Actual Cable Data update

Actual Cable Data update Insertion loss, return loss, charcteristic impedance at worst case condition, propagation delay Coupling attenuation at different ageing conditions Screening attenuation at different ageing conditions Ricky Vernickel (LEONI Kabel GmbH) Supporters: Thomas Müller (Rosenberger) Eric DiBiaso (TE Connectivity) Josef Ohni (MD Elektronik) Bangkok, November, 2018 1

Motivation Actual Cable Data - Update about development progress of different cable types - Requests for additional cable parameter data - Review of actual limit recommendation and new proposal 2

Tested cables Shielded twisted pair (STP) 0.14mm² AWG 26 specified up 5.5 GHz Shielded twisted pair (STP) 0.22mm² AWG 24 specified up to 4.5 GHz PROTOTYPE! Cable length 3.5m (coupling and screening attenuation) Cable length 10m (all other measurements) All cables are bulk cables no extra inliner or connector! 3

Test setup (Coupling attenuation) A standard 4-port vector network analyzer from 1 MHz to 8 GHz was used The tests were performed at room temperature (RT, 23 C) before and after ageing (240h/130 C) The length of each cable is 3.5m (3.0m inside tube) The 3m Rosenberger Bedea measurement tube (triaxial method) was used The measured cables are bulk cable without any inliner or connector 4

a c in db Coupling attenuation data AWG 26 0-20 Length measured: 3m Tube length: 3,5m cond. as delivered Temperature: RT Evaluation window: 36,3 MHz - 1 GHz Coupling attenuation IEEE 802.3ch STP AWG26 Differential-mode coupling attenuation acdiff Limit Condition as delivered (new) at room temperature No limit violation -40 Limit: * mueller_3ch_02a_0518.pdf -60-80 -100-120 1,0 10,0 100,0 1.000,0 f in MHz Plot created with: LEONI coupling attenuation V 1.2 5

a c in db Coupling attenuation data AWG 26 0-20 Length measured: 3m Tube length: 3,5m after short time ageing 240h/130 C Temperature: RT Evaluation window: 36,3 MHz - 1 GHz Coupling attenuation IEEE 802.3ch STP AWG26 Differential-mode coupling attenuation acdiff Limit Condition: after short time ageing (240h/130 C) at room temperature No limit violation -40 Limit: * mueller_3ch_02a_0518.pdf -60-80 -100-120 1,0 10,0 100,0 1.000,0 f in MHz Plot created with: LEONI coupling attenuation V 1.2 So significant change of coupling attenuation for this cable type after ageing. 6

a c in db Coupling attenuation data AWG 24 0-20 Length measured: 3,5m Tube length: 3m cond. as delivered Temperature: RT Evaluation window: 44,1 MHz - 1 GHz Coupling attenuation IEEE 802.3ch STP AWG24 Differential-mode coupling attenuation acdiff Limit Condition as delivered (new) at room temperature No limit violation -40 Limit: * mueller_3ch_02a_0518.pdf -60-80 -100-120 1,0 10,0 100,0 1.000,0 f in MHz Plot created with: LEONI coupling attenuation V 1.2 7

a c in db Coupling attenuation data AWG 24 0-20 Length measured: 3,5m Tube length: 3m after short time ageing 240h/130 C Temperature: RT Evaluation window: 44,1 MHz - 1 GHz Coupling attenuation IEEE 802.3ch STP AWG24 Differential-mode coupling attenuation acdiff Limit Condition: after short time ageing (240h/130 C) at room temperature No limit violation -40 Limit: * mueller_3ch_02a_0518.pdf -60-80 -100-120 1,0 10,0 100,0 1.000,0 f in MHz Plot created with: LEONI coupling attenuation V 1.2 So significant change of coupling attenuation for this cable type after ageing. 8

