Eisspeedway

Talk:Ethernet over twisted pair

Removed Redirection Notice

I have removed the label at the top that states "RJ45" redirects here. For the generic 8P8C modular connector, see 8P8C. For the Registered Jack (RJ) wiring standard, see Registered jack.

Reason: The RJ45 article does not redirect here. CraigMatthews 01:12, 12 Dec 2009 (UTC)

Rename/move to Ethernet over twisted-pair

@Burt Harris: IMHO, that move is in error. "Twisted-pair" is a compound adjective that requires a noun, so it should be either "Ethernet over twisted pair" or "Ethernet over twisted-pair cabling". --Zac67 (talk) 07:41, 28 January 2021 (UTC)[reply]

Zac67 is correct. See Compound modifier. ~Kvng (talk) 15:11, 31 January 2021 (UTC)[reply]

Cat 7

Why it does not have the Cat7 in comparison table? — Preceding unsigned comment added by Daya Kathir (talk • contribs) 12:17, 9 May 2021 (UTC)[reply]

@Daya Kathir: Please try to use the New section feature on top to create a new topic section. Cat 7 isn't listed because there's no Ethernet-over-twisted-pair variant that requires it – it's backwards compatible with Cat 6 but not with Cat 6A. Cat 7 was intended for 10GBASE-T but requires special/exotic termination while the somewhat younger Cat 6A doesn't. Also, Cat 7 is popular in Europe and other countries but virtually non-existent in the US. --Zac67 (talk) 12:59, 9 May 2021 (UTC)[reply]

Probable Error in the table

Hello all.

I beleive that in table "Comparison of twisted-pair-based Ethernet physical transport layers (TP-PHYs)"

There is maybe a mistake: 100BASE-T1 -> Encoding = 4B3B -> I beleive it's 4B3T

Best regards Haarkon (talk) 09:07, 20 September 2024 (UTC)[reply]

IEEE 802.3 calls that "4B/3B", even if it uses ternary symbols – see clause 96.3.3. --Zac67 (talk) 12:34, 20 September 2024 (UTC)[reply]
It is a 2-step process. First 4B3B (no actual coding, just different grouping of bits, the bits themselves remain identical, no change in symbold rate, i.e. somewhat misleading), then 3B2T. Symbol rate (66.66MHz) and bandwidth (33.33MHz) are wrong in the article.
By the way, they are also wrong for 100BAse-TX. The highest fundamental frequency component is 31.25MHz, correct - however lossless reconstruction is not possible with that bandwidth as the intermediate values are affected. 62.5MHz bandwidth are required for perfect reconstruction (which is normally what one would call bandwidth). That's why basically everywhere else (except for those copying from Wikipedia) it's correctly stated that the automotive Ethernet variant 100Base-T1 roughly cuts the required bandwidth in half compared to 100BAse-TX.
Source: e.g. Automotive Ethernet, Cambridge University Press, https://doi.org/10.1017/9781108895248 VehicleEngineer (talk) 13:06, 24 November 2024 (UTC)[reply]
Spurgeon is cited for the whole table but I doubt all this information is available in that source. The table probably was likely expanded beyond what's available in the ref.
I see 66 MHz mentioned as a cable requirement but don't see 33.33 MHz or 66.66 MHz anywhere.
Looks like the table uses a special definition for bandwidth, see note C. ~Kvng (talk) 03:54, 25 November 2024 (UTC)[reply]
3B2T encoding cuts the symbol rate to 2/3 -> that's where the 66.66MHz (and bandwidth 33.33MHz) I mention is coming from.
The table states 75MHz (probably due to 4B3B, but as I mentioned that's misleading), which is simply wrong.
The definition of bandwidth in note C is quite unique (and I don't think it's very helpful): While this is the limit to the main peak in the spectrum, there are further frequency components beyond it, and it is not possible to recover the signal with a cutoff at that frequency (there's a reason why the cables are rated for 100MHz). 193.196.67.18 (talk) 09:04, 25 November 2024 (UTC)[reply]
Do you have access to the Automotive Ethernet source you're suggesting we use to do these updates? ~Kvng (talk) 15:59, 25 November 2024 (UTC)[reply]
Yes. Not sure how to provide it though (it's watermarked for my company). Google does find a copy of the second edition though - Table 4.10, page 149. 2A02:3102:C260:3020:644F:A727:CA5:1D11 (talk) 09:19, 26 November 2024 (UTC)[reply]
It is perfectly fine to cite non-free content as sources. It will be more difficult for readers and other editors to verify but go ahead and make the changes citing that source and I will assume you've done it right. ~Kvng (talk) 13:55, 26 November 2024 (UTC)[reply]

Note #23 leads to a deadlink

"New 802.3bw Ethernet Auto Standard Leaves LVDS Cables in the Dust". 8 April 2016. 147.161.185.105 (talk) 13:39, 21 October 2024 (UTC)[reply]

Quelle: https://en.wikipedia.org/wiki/Talk:Ethernet_over_twisted_pair