Miniature ribbon connectors such as the Champ™ ribbon connector introduce an undesirable level of crosstalk between connector pairs for some applications. As always, this crosstalk results from the specific geometry of the ribbon connector. As applications are pushing to ever higher bandwidths, a number of proposals have been made to reduce the resulting connector crosstalk. Some proposals have attempted to twist components or wires inside the ribbon connector in a unique geometric arrangement to reduce the crosstalk. This solution adds extensively to the manufacturing costs of the ribbon connector, however. Another proposal that has been commercialized is to deviate from the historical pin assignments. The historical pin assignment of tip/ring pairs in the connector, which is referred to as a “standard pinout,” is not optimal from a crosstalk perspective as it creates a large amount of inductive coupling. By re-assigning the pins, the coupling can be made predominantly capacitive. Counter-balancing capacitive coupling is then built into the connector. In addition to adding significant cost, this method creates compatibility problems with a huge base of installed equipment utilizing the historical pin assignments.
In accordance with a non-limiting example, a connector mates to a circuit board at a connector interface. The connector often introduces an undesirable level of crosstalk between pairs. Traces are formed on the circuit board in a “compensation region” that also introduces crosstalk between pairs. The “compensation region” is created in a geometrically controlled fashion such that the crosstalk in the compensation region is of equal magnitude but opposing phase to the crosstalk introduced by the connector. Thus, the overall crosstalk is minimized.
In one example, the compensation region is optimized geometrically to minimize the crosstalk to the nearest neighbor pairs (for example, T3/R3 to T4/R4 and T2/R2). In another example, the geometry is configured to minimize the total crosstalk from all aggressors, resulting in an effective CAT-5 compliant connector from a standard ribbon connector.
A method of forming the circuit board is also set forth.
Other objects, features and advantages of the present invention will become apparent from the detailed description of the invention which follows, when considered in light of the accompanying drawings in which:
The subject matter of this document will now be described more fully hereinafter with reference to the accompanying drawings. This subject matter may, however, be implemented in many different forms and should not be construed as limited to the implementations set forth herein. Rather, examples are provided so that this disclosure will be thorough and complete. Like numbers refer to like elements throughout.
Conductors 122 and 124 form a next nearest neighbor pair of conductors (“NNN” or “next nearest neighbor”) relative to the first pair (e.g., conductors 114 and 116). As used throughout this document, a next nearest neighbor pair of conductors are a pair of conductors that are separated from a given pair of conductors by another pair of conductors. For example, conductors 122 and 124 are considered an NNN relative to the first pair (e.g., conductors 114 and 116) because the conductors 118 and 120 are located between the first pair and the conductors 122 and 124. Similarly, conductors 122 and 124 are an NNN relative to conductors 130 and 132 because the pair of conductors 126 and 128 are located between the pair of conductors 122/124 and the pair of conductors 130/132.
As shown in
As illustrated by
Generally, two pairs of conductors that are closer to each other will experience a higher magnitude of crosstalk than conductors that are farther away from each other. Therefore, with reference to a given conductor pair (e.g., the first pair 114 and 116), the magnitude of crosstalk between an NNP (e.g., 118 and 120) and the given conductor pair will generally be higher than the magnitude of the crosstalk between an NNN (e.g., 122 and 124) and the given conductor pair. Accordingly, the relative relationship (e.g., NNP or NNN) of the given conductor pair to another conductor pair as well as the spacings between the pairs will be a consideration for selecting compensation coupling lengths for the various conductor pairs and/or whether any of the pairs are twisted at a point in the compensation region. As discussed in detail below, it is possible to cancel both near end crosstalk and far end crosstalk using a compensation region similar to that discussed throughout this document.
In accordance with a non-limiting example as shown in the board structures in
The cost of this scheme is limited to the PCB real estate required to implement the compensation region trace structures in the PCB board. These trace structures can be optimized towards Near End Crosstalk (NEXT) cancellation, Far End Crosstalk (FEXT) cancellation, or a compromise between the two. Compensated CAT5-rated ribbon connectors are available at a high relative cost, though they require a non-standard pinout. In accordance with a non-limiting example, the system described herein achieves equivalent performance from a cheaper part while preserving the traditional pair assignments, if desired. It should be understood that the description is not limited to the ribbon connectors, but it can be applied to many different PCB-mounted connectors where the connector itself plus the PCB compensation region is electrically small. Also, in some embodiments, the electrical conduction can be by other than a circuit board trace, for example, by use of wire conductors, as long as the proper geometry is established.
