Field Testing Update

Transcription

1 Field Testing Update Adrian Young Fluke Networks November, 2013 Singapore

2 Objectives for this session Copper field standards update Look at new copper field measurements Fiber field standards update

3 IEC Ed.4 Defines field tester accuracy for ISO/IEC 11801:2010 Level IIe, supports Class D (100 MHz) Level III, supports Class E (250 MHz) Level IIIe, supports Class E A (500 MHz) Level IV, supports Class F (600 MHz) Level V, supports Class F A (1,000 MHz) Status Draft, publication Q2 2014? Work can begin on 2 GHz field testing once published

4 ANSI/TIA-1152 Defines field tester accuracy for ANSI/TIA-568-C.2 Level IIe, supports Category 5e (100 MHz) Level III, supports Category 6 (250 MHz) Level IIIe, supports Category 6A (500 MHz) Status Published September 2009 Opened for review at TIA meeting in Portland (June 2013) Editor: Fluke Networks Work has begun on supporting 2 GHz field testing Other measurements to be considered Resistance, Resistance Unbalance Possible new link definition(s)

5 Resistance Unbalance Specified in IEEE 802.3af and IEEE 802.3at for the successful transmission of Power over Ethernet

6 Resistance Unbalance Difference in DC Resistance between conductors of the same pair 1.87 Ω 1.85 Ω Resistance = 3.7 Ω Resistance Unbalance = 0.02 Ω

7 Resistance Unbalance ISO/IEC 11801: Ω or 3.0% for the Permanent Link (Greater of the two) 0.20 Ω or 3.0% for the Channel (Greater of the two) Not a requirement in IEC field testing standard.yet ANSI/TIA-568-C Ω or 3.0% for the Channel (Greater of the two) Not a requirement in ANSI/TIA-1152 field testing standard yet

8 Copper Cladded Aluminum Cable Limited data from the field suggests this measurement could be used to detect the presence of such cable CCA cable is not permitted in either TIA or ISO

9 Category 8 ANSI/TIA-568-C.2-1 Support proposed for 40GBASE-T Supported Channel length now 30 m Status Working on laboratory fixtures to support 2 GHz testing Need round robin testing between labs to see if there is good correlation of measurements Test plug performance to be decided No fixture available to measure test plugs to 2 GHz Field testers Baseline accuracy demonstrated No Channel measurement accuracy demonstrated No Perm. Link accuracy demonstrated

10 Baseline accuracy ANSI/TIA-568-C.2-1 Accuracy without adapters Adapter development is the real challenge here

11 New FIELD copper measurements

12 Balance Balance is critical for successful transmission Reduces emissions Mitigates external noise sources

13 Balance Balance is critical for successful transmission Reduces emissions Mitigates external noise sources

14 Balance Defined in ANSI/TIA-568-C.2 Channel limits only for TCL and ELTCTL Not a requirement in ANSI/TIA-1152 field testing Defined in ISO/IEC 11801:2010 Channel limits for only TCL and ELTCTL Not a requirement in IEC Ed.4 (draft) field testing Industrial Ethernet group very interested

15 TCL (Transverse Conversion Loss) Transverse Conversion Loss is the ratio (in db) of a common-mode voltage measured on a wire pair relative to a differential-mode voltage applied to the same end of the pair. The TCL value shows you how well the impedances of the pair s conductors are balanced.

16 TCL (Transverse Conversion Loss) A high TCL value means that the impedances of the conductors relative to ground are almost equal. High TCL values correspond to better noise immunity and lower emissions.

17 TCTL (Transverse Conversion Transfer Loss) Transverse Conversion Transfer Loss is the ratio (in db) of a common-mode voltage measured at the far end of a wire pair relative to a differential-mode voltage applied to the near end of the same pair. TCTL values improve with longer links We actually use Equal Level Transverse Conversion Loss (ELTCTL)

18 ELTCTL (Equal Level Transverse Conversion Transfer Loss) Subtracting the effects of insertion loss normalizes the results for length and produces ELTCTL values, which are not dependent on length. Because ELTCTL does not depend on length, it is used instead of TCTL to evaluate cable performance. High ELTCTL values correspond to better noise immunity and lower emissions.

