OCC Design and Manufacture Process

Overview:

A simple patch cord can be the weakest link in a networking system. Connecting hardware designs may be tuned to either the high or low end of the test plug compatibility range, resulting in incompatibility between components and reduced mated plug and jack performance. This directly impacts link and channel performance.

OCC's patented technology has helped create the pathways that make today's high-speed information networks possible. It continues this technical leadership through the creation of the Design Manufacture model, which offers new and more accurate ways of measuring the true performance of components and systems, and of understanding how that affects the total system performance.

OCC's contribution to the standards organizations that govern our industry, along with 19 active patents, 32 active foreign patents, cross-licensed to another 50+, have made OCC a recognized leader in measurement technology. This capability has put OCC at the forefront of standards that directly impact performance limits relevant to the end user.

Crosstalk 101:

Evolutions in cable technology have resulted in improved performance in high-speed data transmission. Category 5e cabling is better than Category 5, and Category 6A has more than five times the bandwidth than that of Category 5e. But what is the difference between two connectors that gives one a better category rating over another?

In category ratings, the primary emphasis is typically assigned to the improvement of Near-End Crosstalk (NEXT). The mated connection of a jack and plug is potentially the greatest source of NEXT within a channel or link, due mostly to the poor NEXT performance of the plug upon which the industry has standardized. Because the plug was designed at a time before digital high speed communications, it possesses physical attributes which create a great deal of NEXT. By itself, the plug would not pass Category 5 connector crosstalk requirements. The plug relies on the jack to compensate or reduce the crosstalk to levels. If left unchecked, this NEXT is enough to prevent high speed data signals from being transmitted properly within the channel or link. The jack's function is to eliminate or cancel as much of the plug NEXT as possible.

To do this, a jack is designed with a NEXT signature that is very close to the opposite NEXT signature of the plug. We call it compensation, and it is the basic design requirement for all connectors that have a mated performance that is better than the individual performance of either of the two parts. In this case, the sum of the parts is greater than the whole. This compensation requirement was invented and patented at OCC, and it revolutionized the connector industry by making possible high speed communication between all types of computers over RJ-45 infrastructures. Any connector having NEXT compensation, uses this method.

An oversimplified explanation of plug compensation can be represented as:
|1 +(-1)| = 0

If the plug is represented by a positive 1. then the connector must be negative 1. When the two are mated together, the result should equal zero. In reality, +1 and -1 are called vectors. These two vectors have the same magnitude (1) but opposite phase (direction). Thus, they cancel each other out.

Unfortunately, because the plug assembly to the cable is done by hand, there can be quite a bit of variation in the plug NEXT performance, plugs tend to range in NEXT performance from a low to a high value and everything in between. This complicates the connector design process, since the connector must work with those plugs which are in the center of the plug limits, as well as those which are right at the limits. Figure 1 illustrates the relationship between plug NEXT on the X-axis, and mated NEXT of the plug and jack on the Y-axis, and shows this relationship with respect to the allowed NEXT limits for categories 5e, 6, and 6A. These categories all have to be backward compatible, so Cat6 and 6A must use plugs with the same or reduced range of Cat5 and 5e. Notice that for any of the three categories, if the plug has more NEXT (at the left) on the horizontal scale, the mated NEXT will have very little margin to the limit line on the vertical scale. Move to the right on the X-axis and the plug NEXT starts to be less, while the mated NEXT starts to get better. This trend continues until the plug NEXT and the connector NEXT reach that magic "1 + (-1)" value, shown where the performance peaks the highest above the limit line. The mated NEXT will not get better than it is at that point for the given connector and plug range. Keep moving to the right, and as the plug NEXT becomes less, the mated NEXT becomes greater. At a point, the plug NEXT and connector NEXT will diverge far enough that they will no longer cancel each other out, and the mated NEXT will fail. An important consideration to remember is that more crosstalk is represented by lower numbers, such as -30, and less crosstalk is represented by higher numbers such as -60.


