feature article
Subscribe Now

Clash of the Tiny Titans

For those of you old enough to remember, “Let’s get small” was a line that got big laughs. Nowadays it’s a serious business plan. Small, light, and power-efficient chips are all the rage. Especially the power-efficient part. Hypermiling hybrid owners have nothing on today’s microprocessor designers, who are wringing every last erg out of every last joule. How’s that for nerd appeal?

Today’s case in point is Microchip’s new NanoWatt XLP microcontrollers. The “NanoWatt XLP” part is just Microchip’s brand name for its power-saving technology; the chips themselves are pretty normal for Microchip. There are 16 new members of this family, twelve of them 8-bitters and four 16-bit devices. You can find the whole lineup at www.microchip.com/XLP.

Microchip has what must be the world’s most successful, most varied, and most confusing product line. The company makes approximately 1.4 zillion different low-end microcontrollers, all beloved by embedded developers around the world. There are 8-bit, 10-bit, 12-bit, and 16-bit subsectors of the product family, and Microchip recently added a high-end 32-bit family based on the MIPS architecture. If, like everyone else, you’re confused by Microchip’s part nomenclature, this isn’t the place to explain it. Call your local Microchip salesman, who’ll be happy to try to make sense of it all. Bring a sandwich.

The short version is, the new 8-bit chips are part of the PIC18F46J series, and the 16-bitters belong to the PIC24F16 line. All 16 new chips are delightfully and tastefully appointed with on-chip RAM, flash memory, A/D converters, timers, UARTs, real-time clock, and assorted serial interfaces. In short, they’re just like most of the zillion other Microchip PIC microcontrollers.

How Low Can You Go?

Power is the XLP chips’ most interesting family trait, so let’s start there. All 16 chips in the family have three (count ’em) different low-power sleep modes. In the lightest of these modes, the onboard real-time clock keeps running so that time-triggered events can run at regular intervals. Shutting off the RTC but keeping the onboard watchdog timer running reduces power a bit more. Finally, deep sleep mode puts the RTC and watchdog to bed while interrupts and I/O pins stay awake, so external events can rouse the chip. In this mode, the chips consume only about 20 nA (that’s 0.000020 mA), or about as much energy as you get from rubbing a cat. Deep-sleep power is almost immeasurably small and could easily be overwhelmed by the leakage from a few resistors or other components. Your microcontroller may no longer be the most power-hungry device in your system.

How does Microchip do it? Well, there are a number of tricks, really, and reducing power to nano-Watt levels requires all of them. The logic is all static, for starters, and doesn’t need to be clocked to retain its state. Internal pull-up and –down resistors all have high impedance, and circuit design minimizes current leakage.

Interestingly, Microchip didn’t use the most obvious – but most expensive – trick of manufacturing with smaller silicon geometry. Instead, the XLP series chips are fabricated in a relatively outdated 0.35-micron process that’s at least three generations behind other chipmakers’ processes. Switching to a more modern 0.18-micron (never mind a leading edge 45-nanometer) process would have slashed power consumption much further, with no changes in chip design, but such fabrication processes are expensive. When your company headquarters says Intel over the door, you can afford 45-nanometer manufacturing because your customers are willing to pay for it. But when you’re servicing a broad and cost-sensitive customer base, it pays to be more modest. Most microprocessor companies discarded their 0.35-micron manufacturing equipment a decade ago. Microchip kept theirs, and they use the fully amortized gear to shave manufacturing cost. That’s part of the secret of the company’s success.

Round Up the Usual Suspects

So what do these chips have in them, apart from really wide silicon? All three have Microchip’s so-called mTouch capacitive-sensing interface. In the wake of the iPhone, touch-sensitive user interfaces have become all the rage, and mTouch makes it simple to hook up and monitor a touch-sensitive screen overlay. Microchip also provides some help with the requisite software, although creating a full-featured user interface is still the designer’s responsibility. On the whole, it’s nice to see touch-screen interfaces simplified (and cost-reduced) to the extent that they’re now available on $1.38 microcontrollers.

The 8-bit members of the family also have re-assignable I/O pins, which does just what it sounds like. Most of the comparator inputs, timer pins, UARTs, and SPI/I2C interfaces can have their pins swapped around and rearranged to suit the programmer’s whim. This is handy when your board layout changes or if you’re trying to shoehorn one of the new XLP devices into the socket originally designed for another PIC microcontroller. Kudos to Microchip for including this unsexy but ever-so-practical feature.

The 16-bit chips come with 512 bytes of EEPROM; not enough to store any code but enough for parameters, user settings, or configuration data. Oddly, these chips don’t have the pin-reassignment feature, which is a shame.

One chip, specifically the PIC18F46J50, also sports a USB 2.0 interface. This is a surprisingly fast and complex interface for a little 8-bit microcontroller that sells for just $1.78 in quantity. A simpler USB 1.x interface would have been adequate, although the additional USB 2.0 features (such as peer-to-peer connections) might be nice to have in some applications. The interface is probably more useful for uploading than for downloading, since the ’J50 chip doesn’t really have anywhere to store incoming data and probably couldn’t swallow it fast enough anyway. As an upload interface, however, it’s a good one to have.

They’re Everywhere!

Power corrupts, but low power is divine. Or something like that. Microchip and its competitors ship billions of new microcontrollers every year, adding to the awesome quantity of microscopic computers embedded around us. Something like 100 million transistors are manufactured each year for every man, woman, and child on the planet. And that’s in addition to the millions of transistors manufactured the year before that, and so on. Microcontrollers are truly ubiquitous, and building them cheaper and more power-efficient makes them more so. Microcontrollers are often cheaper and more reliable than mechanical devices, switches, and dials. They’re certainly more flexible. And they keep engineers like us in business. We have Microchip to thank for that.

Leave a Reply

featured blogs
Jun 22, 2018
A myriad of mechanical and electrical specifications must be considered when selecting the best connector system for your design. An incomplete, first-pass list of considerations include the type of termination, available footprint space, processing and operating temperature...
Jun 22, 2018
You can't finish the board before the schematic, but you want it done pretty much right away, before marketing changes their minds again!...
Jun 22, 2018
Last time I worked for Cadence in the early 2000s, Adriaan Ligtenberg ran methodology services and, in particular, something we called Virtual CAD. The idea of Virtual CAD was to allow companies to outsource their CAD group to Cadence. In effect, we would be the CAD group for...
Jun 7, 2018
If integrating an embedded FPGA (eFPGA) into your ASIC or SoC design strikes you as odd, it shouldn'€™t. ICs have been absorbing almost every component on a circuit board for decades, starting with transistors, resistors, and capacitors '€” then progressing to gates, ALUs...
May 24, 2018
Amazon has apparently had an Echo hiccup of the sort that would give customers bad dreams. It sent a random conversation to a random contact. A couple had installed numerous Alexa-enabled devices in the home. At some point, they had a conversation '€“ as couples are wont to...