10Gbits/s tested at 91m on cat5e

https://www.youtube.com/watch?v=zACK7Aka4Us

Probaby not the most scientific test but for an average Dave test, 5e does pretty well.

Suppose in the same way that you could take a Porsche 911 GT3 RS and take it down to Asda Tyres and put some best of MiC cheapies on it before heading off on a 300+ kph sojourn on the nearest racetrack or autobahn.

...YMMV 🤣

[I didn't watch it.]

I didn't watch it either, as there's no point.

* Much cable sold as "Cat5" or "Cat5e" actually meets higher standards

* The standards allow for a certain amount of margin for successful operation (allowing for induced noise, termination losses, manufacturer variations etc).

Therefore it may well work in one environment and at one point in time, but at much reduced margin. On a different day, or in an environment with more noise, it might fail or give an unacceptable bit error rate.

If you want to install a 10G copper link over 90m and have confidence that it will work *reliably* and *consistently*, you use cable certified to Cat6A. End of story.

In reply to a post by Taras:

Probaby not the most scientific test but for an average Dave test, 5e does pretty well.

Shows that if you just need it between a couple of Mac Mini's in your home lab, its likely you will "get away" with 10 GigE on much older cables.

If you're spending on 10 GigE capable hardware and switching and scrimping on cabling then something else is wrong with you

If you're spending a couple of grand on a couple of Mac Mini's with 10G Ethernet ports and you can't stretch to some half decent network cables....then I'd also say something else is wrong 😎

Then again that's pretty lame/boring/boo hiss, just wouldn't get the faintest eyeball or clicks these days: hey ma watch me get 400G on some old plastic tube I found in the garage...🤣

Edited by Pheasant (Sat 21-Dec-24 17:18:20)

In reply to a post by Pheasant:

Then again that's pretty lame/boring/boo hiss, just wouldn't get the faintest eyeball or clicks these days: hey ma watch me get 400G on some old plastic tube I found in the garage...🤣

Ho Ho Ho. 🎅🎄🎄

Edited by jchamier (Sat 21-Dec-24 17:32:16)

Nice segue into the season of cheer and merriment 😅 Here's a merry Xmas to you too JC 🎅🎄🎅

I assume that the packet loss is off the chart despite negotiating a 10Gb link, hence the speeds dropping to half where they want to be.

Would have been more interesting to take 100m of Cat5e and chopping 5m off it until it certifed.

Edited by jpm (Sat 21-Dec-24 18:23:24)

In reply to a post by jpm:

I assume that the packet loss is off the chart despite negotiating a 10Gb link, hence the speeds dropping to half where they want to be.

Would have been more interesting to take 100m of Cat5e and chopping 5m off it until it certifed.

Potentially as both sides support NBASE-T it's using downshifting to go to 5Gb as 10Gb wasn't viable. The straight 10GBASE-T spec doesn't allow this, the NBASE-T spec that provides the 2.5Gb and 5Gb speeds in between does. If a port only has 1Gb and 10Gb it's running 10GBASE-T, if it has the other two it's NBASE-T. Downshifting will lower rate depending on cable quality.

Then there's Andrews & Arnold's ADSL over wet string...
https://www.revk.uk/2017/12/its-official-adsl-works-...

I guess that's what OR use the blue rope for

Cheers!

Very informative, two commonly asked and/or posted questions on here are, "I've got cat5e cables should I upgade them for my new FTTP connection", and "Do I need ca6 cable for my new FTTP connection", This shows that these cables are not working on the limits, but easily cope.

I agree with the other posters, theres huge flaws in the testing. But i posted it for exactly the reasons you said stated

In reply to a post by XGS_Is_On:

In reply to a post by jpm:

I assume that the packet loss is off the chart despite negotiating a 10Gb link, hence the speeds dropping to half where they want to be.

Would have been more interesting to take 100m of Cat5e and chopping 5m off it until it certifed.

Potentially as both sides support NBASE-T it's using downshifting to go to 5Gb as 10Gb wasn't viable. The straight 10GBASE-T spec doesn't allow this, the NBASE-T spec that provides the 2.5Gb and 5Gb speeds in between does. If a port only has 1Gb and 10Gb it's running 10GBASE-T, if it has the other two it's NBASE-T. Downshifting will lower rate depending on cable quality.

This makes the most sense to me. In my home I mostly have all Cat6 cables now which happily run 10Gb without issues but there's still one place that has an old Cat5 cable. It won't run 10Gb (I've tried) but I bought a 2.5Gb switch recently and it's happily running at those speeds. I have no idea if it'd also run at 5Gb speeds because I don't have a switch that supports this it's either 10Gb negotiates correctly or it goes to 1Gb at least that's my understanding of things.

Edited by aidanh (Sun 22-Dec-24 04:43:19)

2.5GbaseT and 5GbaseT are designed to work at 100m on Cat5e. That's the main reason these standards exist in the first place.

Cat5 was only designed for 100Mbps at 100m, but your "Cat5" cable may well have been manufactured to Cat5e spec, and in any case is likely not anywhere near 100m.

