CAT8 - Frequently Asked Questions

Yes and no.

There are different versions of “CAT8”, which are at different stages of ratification.

The US American version of CAT8 has been ratified in autumn 2016 as ANSI/TIA-568-C.2-1.

The International Standard ISO 11801 3rd edition, containing 2 “CAT8 versions” will be ratified around autumn 2017.

ISO 11801 3rd edition will contain two different “CAT8” performance classes: Class I for cabling with RJ45 connectors, based on Class EA and mostly identical with the American ANSI/TIA Cat8 and Class II which is based on cabling with non-RJ45 connectors and extends the Class FA to 2000 MHz.

Both Class I and II will support Ethernet speeds up 40GBase-T.

In addition to the above, there is a TR ISO11801-9901 since 2014 that describes cabling based on “non-RJ45” systems in support for 40GBase-T. That early version of a first standard for cabling for 40GBase-T will be superseded by the 3rd edition of ISO11801 in 2017.

You also have to differentiate between cabling standards and measurement standards for cabling. The measurements for cabling are defined in a separate set of standards.

The TIA standard 1152-A was ratified in autumn 2016. It specifies the parameters and accuracies required to certify structured cabling based on the ANSI/TIA performance categories up to Cat8.
The equivalent IEC standard 61935-1 5th edition will be ratified in autumn 2017. It specifies the parameters and accuracies required to certify structured cabling based on the ISO/IEC performance classes up to Class II. It also differentiates accuracy limits for measurements up to 2000 MHz between Class I and Class II.

Both standards define how to measure cabling built from components CAT5E up to CAT8.2 and which accuracies need to be met up to 2GHz frequency range.

Currently the WireXpert is 3rd party verified by ETL up to 1GHz Level V based on the currently existing IEC61935-1 ed. 4. As soon as the 5th edition of the IEC standard 61935-1, containing accuracies up to 2GHz, will be ratified, the WireXpert Series will also be 3rd party verified up to 2GHz by ETL. That will guarantee proper accuracies for meeting the requirements from ANSI/TIA and ISO/IEC for measurements up to 2GHz.

Category 8 (ANSI/TIA) / 8.1 (ISO/IEC) connectors with RJ45 interface and Category 8.2 (ISO/IEC only) connectors with “non-RJ45” interface such as TERA, ARJ45, GG45, and MMCPRO can be used for building Structured Cabling up to 2000 MHz bandwidth. Supporting applications like 25/40GBASE-T (IEEE 802.3bq.

As of (April 2017) the WireXpert series supports TERA, ARJ45, GG45, and MMCPRO (testing on ARJ45 and GG45 is done with adaptors for ARJ45) due to availability of components (and cables).

Yes, it is planned to support testing Structured Cabling based on CAT8 (ANSI/TIA) / CAT8.1 (ISO/IEC) as well. Currently no such RJ45 components are commercially available.

Resistance unbalance tests are optional tests per definition by the measurement standards (ANSI/TIA 1152-A and ISO/IEC 11801 3rd ed.) and will at some point be interesting for high power PoE applications up to 100W. At current, the IEEE organisation is in the process to define 100W PoE (IEEE 802.3bt) and the international standards organisation is working the technical report TR29125 for remote powering to support above IEEE 802.3bt. Once those standards become relevant in the field, field testing of resistance unbalance will make sense in applications where high power PoE up to 100W will be used.

It is planned that the WireXpert Series will support resistance unbalance tests.

Unfortunately, there is no IEC standard that addresses the quality of a shield other than the laboratory test for coupling attenuation. There is no IEC field testing method to test the quality of a shield in the field. There is a very basic method described in TIA to test “shield integrity”, however that method does not include any clear descriptions of what to measure nor any accuracies. In fact, the outcome of that measurement is a simple “pass” or “fail” based on the existence of a shield conductor. That can be even produced by a simple wiremapper.

