Electronics will eventually fail. Learn how to protect yourself.
By Paul Stamler
All electronic components will eventually fail. Mortality and deterioration are inevitable, but there are ways to postpone them. We’ll start by looking at the reasons behind equipment failure and then we’ll look at ways you can protect yourself.
Why components fail
Resistors and capacitors
First, let’s look at passive components—resistors and capacitors. The stuff that makes resistors resist is susceptible to chemical changes: reactions with oxygen in the air, or chemicals in the plastic with which they’re encased. These changes alter the value of the resistor over the years; the resistors in an old guitar amp, for example, may be 50% higher in resistance than the label says.
Usually, carbon composition resistors (the ones found in most equipment until a few years ago) change the most; carbon film resistors (which have replaced carbon comps as low-budget choices) change somewhat less; metal film and wire-wound resistors change hardly at all.
Similarly, as paper-and-foil capacitors get old, the paper’s electrical characteristics change, and the cap’s value may change with it, or it may begin to leak DC, a common problem with old tubed amplifiers. This mostly happens to paper capacitors; plastic film caps (Mylar, polypropylene, etc.) seem immune.
Other chemical changes include oxidation of the wires coming out of electronic components. Wires are sometimes tinned with solder, which also oxidizes. So do the solder connections that are supposed to conduct signal into the component. If a solder connection’s a little wonky to begin with, eventually oxidation can make it go intermittent or fail completely.
A fundamental law of chemistry is that the rate of chemical reactions is proportional to temperature. Resistors that are heated change value more quickly than resistors that aren’t; wires and solder joints oxidize faster in a hot box than a cold one. Heat is the enemy.
Passive components can also fail catastrophically. The resistive element of a resistor is typically connected to a pair of end caps which clamp down at either end, conducting electricity from the connecting wires into the element. Every time the resistor changes temperature, it expands or contracts slightly. After many such cycles, one of the end caps may work loose; the resistor will go intermittent, or perhaps quit working at all. Capacitors can do the same thing.
Some film capacitors can also arc; tiny bolts of lightning can punch through the film layer when voltage is applied, poking holes in it which allow more arcing. (Better capacitors use “self-healing” films which are less prone to this behavior.) An arcing capacitor can short out, possibly coupling voltages into places they aren’t supposed to be and blowing other components out in turn.
The type of capacitor usually found in power supplies is the “electrolytic.” The electrolyte that separates the coils of foil that store the charge is a thick paste, and over the years it dries out. The drier it gets, the less it works (and the more it leaks), until the capacitor’s value is far less than it’s supposed to be. If it’s being used to filter a power supply, it will do so less effectively, and the piece of equipment may hummmmmm. More spectacularly, a dried-out electrolytic cap will occasionally explode, showering the area with bits of foil and electrolyte goo. And the hotter the environment, the faster the electrolyte dries out.
Cash in your chips
IC chips are at the heart of most contemporary audio and musical equipment. Analog chips (operational amplifiers, A/D and D/A converters, etc.) include some resistors on the chip surface, and so in theory they are subject to the same problems as discrete resistors. In practice, however, the failures of analog chips, li ke digital chips, tend more toward catastrophe.
Chips, analog or digital, consist of multiple transistors etched into silicon substrates; the various parts of these transistors are connected to the rest of the circuit by thin metal conductors that are evaporated onto the chip after the transistors themselves have been created. The circuit’s inputs and outputs are then soldered to tiny wires (by people with magnifying glasses, steady hands and low wages) which connect to the pins of the IC.
Vulnerable spots include the connections between the transistors (the metallization layer), insulating layers between sections, and the tiny wires connecting the chip to the outside world. When the chip changes temperature, it expands and contracts like anything else; these slight movements can eventually damage the connection between the metallization layer and the transistors, or cause a tiny solder connection to fail. The IC now is open circuited—nothing gets through. The same thing occasionally happens where the pins of the IC or its socket are soldered to the circuit board.
Another failure mechanism is internal. If one part of a transistor has slightly less resistance than its surroundings, it will carry proportionally more current. This additional current heats up the more conductive area, causing its resistance to drop still further. The process continues until so much current is passing through a limited area that it fries. It may go open-circuit (no current passes) or it may short out, passing enough current to blow fuses or damage other parts of the equipment. This happens to both ICs and individual transistors (the ones found in power amplifiers and discrete preamps).
The basic rule is that current causes heat, which causes stress, which causes burnout. (Just as with people.)