In reply to a post by candlerb:

2.5GbaseT and 5GbaseT are designed to work at 100m on Cat5e. That's the main reason these standards exist in the first place.

Wrong. 5GbaseT is specified for 55m over Cat5e and 100m over Cat6.

10Gbps is specified to run over Cat6 as well but with limitations.

The biggest issue is a single Cat5e cable may well support 10Gbps, but putting them in a bundle, wiring them into a patch panel, and expecting them to all work at 10Gbps is an unrealistic expectation. If that is what you want then you are going to need to break out the Cat6a which is designed to work in those sorts of scenarios.

The biggest issue at the moment is the IET is busy incorporating Ethernet into the 18th edition if it might be used for PoE but the point of PoE is that any Ethernet cable can be used for PoE so all Ethernet wiring is now going to come under Wiring Regulations!!!

In reply to a post by jabuzzard:

In reply to a post by candlerb:

2.5GbaseT and 5GbaseT are designed to work at 100m on Cat5e. That's the main reason these standards exist in the first place.

Wrong. 5GbaseT is specified for 55m over Cat5e and 100m over Cat6.

Y'know, that's what I thought too, but someone who ought to know these things pointed me to Wikipedia which says it's 100m over Cat5e.

https://en.wikipedia.org/wiki/2.5GBASE-T_and_5GBASE-T

"2.5GBASE-T and 5GBASE-T can be deployed at a cable length of up to 100 m on Cat 5e or better cables.[6][7]"

But of course, Wikipedia isn't always right.

I thought this would be a good test for perplexity.ai. It tells me that "For 5GBASE-T, Cat5e performance is not officially specified and may vary". One of the references is a Reddit post:
https://www.reddit.com/r/networking/comments/12xq09k...

where one of the commenters says:

"Let's use IEEE 802.3bz and not Cisco's own made up mGig.

2.5G is supported at 100M on CAT5e, but is unspecified at 5G."

If you can find an official standards doc that says 5G definitely works at 55m over Cat5e, that would be great - and you could edit wikipedia too

Of course, you have to pay to access the IEEE standards documents. But I found this, publically available, from 2016 when 802.3bz came out: https://www.ieee802.org/3/bz/public/mar16/diminico_3...

Define a 5 Gb/s PHY for operation over
‒ Up to at least 100m on Class E (Cat6) balanced copper cabling on defined (2) use cases and deployment configurations
‒ Up to 100m on Class D (Cat5e) balanced copper cabling on defined use cases and deployment configurations

In slide 15, it shows two out of three failing test cases for 5G on Cat5e for 90m and 60m bundled cabling.

EDIT: I think I found it. Full 802.3 (7000+ pages!) available from link at footnote 6 of wikipedia page.

126.7.1 Cabling system characteristics
...
b) 5GBASE-T is an ISO/IEC 11801-2002 Class D and ANSI/TIA-568-C.2 Category 5e application, with additional installation requirements and transmission parameters specified in this clause, including extended frequency performance beyond that specified for Class D and Category 5e.
...
A link segment consisting of up to 100 m of Category 6A/Class EA or better will meet the transmission parameters of 126.7 and provide a reliable medium for 5GBASE-T without further qualification.

As I understand it: 5GbaseT signals at 200MHz. Cat5e standard specifies 100MHz bandwidth, but many "Cat5e" installations utilize 200MHz or even 350MHz cable.

Therefore, minimally-compliant Cat5e won't support 5G over 100m, whereas Cat6 supports it without question; but in practice Cat5e might work.

I've not been able to find any reference to the 55m figure, except for the test results slide above which shows it just meeting SNR tolerance at 60m with unbundled patch cords at each end.

Edited by candlerb (Tue 24-Dec-24 09:47:41)

TIA TSB-5021 has guidelines and recommendations but is not a standard and I'm not going to buy it. The figure Mr Buzzard gave will potentially be from hardware vendors.

https://www.arista.com/assets/data/pdf/Cabling-optio... has a little information in it. It's the crosstalk that's the killer.

I came across this graphic a few years ago. I believe it was done around the time 802.3bz was being ratified. I think it neatly captures the spectral bandwidth capabilities of the various Categories of data cabling vs what is required for the various Ethernet standards...

Ethernet - Spectral bandwidths on Cat 3, 5e, 6 and 6a Cabling

Some good points above about cross-talk - in particular what led to the relatively rapid change from Category 5 to Category 5e in the late noughties was the realisation that Alien Crosstalk, particularly Near End Crosstalk (NEXT) and Far End Crosstalk (FEXT) on the channel was crucially significant, especially in a commercial installation where arm-thickness bunches of cabling are par for the course.

Its worth noting that the bandwidth figures being discussed, 100 MHz for Cat5e, 250 MHz for Cat6 and 500 MHz for Cat6a are for the full length of a certified cabling "channel" - a channel in structured cabling is 90 metres of fixed, solid core cabling which terminates on industry standard IDC blocks typically RJ45 (although historically it was the 110 blocks) + 10 metres of "cordage", with cordage being stranded conductor cabling uses for patching and at the work area. Stranded conductor cabling is inferior to solid conductor cabling in term of its data throughput capability, which is why in the TIA/EIA and ISO standards it is limited in length, for certification purposes.

