Blog Signal Chain Basics

eUSB2 brings USB to embedded systems

As the USB ecosystem moves into smaller, lower-power process nodes, key opportunities for embedded USB2 (eUSB2) have arisen. Engineers can use eUSB2 for low-voltage interchip communication or to provide a fully compliant USB 2.0 interface by using an eUSB2 repeater.

eUSB2 supports two different modes of operation:

Native mode provides permanent onboard chip-to-chip communications with direct eUSB2 signal connections, as shown in Figure 1. It provides the benefits of lower input/output (I/O) voltage and power efficiency implemented in the specification supplement, “Embedded USB2 (eUSB2) Physical Layer Supplement to the USB Revision 2.0,” while remaining compliant to USB 2.0 at the protocol layer. eUSB2 also supports longer printed circuit board (PCB) trace length support than many other chip-to-chip interfaces; High-Speed Interchip (HSIC) and USB 2.0 Transceiver Macrocell Interface plus Low-Pin Interface (ULPI) are typically 3 in. to 4 in. at most. Native mode is only supported as an interchip interconnect and isn’t backward-compatible to the physical USB 2.0 specification. Standard USB 2.0 hosts or devices cannot directly connect to a native eUSB2 application.

eUSB2 native mode
Figure 1. Direct eUSB2 signal connections through native mode can be used for onboard chip-to-chip communications.

Repeater mode lets chips that implement eUSB2 connect to standard USB hosts or devices through a separate eUSB2 repeater. An eUSB2 repeater provides level shifting to the 3.3 V physical interface of USB 2.0, as well as translation of single-ended eUSB2 full-speed/low-speed (FS/LS) signaling to standard USB 2.0 FS/LS signaling. Routing a host or device chip with an eUSB2 interface to a standard USB connector through an eUSB2 repeater will support USB 2.0 connections of any speed, as shown in Figure 2.

eUSB2 repeater mode
Figure 2. eUSB2 repeaters can route a host or device chip with an eUSB2 interface to a standard USB connector in order to support for USB 2.0 connections of any speed.

The USB 2.0 interface of an eUSB2 repeater meets all of the signaling, interoperability and backward-compliance requirements of a standard USB 2.0 port and can handle maximum cable lengths of 5 m. An eUSB2 repeater application provides the benefits of lower I/O voltages and power efficiency implemented by the eUSB2 standard for the host or device chip. The ability to place the eUSB2 repeater close to the USB connector eases routing and signal-integrity concerns when using larger PCBs.

eUSB2 repeaters are extremely flexible and can be implemented on hosts, devices or dual-role devices with simple repeater configuration signaling from the eUSB2 host or device.

eUSB2 design considerations
The specification supplement does not impact protocol or application layers, allowing reuse of existing USB software and drivers, but outlines physical layer changes that may impact system design and layout.

eUSB2 applications have many of the same design considerations as a typical USB 2.0 design. You should route eD+ (eUSB2 data+) and eD- (eUSB2 data -) lines as a differential pair, matching trace lengths and preferably routing over a solid ground plane. While the nominal PCB trace length for eUSB2 is 10 inches – which provides more routing flexibility – the output impedance for eUSB2 is 40 Ω with a differential termination impedance of 80 Ω, instead of the 45 Ω/90 Ω required by USB 2.0. To enable easier routing on PCBs with eUSB2 and USB 2.0 traces, the recommended trace differential impedance for an eUSB2 design (native or repeater) is 85 Ω.

While many of the latest innovations in the USB ecosystem revolve around its power delivery capabilities, the eUSB2 specification is agnostic to VBUS. A native mode application will likely have no VBUS component, but a repeater mode application would require the handling of VBUS power provider or consumer capabilities externally to the eUSB2 repeater.

Hub tiers
The USB 2.0 specification allows for up to five tiers of high-speed or FS hubs in a valid bus topology. While applications using that many tiers of hubs are rare, note that when taking into account the USB high-speed start-of-packet synchronization field truncation and end-of-packet dribble, an eUSB2 repeater is equivalent to a single high-speed hub tier. An eUSB2 repeater is equivalent to two FS hub tiers because of jitter budget constraints.

eUSB2 is positioned to help USB continue to be a ubiquitous connection interface even as applications transition to lower-power, smaller process nodes. eUSB2 repeaters enable interoperability for advanced semiconductors, while also being backward-compliant to the existing USB 2.0 ecosystem.

Julie Nirchi is an applications engineer supporting interface products at Texas Instruments in Dallas, Texas. Julie has more than twenty years of engineering experience in interface standards such as USB 2.0, USB 3.0, HDMI, and IEEE 1394. She holds a B.S.E.Elec from the University of Michigan.

0 comments on “eUSB2 brings USB to embedded systems

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.