The Universal Serial Bus (USB) has been the most successful peripheral interface in the history of PCs. USB 2.0 is poised to succeed the first generation and offers particular benefit to the mass storage class of PC peripherals. These products include portable digital audio players, external hard disk drives (HDDs), ZIP drives, CD burners, DVD burners, high-density PCMCIA type-II storage card readers, Magneto-Optical (MO) drives, and others. The original USB standard, USB 1.1, had insufficient bandwidth to attain optimal performance from these types of drive products. With its appreciably higher throughput (480 Mbs), however, USB 2.0 is capable of keeping up with the high bandwidth requirements of these mass-storage peripherals. USB 2.0 has already begun the process of enabling a myriad of high performance mass-storage peripherals.
ATA drive, ATAPI drive: A standard, off-the-shelf drive, incorporating an Advanced Technology (as defined by American National Standards Institute group X3T1Q) Attachment /ATA Packet Interface (ATA/ATAPI) bus and connector as its electrical interface. Examples include HDDs, CD-RW and DVD drives. They normally reside within a PC or Mac, with connectivity provided via an ATA/ATAPI ribbon cable.
USB 2.0 drive: The combination of a standard ATA or ATAPI drive and a USB 2.0 to ATA/ATAPI bridge board, in order to create a fully functional USB 2.0 peripheral drive.
Bridge boards needed
The basis for any USB 2.0 mass storage peripheral (MSP) is a USB 2.0 to ATA/ATAPI bridge board. Peripheral developers wish to get to market quickly, and to do this they rely on using off the shelf ATA/ATAPI drives " so called because the connector they all share is one form or another of an ATA/ATAPI connector. These ATA/ATAPI drives understand and respond to ATA/ATAPI commands, which are presented to them via their ATA/ATAPI connectors, or buses. The USB 2.0 bus, however, neither communicates with nor understands ATA/ATAPI commands. A USB 2.0 to ATA/ATAPI bridge board must act as an instruction translator and data manager between the two buses (Figure 1).
Figure 1: Bridge Board Function
Multiple choices for USB 2.0 drive product developers
Peripheral drive developers have a series of developmental requirements when selecting a USB 2.0 " ATA/ATAPI bridge solution. One important requirement is the ability of the bridge to interoperate flawlessly with a wide variety of ATA and ATAPI drives. Another important requirement is the ability to implement innovative features, including cable-powered products. Innovative features are important because they allow drive developers to differentiate their products against their competition. Selecting the right USB 2.0 bridging solution is vital if the developer is to meet these requirements and get to market quickly. Let's examine these developmental requirements one at a time.
Flawless operation with a wide variety of ATA/ATAPI drives
"Flawless operation" actually has multiple aspects. A USB 2.0 bridge solution must be able to compensate for operational variations inherent in ATAPI drives. Put simply, many ATAPI devices have some sort of non-compliance with the ATAPI specification, and as a result exhibit "operational variations" from what would normally be expected. These variations come in multiple forms and are difficult to predict. One category of variation common in ATAPI drives is timing variations on the ATAPI bus. The width of an ATAPI bus handshaking signal, such as data strobe, may differ based on whether the drive is operating in PIO mode or in UDMA mode. Unless a bridge device has adjustable timing in the ATA/ATAPI physical interface, a bridge board based on that device will not be able to compensate for ATAPI drive timing variations. A drive developer will be forced to select either a different ATAPI drive or a different bridge device. Either way, the developer's time-to-market is slowed.
Another example of an ATAPI drive operational variation is the inconsistent method in which drives communicate their operational status to a USB 2.0 bridge board. While the ATAPI specification defines the use of a status register within the drive, the use of this register has been shown to vary from drive to drive. A specific example is the status register's BUSY bit. ATAPI drive vendors vary in their implementation of this bit's operation and functionality. In a case such as this, unless a bridge device is flexible enough to vary its own operation, based on the BUSY bit's behavior, a likely consequence will be a drive "hang" condition. This is when the USB 2.0 drive stops communicating with the PC or Mac it is attached to and becomes inoperable. A bus reset or rebooting of the PC/Mac is then required if the USB 2.0 drive is to resume normal operation. Clearly this is undesirable from an end user's standpoint, and it may result in user returning the USB 2.0 drive to the store.
