Which USB is right for your application? (Part 3)

Dan Harmon, Texas Instruments

April 08, 2012

Dan Harmon, Texas InstrumentsApril 08, 2012

In 2007, I wrote a two-part series titled “Which USB is Right for your Application” for Planet Analog (Part 1 and Part 2). Since then, several new and different “versions” of USB have been released. In this article, I discuss how they have been deployed in the market in the almost five years since.

Also, I address SuperSpeed USB (USB 3.0) and its impact on the USB market. I would also be remiss if I do not address Thunderbolt™ Technology as well and its impact on PC interfaces moving towards the future.

We will close this article by discussing some recently announced new initiatives from the USB-IF aimed to enhance the end consumer experience with USB.


USB 2.0 revisited

As Brian O’Rourke of In-Stat commented many years ago, USB is the most successful I/O in the history of the PC. This remains the case today after over 15 years of life. As discussed in 2007, USB 2.0 is a specification that defines three speed modes: low-speed, full-speed, and high-speed.

All three speeds are still used extensively today for new products. Low-speed continues to be used for most human-interface devices such as keyboards and mice. Full speed still is used in many industrial, medical, and industrial control applications. Finally, high-speed has been the main interface of choice for most of the new content-rich consumer products such as portable media players, digital still cameras, and the ubiquitous flash drives that I do not know how we ever lived without!

USB high-speed products will be the interface that is most impacted by SuperSpeed USB as even greater content is stored and the desire to enhance the sync-and-go experience of the consumer, but more on that later.

The second most significant development (the release of the USB 3.0 specification being the most significant) in the USB market since my original articles is the use of the USB connector for charging battery-powered portable-consumer products. When first released, the USB specification did not really take into account the potential desire to charge batteries with VBUS that is defined as 5V and 500 mA for a host port or a self-powered hub port.

The USB-IF has released a battery-charging addition to the USB specification with the goal of helping to improve the user experience when it comes to charging their devices via a standard USB connection. The battery charging (BC) specification defined a mechanism for the device to negotiate for additional power, if it, as well as the host (hub) port are BC compliant. On the host side, the specification defines three types of ports to support charging:


Standard downstream port (SDP):

○This is the same as any pre-BC downstream port

○Provides up to 100 mA during enumeration

○Provides up to 500 mA after enumeration for charging with data transfer

Charging downstream port (CDP):

○Provides up to 100 mA during enumeration

○Provides up to 1.5A for charging with data transfer

Dedicated charging port (DCP):

○Referred to as wall adaptor charging mode

○Provides up to 1.5A for charging but with no data transfer

○Provides charging even if the host is turned off.


In addition to the USB-IF BC specification, there have been other initiatives related to battery charging. The Chinese government has mandated the mobile phones must use a micro-USB receptacle for charging. This is called Chinese Telecom Mode (Standard YD/T 1591-2009) as it relates to battery charging. The mandate is intended to eliminate the need for phone providers to ship a new charger (with proprietary connectors) with every phone.

The mandate does not require that a standard USB-cabled connection be used for charging, it requires that phones charge via the VBUS line in the USB connector. It could be a wall adapter with a micro-B USB plug on the end of the wires.

In addition, the European Union issued a directive that took effect in 2011 that essentially says the same as the Chinese mandate, phones must charge via a micro-USB connection. Finally Apple products, such as the iPhone and iPad, have their own detection scheme that allows for even more power to be supplied when they are connected to MAC USB ports.

USB On-the-Go (OTG) not real portable…

Back in 2007, the USB-OTG addendum to the USB 2.0 specification was relatively new and the jury was still out on what type of products would adopt the standard aimed, as the name implies, at portable devices. Since then, very few devices have adopted the OTG as defined in the original specification.

However, a new version of the specification was released that defined a new “class” of devices that the specification defines as embedded host (EH). An EH product essentially is a device that wants to act as both a USB host as well as USB peripheral, but for separate buses/applications. An EH device can have a single B-type receptacle (or captive cable with A plug) to enable it to act as a peripheral to a standard USB host. It can also have one or more A receptacles for servicing attached USB peripherals.