Figure 1: Plug NEXT versus Mated NEXT

It may seem counter-intuitive to see the mated NEXT get worse, while the plug NEXT becomes better, but because of the way the jack compensates for the plug, that is exactly what happens. To put it in terms of the aforementioned formula, if the plug were much better, like 0.2, and the connector didn't change (it can't), then our formula would look like this:
|0.2+(-1)| = 0.8

The difference is 0.8. Closer to 1 than 0 and would theoretically fail in this scenario. Figure 1 illustrates the mated response of a plug and jack that has been tuned for the center of the ANSI/TIA-568-C.2 plug range. This range was chosen based on many plug measurements taken by OCC and others to determine the practical range of plug NEXT. The restriction of plugs to this performance range has allowed for the industry to maintain compatibility with lower category levels as well as compatibility between connectors from different manufacturers. It is possible, and sometimes easier, to design a better connector that works with a different range of plugs; however, such a connector would not be backward compatible with previous generations of components and so would not be compliant to the standard, or possibly even compatible with it. This would mean that one would only be able to use plugs and jacks from the same manufacturer in a system – as though a special USB plug was necessary for each computer that it plugs into.

Testing Performance:

The test equipment used for Category 5 was very generic, but was adapted to specific needs by individual laboratories. In general, the equipment needed would be a vector network analyzer (VNA), and two or more balun transformers. A VNA is a precision piece of equipment that is the core of any telecommunications test lab, and it generally is responsible for measuring all transmission performance parameters specified for a device under test (DUT) especially NEXT. A VNA works by sending out a signal on one transmission path, while a receiver detects the level of that signal on another transmission path. It reports the level detected in decibels, a logarithmic unit indicating the ratio of power received relative to the input power.

Balun transformers are devices which are inserted between the VNA and the DUT to convert the VNA signals to differential signals, and back again. These are necessary because the structured cabling used in these communication systems is based on differential signaling schemes. Differential signals are unique in that there is not a ground or zero reference for the signal being transmitted. Instead, the signal transmitted on a conductor is referenced to a complimentary signal on a second conductor in close proximity.

The main advantage to this signaling method is the great speed with which signals may be transmitted while simultaneously rejecting outside sources of noise. Other items would include connecting cables from the VNA to the baluns, and some sort of fixture to mount the baluns for convenience. Since the ground connection is not present in unshielded transmission cabling, the ground connection might be left "floating" or disconnected from equipment ground. This was found to affect the test results in random ways, making comparisons difficult.

These test setups were not up the task of reaching frequencies much beyond 100 MHz, and the new Category 6 standard being developed proposed an upper bandwidth limit of 250 MHz, more than double that of the Category 5e standards. , The main problem to be resolved to increase the reliable frequency range of testing was with the impedance of test leads used to connect the DUT to the baluns. When sitting in free air, the twisted pair leads would have relatively uncontrolled impedance, causing performance anomalies.

To solve this problem, OCC developed a completely new system for testing connecting hardware. This system consisted of a standardized and practical arrangement for the test baluns, a set of adapters based on printed circuit technology for repeatability, pin and socket interface pins to maintain impedance matching to the test baluns, and a plated pyramid adapter that controlled the test lead impedance and reduced mismatch to the DUT. The fixture also included a set of calibration standards.


Figure 2: OCC Category 6 Test Fixture with Pyramid

These fixtures increased the measurement bandwidth of connecting hardware from 100 MHz to frequencies beyond 250 MHz, making Category 6 performance testing consistent between different labs, and simultaneously improving the testing of Category 5e. When scientists and other industry leaders saw what OCC had developed, they immediately recognized the benefits and adopted it in their own laboratories. When the new Category 6 standard was written, the OCC Category 6 test fixtures were specified as the primary compliance test interface. The pyramid adapter became ubiquitous in labs across the country, and around the globe.