Test setup (Screening attenuation) A standard 4-port vector network analyzer from 1 MHz to 8 GHz was used The tests were performed at room temperature (RT, 23 C) before and after ageing (240h/130 C) The length of each cable is 3.5m (3m inside tube) The 3m Rosenberger Bedea measurement tube (triaxial method) was used (cable treated quasi-coaxial) The measured cables are bulk cable without any inliner or connector 9

a s in db Screening attenuation data AWG 26 0 10 20 30 40 Length measured: 3,5m Tube length: 3 m Temperature: RT Condition: see legend Evaluation window: 193 MHz - 6 GHz Screening attenuation (as) IEEE 802.3ch STP AWG 26 cond. as delivered as minimum screening attenuation in evaluation window STP AWG 26 after short time ageing 240h/130 C as minimum screening attenuation in evaluation window Condition as delivered and after short time ageing (240h/130 C) at room temperature. Minimum requirement of 45 db met (*DenBesten_3ch_03_311018.pdf) 50 52,2 52,1 60 70 80 90 100 10,0 100,0 1.000,0 f in MHz Plot created with: LEONI screening attenuation V 2.2 10

a s in db Screening attenuation data AWG 24 0 10 20 Length measured: 3,5m Tube length: 3 m Temperature: RT Condition: see legend Evaluation window: 212 MHz - 6 GHz Screening attenuation (as) IEEE 802.3ch STP AWG 24 cond. as delivered as minimum screening attenuation in evaluation window STP AWG 24 after short time ageing 240h/130 C as minimum screening attenuation in evaluation window Condition as delivered and after short time ageing (240h/130 C) at room temperature. 30 40 Minimum requirement of 45 db met (*DenBesten_3ch_03_311018.pdf) 50 48,7 51,9 60 70 80 90 100 10,0 100,0 1.000,0 f in MHz Plot created with: LEONI screening attenuation V 2.2 11

Test setup (IL, RL, impedance, propagation delay) A standard 4-port vector network analyzer from 1 MHz to 20 GHz was used. The tests were performed at room temperature (RT, 23 C) and at 105 C at climatic chamber before and after ageing (240h/130 C). The measured length of each cable is 10m and the results for IL were calculated for 11m and 15m respectively in the diagrams and compared to current insertion loss limit. The measured cables are bulk cable without any inliner or connector. Only the worst case condition after ageing (240h/130 C) is displayed in the diagrams for a better overview (insertion loss and return loss). Worst case situation evaluated in presentation *vernickel_3ch_01b_0918 during last face to face meeting in September. 12

Z (TDD11) in Ω Z (TDD11) in Ω Characteristic impedance data for AWG 26 115 110 105 100 95 90 Characteristic Impedance Differential Mode (CIDM), near end IEEE P802.3ch Impedance now more optimized compared to previous presentation. 85 0 1 2 3 4 5 6 7 8 9 10 length in m AWG 26 RT after short time ageing 240h/130 C CIDM 115 CIDM in evaluation window (0,5..1,5m) AWG 26 RT after short time ageing 240h/130 C CIDM, corrected; mean value in evaluation window (0,5..1,5m) = 102,22 Ohm AWG 26 To 105 C after short time ageing 240h/130 C CIDM AWG 26 To 105 C after short time ageing 240h/130 C CIDM, corrected; mean value in evaluation window (0,5..1,5m) = 102,86 Ohm Plot created with: LEONI Plot29 V 2.12 110 105 100 95 90 85 0,5 0,7 0,9 1,1 1,3 1,5 length in m Length measured: 10 m Ring temperatures see legend condition see legend Rise time: 37,6ps The TDR was calibrated using an E-Cal-Kit. Then, a pair of airlines with Z_(DM_Airline )=100Ω was measured. The measured sample impedance is corrected using the measured airline impedance using Z Sample = Z meas Z DMAirline Z Airlinemeas The impedance values are attenuation corrected using a method similar to Open Alliance TC 2, Annex B or Open Alliance TC9, Annex B 13