The intent is for the two buses to be completely independent of each other, to serve two different purposes/applications. This is compared to an OTG device that per the specification can have a single micro-A/B receptacle that, depending on which end of the cable is inserted, will determine if the product is in host or peripheral mode, but only a single USB bus is active at any time.

One example of potential embedded host products would be a printer that connects to a PC on its peripheral bus, but also offer the capability of inserting a digital still camera or flash drive in the A receptacle for direct printing. A second example could be a digital picture frame that can connect to a PC via the peripheral connection for loading images or allow a flash drive to be inserted for download. Similar to USB OTG, EH devices define their targeted peripheral list (TPL) since they may not have the ability to load drivers for unrecognized peripherals connected to their host bus.

Certified wireless USB, a good idea, but…?

As with OTG, Certified Wireless USB had only recently entered the market back in 2007. In the intervening time, this version of USB has seen little traction in the market. If you go to the Integrator’s list on the USB-IF website and search for Wireless USB products, you will find there are 125 certified products (as of 12/8/2011).

For comparison sake, if you search for SuperSpeed USB products in the same list, you will see 285 products have been certified since the release of the USB 3.0 specification in late 2008, over twice as many products in about half the time. I believe this is really the case of what sounds like a good idea that forgot one of the key benefits of a cabled USB connection – the ability to charge portable devices. As discussed above, that was a feature that was not really part of the original USB 2.0 specification and has only really crystallized in the time since the Wireless USB specification was released.


SuperSpeed USB update

For those who have not kept up, the USB 3.0 specification was released to the public in November  2008. The promoter group’s goal was to deliver four key values with SuperSpeed USB:


Increased transfer rate to 5 Gbps for fast sync-n-go applications

Improved bandwidth utilization

Optimized power efficiency

Backwards compatibility with USB 2.0 and USB 2.0 is forward compatible with USB 3.0!


While the adoption rate for SuperSpeed USB in PCs have outpaced that of USB 2.0 when it was released at the turn of the 21st century, as we look towards 2012, this will skyrocket as both AMD and Intel release their chipsets with integrated support for USB 3.0.

So the question is, “What applications care about the benefits?” Looking at the end of Part 2 of the series from 2007, I laid out the case at that time a need for the improved throughput offered by SuperSpeed USB. The amount and type of content being generated is ever increasing. As resolutions for images and movies increase, the individual file sizes are growing significantly. These combine to drive the need for additional storage capacity in the portable devices, and end user desire to have a very high-speed pipe to move this content from their portable device to a centralized storage location or vice-versa. This is what we refer to as fast sync-n-go.

In addition to these portable consumer products, there have been developments in the industrial, medical, and control applications that love the ability to move lots of data into a processing unit for real time response/control. Image acquisition, machine vision, and data acquisition are all examples where customers have implemented SuperSpeed USB outside the consumer products segments.

Another benefit of SuperSpeed USB is the increase in the available peak current from 500 mA to 900 mA. However, a peripheral can only request 900 mA, if it is enumerated at SuperSpeed. If it is connected as a USB 2.0 device, it is still limited to 500 mA maximum peak current. While the peak current is higher with SuperSpeed USB to accommodate the increase in speed from 480 Mbps to 5 Gbps, SuperSpeed USB transfers actually only consume about one-third the power of equivalent USB 2.0 transfers.

The power efficiency improvements written into the specification help accomplish this. The two major contributors to this are the fact that all SuperSpeed transactions are directed and the definition of intermediate low power states for links.

In USB 2.0, transactions are broadcast. This means all downstream ports have to transmit the same information, regardless of the intended target. This also means that all downstream ports and peripherals are consuming power to determine if the data is intended for it or not. With directed transfers, only the host downstream port (and any intervening downstream hub ports that map to the target) to which the peripheral is connected will consume power.

Additionally, only the intended peripheral has to consumer power receiving the data. The second factor is that devices can go into intermediate idle (lower power) states when they are not actively transmitting or receiving transactions. The final significant change is the elimination of device polling. This helps to reduce the overall power footprint of the USB bus while also improving the bus usage efficiency.


A more detailed look at SuperSpeed USB can be found here.

< Previous
Page 1 of 2
Next >

Loading comments...

Most Read

  • No Articles

Most Commented

  • Currently no items