Tubes and stuff
Vacuum tubes go bad, of course—that’s why they’re in sockets, and why Grandpa used to take his tubes to the drugstore tube-checking machine to see if they were still working right. The filaments of tubes are like the filaments of light bulbs (the Russian name for a vacuum tube is “elektronnaya lampa”); they get dim with age, and heat the other tube elements less than they should. They’re also low-resistance when cold, so they pass a lot of current at the moment they’re turned on (unless the current is limited by a voltage regulator). You can see some tubes flare brightly at the moment you turn on the piece of equipment; just as with a light bulb, this is also the moment when the filaments can burn out.
Tubes sometimes get gassy; the noise level can go up, or they can become microphonic (that characteristic “ping-g-g-g” when you tap them with the eraser of a pencil). The cathodes can strip, lowering the gain or power capacity of the tube. Or an internal wire, suffering the same stresses of repeated temperature changes, can come unsoldered—and you can’t get in there to fix it.
Two elements can also short to one another for no apparent reason. If the plate shorts to another part, they can conduct spectacular amounts of current; if you’re lucky, the fuse will blow before the tube takes the rest of the circuit with it.
Switches and controls (pots) oxidize—faster if you live in the city, where the sulfur dioxide and ozone run free. Switch contacts develop an oxide layer which causes distortion or becomes intermittent. The conductive track of a control can oxidize, too; eventually, the sliding part no longer makes proper contact everywhere on the track, and the pot gets noisy or stops working at all.
The same thing can happen to connectors, including the parts that connect to one another inside the boxes themselves. Edge connectors can become remarkably corroded and unreliable. If the pins on ICs and their sockets aren’t plated with gold, they can go the same way.
Electrons rush in
If you have access to an analog VOM (volt-ohm-milliammeter) and a nice big electrolytic capacitor, here’s an experiment. Turn the VOM to the lowest “ohms” setting, connect one of the probes to one lead of the cap, then watch the meter as you touch the other probe to the other lead. The meter will swing momentarily toward the end of the scale indicating zero ohms (a short circuit), then slowly swing back to the other end.
What you’re seeing is the capacitor charging up. A charged capacitor has fairly high resistance to DC, but an uncharged cap is pretty close to a short.
Now think about what happens when you turn on a piece of equipment. The electricity whizzes through the power transformer, passes through the diodes, and hits—a dead short to ground, or at least a very low resistance, in the form of an uncharged power supply capacitor. A low resistance conducts a lot of current, and so the power supply draws a lot of juice from the wall in those few seconds before the capacitors in the power supply are charged. (When you turn on a power amp, you’ll often see the room lights dim for a moment.)
These high currents aren’t good for the gear; they place maximum stress on the transformer and the diodes, not to mention the capacitors themselves. Less obviously, they can cause wild swings in the voltage downstream, possibly blowing other components during that instant of stress.
Spike driver blues
Turn-on isn’t the only time when momentary rushes can wreak havoc with your equipment. In this imperfect world, the AC that comes out of your wall socket is far from constant; a refrigerator or air conditioning compressor switching on or off (even next door) can generate a large pulse of electricity that travels down the line to your studio. When the power company switches extra generators on or off the line, they can cause major jolts; so, of course, can bolts of lightning a few miles down the wire.
Theoretically, the power supplies in your equipment should filter out anything untoward from the AC. In practice, often it ain’t so. Spikes contain a lot of high-frequency energy, and this can couple right around any filter capacitors and regulators and get straight into audio or digital circuits. On a good day, you may hear a click or thump; on a bad day, the equipment bites the big one. Transistors (including the ones in ICs) have definite ideas about how much voltage they want to see; exceed that level and they go br-zap in an instant. Capacitors too.
It’s taken a lot of words, but I think it’s clear that while failure mechanisms such as oxidation and other chemical reactions can damage unused equipment, they are greatly accelerated when the equipment is used. And some failures, such as those caused by expansion/contraction, turn-on stress and spikes, only happen to equipment that’s plugged in and turned on.
“But we have to plug it in and turn it on if we want to use it,” I hear you cry. Well, there’s no need to lie down and wait for the stroke of doom.
Keeping the unavoidable from ruining your life
We started by describing the gory details of equipment failure. Now let’s suggest some ways to protect your gear and avoid the dire consequences of breakdowns.
Keep it cool
The first rule of electronic longevity is to keep it cool. Heat is the enemy. This means that jamming every piece of hot gear you own into a single rack is a bad idea; at the very least, use blank rack panels to space the warmer pieces apart. Use ventilated or open rack cabinets, with side-wall vents and perforated rather than solid tops. And, if possible, put the worst heat generators (tube mic preamps and such) at the top of the stack, so they don’t cook everything else (heat rises).
Power amplifiers are the worst offenders; if possible, don’t mount them in your rack. Instead, reserve a spot on the floor, halfway between your monitors; if your floor is carpeted, put the power amp(s) on a piece of plywood or particle board, to avoid having the carpet impede air flow to the bottom of the amp.
