If you’ve been following the posts on Integration Nation, you’re no doubt familiar with a variety of reasons why a system designer might choose to use (or not use) an integrated analog solution for a given project.
Reducing space and cost are common reasons for integrating. If you have space constraints, integration may help reduce board real estate. If you’re designing a price-sensitive consumer product, integration may allow you to reduce costs compared to a discrete implementation. But in many cases, the main reason for integration is not to reduce size or cost, although either may be a secondary benefit. Instead, the best reason for taking advantage of analog integration is often that when it’s done right, it can help reduce the time you invest in your design.
So, the right chip not only integrates circuit blocks, it also integrates time.
Here’s a simple example: a portion of an AFE signal path consisting of an amplifier, an ADC, and a voltage reference. You can clearly build this function using individual ICs. Selecting each of the blocks -- ensuring they meet your performance requirements, stay within your board area limits, and hit your cost target -- certainly isn’t rocket science, but it can be a time-consuming slog, first through a bunch of parametric selection tables, then through a bunch of data sheets to find the right combination of components.
A “perfect” solution can be elusive. Maybe a critical specification for an otherwise-ideal component has just a typical data sheet value rather than a maximum limit. (So do you keep looking, or try to find a way to change your architecture so that the specification is no longer critical?) Or perhaps what looked like the right op amp won’t settle fast enough when driving the input stage of the ADC you’ve selected. (Do you start searching for a different amplifier, or a different ADC?) Or maybe the voltage reference with the right combination of output voltage, supply current, initial accuracy, and drift exceeds your cost target. (Can you find a way to relax the requirements, or do you start looking for cheaper versions of the other components to keep the overall cost in line?)
Those sorts of dead ends and decision points continue to eat up your time.
Now, this was a simple example involving just three components (ADC, reference, and amplifier) that are easy to find integrated together. When more circuit functions are integrated, the potential development-time benefits increase.
The integrated design -- again, assuming it’s done right -- can eliminate a big chunk of the component search process. Even better, all of the pieces will work together (they’d better -- it would be a major design fail if they didn’t), which should further reduce the hours you put into the project. And a really good integrated design may also take care of some design details that you would have otherwise needed to work out.
A single data sheet can tell you whether the circuit has a high probability of fitting into your design or not. If it’s a good fit, you’ve just saved a lot of time -- a precious commodity in any design environment.