Optimised solutions for key Bluetooth applications

According to BrightCom, Bluetooth devices need to contain embedded stand-alone intelligence, along with software and additional processing power.

The original concept for Bluetooth wireless technology was to implement a simple cable replacement.

However, once the cable is cut, "dumb" devices no longer have access to the intelligence of the host.

To solve this problem, stand-alone intelligence needs to be embedded, along with software and additional processing power, into Bluetooth solutions so that they can provide more than basic Bluetooth communication.

Examples of such intelligence would be a print manager application on a printer, OCR algorithms implemented on a wireless handheld scanning device, and so forth.

On top of that issue, implementing Bluetooth technology in real life applications requires differently optimised solutions that take into account the different power requirement, different performance requirements and different amounts of memory for the various types of devices requiring to be wirelessly networked.

During the early days of Bluetooth, companies were talking about a single Bluetooth device bearing a price tag of $5 yet for which nobody could define what its characteristic and performance would be.

BrightCom Technologies, a silicon and software fabless chip company headquartered in Israel, understood that for a diversified market of Bluetooth devices, the system price will fall down only by customising solutions to the applications not by shrinking the Bluetooth part by brute force.

It has created and pioneered the application processor approach of optimised solutions to meet different market needs that it calls IntelliBLUE application processors.

"Ours is a real world approach that recognises and solves the needs of companies wanting to implement Bluetooth wireless technology", said Yuval Ben-Ze'ev, BrightCom's President and CEO.

"The total solution cost for adding Bluetooth functionality must be kept as low as possible.

This figure has to include all the hardware, all the software and all the, often overlooked, costs of working out how to implement it.

We provide complete, fully integrated, off-the-shelf, system solutions that are easy and quick to implement giving manufacturers a rapid and inexpensive route to market".

Bluetooth as a communications standard is very new to the industry.

As such, the initial implementations of Bluetooth into existing devices are expected to be in the form of add-on devices or "after market", which is a common industry trend in the technological space.

However, Bluetooth poses some new challenges.

The Bluetooth SIG committees, while defining the Bluetooth layers and environment, have opted to make Bluetooth wireless technology the most prevalent industry standard such that it will be used in every future electronic device.

Strongly adhering to the initial mantras of "cable replacement", "simple" and "low cost", they did not choose TCP/IP as the common communication basis, below which Bluetooth could have been defined in layers 1 and 2 only.

Instead, the SIG defined the details of a protocol stack and created the mechanism of "Profiles" step-by-step recipes defining how software applications should use the Bluetooth protocol stack's various layers for a particular usage model.

Essentially, the profiles are a set of application programming interfaces (APIs) that are comprehensive enough for any implementation of the Bluetooth usage models.

They are very complex and touch almost every aspect of the protocol stack and, as a result, applications residing on different appliances, which are written by different software companies and use the same profile definitions, will be able to communicate in a natural, seamless manner.

However, a software programmer wishing to code a Bluetooth-communicating application has to acquire a significant amount of knowledge in communications and protocols.

Such a requirement is naturally inconceivable except during the initial phases of the technology - the phase where only experts are doing some hands-on work with the new technology.

The overwhelming support to Bluetooth and the mass of development teams involved with interfacing Bluetooth to existing and new systems raises the demand for software programmers that are also communications experts.

The scarcity of such programmers is probably one of the major factors contributing to interoperability issues and problems during the first phases of Bluetooth productisation.

A Bluetooth Application Processor is basically a chip device that performs two main tasks: it processes the Bluetooth Baseband and protocol stack software including Profiles, and is targeted to a specific application by means of supplying the hardware interfaces and the CPU power to run a software application if required.

Running all the Bluetooth software and applications in the application processor means that no changes have to be done to any existing software programmes and that no addition resources are required from the host processor, removing the possible need for a more powerful processor to cope with the additional demands or the expense of including a processor if one does not already exist in the device.

The application processor approach gives system stability and independence as existing operating systems and applications, running on the host CPU, will be separated from the protocol stack and have less interference, less portability and porting problems, enabling Bluetooth functionality to be easily and quickly added to products.

An interesting consideration, on which BrightCom had put emphasis, was how to ease the task of application development and reduce the level of communications expertise required from a software application developer.

The result of this was the BrightAPI, a programmer-friendly application programming interface that allows the programmer to focus on application development rather than dive into the depth of Bluetooth protocol stack.

All the OEM has to do is write to the API in plain C code or other high level languages (if running under an existing OS), which requires no knowledge of Bluetooth wireless technology to do.

For comparison, other baseband chips are usually supplied with software that only goes up to the host control interface.

BrightAPI is basically a set of function calls, which address the various profiles in such a way that can be useful for the application programmer.

Using the BrightAPI, a software developer can master the Bluetooth system within hours compared to weeks or months if using the raw API of the Bluetooth profiles.

BrightCom has created a platform approach for its IntelliBLUE chips that divides into three families to address the three key volume areas for Bluetooth applications that it wishes to address, ie networking devices, network access and audio.

For wireless networking between PCs, PDAs and peripherals, the company has the BIC2101, which has a built-in USB and fast serial interfaces that makes it very easy for manufacturers to drop into existing designs by simply attaching to the USB bus within the product.

First silicon is due shortly.

The BIC2101 is priced at under $10 in volume quantities, which include the application processor IC with all baseband functionality and application processing, software drivers for UART, USB and PCM, the complete Bluetooth protocol stack and the BrightAPI with Bluetooth profiles support.

For network access to larger, wired networks such as to the Web or a company's in-house, hard-wired LAN, BrightCom has the BIC2301 that has a built-in digital signal processor and Ethernet interface.

First silicon is due Q4/2001.

For audio links to loudspeakers and stereo headsets, there will be the BIC2201 that will have first silicon early 2002.

OEM's have complete freedom to use virtually any RF chip with the BrightCom application processor as it has a programmable RF interface enabling OEMs to select and change RF chips according to availability, price and performance specifications to suit their current needs.

Back to top