IL in db/-11m Insertion loss data for 11m AWG 26 Insertion Loss (IL) IEEE P802.3ch Calculated for 11m 0-5 -10-15 -20-25 -30-35 AWG 26 RT after short time ageing 240h/130 C IL SDD12 Worst case after ageing displayed only. (see presentation vernickel_3ch_01b_0918.pdf) No limit violation of actual limit. New proposal for a stronger limit: ILdBf 0.0025 f+0.35 f+1/ f f is the frequency in MHz: 5 5500-40 AWG 26 To 105 C after short time ageing 240h/130 C IL SDD12-45 New proposal Current Limit -50 0 1000 2000 3000 4000 5000 f in MHz Plot created with: LEONI Plot29 V 2.12 Note: S DD21 and S DD12 usually overlap in a w ide range. This limit would allow 4dB (@ 3 GHz) margin for connectors/inliner, measurement uncertainty and production variations for the worst case after ageing at 105 C. Current Limit: *DiBiaso_3ch_01_0718.pdf ILdBf 0.002 f+0.45 f+1/ f 14

-RL in db Return loss data AWG 26 0-10 -20-30 -40 Return Loss (RL) IEEE P802.3ch No limit violation Gated measurement: Influence of measurement adapter compensated Pure cable data -50-60 -70-80 AWG 26 RT after short time ageing 240h/130 C RL SDD11 AWG 26 To 105 C after short time ageing 240h/130 C RL SDD11 RL (IL3GHz >20dB) RL (10 db<il3ghz <20dB) RL (IL3GHz <10dB) 0 1000 2000 3000 4000 5000 f in MHz RL Limit: * Farjad_3ch_01b_0118.pdf Plot created with: LEONI Plot29 V 2.12 15

Lz in ns/1m Propagation delay data AWG 26 8 7 Propagation delay IEEE P802.3ch Propagation delay always better than the required 6,27 ns/m in all conditions. 6 5 4 AWG 26 RT cond. as delivered SDD21 Lz AWG 26 To 105 C cond. as delivered SDD21 Lz 3 AWG 26 RT after short time ageing 240h/130 C SDD21 Lz AWG 26 To 105 C after short time ageing 240h/130 C SDD21 Lz Limit 6.27 ns/m 2 0,1 1 10 100 1000 f in MHz Plot created with: LEONI Plot29 V 2.12 16

Z (TDD11) in Ω Z (TDD11) in Ω Characteristic impedance data for AWG 24 115 110 105 100 95 90 Characteristic Impedance Differential Mode (CIDM), near end IEEE P802.3ch Impedance of AWG 24 PROTOTYPE needs more improvement which will lead to an increase of insertion loss (simulation data added to the insertion loss diagram). 85 0 1 2 3 4 5 6 7 8 9 10 length in m AWG 24 RT after short time ageing 240h/130 C CIDM 115 CIDM in evaluation window (0,5..1,5m) Plot created with: LEONI Plot29 V 2.12 110 105 AWG 24 RT after short time ageing 240h/130 C CIDM, corrected; mean value in evaluation window (0,5..1,5m) = 109,05 Ohm AWG 24 To 105 C after short time ageing 240h/130 C CIDM 100 95 90 AWG 24 To 105 C after short time ageing 240h/130 C CIDM, corrected; mean value in evaluation window (0,5..1,5m) = 109,44 Ohm 85 0,5 0,7 0,9 1,1 1,3 1,5 length in m Length measured: 10 m Ring temperatures see legend condition see legend Rise time: 37,6ps The TDR was calibrated using an E-Cal-Kit. Then, a pair of airlines with Z_(DM_Airline )=100Ω was measured. The measured sample impedance is corrected using the measured airline impedance using Z Sample = Z meas Z DMAirline Z Airlinemeas The impedance values are attenuation corrected using a method similar to Open Alliance TC 2, Annex B or Open Alliance TC9, Annex B 17