Power it up gently
Second rule: always use a surge absorber on everything in your studio—and don’t skimp. The better surge protectors, like the Monster Cable HT-800 I use ($79.99 retail at www.monstercable.com), do an excellent job of blocking line spikes, but they also filter out radio-frequency interference (RFI), making your equipment actually sound better. Hardware-store surge protectors won’t do that.
A good surge protector also minimizes the surge generated when you turn on the equipment, by absorbing the momentary current spike. It’s a good idea, incidentally, to avoid using the switch on a surge absorber or power strip as a master on/off switch for your system; if you turn on each piece of gear separately, using its own power switch, you avoid drawing a whopper surge as all the power supply caps try to charge at once. If any of your equipment includes a “Standby” switch, use it. Often letting the unit warm up in standby significantly reduces the stress on components.
If possible, your studio should run from its own house circuit, one to which nothing else is attached. Ideally, this circuit should be on the opposite leg of the breaker box from the ones powering the refrigerator, washing machine, air conditioners, etc.
Clean it up
It’s also good to clean contacts regularly—switches, jacks, edge connectors, pots. Over the years I’ve mentioned various products from Caig (www.caig.com)—DeoxIt® and PreservIt® for proletarian contacts and pots, ProGold® for gold-plated connectors. They really work; you’d be surprised how black a Q-tip gets when you clean old XLRs or rotary switches. I mostly buy the harder-to-find small bottles of liquid rather than aerosol cans; the aerosol tends to be messier. (On the other hand, sometimes the only way to clean a hard-to-reach pot or switch is to poke the snoot from a spray-can up and schpritz away, so it’s a good idea to keep both forms around.) Caig’s newest product, CaiLube®, is designed to clean and lubricate conductive-plastic pots without damage (it works on other types, too).
An un-obvious use for contact cleaners is to clean AC plugs. They get tarnished too, and the resistance can go up enough to generate significant heat as current passes through. I once melted the end of an extension cord; since then I clean AC plugs religiously.
Oh, yes: you should never allow smoking in your studio. Ever. Tobacco smoke leaves a greasy, corrosive film on everything, including contacts, tape heads, and microphone diaphragms, and it will significantly diminish the longevity of your equipment—and your lungs.
Orphans of the storm
There’s a saying: “If you find something you really like, buy a lifetime supply; they’ll discontinue it next year.” Having spent a day in fruitless quest of the arch supports I’ve come to rely on, I can testify to the dismal truth embodied in that saying; likewise when I search in vain for more of the only fingerpicks I like, or a bottle of Bonsai low-salt soy sauce.
By law, manufacturers are supposed to offer parts support for a certain number of years after a product is discontinued. That doesn’t help much, though, if the company’s owners decide to retire after forty years in the business (as the Sugden people did, a month after I bought my turntable), or if they get over-extended and go bellyup.com, or are bought and dismantled by the Megatherium Trust, or the treasurer absquatulates to a place with no extradition treaty.
For older gear, this isn’t always a big problem. Resistors can be replaced with the same value, often with higher quality (e.g., moving from carbon comps to metal films); ditto capacitors. Transistors, at least from domestically-produced equipment, are fairly standard, and there are substitution guides that will point you to an equivalent replacement most of the time. It helps a lot to have a circuit diagram.
Tube gear, surprisingly enough, is perhaps the most resistant to being orphaned. Because almost everything in tube equipment is an off-the-shelf standard part, a few hours with the catalogs can usually get you (or your service person) the components you need. The main exceptions are transformers, and they don’t often go bad anyway (if they do, you can almost always get a usable replacement, often better than the original).
With a few exceptions (notably the 4136 and TL075 quad op amps), analog ICs are reasonably easy to replace too; even if the original is no longer made, something with the same pinout and similar or superior performance is usually available, provided you know what the original chip was. (A special corner in Hell is reserved for manufacturers who painted over or ground away the part numbers on ICs to prevent users from buying their own replacements.) The articles I wrote a few years ago on upgrading parts (“Clean Up Your Gear”, Recording March/April/May 1996) can serve as a useful guide.
Mike Metlay, my synth-head Associate Editor, has asked that I mention one other potential trouble spot in replacing ICs: the oscillator, filter, and amplifier chips used in many early analog synthesizers, made by Curtis Electronics and SSM, are dangerously rare and getting rarer by the year. If you buy an old synth, get a complete schematic and surf the web for suppliers of the chips you might need. And be very wary of synths whose manufacturers are destined for my special corner of Hell; replacing their “special” ICs may be literally impossible.
