Do you ever have to ensure that an exact amount of time passes between two tasks in your microcontroller code? Do you know what’s the difference between precision and accuracy? Today, [Jim Mack] tells us about pushing timers and interrupts to their limits when it comes to managing time, while keeping it applicable to an ever-popular ATMega328P target! Every now and then, someone decides to push the frontiers of what’s possible on a given platform, and today’s rules is coding within constraints of an Arduino environment. However, you should check [Jim]’s post out even if you use Arduino as a swearword – purely for all of the theoretical insights laid out, accompanied by hardware-accurate examples!
This will be useful to any hacker looking to implement, say, motor encoder readings, signal frequency calculations, or build a gadget processing or modifying audio in real time. To give you a sample of this article, [Jim] starts by introducing us to distinctions between precision and accuracy, and then presents us with a seemingly simple task – creating exactly 2400 interrupts a second. As much as it might look straightforward, problems quickly arise when clock crystal frequency doesn’t cleanly divide by the sampling frequency that you have to pick for your application! This is just a taste of all the examples of hidden complexity presented, and they’re accompanied with solutions you can use when you eventually encounter one of these examples in your hacker pursuits. In the end, [Jim] concludes with links to other sources you can study if you ever need to dig deeper into this topic.
Keeping our projects true to the passage of time can be an issue, and we’ve been at it for ages – calibrating your RC oscillator is a rite of passage for any ATTiny project. If you ever decide to have an interrupt peripheral help you with timing issues, we’ve gone in-depth on that topic in the past, with a three-part series describing the benefits, the drawbacks and the edgecases of interrupts. Going for a more modern target? Our piece on using interrupts with STM32 is a great path for trying out tools of the modern age.
This napkinCAD sketch shows how [SM6VFZ] implemented the boost regulator in the ATtiny214. The AND gate is formed using one of the CCL LUT’s. The first “timer 1” on the left, connected to one input of the AND gate, is free running and set at 33 kHz. The analog comparator compares the boosted output voltage against an internally generated reference voltage derived from the DAC. The output of the comparator then “gates” timer 1 signal to trigger the second “timer 2” — which is a mono-shot timer set to max out at 15 us. This makes sure there is enough time left for the inductor to completely release its energy before the next cycle starts. You can check out the code that [SM6VFZ] used to built this prototype, and his generous amounts of commenting makes it easy to figure out how it works.