A regular ribbon connector can be mounted on the circuit board and the geometry of the traces in the PCB are arranged to cancel the crosstalk in the ribbon connector. In an example, the traces are geometrically arranged by twisting some pairs at the connector to PCB interface. The twisting action may cause the flux lines in the PCB compensation region to be 180 degrees out of phase with the flux lines from the nearest neighbor pair in the connector. If the magnitudes of the inductive coupling are made the same, the inductive coupling is cancelled. The twisting action also allows the capacitive coupling to be balanced as well if the pair-to-pair spacing and the coupling distance are configured correctly. If the capacitance is balanced between each of the wires in the coupled pairs, capacitive coupling will be limited to the common mode, which is ignored by differential communication systems (this presumes good longitudinal balance—a hallmark of differential systems). Cancelling the flux and balancing the capacitance is controlled via the vertical and horizontal separations of the trace structures along with the length of the compensation region. The combination of the horizontal and vertical separation optimizes the induction and capacitance for crosstalk cancellation, creating a reduced crosstalk interface at the end of the compensation region on the PCB. No special manufacturing of a ribbon connector or other connector is required with this technique. Any off-the-shelf (OTS) components that are electrically small can be compensated for using this technique, and it is possible to have 20 dB improvement using special geometries. Example geometric configurations are discussed in more detail with reference to the figures that follow.
While the results of
When the compensation region is designed to effectively cancel the crosstalk between nearest neighbors, other couplings may become dominant. For example, the coupling to the NNN may become stronger than the cancelled NNP coupling. This is demonstrated in
For example, conductor pair 410 has the first fill pattern along an entire length (e.g., 3.2 inches) of the compensation region. Thus, for the entire length of the compensation region, the conductor pair marked 410 is a pair of conductors having a tip conductor on the top surface of the substrate and a ring conductor on the bottom surface of the substrate.
Meanwhile, conductor pair 430 has a top left to bottom right fill pattern, referred to as a second fill pattern, indicating that the conductor pair 430, which is an NNN relative to the conductor pair 410, visually represents a pair of conductors having a ring conductor on the top surface of the substrate and a tip conductor on the bottom surface of the substrate. As such, the crosstalk coupling between the conductor pair 410 and the conductor pair 430 is inverted (e.g., 180 degrees out of phase or in antiphase) relative to the crosstalk between these conductor pairs that is caused by the connector (“connector crosstalk”). Accordingly, the connector crosstalk between the conductor pair 410 and the conductor pair 430 is reduced and/or cancelled over the entire length of the compensation region.
Similar crosstalk cancellation occurs between the conductor pair 430 and the conductor pair 450 because the crosstalk coupling between the conductor pairs 430 and 450, which are next nearest neighbors, is similarly inverted relative to the connector crosstalk between conductor pair 430 and conductor pair 450. Connector crosstalk between other sets of next nearest neighbors can similarly be reduced and/or cancelled by using a similar layout.
As noted above, the distance between pairs of conductors affects the magnitude of crosstalk coupling between the pairs. Therefore, in
In some implementations, the spacing between the conductor pair 410 and a first portion of conductor pair 420 (e.g., the portion of conductor pair 420 formed over a first portion of the compensation region 402 that extends 1.6 inches) can be selected so that the connector crosstalk between the conductor pairs 410 and 420 can be reduced and/or cancelled based on the crosstalk coupling between the first portion of the conductor pair 420 and the conductor pair 410. For example, as illustrated by
A twist point (shown in more detail with reference to
Meanwhile, the tip conductor of the conductor pair 420 is located on the bottom of the substrate in the first portion of the compensation region, and is routed to the top of the substrate at the twist point. In some implementations, the conductors are routed through the substrate using vias (e.g., a separate via for each of the tip conductor and the trace conductor). After the twist of the conductors at the twist point, the tip conductor of conductor pair 420 remains on top over the second portion of the compensation region (e.g., the 1.6 inch portion following the twist point).