Even more painfully, a lot of the gear we have come to love in the last few years doesn’t use stock parts. The rise of the ASIC (Application-Specific IC) has meant that many pieces of equipment are based around custom-made integrated circuits that aren’t used for anything else, and often aren’t sold to anyone but the equipment’s manufacturer. This leaves you stuck with paying whatever the manufacturer wants to charge when the IC goes bad—or you’re S.O.L. if they go under.
Even worse, a lot of equipment is no longer built with serviceable parts. Because compact designs require high density, the manufacturers use surface-mounted parts, very tiny, almost impossible to desolder and replace. (They’re usually built by robots, in case you wonder how the parts get on the boards in the first place.) Usually the standard repair is to replace an entire board rather than individual parts. This increases the price, of course. It also makes the parts problem worse; if a company goes under, you might, by waving a fistful of cash, persuade the chip manufacturer to sell you a replacement IC, but no one save the equipment’s maker is likely to have a circuit board.
In this unpleasant new world, it’s worth taking a bit of defensive action. Some equipment isn’t worth it; any piece of your computer (except for that miserable little fan) is likely to be obsolescent way before it breaks. (Note added while revising: I take that back, having just put in a new power supply. Still applies to most of the machine, though.)
But some equipment isn’t just equipment. Take Mike’s first analog synth—he’s put 18 years into learning its quirks, finding ways to make it sit up and do tricks, integrating it into his musical life in the same way as my battered but beloved old guitar.
Something you couldn’t work comfortably without is worth taking extra steps to preserve, over and above the longevity moves I’ve already outlined. So I offer some suggestions that are not particularly cost-effective, but may save your psyche one hot day.
First, if there’s a piece of equipment you really like, you can buy two of them. Put one in a cool room with controlled, low humidity; turn it on every few months, but otherwise leave it unplugged and unused. Some parts will still deteriorate, but at least there will be few heat- or turn-on-related problems.
Okay, that may not be realistic. So here’s Plan B: Go to a manufacturer’s service center and persuade them, using whatever tactics are necessary, to sell you any proprietary parts or replacement boards you might need if the piece shoots craps. Also persuade them to sell you a service manual; without it, you or a future service tech may be poking in the dark. Store the board and parts in that cool, low-humidity room.
What do you do when you learn that the maker of your beloved Ultra Frammistat just crashed and burned? If you haven’t already done so, implement Plan B; run, do not walk, to the service center and get the parts before the next bloke snaffles them.
Should that fail, if you really can’t live without the beast, head for Ebay or the equivalent. A used Frammistat is better than none at all, and the guts may still be healthy even if the peripheral bits are showing their age.
Do it yourself—or not?
How much work should you do on your equipment? That depends, really, on how much experience you have working on electronic gear, and how much sweat equity you’re willing to put into it. If you’re not experienced in handling a soldering iron (or if you automatically think “gun” when you think of soldering), you should let a professional do the job; it’s awfully easy to cause permanent damage to a possibly irreplaceable board or component.
But if you’re an old hand at electronic construction and repair, you already know you can do it. One caution: store all spare ICs, transistors and circuit boards in the conductive-plastic antistatic bags designed for the purpose, install ICs using the special tools made for the purpose, and be sure to wear an antistatic strap that’s grounded to the circuit before beginning work. If you don’t know what I’m talking about here, you shouldn’t work on solid-state stuff.
Perhaps it’s obvious, but make sure any gear you work on has been unplugged for at least a couple of minutes before you start poking around. Most equipment has 117V AC floating around inside, unless it’s powered by a wall wart, and tube gear can have upwards of 400V in exposed places, or stored in capacitors, and that can kill you dead as a mackerel. Use caution.
When should you throw in the towel? Unfortunately, there’s no clear answer—it depends how many bits are failing, how hard replacements are to find, how much you cherish the piece, and how much time and money you’re willing to invest in rehab. One good rule of thumb: if you find yourself hesitating to take the gear to a gig or schedule it for an important session, for fear it may prove unreliable, it’s probably time to put it out to pasture.
One more thing
This may not be welcome advice, but it’s unfortunately true: you’re usually better off buying higher-priced gear. Cheap equipment had to have its costs cut someplace, and usually manufacturers do that with cheaper parts that have lower voltage and temperature ratings; these are more likely to fail. Cheap gear also uses smaller power transformers, which generate more heat; if the transformer is inside the chassis, this can make the whole box run hotter.
Maybe nothing lasts forever—entropy still rules—but with some care you can keep the music flowing for years and years. The rules (not unique to audio) are straightforward: Use good stuff, use protection, keep it clean, and keep it cool.
Paul J. Stamler is a recording engineer, folk musician, and student of the Zen Of Electronics—”the quietest capacitor is no capacitor”— in the St. Louis area.