|
|
Going to take inspo from @jabuzzard's posting (reproduced below) in the Cheap-n-cheerfiul 10 Gigabit Ethernet switch thread, and spin this into a topic / thread all of it's own as it probably deserves separate discussion:
10Mbps Ethernet is not going anywhere If anything, it is going to become even more commonplace than it is currently as the march of Ethernet as the universal communications bus continues with the X Base-T1 standards. So for example 802.3cg-2019 (aka 10BaseT1L) which allows you to do 10Mbps full duplex over a single twisted pair out to 1000m with PoDL providing up to 50W of power.
On a practical point 100Mbps is more than enough for any streaming device, VOIP or camera, and for VOIP and streaming audio then 10Mbps is more than adequate. The bit rate on a CD is only 1.5Mbps.
The point is I could offer a 10Base-T1L video camera with power delivery that can send a H264 1080p picture down 1000m of single twisted pair. Why would I be hanging my head in shame? I can imagine that there is a sizable market for this sort of product. Not having intermediate network switches in my CCTV system is a security win in my books. Would be interesting if a standard port Ethernet port could auto negotiate to 10Base-T1L
Even a video streaming device does not need a 1Gbps connection, with 100Mbps being more than adequate. Typically a 4k h.264 stream only needs about 35Mbps or about 50Mbps if it is HDR. You only need 1Gbps if you have an 8k stream but very few people are doing 8k streams.
So here we go....
|
|
|
What is Single Pair Ethernet.
SINGLE PAIR ETHERNET - STANDARDIZATION LANDSCAPE
"With Single Pair Ethernet, a new physical layer is entering the world of Ethernet communication. Born out of the automotive industry, SPE is also developing into a key technology for the coming steps of digitization in industrial automation. Ethernet IP can finally be transmitted continuously from the cloud to the field level via just one pair of copper wires."
IEEE 802.3 ETHERNET PROTOCOLS
This diagram shows the published SPE protocols with their associated possible transmission lengths and essential details:
- 10 Mbps. 1000 metres point-to-point. IEEE 802.3cg 10Base-T1
- 100 Mbps. 40/15 metres point-to-point. IEEE 802.3bw 100Base-T1
- 1000 Mbps. 40/15 metres point-to-point. IEEE 802.3bp 1000Base-T1
- 2.5/5/10 Gbps. 15 metres point-to-point. IEEE 802.3ch MultiGigBase-T1
Edited by Pheasant (Sat 23-Nov-24 21:57:10)
|
|
|
...and to that:
The 802.3dg working group will close the first gap in the transmission length with 100BASE-T1L, a new protocol for 100 Mbit/s over at least 500 m transmission length. As of January 2024, final results are expected by the end of 2024 and the location should be published in mid-2025. Information on the IEEE802.3 projects can be found here:
www.ieee802.org/3/index.html.
|
|
Register (or login) on our website and you will not see this ad.
|
|
|
|
|
|
|
...and to that:
The 802.3dg working group will close the first gap in the transmission length with 100BASE-T1L, a new protocol for 100 Mbit/s over at least 500 m transmission length. As of January 2024, final results are expected by the end of 2024 and the location should be published in mid-2025. Information on the IEEE802.3 projects can be found here:
www.ieee802.org/3/index.html.
That would be insanely useful for remote displays and other remote ethernet toys!
also i may have read that as 100Gbit/s initially 🙈
|
|
|
I hadn't spotted that standard. 100Mbps out to 500m over single pair with 50W PoDL is even bigger than 10Mbps out to 1000m. Even in a domestic market 100Mbps is ample for any streaming for some considerable time to come and a single twisted pair cable can be a lot more discrete for running around a room tucked in under the edge of carpet etc.
It's the march of all conquering Ethernet. Back in the day there was Token-Ring, FDDI, ATM, X.25, Frame Relay, ARCnet etc. all now confined to the dustbin of history. It's now coming after CAN bus and RS485
I wonder if they can squeeze 1Gbps out to say 100m over a single pair. That would make wiring up WiFi access points to provide gold standard WiFi coverage in a home a lot easier with that much thinner cable.
|
|
|
Yes there's an important factor about needing shielded twisted pair to go beyond the 15 metre runs which people should consider if they were hoping to re-use any old existing wiring or in an industrial application.
For that distance makes sense why it would not be a limitation in automotive use though.
However for folk hoping for a new way to place 1 to 4 devices at the end of an entire standard boxed reel (305m) of UTP they may be disappointed.
prlzx on Zen: FTTC (VDSL) at ~40Mbps / 10Mbps
with IP4/6 (no v6? - not true Internet)
Edited by prlzx (Sun 24-Nov-24 15:13:33)
|
|
|
Yes, perhaps I should have highlighted that in my OPs.
From what I can see SPE is being driven commercially by a conglomerate of continental industrial cabling companies. In Europe shielded twisted pair is the de facto commercial structured cabling standard, especially in Germany and more especially for industrial cabling settings where SPE is looking like a good fit.
This is counter to what “we” generally install here in the UK, Australia and the USA which is predominantly UTP cabling, ultimately derived from telephone cabling which morphed from Cat 3 onwards / and driven by AT&T then Lucent ultimately once they were spun out of the Baby Bell motherships in the 80’s and happened to coincide with the march of Ethernet from coax to twisted pair. The rest as they as is history.
Edited by Pheasant (Sun 24-Nov-24 17:04:41)
|
|
|
|
SPE is primarily aimed at the Industrial Internet of Things to replace a variety of legacy industrial communication protocols over existing cabling as that is typically the most expensive item to replace when retrofitting upgrades.
Other uses are short range (a few tens of metres at most) high speed data links in automotive and industrial robotics as cabling is simplifed, takes up less space and weighs less.
It was never intended to provide long-distance ethernet connectivity over phone-grade cabling, and i wouldn't expect the demand to be sufficient for standards and chips to be developed. Where 75 ohm coax exists for CCTV or TV distribution there are a number of ethernet-over-coax products, and for 50 ohm coax there is always the original 10Base2 (upto 185m) and 10Base5 (upto 500m) standards.
|
|
|
|
Whilst that is true, and I agree that it was likely "vanquishment" of vehicle-centric CAN bus and industrial RS422/485/BACnet/Modbus etc etc. was probably a key reasons for SPE establishment - I think like many accidental or shall we say 'unintended consequences' that seem to be apparent in computing and networking - we will find things like SPE gradually and sometimes surprisingly creeping into other areas that the original designers probably didn't quite or never foresaw...
Perhaps if SPE gets a proper toe hold over time - some off the skew towards the STP/shielded-centric cabling distances, especially at the more modest throughputs, will be done away with.
I suppose never say never with coax. When we were ripping up trading floors thick with IBM Type 1 in the early nineties to replace with new-fangled (unshielded) Category 5 and then ooooh-ahh 5e in the nineties, we never saw that Category 7, 8 etc would bring be the return of shielded cabling. Nothing new under the sun!
|