The Thirteen Cases: More on Flawless Operation
The "BUSY bit" problem is a specific example of a bug that falls under the category of "The Thirteen Cases". The Thirteen Cases are a subset of the USB Mass Storage Class specification and are discussed in detail in chapter six of that document. The document specifies what is to happen under all possible permutations involving drive/host communication involving the transfer of data. It is important that USB to ATA/ATAPI bridge devices employ proper error handling procedures whenever a host-drive mismatch is present. These errors can happen when the host and drive "disagree" on either the direction of data transfer or the amount of data to be transferred.
Proper bridge implementations result in the thirteen cases category of errors being transparent to the drive developer and thus to the end user. Improper implementations, however, can result in drive hang, data loss and data corruption symptoms. Clearly, proper implementation of the thirteen cases within a bridge device is vitally important to USB 2.0 peripheral vendors making bridge device choices.
Innovation requires a flexible, firmware-driven architecture
Drive developers who wish to differentiate themselves by developing innovative USB 2.0 mass storage peripherals will be well served by seeking out a firmware-driven device. A firmware-driven device, coupled with plentiful general purpose I/O, allows drive developers to leverage their unique skill sets and market perspectives during product development. The results are novel and inventive drive products and the success in the market of manufacturers providing those drive products.
Innovative Solutions for Portable Digital Audio and Video Players
HDD-based portable digital audio players are an increasingly popular product category. While audio players have historically relied on flash memory technology for audio file storage, the use of small form factor HDDs for audio file storage is increasingly popular. The reasoning behind this migration is simple - song-storage capacity. Whereas a traditional, flash-based audio player could store 10 to 20 songs, a HDD-based player can store thousands. Apple's iPod was the first widely known portable audio player that used a HDD, making the iPod's slogan, "1000 songs in your pocket," literally true. People have begun to carry their entire song libraries around with them.
With the increased capacity of HDD-based audio players came the necessity of having a high data throughput connection to the PC or the Mac. Whereas Apple's iPod utilized the "fast wire" characteristics of the 1394a "Firewire" high-speed connection, other firms have begun utilizing USB 2.0 as their high-speed connectivity solution. As with USB 2.0 peripheral HDDs, this application also requires a USB 2.0 to ATA bridge solution. That said, portable audio players are more complex than peripheral HDDs, and as a result, the USB 2.0-based audio player developers require additional functionality in that bridge solution, over and above that needed for a USB 2.0 HDD alone.
Bridge solutions for HDD-based portable audio players need to be "intelligent." That is, they need to be programmable, with a firmware-driven architecture. The best way to accomplish this is via an embedded processor within the bridge device (USB 2.0 to ATA). Embedding the processor within the bridge device offers the benefits of saving both board space as well as cost. Additionally, I/O channels between the DSP and the intelligent bridge need to be present in order to allow inter-device communication and coordination to occur. This can be accomplished with general purpose I/Os (GPIO), with a serial bus, or a combination of the two. Let's take a look at why communication between the bridge device and DSP is important in a portable audio player.
Innovative audio player developers are taking advantage of the gigabytes of storage capacity available in their internal HDDs. With a plethora of data storage space available and the desire to differentiate their product offering from their competitors, developers are adding non-audio functions and capabilities to their audio players. Many of these new functions are normally associated with personal data accessories (PDAs). By adding functions such as calendars, to-do lists, contact lists and expense entries, HDD-based audio players now function as fully featured PDAs as well. Additionally, developers are also augmenting the traditional capability by adding functions such as play list synchronization.
A key requirement in the implementation of these additional features is the ability for the bridge and the DSP to communicate freely and efficiently. While much of the movement of data continues to be between the PC and the HDD, information about this data must be communicated to the DSP from the intelligent bridge device. This is accomplished in a straightforward fashion through customization of the bridge's firmware. Of course, if the bridge does not have an integrated processor -- if it is not an intelligent, programmable bridge device -- this customization will not be possible.
Figure 2: USB 2.0 HDD-based Portable Audio Player
Portable Video Players
Portable video players are an emerging product segment and can be considered a close relative to portable audio players. They have many similar product requirements. Because the video files they deal with are orders of magnitude larger than audio files, video players are essentially required to make use of HDDs for video file storage. Portable video players value an intelligent bridge for the exact same reasons that portable audio players do.
