This circuit is a great way to transmit your video without having to use wired connections. By
using this circuit you will be able to transmit your video without any lag
PCB layout and channel design for optical transceivers is actually based on high-speed PCB design.
Indeed, all aspects of this process must be considered, especially for very high data rates. Those that reach 400 Gbps over 10 lanes (i.e. 40 Gbps per lane!) are possible over long distances by deploying suitable topology and routing techniques.
Optical Transceiver Topology and Routing Challenges
Routing between chips, or between a chip and an optical transceiver, at high network speeds requires consideration of a number of high-speed design rules, both for an individual transceiver and for the backplane that connects several of them.
Some particularly important aspects of this type of design manifest themselves at very high Gbps data rates, which we will discuss in more detail below.
Designing a typical interconnect between a chip and a transceiver, or between two chips, as part of high-speed channel design, therefore requires consideration of the following:
PCB substrate material
The substrate material determines the effective dielectric constant of the board and its loss tangent. The trace impedance increases as the dielectric constant of the substrate decreases.
The geometry of a given trace must therefore be modified in order to ensure that the impedance of the trace takes on a consistent value over the entire interconnection.
Dispersion in the substrate causes different harmonics that comprise a digital signal moving at different speeds to move, causing signal distortion and spreading.
This increases phase jitter at the receiver. Therefore, you should choose a substrate material with a flat dielectric constant at frequencies between that of the signal repetition and the corner frequency.
The substrate should also have low losses. Note that it is not always possible to satisfy these two requirements simultaneously in each frequency band.
Manufacturing Considerations
At the very fast signal rise times required for high-speed networks, impedance discontinuities must be minimized throughout the board. This means that the use of vias should be kept to a minimum on high-speed interconnects.
The impedance of a given trace can vary due to variations in surface roughness and geometry, which can lead to signal integrity issues that contribute to jitter.
You must also take into account another element relating to surface roughness. This is because at very high speeds, the input/output current of a track tends to deposit near the edge of a copper conductor due to the skin effect, leading to increased losses. resistive.
Copper conductors can be electroplated or pressed and rolled. This latter process tends to produce conductors with smoother surfaces; we therefore recommend that you exploit it in order to reduce the resistive losses in an interconnection.
The stacking of layers
Routing recommendations for Ethernet over copper are typically implemented on 2 or 4 layer PCBs with power and ground islands.
In high-speed PCBs for optical transceivers, the designation of high-speed signal layers within the stack directly affects signal performance.
Boards that include one or more BGA-mounted FPGAs typically use stacks of at least 6 layers, as this helps provide the necessary number of signal layers for escape routing from the BGA.
Stripline routing at Gbps and faster transmission speeds is known to lead to lower losses than microstrip routing.
Also, it will inevitably be used to escape a high pin density FPGA or other controller.
When routed between two conductive planes, stripline traces have some natural immunity to external electromagnetic interference. However, a thicker dielectric is required to achieve a given controlled impedance value and the vias must be uused at the PHY, MAC, and transceiver connections.
All vias placed on these high-speed interconnects must be drilled out to avoid stub resonance.
Jitter and Routing
The most complex part of setting up an optical transceiver is not necessarily the data transfer rate, but rather the rise time of the converted electrical signals. It is the limiting factor that determines the impact of high-speed signaling effects in any PCB.
As the data rate increases, the signal rise time must also decrease. In the field of telecommunications, we often speak of the unit interval (UI), which can refer to the duration of existence of a given symbol in a data stream. At 50 Gbps in a single lane, the UI only designates the inverse of the data rate, which is 20 ps/baud.
Jitter is only one determining factor of bit error rates. However, to maintain data integrity at a bit error rate below a certain maximum, it must be maintained below a certain allowable margin.
This margin is generally expressed in proportion to the UI. For example, a jitter margin of 0.05 UI corresponds to a maximum jitter of 2 ps in a 25 Gbps channel (UI = 40 ps/baud). Jitter must therefore be addressed at the chip level, as it requires extremely stable transmission, as well as at the PCB level, with suitable layout and manufacturing.
Differential signaling is typically used because it provides common-mode noise immunity and reduces inductive crosstalk between traces.
By placing a ground plane as close to the surface layer as possible, you'll have a better chance of eliminating crosstalk and electromagnetic interference.
The jitter margin will also determine the limit of length offset allowed between each end of a differential pair. This offset, combined with jitter, results in successive offsets for signals traveling over an interconnect.
Given the very fast rise times used in Gbps and Fast Ethernet, including Ethernet over fiber, the interconnects between the transceiver and the chip, or between two chips, must be very short.
Otherwise, the behavior of the transmission line will be easily noticed and will corrupt your signals. These lines must therefore be terminated and/or their impedance must be matched to prevent signal reflection.
With modulation schemes such as 4PAM, a high reflection at the signal level is likely to create a significant increase in FEC due to the staircase response in digital signals due to repeated reflections.
Controlled impedance routing is therefore an essential issue, because it makes it possible to reduce the number of impedance matching networks required on the board.
Altium Designer's routing and simulation tools are designed to help you design PCBs for almost any application. High-speed design and simulation tools are ideal for optical transceiver layouts, and data management and documentation tools can help you prepare for manufacturing.
Contact us or download a free trial if you want to learn more about Altium Designer. This gives you access to the industry's best routing, simulation, and data management tools in a single program. Contact an Altium expert today to learn more.
