Zero is zero, … consumption

16-1

The ability to monitor in real-time very large meshes of sensors has awaken the hunger for more and more data: more traffic sensors, more weather stations, more customer habits trackers, etc. All this reminds me of the escalate of the hardware performance after the software demands in the decades before this millennium. But.. what did we learn from that?

It may be natural to ask for more when we think that there is a lot to share. However, the Internet of Things has been born under a very stringent energy budget: to run economically (and ecologically) sustainable meshes of sensors, we humans (let’s say maintenance services) need to revisit each hardware node as little as possible – ideally, not to revisit anything. On these grounds, the critical question is how to store energy for feeding the devices for longer and longer missions. While we do recharge our smartphones every night, those little gadgets we left out there into the wild can’t.

b19

This new energy factor was less relevant (or absent at all) during the technological race in the last decades: the power outlet was just there. If not, fossil fuels were next door. Recently, hardware has been preparing for this challenge by reducing the operating voltage, optimizing idle periods consumption, using software that addresses resources more carefully, etc. The arena is however open for crazy new ideas – they are indeed equally welcome and needed. Could we recycle some old ideas to reach nearly “zero consumption”?

Let’s call it Analog passivation. imagine you just want to count very small pulses that occur spontaneously. A real-time approach may use a very fast computer to monitor the input signal in high frequencies. This device, no matter how “power efficient” it is, will always be switched on and reading every millisecond (most of the time is processing background noise). Alternatively, If we had a passive block that incrementally counts +1 and stores this value in a non-volatile manner, i.e. a piezoelectric charging a capacitor, we could come back each “second”, read the value, reset it to zero, and go back to sleep. In this case, the electronics is most of the time sleeping. It will consume energy only during the data reading and transmission periods which can be scheduled to be in the seconds, minutes, or… even the daily range. For example: if you want to monitor traffic density, you probably need just three colors (red, orange and green) to describe three states (collapsed, dense, fluent). More colors are reallly unnecessary. Therefore, if you read that imaginary capacitor connected to a piezoelectric cable each one minute and you get a large charge (voltage) value on it, you may infer which color it is.

Therefore, one active branch of modern electronics will be the development of passive blocks that implement very basic functions. Ironically, there is nothing technologically new to discover and, perhaps, we should start scavenging and rediscovering documents from the 70s again! In the linked YouTube talk, we address how spintronics is a potential technology that can improve and enrich the “classic capacitors”. If you are curious about how this has been already implemented, visit igsresearch.com/piezo

Thanks for reading,
Xavi Marti