ETHERNET PROTOCOL

Ethernet Protocol

Article by:
Date Published:
Last Modified:
WarningThis page is in notes format, and may not be of the same quality as other pages on this site.

1. Overview

2. History

Originally, ethernet used a single coax cable (10base2).

3. Components

3.1. Connector

RJ45 jack. Connectors which integrate the magnetics are called integrated connector modules (ICMs).

3.2. Magnetics

Magnetics are part of the Ethernet specification for all xBASE-T networks (which includes pretty much every Ethernet PHY standard except for the first StarLAN-1, StarLAN-10 and LattisNet).

There are two options for magnetics, either:

  1. Buy a ethernet magnetics component for your PCB.

  2. Buy a jack (connector) which already has the magnetics integrated into it. Connectors which integrate the magnetics are called integrated connector modules (ICMs).

3.3. Physical Layer (PHY)

The PHY is the circuitry which drives the magnetics.

3.4. MAC

4. Standards

4.1. 10BASE-T

Used Cat 3 cable.

4.2. 10BASE-T1L

Designed for automotive and IoT applications.

4.3. 10BASE-T1S

Designed for automotive and IoT applications. Allows for a multidrop architecture.

5. Ethernet Jacks

Some ethernet jacks come with the magnetics already in-built, saving you the trouble of including them yourself on the mounted-to PCB

6. Power Over Ethernet (PoE)

Power over Ethernet (PoE) is the term used to describe ethernet cables/systems that also provide a power source over the same cable as the data. Typically 44-57VDC (48V is a common choice) is passed over one or more of the twisted pairs in most ethernet cables. Data and power can be transmitted along the same twisted pairs as long as proper transformer-based extraction (center-tapped isolation transformers) of the AC data signal and DC power is used. The early PoE standards could provide around 10W of power, but the newer standards allow up to approx. 75W.

Power sourcing equipment (PSE) are the devices that provide power onto the ethernet bus. The powered device (PD) is the device on the other end of the cable which is consuming the power. The PDs can be up to 100m away from the PSE (as defined by the existing Ethernet standard, this is unchanged with the introduction PoE).

A PoE injector is a stand-alone device used to add ("inject") DC power into an existing Ethernet cable.

The assured power is the amount of power available to the powered device at the end of cable. This is less than the power outputted by the PSE due to the energy lost in the impedance of the cabling.

NameIEEE standardYearPower to PDMax. power per portEnergized pairsSupported devices

PoE

IEEE 802.3af (Type 1)

2003

12.95 W

15.4 W

2-pair

VoIP phones, WAP

PoE+

IEEE 802.3at (Type 2)

2009

25.5 W

30 W

2-pair

PoE++ (UPoE)

IEEE 802.3bt (Type 3)

51 W

60 W

4-pair

PoE++

IEEE 802.3bt (Type 4)

71.3 W

100 W

4-pair

Tilt/pan/zoom (PTZ) security cameras.

PoE++ is also known as 4PPoE. UPoE is an acronym for Ultra Power over Ethernet.

ltc4267 poe regulator typical application
Figure 1. Typical application schematic for the Analog Devices LTC4267 PoE regulator IC. Image retrieved 2021-08-26 from https://www.analog.com/media/en/technical-documentation/data-sheets/4267fc.pdf.

Signature resistance of 25kR.

7. EtherCAT

EtherCAT is a communication protocol with it’s physical layer following the Ethernet standard in IEEE 802.3 (the magnetics and PHY are the same). However, the similarity somewhat stops there, it uses a novel technique where:

  1. Each device is daisy chained to the next via Ethernet cable.

  2. The master emits a Ethernet frame (telegram).

  3. Each slave device reads the data addressed to it "on the fly", and inserts it’s own data into the frame "on the fly".

  4. Propagation times between slaves are only limited by hardware propagation delays.

The master node can be implemented using any device with standard Ethernet capability, the same magnetics, PHY and MAC is used. However, whilst the slave node use the same magnetics and PHY, the MAC is replaced by a special EtherCAT Slave Controller (ESC) which can process frames on the fly.

8.1. WIZnet W5xxx Family

The WIZnet W5xxx family of serial-to-ethernet ICs is very popular in the maker community. This family of ICs is used by the Arduino Ethernet board are Arduino Ethernet Shield.

  • W5200: This chip implements the PHY, the TCP/IP stack (fully hardwired), and the 10/100 MAC Ethernet MAC, in a QFN-48 package. It uses the SPI Protocol to talk to a microcontroller. It’s power save features include power-down mode and WOL (wake on LAN). It runs of 3.3V but has 5V I/O tolerance.

  • W5500: Supports up to 8 independent sockets (i.e. 8 different connections to different ports). Contains a 10BaseT/100BaseTX PHY. Like the W5200 is runs of 3.3V but has 5V I/O tolerance.

    wiznet w5500 ic photo
    Figure 2. Image of the WIZnet W5500 serial-to-ethernet IC. Image from https://www.wiznet.io/product-item/w5500/.

8.2. Microchip LAN867x Family

Microchip’s LAN8670, LAN8671 and LAN8672 are ethernet PHY ICs that use 10BASE-T1S, allowing a multidrop architecture of "at least 8 nodes and a minimum 25m of length".


Authors

Twitter

Related Content:

Tags:

comments powered by Disqus