refer to the leakage or “bleeding” of sound from one instrument into a microphone placed in front of another musical instrument or singer. A common example is the leakage of the high-pitched, heavily-amplified sound of the lead guitar into the microphones for other instruments. Note that this is nearly always an acoustic effect, not electrical.

In stereo audio reproduction crosstalk can refer to signal leaking across from one program channel to another. This is an electrical effect and can be quantified with a crosstalk measurement.

In Full-field Optical Coherence Tomography, “crosstalk” refers to the phenomenon that due to highly scattering objects, multiple scattered photons reach the image plane and generate coherent signal after traveling a pathlength that matchs that of the sample depth within a coherence length.

In Stereoscopic 3D Displays, “crosstalk” refers to the incomplete isolation of the left and right image channels so that one leaks or bleeds into the other – like a double exposure. In this area, crosstalk and ghosting are often used interchangeably, however crosstalk is a physical entity and can be objectively measured, whereas ghosting is a subjective term and refers to the perception of crosstalk.

 

Return Loss

In telecommunications, return loss or reflection loss is the loss of signal power resulting from the reflection caused at a discontinuity in a transmission line or optical fiber. This discontinuity can be a mismatch with the terminating load or with a device inserted in the line. It is usually expressed as a ratio in decibels (dB);

RL (dB) = 10 log Pi/Pr

where RL(dB) is the return loss in dB, P_i is the incident power and P_r is the reflected power.

Two lines or devices are well matched if the return loss is high. A high return loss is therefore desirable as it results in a lower insertion loss. Return loss may be given a minus sign, see below.

Sign

Properly, loss quantities, when expressed in decibels, should be positive numbers However, return loss has historically been expressed as a negative number, and this convention is still widely found in the literature.

Taking the ratio of reflected to incident power results in a negative sign for return loss;

RL’ (dB) = 10 log Pr/Pi

where RL’(dB) is the negative of RL(dB).

Caution is required when discussing increasing or decreasing return loss since these terms strictly have the opposite meaning when return loss is defined as a negative quantity.

Electrical

In metallic conductor systems, reflections of a signal traveling down a conductor can occur at a discontinuity or impedance mismatch. The ratio of the amplitude of the reflected wave V_r to the amplitude of the incident wave V_i is known as the reflection coefficient Γ.

Γ = Vr/Vi

When the source and load impedances are known values, the reflection coefficient is given by

Γ = (ZL – Zs) / (ZL + Zs)

where Z_S is the impedance toward the source and Z_L is the impedance toward the load.

Return loss is the negative of the magnitude of the reflection coefficient in dB. Power is proportional to the square of the voltage.

When the actual transmitted (incident) power and the reflected power are known (i.e. through measurements and/or calculations), then the return loss in dB can be calculated as the difference between the incident power P_i (in dBm) and the reflected power P_r (in dBm),

RL (dB) = Pi (dBm) – Pr (dBm)

Return loss is identified with the S-parameter S₁₁ from two-port network theory.

Optical

In an optical fiber, the loss that takes place at any discontinuity of refractive index, especially at an air-glass interface such as a fiber endface, at which a fraction of the optical signal is reflected back toward the source. This reflection phenomenon is also called “Fresnel reflection loss,” or simply “Fresnel loss.”

Fiber optic transmission systems use lasers to transmit signals over optical fiber, and a high optical return loss (ORL) can cause the laser to stop transmitting correctly. The measurement of ORL is becoming more important in the characterization of optical networks as the use of wavelength-division multiplexing increases. These systems use lasers that have a lower tolerance for ORL, and introduce elements into the network that are located in close proximity to the laser.

ORL (dB) = 10 log Pi/Pr

where is the reflected power and is the incident, or input, power.

Mismatch Loss

Mismatch loss in transmission line theory is the amount of power expressed in decibels that will not be available on the output due to impedance mismatches and reflections. A transmission line that is properly terminated, that is, terminated with the same impedance as that of the characteristic impedance of the transmission line, will have no reflections and therefore no mismatch loss. Mismatch loss represents the amount of power wasted in the system. It can also be thought of as the amount of power gained if the system was perfectly matched. Impedance matching is an important part of RF system design; however, in practice there will likely be some degree of mismatch loss. In real systems, relatively little loss is due to mismatch loss and is often on the order of 1dB.

Calculation

Mismatch loss (ML) is the ratio of incident power to the difference between incident and reflected power:

Figure. Simple circuit showing characteristic impedance Z_o and the load impedance Z_L. In a perfectly matched system Z_L=Z_o, and there is no mismatch loss.

ML (dB) = 10 log (Pi – Pr)/Pi

Pr = Pi – Pd

where

• P_i = incident power
• P_r = reflected power
• P_d = delivered power (also called the accepted power)

In terms of the voltage standing wave ratio (VSWR):

ML (dB) = 10 log [ 1 – { ( VSWR – 1 ) / ( VSWR + 1 ) } ² ]

Sources of Mismatch Loss

Any component of the transmission line that has an input and output will contribute to the overall mismatch loss of the system. For example, in mixers mismatch loss occurs when there is an impedance mismatch between the RF port and IF port of the mixer. This is one of the principal reasons for losses in mixers. Likewise, a large amount of the loss in amplifiers comes from the mismatch between the input and output. Consequently, not all of the available power generated by the amplifier gets transferred to the load. This is most important in antenna systems where mismatch loss in the transmitting and receiving antenna directly contributes to the losses the system—including the system noise figure. Other common RF system components such as filters, attenuators, splitters, and combiners will generate some amount of mismatch loss. While completely eliminating mismatch loss in these components is near impossible, mismatch loss contributions by each component can be minimized by selecting quality components for use in a well designed system.

Mismatch Error

If there are two or more components in cascade as is often the case, the resultant mismatch loss is not only due to the mismatches from the individual components, but also from how the reflections from each component combine with each other. The overall mismatch loss cannot be calculated by just adding up the individual loss contributions from each component. The difference between the sum of the mismatch loss in each component and total mismatch loss due to the interactions of the reflections is known as mismatch error. Depending on how the multiple reflections combine, the overall system loss may be lower or higher than the sum of the mismatch loss from each component. Mismatch error occurs in pairs as the signal reflects off of each mismatched component. So for the example in Figure 3, there are mismatch errors generated by each pair of components. The mismatch uncertainty increases as the frequency increases, and in wide-band applications. The phasing of the reflections makes it particularly harder to model.

Insertion Loss

In telecommunications, insertion loss is the loss of signal power resulting from the insertion of a device in a transmission line or optical fiber and is usually expressed in decibels (dB).

If the power transmitted to the load before insertion is P_T and the power received by the load after insertion is P_R, then the insertion loss in dB is given by,

10 log (Pt/Pr)

Electronic Filters

Insertion loss is a figure of merit for an electronic filter and this data is generally specified with a filter. Insertion loss is defined as a ratio of the signal level in a test configuration without the filter installed (|V₁|) to the signal level with the filter installed (|V₂|). This ratio is described in dB by the following equation:

Insertion Loss (dB) = 10 log (lVinl/lVoutl)²

Link with Scattering Parameters

In case the two measurement ports use the same reference impedance, the insertion loss (IL) is defined as

IL = =20 log lS21l dB

and not, as often mistakenly thought, by:

IL = 10 log [ lS21l² / ( 1 – lS11l² ) ] dB

It is the extra loss produced by the introduction of the DUT between the 2 reference planes of the measurement. Notice that the extra loss can be introduced by intrinsic loss in the DUT and/or mismatch. In case of extra loss the insertion loss is defined to be positive.