KEY PRESENTATIONS ON 802.5 REVISED PHY STANDARD

Prepared by Jim Carlo

IEEE 802.5 Vice Chairman (9-Nov-1992)

 

The 802.5 UTP development effort has run in parallel with the 802.5 PHY maintenance effort since 1990. The PAR for UTP, with Alan Marshall as Rapporteur, was approved based on preliminary experimental results by Tricia Hill. Bob Love, Rapporteur of the PHY maintenance document, directed the committee effort on revising the STP Section. In May, 1992, the committee decided to combine both UTP and STP into the single Revised 802.5-1993 Document. The papers described here are just a small portion of the presentations made before this committee to help develop the standard. These papers are working documents and thus may contain errors or information that has been subsequently updated. Nevertheless, these key papers still provide vital insight into understanding changes and rationale for 802.5-1993.

This document has been placed in electronic format for readability by Jim Carlo (31-Oct-1998)

 

ACCUMULATED PHASE SLOPE AND INTEROPERABILITY

The first paper on the importance of Accumulated Phase Slope with regards to interoperability was made by Dave Pearce at the 7/90 meeting. Data showed the importance of evaluating certain frames (with large numbers of ones and zeros) in testing a token ring system. The data showed a phase slope larger than anticipated with a simple PLL cascaded model. At the next meeting in October, both Kathy Otis and Dave Pearce introduced the concept of a Net Delay Time , defined as the difference between the bit delay time of the node and the phase delay time of the node. These new models correctly predicted that phase slope buildup, but at the same time pointed out that unless phase slope was specified in the IEEE 802.5 standard, interoperability was not achievable. The models were reduced to the simple formula below:

Phase Slope Gradient = (K * mFCJ)/ (1 - K*T)

where K is the 3dB bandwidth of the PLL, mFCJ the filtered correlated jitter and T the Net Delay Time. A paper by Takasaki appeared in the literature in detailing this same effect, derived for a different application.

Dave Pearce presented a simple model for the buildup of the accumulated phase slope with a system of two heterogeneous stations and showed that as long as K*T <1 for both stations, then the accumulated phase slope buildup was not greater than any of the slope buildups in the homogeneous case. Kathy Otis later summarized these results in a formal paper on Jitter Accumulation, while Leo Staschover provided a summary tutorial on Correlated Jitter Accumulation. James Harrer provided some digital filtering techniques that could filter the discrete data from the phase slope data to simulate a low pass single pole filter.

In parallel with the developments in jitter accumulation, the definition of interoperability by defining a clear boundary between jitter tolerance and accumulated phase slope was proposed by Jim Carlo. Jitter tolerance was required to be greater than accumulated phase slope, and John Creigh provided initial starting point values of these parameters that would also allow the installed base of equipment to interoperate with new stations. Jim Carlo provided a tutorial on PLL modeling equations and Dave Pearce developed a more complete model for determining the margins between accumulated phase slope and jitter tolerance.

More detailed investigation of accumulated phase slope by John Creigh detailed the effects of Jitter Peaking if the PLL damping factor is too large. Jitter Peaking is a weak accumulation effect, showing up only after many nodes.

 

CHANNEL CHARACTERISTICS

A major effort of the committee was spent analyzing UTP cable characteristics and their effects on receiver performance. Tricia Hill presented the effect of NEXT on reduction in receiver eye opening. Ivan Reede presented a tutorial on modeling the effects of attenuation, crosstalk and reflections on digital pulse distortion. Luc Adriaenssens presented equalization issues for DIW and Cat-5 cable. Luc Adriaenssens presented data on reflections at cable impedance mismatches, which determined that a single impedance from 85 ohm to 135 ohm (120 ohm cable) for cord and cable could not be accommodated due to high pattern jitter effects. Fabrice Autrique presented data on reflections at cable impedance mismatches which showed that a homogenous 120 ohm system may support token ring operation. Dave Pearce developed a statistical model for NEXT effects on alignment error and Luc Adriaenssens provided data on NEXT distribution for DIW and Cat-5 cable. Simon Cox proposed a technique for measuring the effects of NEXT on individual token ring stations.

In addition to cable loss issues, John Creigh presented flat loss for the installed base of token ring concentrators with a proposed allocation of 5.0 dB flat loss. Bob Conte subsequently showed data that 2.0 dB flat loss requirement was acceptable. These values form the basis for the test channels specified in the Receiver Section of the 802.5-1993. Bob Love presented an example of wiring guidelines for different cable type, concentrator type and node count, which formed the basis for the cable design guidelines found in the Annex of the Revised 802.5 Standard.

 

TRANSMITTER

John Creigh provided an initial proposal for tightening the transmitter specification in order to provide control of correlated jitter buildup. This initial paper presented filter poles and zeros and a filter example. Later, the committee decided to specify the transmit waveform masks for zeros, ones and the initial burst_3 in an SDEL and place the filter poles in the Annex.