Soon after the approval of the Category 6 standard of ANSI/TIA/EIA-568-B.2-1, work began on the next level of performance: Category 6A. Once again, there was an increase in bandwidth with the upper range being specified up to 500 MHz. The increase in frequency brought new challenges in measurement accuracy. While the pyramid adapters worked well for their design intent, the new challenges of measuring out to 500 MHz required a still more refined approach. Once again, OCC met these challenges with an innovative new idea.

Instead of routing long test leads from the balun interface to the DUT, OCC created an engineered signal path between the balun and DUT that could be error corrected using existing calibration technologies. The length of test leads, and associated impedance variation was reduced from about 2.5" to less than 1/4". This novel approach made it possible to shrink the pyramid adapter into the very small DUT interface board shown in Figure 3.


Figure 3: OCC 6A DUT Adapter

For a size comparison of the new C6A DUT Interface with the old pyramid adapter, see Figure 4. The new form factor allows for almost zero length breakout of connections from the DUT to the interface, keeping mismatches to a minimum.


Figure 4: OCC C6A DUT Adapter Vs. Pyramid Adapter

Test Plugs:

The specified NEXT range of test plugs is not chosen at random. The values represent the performance of plugs found on quality patch cords. Also, the industry has been careful to specify values which maintain backwards compatibility to previous standards. This can be seen again in Figure 1. Notice how the Category 6 connector mated with a plug in the 5e (but not 6) range will still pass 5e limits.)

The work done to obtain these values was not trivial. Historically, there were special problems in trying to characterize a plug. A plug's performance could not be measured directlywithout connecting it to some device. The only device available to mate to an RJ-45 plug was typically an RJ-45 jack, and the jack will always contribute something to the measurement results, so the plug itself was still an unknown.

The earliest method of measuring a plug without any influences was to terminate the plug leads in the system impedance right at the balun output. This was known as the TOC (Terminated Open Circuit) method. The TOC results, however did not correlate well with the mated plug and jack results , because it did not measure all of the crosstalk within the plug (see Figure 5). Using this method, one lab could qualify a plug at a certain NEXT value, while another lab might qualify a different plug at that same NEXT value, however, the measured crosstalk of the same connector mated with either of these two plugs would be different.

The second method developed by the TIA was the de-embedding method, in which the plug would be measured while mated to a jack having a known signature. That signature would then be subtracted from the mated result, leaving only the NEXT of the plug. This method was better than the TOC method but still yielded different results between labs. Other drawbacks to the de-embedding method included its complexity and difficulty when characterizing other plug performance factors such as far-end crosstalk (FEXT) and return loss (RL). Still, this was a solid improvement over the TOC method.


Figure 5: Terminated Open Circuit Test Method

Seeing a need for improved measurement techniques and fixtures, OCC developed the industry's first standardized means for testing plugs and connecting hardware. The Direct Plug Measurement Fixture (DPMF) was created for the Cat6 test fixtures, and was designed to connect a plug to the test fixture while having almost zero impact on the plugs performance, allowing the plug to be characterized for all critical parameters, including FEXT and RL. Figure 6 shows a plug mounted to a pyramid, and being inserted into a DPMF. The DPMF was later improved to mate with the Cat6A test fixtures. Figure 7 shows a plug mounted, and being tested with a Cat6A compatible Direct Plug Measurement Fixture 2 (DPMF-2).


Figure 6: Original Direct Plug Measurement Fixture


Figure 7: Direct Plug Measurement Fixture 2

With the creation of the Cat6A standard (ANSI/TIA/EIA-568-B.2-10), OCC's DPMF-2 became the standard method for test plug characterization.

Patch Cord Testing:

The premise behind a test plug is that it should represent the plug on the end of a well made patch cord. Conversely, the patch cord plug should fall within the range of requirements of the test plug. If the patch cord plug falls outside of the range of the test plug, the mated connection will not meet the performance requirements.