This circuit is a great way to transmit your video without having to use wired connections. By
using this circuit you will be able to transmit your video without any lag
PCB layout and channel design for optical transceivers is actually based on high-speed PCB design.
Indeed, all aspects of this process must be considered, especially for very high data rates. Those that reach 400 Gbps over 10 lanes (i.e. 40 Gbps per lane!) are possible over long distances by deploying suitable topology and routing techniques.
Optical Transceiver Topology and Routing Challenges
Routing between chips, or between a chip and an optical transceiver, at high network speeds requires consideration of a number of high-speed design rules, both for an individual transceiver and for the backplane that connects several of them.
Some particularly important aspects of this type of design manifest themselves at very high Gbps data rates, which we will discuss in more detail below.
Designing a typical interconnect between a chip and a transceiver, or between two chips, as part of high-speed channel design, therefore requires consideration of the following:
PCB substrate material
The substrate material determines the effective dielectric constant of the board and its loss tangent. The trace impedance increases as the dielectric constant of the substrate decreases.
The geometry of a given trace must therefore be modified in order to ensure that the impedance of the trace takes on a consistent value over the entire interconnection.
Dispersion in the substrate causes different harmonics that comprise a digital signal moving at different speeds to move, causing signal distortion and spreading.
This increases phase jitter at the receiver. Therefore, you should choose a substrate material with a flat dielectric constant at frequencies between that of the signal repetition and the corner frequency.
The substrate should also have low losses. Note that it is not always possible to satisfy these two requirements simultaneously in each frequency band.
Manufacturing Considerations
At the very fast signal rise times required for high-speed networks, impedance discontinuities must be minimized throughout the board. This means that the use of vias should be kept to a minimum on high-speed interconnects.
The impedance of a given trace can vary due to variations in surface roughness and geometry, which can lead to signal integrity issues that contribute to jitter.
You must also take into account another element relating to surface roughness. This is because at very high speeds, the input/output current of a track tends to deposit near the edge of a copper conductor due to the skin effect, leading to increased losses. resistive.
Copper conductors can be electroplated or pressed and rolled. This latter process tends to produce conductors with smoother surfaces; we therefore recommend that you exploit it in order to reduce the resistive losses in an interconnection.
The stacking of layers
Routing recommendations for Ethernet over copper are typically implemented on 2 or 4 layer PCBs with power and ground islands.
In high-speed PCBs for optical transceivers, the designation of high-speed signal layers within the stack directly affects signal performance.
Boards that include one or more BGA-mounted FPGAs typically use stacks of at least 6 layers, as this helps provide the necessary number of signal layers for escape routing from the BGA.
Stripline routing at Gbps and faster transmission speeds is known to lead to lower losses than microstrip routing.
Also, it will inevitably be used to escape a high pin density FPGA or other controller.
When routed between two conductive planes, stripline traces have some natural immunity to external electromagnetic interference. However, a thicker dielectric is required to achieve a given controlled impedance value and the vias must be uused at the PHY, MAC, and transceiver connections.
All vias placed on these high-speed interconnects must be drilled out to avoid stub resonance.
Jitter and Routing
The most complex part of setting up an optical transceiver is not necessarily the data transfer rate, but rather the rise time of the converted electrical signals. It is the limiting factor that determines the impact of high-speed signaling effects in any PCB.
As the data rate increases, the signal rise time must also decrease. In the field of telecommunications, we often speak of the unit interval (UI), which can refer to the duration of existence of a given symbol in a data stream. At 50 Gbps in a single lane, the UI only designates the inverse of the data rate, which is 20 ps/baud.
Jitter is only one determining factor of bit error rates. However, to maintain data integrity at a bit error rate below a certain maximum, it must be maintained below a certain allowable margin.
This margin is generally expressed in proportion to the UI. For example, a jitter margin of 0.05 UI corresponds to a maximum jitter of 2 ps in a 25 Gbps channel (UI = 40 ps/baud). Jitter must therefore be addressed at the chip level, as it requires extremely stable transmission, as well as at the PCB level, with suitable layout and manufacturing.
Differential signaling is typically used because it provides common-mode noise immunity and reduces inductive crosstalk between traces.
By placing a ground plane as close to the surface layer as possible, you'll have a better chance of eliminating crosstalk and electromagnetic interference.
The jitter margin will also determine the limit of length offset allowed between each end of a differential pair. This offset, combined with jitter, results in successive offsets for signals traveling over an interconnect.
Given the very fast rise times used in Gbps and Fast Ethernet, including Ethernet over fiber, the interconnects between the transceiver and the chip, or between two chips, must be very short.
Otherwise, the behavior of the transmission line will be easily noticed and will corrupt your signals. These lines must therefore be terminated and/or their impedance must be matched to prevent signal reflection.
With modulation schemes such as 4PAM, a high reflection at the signal level is likely to create a significant increase in FEC due to the staircase response in digital signals due to repeated reflections.
Controlled impedance routing is therefore an essential issue, because it makes it possible to reduce the number of impedance matching networks required on the board.
Altium Designer's routing and simulation tools are designed to help you design PCBs for almost any application. High-speed design and simulation tools are ideal for optical transceiver layouts, and data management and documentation tools can help you prepare for manufacturing.
Contact us or download a free trial if you want to learn more about Altium Designer. This gives you access to the industry's best routing, simulation, and data management tools in a single program. Contact an Altium expert today to learn more.
No comments:
Post a Comment