Monday, June 11, 2012

The Least You Should Know – Electronics 1

Electronics looks like a giant field to me, perhaps because it's where I make a living and I see more subtleties and nuances than an outsider.  Still, it's a really important field and keeping electronics going (or getting it going again) may well affect our quality of life.

Modern electronics can be broadly divided into two domains: analog and digital.  Analog systems deal with continuously varying voltages or currents.  They don't have to be continuous in time; pulses are often used in the analog domain.  Digital systems deal with voltages or currents that usually occur in only two levels, (on and off or high and low) and use laws of logic to process these signals (you may know the term "ones and zeroes").  There are exceptions to the “two levels” rule of thumb for digital systems that I'll get to down the road. 

Digital is the darling of the modern world; it's hyped in art and song, and systems can be sold simply by adding the word “Digital” to the package.  You've seen “Digital Ready” headphones?  The headphones apply varying voltages to moving coils: continuously varying voltages are the definition of analog.  The world is still changing daily due to the relentless advancement of microprocessors, and the increase in capabilities that they bring.  The “Moore” of Moore's Law is Gordon Moore, founder and retired president of Intel, who observed long ago that the number of transistors in a processor was doubling every two years and predicted that would continue.  Although this can't go on forever, it has remained a remarkably reliable guideline.  We can't keep making transistors smaller and smaller, though.  Eventually, transistors would have to be made of single atoms (they're working on it...), then subatomic particles, and then... well, I have no idea.  Individual quarks?

But...what's a transistor?  Patience, please.  We'll get there.

One of the fundamental building blocks is the diode, a component with two terminals which only allows current to flow in one direction.  The name has two halves: di-, meaning two and -ode from the old term “electrode”.  The diode first appeared as a circuit element long ago, as the simplest vacuum tube, but semiconductor diodes have replaced them for virtually all uses.  Semiconductor diodes – almost always silicon – are faster, lower loss, smaller and cheaper.  An overview of diodes can be found here.  Owners of vacuum tube gear will often modify it to use silicon diodes instead of tubes; when they do so, they get a higher voltage out of the diodes and they need to get rid of the extra voltage to keep from damaging the old equipment.  Because they conduct in one direction, you can easily test them for function using an ohm meter: one orientation of the leads will give a low resistance, and the other direction will be a high resistance. 

Diodes are rated in the amount of reverse voltage they can stand, reverse or PIV (peak inverse voltage), in their switching speed, and in the amount of forward current.  Some have ratings for their forward voltage drop at the specified current.  High current diodes tend to be packaged with large screws or bolts as their terminals, instead of wires.  Diodes are used as switches, blocks to signals going in the "wrong" direction, "rectifying" or turning AC into DC, and special types of diodes are used for a variety of other tasks.  The best known is probably the Light Emitting Diode, or LED. 
Semiconductor might seem like a bizarre term itself, but simply means a material that isn't very good as a conductor or an insulator by itself.  The first one widely used was Germanium, (atomic symbol Ge) but silicon (Si), found in sand around the world, has largely supplanted Ge.  By mixing in atoms of other elements (“doping”), silicon can be made to have an excess of electrons (N-type) or a scarcity of electrons (P-type) compared to intrinsic silicon.  This allows semiconductor designers to customize the conductivity and other properties of the material.  When a piece of N-type silicon is being grown, and the dopant changed to produce P-type material, a PN junction is formed.  A PN junction is a diode, but it's also the basis of the vast majority of semiconductors.  A lot of semiconductor literature talks about the flow of electrons and holes – a place where an electron should be, but isn't in P-type material.  I find the concept of hole flow bothers some people intensely, but not others.  Don't let it bother you.

The next most common vacuum tube, and the next one invented, was called the triode.  It has (surprise!) three electrodes, and the semiconductor equivalent of a triode is a transistor.  A transistor is a three layer device, which creates a pair of junctions on either side of the middle element, either PNP or (most commonly used) NPN.  These layers are called the Emitter, Base and Collector.  They are not usually symmetrical in size, with the collector being the largest of the three.  All three regions have a terminal connected to them. The base is usually the thinnest layer so that electrons speeding from one N region to the other can sometimes reach it without being absorbed in the P layer.  This is called a bipolar junction transistor, or BJT, and while they are economically very important, they aren't the only kind and not the most important kind.
(basic NPN (left) and PNP (right) BJT symbols)
There is another type of transistor that works differently.  In the Field Effect Transistor, or FET, instead of layers, there is a channel connected at both ends – called the source and drain (you can think of something liquid flowing in the source and sloshing out the drain).  FET is pronounced with a short E.  Somewhere between the source and drain is a “dab” of opposite type material called the gate.  The electric field created by voltage on the gate can allow full current flow, or it can reduce or stop (“pinch off”) the flow in the channel.  There are N-channel and P-channel FETs; junction FETs and MOSFETs, ("Moss  FETs") where the gate is a Metal Oxide layer, and MESFETs ("Mess FETS") where the gate is just a metalized region.  For years, the highest performance receivers were based on a FET made from a non-silicon semiconductor, Gallium Arsenide, or GaAs (pronounced gas), called GaAsFETs.  And my personal favorite these days, the PHEMT – the pseudomorphic high electron mobility transistor, a new type of FET with excellent performance in the UHF and microwave bands (Pee-hemt). 

What makes a transistor so important?  All of them are capable of controlling the flow of current in a circuit, and can be used as either a switch or an amplifier, depending how they're configured in that circuit.  A microprocessor, for all its glitz and glamor, is just a huge collection of switches.  With millions of transistors in every microprocessor, and more in every generation since before the first PC, humanity has probably produced more FETs than anything our species has made – even screws or rivets.  I have no idea how to prove or disprove that. 

Analog systems are all around you.  Whether guitar amplifier or iPod, the audio we listen to is analog.  The image on your TV - yes, even digital TV - is analog changes in brightness of a light source.  Remember the smoothly varying sine wave in the AC post?  As the frequency goes up, the AC goes from sub-audible to audio, to ultrasound, through radio, and on and on; there is no upper frequency limit.  All analog.  So let me briefly get into analog electronics and go over some common uses. 

The two most common analog systems, and if you're reading this online you have at least one of these in your house, are power supplies and audio systems.  Power supplies most commonly turn AC from the power grid into DC to run electronic systems, like your computer, but also will turn DC from a battery bank into AC to use anywhere AC is needed (called an inverter – for some reason).  Power supplies run the gamut from exceptionally straightforward and simple to very complex systems that require experts to analyze.  All of the components we've introduced along the way, resistors, capacitors and inductors, play their part in analog systems, along with the diodes and transistors we’ve just introduced.

We'll get into power supplies next.

1 comment:

Borepatch said...

Man, this takes me back.