Technical Tips | Leveraging Single-Pair Ethernet in Video Surveillance Applications

The continued demand for high-resolution images and videos has driven improvements and innovations in image sensors, increasing resolutions by about tenfold from 320 x 240 pixels to 4,096 x 2,160 pixels and beyond. The increase in pixel count also means more data needs to be transferred from the image sensor to the video processor via the MIPI interface. To support high-quality video transmission, the Ethernet physical layer (PHY) in IP cameras also needs to increase from 10Mbps to 1Gbps. For this reason, the ability of the video processor to run video compression algorithms becomes important, which can minimize the data transmitted over the Ethernet cable.

A 4K resolution image sensor running at 30fps produces an uncompressed video data rate of 9.56Gbps. Although MIPI is designed to support rates this high or even higher, transmitting data at this rate over Ethernet and storing uncompressed high-resolution video is not economically feasible (it requires a large storage space ). Using efficient video compression algorithms such as H.265, the data rate requirement can be reduced to less than 10Mbps even when compressed to medium quality using a 4K image sensor . While image sensor companies try to create higher resolution sensors, standards bodies such as the International Electrotechnical Commission, the International Organization for Standardization, and the International Telecommunication Union are working on video compression algorithms that will limit video data rates over Ethernet to less than 10Mbps in certain operating scenarios.

Standard Ethernet interfaces in IP cameras are limited by specification to only 100m of cable transmission distance; however, there is a new technology that can increase the minimum cable length to 1,000m. The distance from the IP camera to the network video recorder can be 1km or more, and connecting over such distances via standard Ethernet requires the use of repeaters or fiber optic cables. An alternative is to use coaxial cable (RG-59) to reach longer distances, but this requires the use of passive adapters to convert the Ethernet signal from CAT 5e to coaxial cable and vice versa. The cost of coaxial cable per 100m is often higher than the cost of standard Ethernet cable.

Recently, the Institute of Electrical and Electronics Engineers (IEEE) defined a new Ethernet standard, IEEE 802.3cg, to enable 10Mbps operation and associated power delivery over a single pair of balanced conductors. More specifically, 10BASE-T1L: IEEE 802.3 PHY specification for 10Mbps Ethernet LAN over a single pair of balanced conductors with a cable transmission distance of at least 1,000m (longer transmission distances can be achieved using 18 AWG wire for point-to-point connections). Since a single-pair cable can now support both data transmission and power delivery, adopting IEEE 802.3cg can provide significant cost savings and is easier to install in video surveillance applications.

The 10BASE-T1L PHY uses full-duplex communication on a single pair of balanced conductors with an effective data rate of 10Mbps in each direction simultaneously . The 10BASE-T1L PHY uses three-level pulse amplitude modulation (PAM3) to transmit at 7.5 megabaud on the link segment . A 33-bit scrambler helps improve electromagnetic compatibility. The MII transmit data (TXD<3:0>) is encoded using 4B3T encoding (i.e., 4 binary to 3 ternary) so that the running average (DC reference) of the transmitted PAM3 symbols does not exceed the range. The transmitter output voltage of the 10BASE-T1L PHY is set to 1.0Vpp or 2.4Vpp differential using the management data input/output interface, which helps achieve longer communication distances on different cables.

The DP83TD510E is an ultra-long reach PHY transceiver that complies with the IEEE 802.3cg 10Base-T1L specification. The device is powered by a single 3.3V supply and supports 2.4V p2p and 1V p2p voltage modes as defined by the IEEE 802.3cg 10Base-T1L specification. The PHY has a very low noise coupled receiver architecture that supports cable lengths up to 2,000m. The device provides media independent interface (MII), simplified MII, simplified Gigabit MII and RMII low power 5MHz master modes connected to the processor's MAC. DP83TD510E diagnostic tools include time domain reflectometry (TDR), active link cable diagnostics (ALCD), signal quality indicators (SQI), multiple loopback interfaces and an integrated PRBS packet generator, simplifying debugging during development and field fault condition detection.

Single-pair Ethernet (SPE) networks also support Power over Data Lines (PoDL) along the same single-pair cable through a low-pass filter in video surveillance applications, as shown in Figure 2.

Table 1 lists the various power classes supported by the IEEE 802.3cg standard. The maximum power that can be delivered to the load is 52W, defined as class 15. IEEE 802.3bu covers power classes below 10.

Classes 8 and 9 (48V regulated power devices) or 14 and 15 (50V to 58V maximum) can support the power levels required by IP cameras, which may require up to 52W of power to operate. This power is sufficient to support most camera systems, even those with built-in heaters. For buildings that need to upgrade, an intermediate solution can use a standard Ethernet to SPE converter. Figure 3 provides an example of the connection of an IP camera system.

Future IP camera products are expected to support SPE for easier installation, and network video recorders will also provide power device ports. Higher compression video through a simplified SPE network can achieve better surveillance without adding complexity and cost.

For more online technical support, please visit the TI E2E™ Chinese support forum (e2echina.ti.com).