After the twist, the crosstalk coupling between the conductor pair 420 and the conductor pair 430 is inverted relative to the connector crosstalk between the conductor pair 420 and the conductor pair 430. Therefore, over the second portion of the compensation region 404, the crosstalk coupling between the conductor pair 420 and the conductor pair 430 can reduce the connector crosstalk coupling between these conductor pairs in a similar manner to that discussed above with reference to the reduction of connector crosstalk between the conductor pairs 410 and 420.
A similar conductor pair pattern can be repeated as shown in
Conductors 502 and 504 are a single pair of conductors (also referred to as a conductor pair), as are each of conductors 506 and 508, 510 and 512, and 514 and 516. The conductor pair 506 and 508 are a nearest neighbor pair relative to the conductor pair 502 and 504, and the conductor pair 510 and 512 are a next nearest neighbor relative to the conductor pair 502 and 504.
The conductor pair 502 and 504 has a tip conductor on a top surface of the substrate 518 and a ring trace on a bottom surface of the substrate 518. The conductor pair 510 and 512, which is a next nearest neighbor of the conductor pair 502 and 504, has a ring conductor on the top surface of the substrate 518 and a tip conductor on the bottom surface of the substrate 518. As such, the conductor pair 510 and 512 is twisted relative to an orientation of the corresponding tip conductor and ring conductor of the connector interface.
For example, as illustrated by
The conductor pair 502 and 504 remain untwisted over the entire length of the compensation region, while the conductor pair 510 and 512, which is a next nearest neighbor pair relative to the conductor pair 502 and 504, maintains its twisted configuration over the length of the compensation region. As discussed above, the untwisted/twisted configuration between the conductor pair 502 and 504 and its next nearest neighbor pair 510 and 512 results in an antiphase crosstalk coupling (e.g., 180 degree crosstalk phase shift) between the conductor pairs 502/504 and 510/512 relative to the crosstalk between these pairs caused by the connector. The antiphase crosstalk coupling destructively interferes with the crosstalk caused by the connector, such that the crosstalk caused by the connector between these next nearest neighbor pairs is reduced and/or cancelled over the length of the compensation region.
As noted above, the conductor pair 506 and 508 are twisted at the first end of the compensation region. The orientation of the conductors 506 and 508 is maintained between the first end of the compensation region to the twist point 520. At the twist point, the conductor pair 506 and 508 are again twisted (e.g., through vias) so that the orientation of the tip conductor and ring conductor are swapped. That is, at the twist point 520, the ring conductor 506 is routed through a via from the top side of the substrate 518 to the bottom side of the substrate 518, while the tip conductor 508 is routed through another via from the bottom side of the substrate 518 to the top side of the substrate 518. This orientation of the ring conductor 506 and tip conductor 508 is then maintained over a second portion of the compensation region that extends from the twist point 520 to a second end 522 of the compensation region.
As illustrated by
The conductor pair 514/516 has a similar configuration as the conductor pair 506/508, but the tip/ring orientation of the conductor pair 514/516 is inverted relative to the tip/ring orientation of the conductor pair 506/508. For example, in the first portion of the compensation region, the tip conductor 514 on the top of the substrate 518, while the tip conductor 508 is located on the bottom of the substrate, and the ring conductor 516 is located on the bottom of the substrate 518, while the ring conductor 506 is located on the top of the substrate. In the second portion of the compensation region (e.g., from the twist points 520 and 524 to the second end 522 of the compensation region), the tip conductor 514 is located on the bottom of the substrate 518, while the tip conductor 508 is located on the top of the substrate 518, and the ring conductor 516 is located on the top of the substrate 518, while the ring conductor 506 is located on the bottom of the substrate 518. In this way, pair 506/508 is inverted relative to next nearest neighbor pair 514/516 over the entire length of the compensation region.
