10GBASE-KR Transmitter Compliance Methodology Proposal. Robert Brink Agere Systems May 13, 2005

Size: px

Start display at page:

Download "10GBASE-KR Transmitter Compliance Methodology Proposal. Robert Brink Agere Systems May 13, 2005"

Penelope Singleton
5 years ago
Views:

1 0GBASE-KR Transmitter Compliance Methodology Proposal Robert Brink Agere Systems May 3, 2005

2 Scope and Purpose Deficiencies of existing transmit template compliance methods are discussed. A proposal for a new compliance testing method capable of verifying a 0GBASE-KR transmitter is presented. Justification for tap range and tap resolution requirements will be presented separately. 2 May 7, 2005 (r.3)

3 Agenda Transmitter Compliance Testing Wish List Template Testing Discussion Equalization Ratio Testing Transmitter model Transmit Equalizer Signal Shaping Transmit Equalizer Solution Space Proposed Specification Methodology Conclusions 3 May 7, 2005 (r.3)

4 Transmitter Compliance Testing Wish List Simple Easy to implement with existing test equipment Repeatable Should test transmitter, not test equipment Fast Minimize time required for compliance testing Definitive Should clearly identify a compliant transmitter 4 May 7, 2005 (r.3)

5 Template Testing Discussion Normalized Amplitude 0 - Bounds rise/fall times, jitter (to some extent), over-/ undershoot and ringing, and relative amplitude. Bounds a single setting of pre- / de-emphasis (with tolerances). Documented as two voltage-time sequences (upper and lower bound). 5 May 7, 2005 (r.3)

6 Template Testing Discussion (continued) Difficult to guarantee across all possible process, voltage, and temperature conditions Masks are generally relaxed to the point that they become meaningless If mask is tight enough to guarantee a given level of equalization most designs will fail somewhere else Susceptible to instrumentation noise, calibration and capability Each template specifies only one equalization setting Variable transmit equalization will require multiple templates Multiple templates will need to be used in characterization testing Difficult to implement and document An algorithmic method may alleviate this problem. 6 May 7, 2005 (r.3)

7 Transmitter Model F(z) is the 3-tap transmit FIR m f m = T is the symbol period A is the peak differential output amplitude x k F(z) t kt Π T A ( f H t ) H 0 ( f ) H t (f) is the transmit pulse shaping filter f 0 = f f 7 May 7, 2005 (r.3)

8 Transmit Equalizer Signal Shaping A f f 0 f f f 0 f V pst 0 f f 0 f V pre A T f f 0 f V pre = A( f f0 + f ) V pst = A( f + f0 + f ) = A( f + f + f ) 0 T NOTE: By convention, f and f - are always negative and f 0 is always positive. f + f0 f = 8 May 7, 2005 (r.3)

9 Transmit Equalizer Solution Space INVALID STATES f - : 8 steps from 0 to 7/40 f : 6 steps from 0 to 5/40 : minimum value should be no less than /0 Pre-Cursor (f ) = 0 = 0 VERIFY VERIFY EACH EACH TRANSMITTER TRANSMITTER STATE STATE 8 VALID STATES Post-Cursor (f ) May 7, 2005 (r.3)

10 Proposed Specification Methodology V pst A 0 e ss Test pattern with equalization For each specified equalization setting, bound the following two ratios: R pre V = V pre ss V R pst = V pre ss A V pre t 0 T t Clause 49 square wave test pattern (recommend n > 8), no equalization Definitions A = peak transmit differential output amplitude T = symbol period t 0 = zero-crossing point of the rising edge of the AC-coupled signal t = zero-crossing point of the falling edge of the AC-coupled signal V pre = peak voltage measured in the interval t 0 T to t 0 V pst = peak voltage measured in the interval t 0 to t 0 +T = steady-state voltage measured as the average voltage in the interval t 0 +2T to t 2T = steady-state error measured as the deviation from in the interval t 0 +2T to t 2T e ss 0 May 7, 2005 (r.3)

11 Notes Use equalizer off setting and clause 49 waveform to validate rise time, jitter, and amplitude requirements. May also want to check peak-peak output amplitude with 00 pattern (n = ). May 7, 2005 (r.3)

12 Transmitter State Definition 8-bit encoding Pre- and post-cursor values encoded as N x step size (for example, 0.025) 7 0 reserved or f sign bit f f - 2 May 7, 2005 (r.3)

13 Proposed Compliance Values State 0 R pre [db] (min) (max) R pst [db] (min) (max) R pre [V/V] (min) (max) R pst [V/V] (min) (max) Refer Refer to to spreadsheet spreadsheet for for a a complete complete list list of of values. values. 3 May 7, 2005 (r.3)

14 Conclusions An algorithmic method for compliance testing the 0GBASE-KR transmit equalizer across settings compatible with the channels of interest is described. The described method has the advantages of being algorithmic and therefore easily implemented and documented. 4 May 7, 2005 (r.3)

15 Back-Up

16 Tolerance Analysis V pre = A( f f0 + f ) V pst = A( f + f + f ) 0 V pre = A ( 2( f + f )) ( + 2( f + )) = A f ( f + f + ) = A f 0 V pst = A( + 2( f + f ) + ( f + f )) 2 ( R ) min = pre ( R ) max = pre max max min min ( V ) pre ( V ) ss ( V ) pre ( V ) ss ( R ) max = pst ( R ) min = pst max min min max ( V ) pst ( V ) ss ( V ) pst ( V ) ss min max ( ) = A( 2 f f ) V pre 2 ( ) = A( 2 f + f ) V pre 2 min max ( ) = A( + 2 f f ) V pst 2 ( ) = A( + 2 f + f ) V pst 2 min max ( ) = A( + 2( f + f ) ( f + f )) 2 ( ) = A( + 2( f + f ) + ( f + f )) 2 6 May 7, 2005 (r.3)

Issues with sampling time and jitter in Annex 93A. Adam Healey IEEE P802.3bj Task Force May 2013

Issues with sampling time and jitter in Annex 93A Adam Healey IEEE P802.3bj Task Force May 2013 Part 1: Jitter (comment #157) 2 Treatment of jitter in COM Draft 2.0 h (0) (t s ) slope h(0) (t s ) 1 UI