Cable-powered drives, defined as USB 2.0 drives that are powered through the USB cable and do not need to be "plugged in" to a wall socket, are very desirable to the end user. Besides offering the end user a simple use model (just plug it in), cable powering also saves money by eliminating the need for a "brick," or wall power adapter.
One criticism of the USB bus specification has been that the cable itself is capable of providing peripherals a maximum of 500 mA of current. As most ATA and ATAPI drives require more than 500 mA, developing cable powered drives has been problematic. Additionally, the USB specification requires any cable-powered device to draw less than 100 mA during enumeration and less than 500uA during the "suspend" power saving mode of operation. A bridge device that meets these power criteria will be essential to drive developers seeking to manufacture cable-powered drives.
In those instances where a developer wishes to utilize a drive requiring power levels the USB cable is not capable of supplying, smart-battery assisted solutions are available. Under this scenario, a rechargeable battery, such as Lithium-Ion, is co-located with the drive, along with smart-battery management devices. The battery will provide a large portion of the drive's operating current. Low current consumption by the bridge board is still vital to this application, however, as sufficient free current must be available to the smart-battery management devices for battery recharging to occur. The alternative is to risk a dead battery and possible data loss or corruption.
Developers need to be able to maximize power transfer from the cable to the battery when implementing battery recharge via the USB cable because the USB cable can supply only 2.5 watts of power (500 mA at 5V). Knowing this, the developer will strive to minimize the power consumption of all devices that are not directly involved with the battery recharge function. The logical approach during battery recharge is to power down the HDD and the DSP. An intelligent bridge device, operating from firmware optimized for battery recharge applications, will power down both the DSP and the HDD. It will also tri-state its own ATA interface, thus reaching a low power usage state optimized for recharging the battery. While the battery is recharging, the bridge device remains free to monitor both the recharge process, as well as the USB bus. Should bus traffic pertaining to the audio player reappear, the bridge will then re-enable the DSP and the HDD.
Figure 3: Smart Bridge Enabled Battery Management
Texas Instruments TUSB6250 - Optimized Solution for Cable Powered Applications
Texas Instruments recently announced the TUSB6250 as the optimal solution for these applications. The device enables innovative, flexible solutions via the integrated microcontroller. This firmware-based architecture will allow the drive developer to address the "operational variations" of the ATAPI drives. It will flawlessly address and adapt to handle the "13 case" scenarios specified in the USB Mass Storage Specification. With its sixteen general-purpose I/O pins, the developer has the flexibility to market new and differentiated solutions. They also are able to interface to additional system chips, such as a DSP, in an audio or video player.
For developers interested in developing cable-powered solutions, the TUSB6250 requires less than 80mA at 3.3 volts. Several low-power HDDs are commercially available which, when coupled with a TUSB6250-based bridge board, can draw power exclusively from the USB cable. The other key USB specifications, which a drive developer must meet to achieve USB Implementers Forum certification, are enumeration current (100 mA) and suspend mode current (500 uA). These parameters are also met by bridge boards utilizing the TUSB6250. The intelligence and low power of the TUSB6250 combine to make it an excellent solution for any device which will implement a battery and power management to charge the battery off of the USB cable.
The innovative architecture coupled with the extremely low power consumption of the Texas Instruments TUSB6250 make it the optimal solution for any of the existing external USB storage peripheral applications as well as for new emerging end-equipment that utilize high capacity storage, battery power, and high-speed data transfer.
TI Smart Battery Management
TI Digital audio/video devices
With more than 16 years of TI experience, Dan Harmon began his present position as Texas Instruments' Strategic Marketing Manager for Catalog Interface Solution Products in January of 2000. He joined TI right out of college in 1986, hired into the Defense business unit (DSEG). He worked as a systems engineer on night vision FLIR systems for three and a half years before transferring to the Semiconductor group working on CCD imaging products. In that capacity, Dan worked as a camera design engineer for five years before assuming Product Marketing Engineer responsibilities for the CCD products in April 1996. In March of 1998 he transferred to the Multimedia Marketing group and then to Connectivity in 2000. He graduated from the University of Dayton with a BSEE in 1986 and from The University of Texas-Arlington with an MSEE in December 1990.