A real shield integrity test according to IEC can only be done by measuring coupling attenuation of a cabling system.

Yes. At current (April 2017) The WireXpert 500 is the only 500MHz CAT6A/Class EA tester that can be upgraded to certify CAT8 and Class I & II Structured Cabling. To ensure accurate measurements, the unit needs to be returned to an authorized Softing service center for upgrade. Softing will perform the necessary upgrade steps and will also factory calibrate the unit up to 2500MHz.

In fact, this makes the WireXpert 500 the most future-proof cabling certifier on the market today. Users that do not need more than CAT6A (ANSI/TIA) or Class EA (ISO/IEC) today and are still unsure about the future of “CAT8”, can today buy a WireXpert 500 and maybe in future, whenever that capability is needed, upgrade the device to 2500MHz frequency range. This allows also to support upcoming plans for 2500MHz Structured Cabling to support possibly 100Gbit Ethernet over Twisted Pair copper cabling.

There are two different aspects to this question:

A1: Cabling:

Due to the length limitations of 30m of CAT8/Class I & II channels, it is mostly a data center application where the full 100m reach of traditional LAN copper cabling is not needed.

Backbone cabling and cabling for access points could also become an interesting niche application in regular enterprise buildings. Many building backbones are less than 30 meter and many access points under the room ceiling do not need long cabling lengths. It is foreseeable that speeds of access points will in future exceed 10Gbit/s and thus will need better cabling than CAT6A.

Also certain applications for Homes might be a driver for that high speed structured cabling. Since typical length in homes are in the same range as in data centers, “CAT8” might become also a future-proof cabling alternative. Especially the aspect of keeping that cabling for long or ever in a home, the highest available performance is appreciated.

A2: Ethernet speeds:

The most commonly quoted application is 40GBase-T.

There is a relatively newer move in[MK1]  IEEE to also create a 25GBase-T ethernet speed.

It turns out that 25GBase-T actually will find more supporters than 40GBase-T.

Why is that?

On one hand cost. Transceivers for 25GBase-T can be produced at lower cost than 40Gbase-T

Secondly, speed aggregation. In data centers, the next higher speed is 100Gbit/s Ethernet, typically on fiber cabling. It is simply easier to aggregate 4 x 25Gbit/s into one 100Gbit/s line that aggregating 40Gbit/s lines into 100Gbit/s lines.

Absolutely not. 40 Gbit/s strictly will require “CAT8” cabling performance up to 2000MHz. There will be a new TR ISO 11801-5 that will define migration up to 25GBit/s for installed CAT6A, CAT7 and CAT7_A cabling. However, this TR will define that the installed cabling, even CAT6A/Class E_A, will have to be measured again up to 1250MHz. The TR will also include very short length limitation for installed cabling.

It is also not allowed to extrapolate a measurement up to 2000MHz that had only been performed up to 500MHz.

Recertification means to post-process measurement data with a PC software to apply a different set of limits.

It is legitimate to recertify to a new standard if both old and new standard use the same physical measurement mode based on Permanent Link or Channel Link topology. For example re-certifying a CAT6A (ANSI/TIA) Permanent link measurement to ISO Class EA (ISO/IEC) Permanent Link limits is legitimate.

It is not legitimate to recertify a link that had been measured with a lower frequency range than the standard that should be used for recertification if the additional required data points are not available. It is also not legitimate to recertify to a standard that is based on a different measurement mode. For example, if the original measurement has been performed in Channel Link mode, it is not allowed to recertify to Permanent Link limits since the wrong measurement cables had been used and the measured data also is not sufficient to recalculate the measurements. In Channel Link mode, patch cables are used for testing and are included in the Length measurement value, too. In Permanent Link mode, the tester uses a special Permanent Link adaptors (test cords not included in Length measurement) with totally different electrical properties than a regular patch cable, therefore a Permanent Link test cannot be later on the PC derived from a Channel Link test.