19 Shield integrity ALL wire mappers would think the shield was connected Grounding of two racks makes this happen Poor shielded terminations can result in poor Alien Crosstalk

20 Shield integrity An incorrectly terminated shield can adversely affect transmission properties such as Alien Crosstalk The data below was taken from a data center where the shield was terminated incorrectly Better Worse The traditional Wire Map passed the shield shield

21 New Wire Map approach Shield Integrity Use a Time Domain Reflectometer test to see if the shield is truly connected by applying a differential signal between the shield and all four pairs

22 Shield integrity Even when the two connectors are grounded together

23 Shield integrity Traditional Wire Mappers will show this connected

24 Modified Single Connector Permanent Link IP camera and Wireless Access Point links are often terminated with an RJ45 plug Standards participants looking to define this link definition which is already found in BICSI but no test guidelines are given

25 Modified Single Connector Permanent Link Recommended method at this time is to use a Permanent Link Adapter at one end and a Channel Adapter at the other Use Permanent Link test limits Channel Adapter Remote RJ45 plug excluded from measurement Perm. Link Adapter

26 Modified Single Connector Permanent Link Using time domain techniques, it is possible to see if the RJ45 plug is bad but there is no standard to support this

27 Fiber Standards Update

28 ISO/IEC (2006) This standard us under review Two methods to measuring a Permanent Link 1 Jumper reference Accepted as the measurement with the least uncertainty 3 Jumper reference Required for legacy test equipment that has fixed input ports Users often end up with negative loss readings Negative loss readings rejected for warranty application Measurement uncertainty in the field is unacceptable Too many negative loss readings! There was a proposal to a variation of the 3 Jumper Reference but it was rejected over concerns of a change in reference planes

29 ANSI/TIA B Title: Optical Power Loss Measurements of Installed Multimode Fiber Cable Plant Localized changes being made to only require Encircled Flux for OM3/OM4 installations Was talk of bring back Coupled Power Ratio (Rejected) Coupled Power Ratio has a measurement uncertainty of 40% Use of expensive and difficult to get hold of launch conditioners driving this change Still recommends Encircled Flux compliant sources for other multimode fibre types

30 ANSI/TIA B No option to use a Vertical Cavity Surface Emitting Laser (VCSEL) as a source, previous A did allow them VCSELs have a spectral width 0.47 nm Measurement uncertainty associated with VCSELs is unacceptable with todays low loss components Cabling vendors rejecting these results too

31 ANSI/TIA C (In progress) ANSI/TIA B = Adoption of IEC ANSI/TIA C = Adaptation of IEC ANSI/TIA C Make 50 µm 850 nm normative for Encircled Flux All other cases informative for Encircled Flux Proposal for HOML for informative cases accepted HOML = Higher Order Modal Loss Experiments by FNET proved concept Produces a field verification method that anyone could do Not a substitute for Encircled Flux 20% confidence factor

32 HOML Procedure Dual wavelength light source 50 µm or 62.5 µm TRC P 0 STEP 1: HOML with TRC, 50 µm or 62.5 µm versions Dual wavelength light source 50 µm or 62.5 µm TRC 5 Turn, 17 mm/22 mm P 1 STEP 2: HOML with simple 5-turn mandrel, 50 µm or 62.5 µm versions HOML is the difference in power comparing 0 turns with 5 turns (P 0 -P 1 ) 32

33 TIA-TSB-4979 Title: Practical Considerations for Implementation of Multimode Launch Conditions in the Field Gives two options on being Encircled Flux compliant Use a launch conditioner User a test cord tuned to an Encircled Flux compliant source Encircled Flux compliant sources reduce measurement uncertainty due to launch conditions from 40%* to 10% * Some experts are suggesting this is closer to 60% based on their testing

34 TIA-TSB-4979 Method 1: Launch Conditioner Makes any LED source Encircled Flux compliant User complain of bulkiness Cost concerns Can be difficult to get hold of

35 TIA-TSB-4979 Method 2: Matched source and test reference cord EF EF Tester vendor specific More elegant Significantly cheaper

36 Is EF needed? Yes Data Centres are multi-connector environments The use of low loss components is required If the consultant or cabling vendor specifies a loss budget (limit) based on these numbers, you may not achieve it with a non EF solution

37 Available online BICSI News

38 Questions? Thank you for your time