Signetics released the venerable NE555 timer in 1971. It has been used in thousands of designs and has been the subject of at least that many articles (see “The 555: Best IC Ever Or Obsolete Anachronism?”). Communication Editor Lou Frenzel’s recent article about Semtech’s SX8122 and the 555 prompted me to examine how timing solutions have changed over time (see “And You Thought The 555 Timer Was Dead?”).
RC Timing
When I first worked with the 555, the microcontrollers (Fig. 1) of the day included Intel’s 8-bit 8741 in a large 40-pin package versus an eight-pin dual-inline package (DIP). Back then, adding a couple of discrete components to an eight-pin DIP was a simple way to build a range of functions like an astable multivibrator (Fig. 2). This simple circuit provided a clock that could be used to drive shift registers and other logic that was typically available at the time as well.
Semtech’s SX8122 uses a single capacitor and resistor to tune its timing (see “Analog Timer Advances On Venerable 555”). The sub-1-V analog part is designed for low-voltage, battery-operated small appliances.
The question I had was whether any microcontrollers could come close to the chips’ low price, $0.36, and also meet the space and power requirements. I didn’t doubt that a micro solution could provide significantly more functionality.
As with most designs, there isn’t just one solution, and the best solution depends on a range of constraints for a particular application. There are many chips to choose from thanks to Atmel, Freescale, Microchip, Renesas, Silicon Labs, STMicroelectronics, Texas Instruments, Zilog, and other vendors that can use a single battery for power and take up as little space as 2 by 2 mm. That would almost fit between two pins of the 8741!
On the voltage side, Texas Instruments’ 16-bit, 4-MHz MSP430L092 is a true 0.9-V ROM-based microcontroller (see “Micro Really Needs Just 0.9 V”). Many of the other flash parts work on a wide range of voltages.
Most tiny micros come with a built-in clock, so generating any type of waveform is a trivial exercise. Many have analog-to-digital converters (ADCs) allowing an application to track analog inputs, not just a crude digital signal.
Micros Too Complex?
So why wouldn’t a tiny micro be the first choice for developers? Assuming price and power requirements are met, micros have one additional requirement over an RC design—programming.
For me, the answer is easy. But then again, I’ve programmed in assembler. Writing a C program to bit bang a complex timing sequence is a straightforward task, but only if C is in your toolbox. If not, a simple circuit is often “easier” to implement. There are other considerations with micros, such as brownout conditions and startup time.
On the other hand, the actions that can be performed by a micro are nearly impossible with a simple analog circuit. Likewise, tuning an application can be easier for some compared to determining how an analog circuit will perform.
So what will you pick for your next design?