In practical terms most cabling installed is rarely at the full channel length. The average on most installs is typically around the 40 to 50 metre mark usually. Thus the overall bandwidth-distance product (spectral bandwidth x length of the permanent cabling0 is more favourable at shorter cabling lengths - and such cabling can typically sustain higher actual throughputs than at the full certified channel length of 100m.

There are a couple of interesting articles online from Fluke Networks which talk in some more detail about the particulars on implementations of 802.3bz for 2.5GbaseT and 5GBaseT on pre-existing Cat5e and Cat 6 certified cabling installations. I'll see if I can find them.

In reply to a post by candlerb:

I've not been able to find any reference to the 55m figure, except for the test results slide above which shows it just meeting SNR tolerance at 60m with unbundled patch cords at each end.

Early vendors of 10GBaseT PHY's used to quote around the 30 metre mark for Cat6 and as their PHY's improved over the years these figures crept up. I've seen figures of 55-60m on Cat6 for 10GBaseT.

It's definitely a vendor (PHY) based figure though. There is nothing defined in any IEEE, ISO/IEC or TIA/EIA standard to this effect.

Edit: I was wrong!!

The reference to 10GBaseT @ 55 metres on Cat6 is apparently found in Clause 55 (Link Segment Specifications) in the 802.3an spec.

See page 24 in the attached, "Why 55 Meters on Class E UTP?"

https://www.ethernetalliance.org/wp-content/uploads/...

Edited by Pheasant (Tue 24-Dec-24 16:10:36)

Found it! Here is the article from Fluke Networks.

Will My Existing Cable Plant Support 2.5/5GBASE-T? from January 2018.

N-BASET Performance and Cabling Guidelines (Ethernet Alliance / N-BASET Alliance, August 2016)

Final point, and then I'll feck off and go and drink some mulled wine and eat some mince pies.

There is a huge amount of variation in not only the quality of the cabling (and connectors) on structured cabling out there. There is also a vast chasm in how well these are installed.

In a commercial installation environment - these factors are de-risked by the following:

- Having correctly trained structured cabling designers and installation engineers that install cabling systems using reputable well known branded products that all match, the cabling, the connectors etc. In most cases these are inspected by the cabling manufacturer and warrantied for a period of 20 or 25 years.

- Following manufacturers guidelines for installation and adhering to industry standard like here in Europe/UK its ISO/IEC 11801 (the is. where cabling "Classes" come from) or in the states its ANSI/EIA-TIA 568 family of standard (this is where cabling "Categories" come from)

- Performance testing each and every permanent link (or sometimes channel), actually meets ClassE/Cat6A, Class D/Cat5e etc etc using a certified and correctly calibrated cabling tester like a Fluke Networks DSX-series cable analyser

In reply to a post by burble:

Very informative, two commonly asked and/or posted questions on here are, "I've got cat5e cables should I upgade them for my new FTTP connection", and "Do I need ca6 cable for my new FTTP connection", This shows that these cables are not working on the limits, but easily cope.

Not necessarily at all. It's extremely variable and not at all compliant to *any* standards. So if you're lucky enough to get a link to stay up, its doesn't mean every link will.

This is where these "wow look" videos sucker people.

Stick with the standards. That's what there for. Not this amateur hour garbage.

In reply to a post by XGS_Is_On:

In reply to a post by jpm:

I assume that the packet loss is off the chart despite negotiating a 10Gb link, hence the speeds dropping to half where they want to be.

Would have been more interesting to take 100m of Cat5e and chopping 5m off it until it certifed.

Potentially as both sides support NBASE-T it's using downshifting to go to 5Gb as 10Gb wasn't viable. The straight 10GBASE-T spec doesn't allow this, the NBASE-T spec that provides the 2.5Gb and 5Gb speeds in between does. If a port only has 1Gb and 10Gb it's running 10GBASE-T, if it has the other two it's NBASE-T. Downshifting will lower rate depending on cable quality.

Indeed. The original NBASE-T Downshift protocol is described here:

https://archive.nbaset.ethernetalliance.org/wp-conte...

In reply to a post by Pheasant:

In reply to a post by burble:

Very informative, two commonly asked and/or posted questions on here are, "I've got cat5e cables should I upgade them for my new FTTP connection", and "Do I need ca6 cable for my new FTTP connection", This shows that these cables are not working on the limits, but easily cope.

Not necessarily at all. It's extremely variable and not at all compliant to *any* standards. So if you're lucky enough to get a link to stay up, its doesn't mean every link will.

This is where these "wow look" videos sucker people.

Stick with the standards. That's what there for. Not this amateur hour garbage.

FYI, 'these' in this case where the cables as tested.
As for the testing done in this video, I refer to your earlier post, to test the cables the equipment used was beyond the specs for cat5e, to test them with inferior equipment would be pointless, same as with tyres, to test a car tyre with a car incapable of challenging them would be pointless, you use a vehicle which can test the tyres and not the car.
I did jump through much of the video, so forgive me if I missed any part of it where it was suggested the cables should be used for anyone wanting 10G.