“Electrosmog”?
From time to time new words are coined in the tech world, and some stick like glue while others evaporate like old political jokes. And often these terms are formed not by the tech community, but by those outside of it. Enter Electrosmog.
Earlier today, I recalled a prescient conversation of the mid-1990’s that I had with a procurement person about how engineers select components. There in our San Jose R&D office on Brokaw, deep in the belly of Silicon Valley, she felt that engineers she worked with would only design in a component if they could order it from Digi-Key. With such a rich world of options out there, why, she asked, would otherwise good engineers trap themselves in the Digi-Key box?
Well here we are over 15 years later, and the answer is obvious. Digi-Key realized that the fastest way to get engineers to use them was to make the selection and ordering process easy. Digi-Key provided engineers with the links they need, and the late-in-the-day shipping deadline allowed them to design during Happy Hour and still have presents by 8am the next morning from the friendly brown truck.
The problem, though, is that we begin to have a monoculture in our designs. Like eating wheat, corn, and oats, we ignore kamut, quinoa, and millet. All are perfectly good, and some of those less common ones might actually be a smarter choice. But when we as engineers can go the easy path (nearly one-click shopping at Digi-Key), it’s easy to ignore the other options out there.
My challenge to the design community is to think outside the Digi-Key box. I know you’re all in love with Atmel, Microchip, NXP, TI, and a number of ARM-based processors for your embedded designs, but let’s not forget the broader landscape. Heard of Holtek for example? Yes, they’re on Digi-Key, but their processor line is not, so you might not even know about it.
If you search for “microcontroller” on Digi-Key’s site, they return 29,867 options from 27 different companies. Whoa. Ask yourself, how many of these companies have you actually considered for a design? How many do you have in your toolkit of tricks right now? Missing anything?
Maybe the kamut of processors out there has just what you need in your next design.
In a recent interview, one of the tasks for the candidate was to give the diode names for various schematic symbols. Some were easy and others were a bit more obscure. And to make matters more interesting, there are multiple symbols for some. Below are a few examples, but know that there are many more not mentioned here.
In Baby’s First Diode Book you just get one diode, this one:
All it does is conduct one way and not the other. That’s easy. And to a first order this gets one through a lot of design problems. In the discipline, we call this the “ideal diode”. Alas, that’s vastly oversimplified and overly limiting. And you can’t buy one.
Turns out that the diode world is quite rich and complicated. Let’s introduce some of the family members, roughly in order of reverse obscurity. First off is what we already know:
This can be a switching diode, fast switching diode, or rectifier. And it comes in silicon (mostly) or germanium (rarely nowadays). The point is, this diode symbol tells you a little bit, but not everything you need to know. This diode’s little secret? It’s light sensitive if you get it in a glass package. The most common flavor is the 1N914 of old and the 1N4148 of today. In the diode family tree, switching diodes are often the lowest cost.
If you need something faster or lower forward voltage drop (Vf), then try this Schottky diode:
For simplicity, sometimes the last tail of the curls is dropped, but the above symbol is best. Schottky diodes are blazingly fast switchers and have lower Vf than the typical switching diodes, but they are generally more expensive and have higher reverse leakage currents. High-current flavors of this part include the 1N5817.
Maybe switching isn’t your goal, but rather regulation. Enter the zener diode:
This one is all about conducting backward, the mode that diodes aren’t supposed to go into. Normall,y. current flows in the A –> K direction, but in reverse (K –> A) the idea is that the diode blocks current flow until the voltage exceeds a certain level. Thus they make excellent voltage regulators, holding the voltage at a set level. Or they work well at protecting processor input pins as they can clamp the voltage at levels safe for downstream circuits. Unfortunately, these are not ideal, and the voltage where they begin to conduct is mushy. This means that they start conducting a little bit at first and then conduct progressively more as the voltage increases. Oh, and their knee voltage is temperature dependent, but with one little secret: for some parts the voltage goes up with temperature, and for some the voltage goes down. But for a few magic values, the temperature coefficient (Tc) is near zero. Pick these values if your application needs stability across temperature.
A special variant of zener diodes is the transient voltage suppression (TVS) diode. These are built for speed and for power. Many of these can absorb pulses of 1500W and turn on in a mere picosecond – the perfect solution for crushing ESD impulses.
Okay, these were the common ones, but let’s get into some more interesting examples. Need a variable capacitor? A diode’s your part for the job, though it needs to be a varactor diode:
Vary the bias voltage and the capacitance changes. Look for these in tuning circuits of radio receivers. These poor diodes are almost always operated reverse-biased so they never get to forward conduct as their non-zener brethren generally do.
Need a double-zener that avalanches? Try a DIAC:
This is not really a diode in that it doesn’t conduct differently in different directions. It’s actually a member of the thyristor family, another rich family of semiconductors that many electrical engineers rarely use in their designs. DIACs stay mostly off until the voltage reaches a threshold, then the device becomes conductive, passing current indefinitely until the current drops below a specified holding current. The most common place to find these components is in conjunction with TRIACs in lamp dimmers.
Read more about the family here. Engineers should know about SCRs and TRIACs at a minimum if they ever need to control AC currents.
As mentioned at the start, there are many other members of the broad diode family. Not even covered here are the diodes designed to emit light, or be sensitive to light, or those made from cat whiskers.
Still, my favorite diode has to be this one released by Signetics, the Noise Emitting Diode (NED). This device is designed to be connected to +1000VDC and emit a noise. Once.
Electrosmog 