The enlarged view 604 also shows that the orientation of the conductors 506 and 508 at the second end of the compensation region is inverted relative to the orientation of these same conductors at the first end of the compensation region. This inverted orientation is due to the twist of the conductors 506 and 508 at the twist point 520 of
Note that the spacings shown in
The conductor pairs over which the signals are received can include three or more different conductor pairs. For example, first signals can be received over a first conductor pair, while second signals and third signals can be respectively received over a second conductor pair and a third conductor pair. For example, the first conductor pair (“first pair”) can be the conductor pair 502/504 of
Crosstalk coupling between the first pair and the NNN are inverted over a given length of a compensation region (1204). In some implementations, the crosstalk coupling between the first pair and the NNN is inverted over an entire length of the compensation region. The crosstalk coupling between the first pair and the NNN can be inverted, for example, by twisting the NNN pair (e.g., relative to the first pair) at a first end of the compensation region.
Twisting the NNN relative to the first pair can be achieved by inverting the tip/ring orientation of the NNN relative to the tip/ring orientation of the first pair. For example, as illustrated in
Twisting the NNN pair relative to the first pair changes the polarity of the crosstalk coupling between the NNN and the first pair, which results in inverted crosstalk coupling between the NNN and the first pair (e.g., relative to the connector induced crosstalk distortion). The inverted crosstalk coupling has an antiphase relative to the connector induced crosstalk distortion, and can be maintained over the entire length of the compensation region so that the connector induced crosstalk distortion can be cancelled.
Crosstalk coupling between the first pair and an NNP is inverted (1206). In some implementations, the crosstalk coupling between the first pair and the NNP is inverted over a first portion of the compensation region. The first portion of the compensation region can be, for example, a length of the compensation region that is less than the entire length of the compensation region. For example, in
In some implementations, the crosstalk coupling between the first pair and the NNP can be inverted by twisting the NNP relative to the first pair. Twisting the NNP relative to the first pair can be achieved, for example, by inverting the tip/ring orientation of the NNP relative to the tip/ring orientation of the first pair. For example, as illustrated in
Twisting the NNP relative to the first pair changes the polarity of the crosstalk coupling between the NNP and the first pair, which results in inverted crosstalk coupling between the NNP and the first pair (e.g., relative to the connector induced crosstalk distortion). The inverted crosstalk coupling may cancel the connector induced crosstalk distortion over the first portion of the compensation region.
Crosstalk coupling between the NNP and the NNN is inverted (1208). In some implementations, the crosstalk coupling between the NNP and the NNN is inverted over a second portion of the compensation region. The second portion of the compensation region can be, for example, a length of the compensation region that is less than the entire length of the compensation region. For example, in
In some implementations, the crosstalk coupling between the NNP and the NNN can be inverted by twisting the NNP relative to the NNN. Twisting the NNP relative to the NNN can be achieved, for example, by inverting the tip/ring orientation of the NNP relative to the tip/ring orientation of the NNN. For example, as illustrated in
Twisting the NNP relative to the NNN changes the polarity of the crosstalk coupling between the NNP and the NNN, which results in inverted crosstalk coupling between the NNP and the NNN (e.g., relative to the connector induced crosstalk distortion). The inverted crosstalk coupling has an antiphase relative to the connector induced crosstalk distortion, and therefore, may cancel the connector induced crosstalk distortion over the second portion of the compensation region.
Many modifications and other embodiments of the invention will come to the mind of one skilled in the art having the benefit of the teachings presented in the foregoing descriptions and the associated drawings. Therefore, it is understood that the invention is not to be limited to the specific embodiments disclosed, and that modifications and embodiments are intended to be included within the scope of the appended claims.
Number | Name | Date | Kind |
---|---|---|---|
4720690 | Popek | Jan 1988 | A |
5397862 | Bockelman | Mar 1995 | A |
5430247 | Bockelman | Jul 1995 | A |
6057512 | Noda | May 2000 | A |
6420778 | Sinyansky | Jul 2002 | B1 |
6617939 | Vermeersch | Sep 2003 | B1 |
8357013 | Arai | Jan 2013 | B2 |
20050202697 | Caveney | Sep 2005 | A1 |
20140184350 | Howard | Jul 2014 | A1 |