 

FURTHER WORK

Several papers were presented describing methods for modifying the single station PLL/Concentrator architecture to provide enhanced token ring performance. Alan Marshall presented "jitter buster" technique using a dual PLL (one at high bandwidth and one at low bandwidth). Julio Guerra and Peter Willams presented alternate approaches using an active retimed PLL at each concentrator port. Thomas Coradetti, Carl Hayssen, Ivan Reede and Gunnar Forsberg, Dave Pearce and Kathy Otis also presented schemes for jitter attenuation using specific concentrators. These schemes should provide future standardization efforts with multiple approaches to consider for future extensions to the standard.

 

REFERENCES

Hill, Tricia, "A Performance Study: 16 Mb/s Token Ring on UTP", Synoptics, P802.5/90/36.

Pearce, Dave, "Accumulated Phase Slope Testing", Madge Networks, 7/90.

Otis, Kathy, "Cascaded PLL Model", Chipcom, 10/90.

Pearce, Dave, "Correlated Jitter Build-up in 16 MBit/s Token Ring",Madge Networks, 11/90.

Takasaki, Yoshitaka, "Alignment Jitter Accumulation in a Chain of Processing Node Regenerators", Trans of IEIIE, Vol.73. No.10,1990.

Pearce, Dave, "Heterogeneous Phase Slope Build-up", Madge Networks, 10/92.

Otis, Kathy, "Jitter Accumulation in Token ring UTP Networks", Chipcom, 802.5N/91/7-09.

Staschover, Leo, "Correlated Jitter Accumulation in a Token Ring Network", North Hills Electronics, 802.5N/91/11-07.

Harrer, James, "Digital Filtering of Measured Data for Token ring Applications", Pulse Engineering, 802.5L/91/05-05.

Carlo, Jim, "Interoperability Clarification", Texas Instruments, 11/12/90.

Creigh, John, "Interoperability Goals for T-R Standard", IBM, 3/12/91.

Carlo, Jim, "Simplified PLL Equations and Tutorial", Texas Instruments, 7/8/91.

Pearce, Dave, "ALLSIM - Models and Assumptions", Madge Networks, P802.5/92/3-14.

Creigh, John "A new look at Jitter Peaking Overshoot", IBM, 802.5/91/11-16.

Hill, Tricia, "Next Effect on Link Performance in UTP Token ring Network", Synoptics, P802.5/90/12.

Reede, Ivan,(Part A and Part B) "Baseband Performance of Twisted Pair Transmission Lines for the Transmission of Digital Pulses", Mux Lab,802.5N/91/05-01.

Adriaenssens, Luc, "Equalization issues for DIW and Category-5 UTP cable", AT&T-Atanta, 802.5/91/11-06.

Adriaenssens, Luc, "Reflection Jitter Calculations for Category-5 and 120 Ohm Cables", AT&T-Atlanta, 802.5/92/3-16.

Autrique, Fabrice, "Jitter, Reflections, Next", 802.5/92/5-03.

Pearce, Dave, "Effects of NEXT on Alignment Error in Token-Ring", Madge, P802.5/91/11-04.

Pearce, Dave, "Effects of NEXT on Alignment Error in Token-Ring:Part Two", Madge, 1991.

Adriaenssens, Luc, "NEXT Distribution for DIW and Category-5 UTP Cable", AT&T-Atlanta, 802.5N/92/1-01.

Cox, Simon, "Proposed Text for the Crosstalk Tolerance Specification Section", BT, P802.5/92/3-09.

Creigh, John, "A Proposal for Improving the IEEE 802.5 Test Channel Flat Loss Specification", IBM, 802.5N/91/3-27.

Conte, Bob, "An Analysis of MAU Flat Loss", AT&T Bell Labs, 802.5N/91/7-08.

Love, Bob, "An Example of Wiring Guidelines", IBM,P802.5N/92/05-01.

Creigh, John, "A Proposal for Tightening the IEEE 802.5 Alternate Transmitter Specification", IBM, 802.5N/91/3-28.

Marshall, Alan, "802.5 UTP Jitter Reduction System", Proteon, 802.5-N/90/39.

Guerra, Julio; Love, Robert; Smith, Robert, "An Approach for Supporting 16 Mbit/s IEEE 802.5 Stations on Unshielded Twisted Pair (UTP), 802.5/91/11-05.

Williams, Peter, "PLL-Per-Port Active Concentrator Implementation Advantages for IEEE 802.5 Networks", Star-Tek, P802.5N/92/03-06.

Coradetti, Thomas; Szczepanek, Andre; "Proposed Method for Implementation of 1000 Station Token-Rings (and a solution for several thorny jitter related problems)", NCR/TI, P802.5/90/29.

Hayssen, Carl," A Technique for Jitter Build Up Elimination on IEEE 802.5 Token Ring", Ungermann-Bass, 802.5/91/3-23.

Reede, Ivan, "Accelerated Token Ring Physical Layer", Cabletron, P802.5/92/3-04.

Forsberg, Gunnar,"Feasibility Study of Active Concentrator w/Phase Slope attenuation and Non-Retimed Ports", Ericsson, P802.5/92/9-01.

Pearce, Dave, "The Advantages and Need for Filtering Ringhubs", Madge, P802.5/92/9-04.

Otis, Kathy, "Simulation Results and Performance Comparison of Different HUB Schemes", Chipcom, P802.5/92/9-03.