I come from a generation that didn’t do much radio. I’ve met relatively few other RF Engineers around my age. Most are older than me, though at work we frequently mentor new RF Engineers from scratch. When I started in radio you had to build a system from building blocks – develop a system gain and noise figure plan, then choose the right parts to accomplish the job. You could optimize each part of the system for the application you were targeting. It was an immensely satisfying cost-benefit analysis exercise where ingenuity and creativity could make a measureable difference in performance and costs.

Then came Chipcon. There were some other single-chip radio products prior to Chipcon, but Chipcon was the first single-chip radio with good RF performance AND an internal dedicated bit sync and byte sync. Something has to actually LISTEN to the radio of course and before Chipcon, it was dedicated logic or a microcontroller. The usual method was a uC but that of course would involve the CPU running all the time, so this was difficult for anything battery powered.

Now there are a whole slew of companies offering perfectly adequate single chip radio ICs every bit as good or better than those first Chipcon products, now with that uC built-in. They’re just not very interesting. All the interesting stuff is done at the IC level so there’s not much a board level designer can do to affect cost or performance. Frankly, they’re boring. From the end customer’s point of view, they’re inexpensive, high performance, widely available, small and use less power than discrete radios.

We could still build a better performing radio out of custom selected system elements, especially the filters. But frankly, any decently executed modern RFIC would be only 3dB worse in performance but 10dB better in cost. Sometimes you need something weird, something that doesn’t have a big enough market to justify the right sort of RFIC, and that’s where the fun is.

The problems however haven’t changed. The little cheap radios have the same board and system level implementation problems that radios have always had. As long as the problems that could be created inside the RFIC are dealt with inside the RFIC, things are pretty much a wash. You still need the same equipment to test the thing and you still have to do all the same steps which means that while the BOM can be smaller, the development time isn’t much different. Radio development isn’t about actually designing the electronics; it’s about testing what you designed.