While the cable used in the cord construction is important, the plug design and the skill of the operator terminating the plug to the cable are the keys to the overall ability of the patch cord to qualify to any category level. There are many patch cords sold today that do not meet the component requirements of the TIA standards. Many may not know that the standards have a special test for qualifying the performance of patch cords as a separate component. A look at the cords being offered by some well known national retailers reveals a lack of effort has been put into controlling the assembly of the product, and the category rating printed on the package typically only reflects the category rating of the cable used to build it. How can the performance of a plug on a patch cord be verified when the only way to measure a plug by itself is to cut it off of the cable and connect it to the test fixture using a Direct Plug Measurement Fixture?

The answer lies in the relationships that exist between test fixtures, test plugs, and mated NEXT. OCC developed the test fixture to characterize a mated connection and a centered test plug, and OCC developed a jack to match perfectly with that centered test plug. Because of this, the jack becomes more than just a jack, it can be used to qualify patch cords. TIA created special requirements for a patch cord test adapter with this in mind. They then created test limits based upon the performance of two mated connections and the cable in between. This is the component patch cord test. It qualifies patch cords based upon their mated performance with qualified (centered) jacks. If the jacks are properly centered, and the plugs are properly centered, the mated crosstalk of the connections passes the connector component requirements and (with the addition of the cable crosstalk) the patch cord then passes its component requirements.

A Patch Cord Test Adapter is a connector that meets the requirements of ANSI/TIA-568-C.2, Clause C.5.2, Test fixturing for modular cords. This clause states that, to qualify as a modular cord test head, the connectors mated NEXT must be at a minimum near the center of the test plug range. It must also meet more stringent requirements for FEXT cancellation and Impedance matching than the specifications given for ordinary category level connectors.
Because of the relationship between mated NEXT, and plug NEXT, a patch cord tested with these adapters will pass the component limits when the plugs are well controlled, or it will fail if they are not.

Summary

OCC's Design Manufacture model offers a closed loop development cycle for the products which it manufactures. This cycle relates directly back to standards and laboratory testing as shown in Figure 8.


Figure 8: OCC Design Manufacture Process

The benefits to the end user include:

  • Mechanical Ruggedness: The design intent of many connectors is to meet the requirements of a structured cabling install. The typical connector needs only to survive the installation, and the occasional changes that may occur within the day to day operation of a business. These connectors are rarely stressed, and because quality costs money, they may only meet the minimum requirements for reliability and performance. OCC connectors go beyond installation survival, because every single jack is designed to be used as tool in constant use as a test adapter. They have the quality built in to withstand the rigors of a production line test environment. And because OCC jacks are developed as a reference connector, they also meet the tighter performance requirements given in the standards for FEXT and RL. No other company has a need to build this into their connectors, but OCC does to ensure the end user has the best infrastructure possible.
  • Installation Flexibility: Challenging installations are no longer a concern with OCC's tightly controlled Design Manufacture model. Short cabling runs which are usually the most difficult to qualify for near end crosstalk and return loss due to the close proximity of the connectors to each other are no problem. OCC products have been third party tested1 to demonstrate standards compliance even in 6ft long, 3 connector links. Three connector links which are that short are rare, but OCC tests the absolute worst case scenario to give the customer peace of mind that their installation will work where it must. And, because OCC's Design Manufacture model means component compliance, OCC products are compatible with any other component compliant products that might already be installed, giving the customer the flexibility to install the products they need without being limited to the offerings of only one company.

Certified Jacks:

OCC now offers Certified Centered jacks to its customers, which meet the rigorous requirements applicable to patch cord test adapters. The Certified Centered jack is a similar product to a patch cord test adapter in that it has been qualified by a technician in our ISO 9001 facility using a fully characterized test plug and vector network analyzer to ensure peak performance. The only difference in a Certified Centered jack and patch cord test adapter is that the Certified jack can be ordered in any color required. With a Certified Centered jack, the customer can be assured of top interface performance.

1 See ETL Intertek reports 100226050CRT-001b and 100226050CRT-001d