IL in db/-15m Insertion loss data for 15m AWG 24 PROTOTYPE 0-5 Insertion Loss (IL) IEEE P802.3ch Calculated for 15m Worst case after ageing displayed only. (see presentation vernickel_3ch_01b_0918.pdf) -10-15 -20-25 No limit violation of actual limit. The improvement of the impedance will lead to an increase of approx. 1.5 db @ 3 GHz acc. to simulation data. -30-35 -40-45 AWG 24 RT after short time ageing 240h/130 C IL SDD12 AWG 24 SDD12 simulation RT 23 C after short time ageing (1GHz-4.5GHz) corrected impedance (100Ω) AWG 24 To 105 C after short time ageing 240h/130 C IL SDD12 AWG 24 SDD12 simulation To 105 C after short time ageing (1GHz-4.5GHz) corrected impedance (100Ω) New proposal New proposal for a stronger limit: ILdBf 0.0025 f+0.35 f+1/ f f is the frequency in MHz: 5 4500 Current Limit -50 0 1000 2000 3000 4000 f in MHz Plot created with: LEONI Plot29 V 2.12 Note: S DD21 and S DD12 usually overlap in a w ide range. Current Limit: *DiBiaso_3ch_01_0718.pdf This limit would allow 5.5dB (@ 3 GHz) margin for connectors/inliner, measurement uncertainty and production variations for the worst case after ageing at 105 C. ILdBf 0.002 f+0.45 f+1/ f 18

-RL in db Return loss data AWG 24 0-10 -20-30 -40-50 Return Loss (RL) IEEE P802.3ch No limit violation Gated measurement: Influence of measurement adapter compensated Pure cable data RL Limit: * Farjad_3ch_01b_0118.pdf -60-70 -80 AWG 24 RT after short time ageing 240h/130 C RL SDD11 AWG 24 To 105 C after short time ageing 240h/130 C RL SDD11 RL (IL3GHz >20dB) RL (10 db<il3ghz <20dB) RL (IL3GHz <10dB) 0 1000 2000 3000 4000 5000 f in MHz Plot created with: LEONI Plot29 V 2.12 19

Lz in ns/1m Propagation delay data AWG 24 8 7 Propagation delay IEEE P802.3ch Propagation delay always better than the required 6,27 ns/m in all conditions. 6 5 4 AWG 24 RT cond. as delivered SDD21 Lz AWG 24 To 105 C cond. as delivered SDD21 Lz 3 AWG 24 RT after short time ageing 240h/130 C SDD21 Lz AWG 24 To 105 C after short time ageing 240h/130 C SDD21 Lz Limit 6.27 ns/m 2 0,1 1 10 100 1000 f in MHz Plot created with: LEONI Plot29 V 2.12 20

Conclusions Proposal (* mueller_3ch_02a_0518.pdf) for coupling attenuation is acceptable for both cable types. A screening attenuation limit of 45dB is acceptable for both cable types. Propagation delay within requirements in all conditions. For 15m and 11m link length no limit violation for current limit with both cable types. 21

Conclusions Actual cable data would allow a stronger limit for insertion loss. New proposal for new insertion loss limit: ILdBf 0.0025 f+0.35 f+1/ f f is the frequency in MHz: 5 5500 (4500 AWG24) 22

Future discussion It is highly recommend to use also STP cable for 2.5 GB/s: - discussions about the use of STP cables also for 1 GB/s, due to EMI problems with unshielded cables in critical installation positions - it is very difficult to connect a cable just with conductive foil to a connector so a normal braiding is required/recommended - Relative cost difference might not be significant enough to justify the development of a complete new connector and/or cable type. Proposal: For 2.5GB/s use the same limit as for 5 and 10 GB/s but different frequency range. Proposal 1 GHz for insertion and return loss. 23

Thank you!!! 24