Paying closer attention to timing in electronic designs

Paying closer attention to timing in electronic designs

New configurable oscillator technology offers greater flexibility for designers

BY DAVID MEANEY
Fox Electronics
www.foxonline.com

We have all experienced, at some point in our lives, how “timing is everything.” Yet when it comes to implementing timing control measured down to a millionth of a second (1 MHz) or less, why would so many designers leave oscillator selection to the last minute?

As demands in electronic designs are trending toward more timing-critical, compact and energy-efficient performance, smart designers are looking for every advantage they can get. One such timing advantage is the evolution of configurable oscillator design. Another is a calendar.

That’s because today’s configurable oscillator designs do more than just deliver highly accurate, low-jitter/low-noise, frequency-specific components in less than two weeks. They also offer potential rewards of improved performance and cost-efficiency for designers who consider their attributes earlier in the design cycle.

Changing technology offers broader flexibility

Increasingly sophisticated demands – ranging from the precise, high frequencies needed for stability-critical apps such as SONET or Ethernet networking down to more compact, low-power solutions for mobile wireless devices are pushing oscillator requirements beyond the limits of commonly stocked off-the-shelf components.

Fortunately, new configurable oscillator technology offers greater flexibility for designers even if they wait until the proverbial last minute to specify them. They combine the best of a high-Q quartz reference source plus the efficiency of very sophisticated ASICs in a single package, able to deliver the ideal custom engineered frequency in a matter of days (see box below, “Building block approach creates greater flexibility”).

But improvements in performance, quicker availability and a more affordable cost for the final component are just the start.

Configurable oscillators available in different footprints, power levels and output buffers share common design principles that make it easy to assemble unique solutions Even though configurable oscillators can be delivered quickly if left to the end of the design cycle, there are distinct advantages to evaluating them earlier. That is because, by the nature of their modular design, configurable oscillators offer added options for those who take time to appreciate their versatility at the beginning of the design cycle even when working with cut-and-paste block designs.

Six advantages of earlier oscillator design-in

Specifying silicon is traditionally the first stage in the design cycle. And since each chipset typically comes with its own specs and designs for the supporting circuitry “within the block” for that particular chip, it is easy to use cut-and-paste engineering to build toward the final design.

Of course, that assumes that all the details of implementation within each design block will mesh together perfectly as you get closer to production. Experience, however, shows that is not always the case.

And even if that approach works from a functional perspective, there are still added efficiencies to be gained from investigating enhanced performance, spacing, power savings and business economies earlier in the cycle. They include the following:

1. Implementing cleaner designs with pinpoint specifications: Without being limited to standard fixed frequencies, configurable oscillators enable product designers to have the component tweaked to match their design instead of the other way around. But pinpointing the specific frequency that is best for your design is just the beginning. The ability to choose among different ASIC components for final implementation can also improve stability in overall performance.

2. Upgrading performance or efficiency: One example of this is configuring terminations to keep traces between the oscillator and the chip as short as practical, to help minimize RF interference and harmonics in the application.

3. Responding to “surprises”: This can involve maintaining your flexibility to adjust to design changes involving stability, tolerance or power supply requirements, at any time. It could also include changing the form factor within the allowable limits of the application (e.g. allowing for appropriate heat dissipation in higher power, higher frequency designs).

And if the oscillator specified on the chip manufacturer’s bill of materials uses a non-standard pinout (or if you are replacing one in the process of upgrading an existing design), a configurable oscillator’s ability to re-map pinouts makes it a readily compatible alternative.

4. Relieving inventory concerns: Keeping a sufficient supply of appropriately configured oscillators in the pipeline to meet anticipated production demands is not necessarily an electronic design consideration, but it still impacts your organization financially.

Because modular design allows the configurable oscillator manufacturer to maintain just the raw building block components in their inventory ready to assemble and ship on demand there is less of an inventory investment sitting on your shelves or theirs for long periods of time. It also means less manpower is required to manage inventory levels within your organization.

5. Avoiding “sole-source” pitfalls: Concerns about maintaining consistent sources for reliable oscillator options, without worrying about putting all your eggs in one supplier’s basket, have changed, thanks to the added flexibility of modular configurable oscillator technology.

First, the modular design allows “upchain” components to be substituted in the extreme event that a first-choice component is not available. And if the same functionality can be configured in multiple footprint options, planning ahead to accommodate the use of either size in your design can create greater flexibility.

Second, the performance of any substitute part or oscillator package remains consistent, because all are deliberately made to be compatible across the manufacturer’s overall product range.

6. Maintaining all your options, with minimal lead times: Once you have devoted up-front consideration to all factors and options of configurable oscillator specification, you retain the option to wait until just prior to the production run before locking in your final purchase decision. This allows you to adapt to subsequent changes in product design and still maintain optimal oscillator design and performance.

A better ending starts with a better beginning

So just how early in your design cycle is the right time to take advantage of the benefits of configurable oscillator design?

Take the case of a nine-month design cycle as an example. Chipsets will be evaluated, selected and implemented into the master design within the first months. You should start considering the critical timing devices that will control those chips immediately after that within the first third to the first half of the total design cycle (in this case, four-and-a-half to six months ahead of the build phase). This still leaves time to solve previously hidden problems, improve performance characteristics or plan for added economies and efficiencies in the final build all without needing to commit to the final order until the very last weeks before production. ■

Building block approach creates greater flexibility

A configurable oscillator starts with a quartz crystal reference. But instead of using a conventional integer phase-locked loop (PLL) that covers a wide frequency range, one of several fractional-N PLLs is chosen based on what frequency band the final frequency needs to be. This fractional-N PLL doesn't divide the reference frequency, and therefore eliminates the problems of the 20 logN rule.

To compensate for the spurs created by the irregular divisors of fractional-N PLLs, a fourth-order delta-sigma modulator (DSM) block is added to lower the overall amplitude of the spurs by spreading out the spurs to different spots over time. The block creates a mirror image of the oscillator's noise, effectively canceling it out – much like the principle of noise-canceling headphones.

The final step adds one of three types of output buffers, depending on the needs of the application.

The combination of these configurable components has enhanced the versatility of quartz-based devices without compromising their quality. The latest generation of this technology has demonstrated reduced jitter (in the range of 0.3 to 0.5 ps), frequency ranges up to 1.350 GHz, stabilities as tight as ±25 ppm, voltage requirements down to 1.8 V, plus better phase noise to meet OC-192 requirements. ■