What is the real limit for a twisted-pair network segment?

Question

Computer networks

It is written in textbooks - no more than 100 m. But any technically competent person understands that this cannot be - it works on 100 meters, but does not work on 101. The decline in performance will certainly be smooth. What parameters of the transmission line affect this? Weakening of the signal strength in the line (to such a state that the receiver ceases to recognize it)? Rate of failures and packet loss? Or something else?

It is best to have a graph that shows how depends on the length of the segment - what? Transmission speed? Failure? Such a schedule would help in cases where 100 meters is not enough, and you have to (albeit slightly) exceed this limit. The need for this arises quite often.

Answer 1

The standard is about the fact that, subject to the specified conditions, the network will work with such and such parameters, manufacturers check and strive for the specified parameters to be issued, everything above is a lottery that depends on a bunch of reasons, including the type of equipment at the ends, the actual condition and composition of the cable, and the main thing is that it is stretched in parallel with this cable (for example, the presence of power lines or other cables nearby)
. 110).
The most annoying thing is that if the network twisted in the bay in your room rises at great lengths, then this will not necessarily work after installing the same line with the same equipment.
What do you mean it won't start? - the desired mode of the 1000/100Mbps network card will simply not turn on, this happens abruptly, it may turn on at the maximum dimensions of a few meters, but there will be errors on the line.
ps it seems that optics gives greater (convenient) predictability, it has no interference from the side

Answer 2

It was all discussed at NAG.
On a good twisted pair cable, with good connectors, with specially selected cards, the distance is up to 180 meters at 10 megabits.

Answer 3

open the tutorial and read how ethernet works. How exactly is the packet transmitted, what is the preamble, etc. And much will become clear.

Answer 4

100 m (300 ft) is the range you can use when designing while still referring to the standard. Works 100BaseT and at 200 meters if the conductors in the cable are 23AWG and above, copper, with uniform insulation. Works lossless. But you will not prove this until you mount the projected line, which means that your project is simply obliged to wrap up any decent examination. Eastern manufacturers often use tricks for their consumer equipment (especially for video surveillance):
1. Declare their own improved version of Ethernet;
2. Force port to 10BaseTX;
3. Apply PoE midspan 56 VDC, put this voltage on the port as a priority
...
Profit!
Your link will start at 160 m and you can indicate in the project the use of a special Chinese technology for confident passing the examination.

Answer 5

Depends on what devices are connected to the cable, the quality of its insulation, interference.
The standard defines what is the minimum maximum cable length that can be used without experiencing problems.
And Ethernet is a digital protocol, so if there's too much noise, it will turn off, completely.

Answer 6

Personally, I have extended working links with a length of more than 150 mb. But there are details here.
1. It all depends on the cable category!
2. In case of losses, you can always try to run the interfaces over half duplex.
It doesn’t sound strange, but it even matters how exactly the cable will be laid ... That is, either it will be stretched, or it will go in a bunch of zigzags, be twisted in places.

Answer 7

There was such a Soviet military cable called P-296. At the dawn of the Internet, they made a link on it to the entire bay - 500 meters at 10 megabits.