Common faults

 

Audio amplifiers are usually subjected to more stress than the majority of other audio equipment, they have to deliver far higher amounts of power and hence consume more power and create more heat. They often lead a hard life and thermal stresses on components are the largest reason for failure, over time stress can cause breaks in conductive traces (PCB tracks), breaks in transformer windings, cause oddball intermittent faults which are heavily reliant on temperature and hence can be a nightmare to trace and fix.

The largest enemy of electronics is heat. No matter what fault is caused, part of the actual failure mechanism will often be excess creation of heat. Heat not only degrades some materials, it also changes resistance of any conducting material and hence the operating characteristics of any component, whether it's something as simple as a connecting wire or as complex as an IC. Keeping everything cool is important, this especially goes for large amplifiers and receivers.

 

Mains transformers

Mains transformers in good quality audio gear very rarely cause problems unless they are abused - transformers are inherently highly efficient components (95% efficiency isn't uncommon) and so waste little power as heat, transformers in higher end equipment tend to be oversized and so are unlikely to suffer enough stress to damage them during normal use.

That said, transformers can fail - although it tends to happen more in lower end, modern audio equipment, where things aren't so heavily engineered and so parts are run nearer their limits. Most of the time it's easy to diagnose a damaged transformer, because they usually go totally open circuit and will kill a circuit stone dead - because there will be no voltage coming out of the other end of them. I guess a transformer could also go 'out of spec' because of damaged or shorted windings, although I'm yet to come across this problem in a well designed piece of audio gear.

If a transformer gets overloaded it may 'saturate' and under this condition will create a lot of heat (mostly because under these conditions the magnetic effect breaks down and it becomes essentially a large wirewound resistor) - most modern transformers have thermal fuses fitted into their windings and these will operate if the windings get dangerously hot (most thermal fuses I've seen operate around 150°C). These fuses are basically a small glass tube containing a pair of sprung contacts held together by a small wax pellet, at a set temperature the wax melts and the contacts spring apart. These fuses can't be reset, the only remedy is to delve into the transformer and replace them.

You CAN replace thermal fuses most of the time, but remember that the transformer probably isn't designed to be taken apart and sometimes this takes a lot of ingenuity. I've salvaged a lot of mains transformers like this over the years - but you NEED to have a good understanding of what you're doing to ensure that it is indeed the fuse at fault (basic ohmmeter tests on both sides of any fuses), also that you're buying the correct type of fuse, that you don't damage the insulation on the windings when you're working on the transformer, also to make sure that any shims and fixing screws or shielding cans go back exactly the same way as they were before you started. Another thing to mention is that the transformer windings are probably pumped full of shellac (to prevent noise and vibration) and that means you'll probably have to carefully dig out the fuses without causing any damage.

Also you need to ensure that you don't inadvertently wire it up backwards when you're done (easier mistake than you might think) - this may sound obvious but if you do work on a transformer, ALWAYS do resistance checks afterwards and then test it on the bench to make sure all voltages make sense before you connect it back up to any circuitry. The reason I stress this is that by connecting a step-down transformer backwards, you create a step-up transformer and although it probably won't work very well like that and would likely get very hot and buzz loudly, you'd still get a serious amount of voltage out of the other end, and that voltage could seriously hurt you.

Finally, if you're faced with an open circuit transformer, you need to ask yourself why it has failed - is it the cause or just a symptom? A shorted output transistor, or a short anywhere else on the supply lines, would pull large amounts of current and although fuses should blow first, there's a possibility that the transformer was the weak link. That said, transformers can genuinely go open circuit just because somebody liked their music loud and it got too hot and popped its thermal fuse.

 

Capacitors

Capacitors (apart from electrolytics, which I cover in great detail here) aren't usually the cause of problems. Anything used on the phono stages and possibly the tone filters will likely be polystyrene, these are very delicate regarding temperature but as they're only used in filter networks they get an easy life and very rarely cause any problems if left alone. The same goes for other capacitors - tantalums only get used on power supplies or as DC coupling on some of the lower model receivers, these can potentially fail but rarely cause me any problems, even if I'm dubious about them being allowed anywhere near audio signals (because of their very non-audio performance). Ceramic capacitors only get used as RF filters and bandwidth limiters across transistors and unless they get damaged mechanically, they can be left alone. That leaves the mylar capacitors used in some of the filter networks, these again are highly reliable if left alone.

 

Resistors

Resistors, for the most part, are usually very reliable - as long as they're not run too hot that is, either through design or because of a fault condition. A heat damaged resistor will usually go up in value or just break completely and go open circuit. The few places where I have a problem with this in vintage Pioneers is the bias resistors for some of the voltage regulators or in the driver stages of the SA-8800/SA-9800 amplifiers - in both cases the resistors go high in value and cause myriads of problems, and it's a fault that isn't always easy to trace unless you test them out of circuit.

I've also had emitter resistors on power amplifiers go open circuit, but that's to be expected - they handle lots of current and are wirewound, so sometimes the fine wire element just 'breaks'. Still, a broken emitter resistor is an easy and cheap fix.

Another thing to mention here is that if you come across a resistor in a vintage Pioneer which is getting very hot, it might be indicative of a fault - it's unusual for any components in vintage Pioneers to be run at more than 60% of their maximum ratings, measuring the voltage across it and doing simple P=V²/R will give you an idea how hard it's being run.

Finally, film type resistors which have been getting hot for a while will discolour - they usually develop a characteristic dark band around the middle; if you see any like this, it's a good idea not only to replace them, but also to check voltages in that area against the schematic to make sure there's not a fault somewhere.

 

Semiconductors

Semiconductor devices commonly fail short circuit at the sorts of voltages they usually see in solid state audio gear (with transistors, a collector-emitter short circuit is the most common in my experience with audio amplifiers, diodes only have two terminals so would just go dead short anode to cathode).

New output transistors in a Pioneer SA-9500 II.

Amplifier driver or output transistors going short on the average amplifier will usually just slam the speaker output near to one of the voltage rails, hopefully either blowing the supply fuses or preventing the loudspeaker relays from energizing. Fault-finding shorted outputs is pretty simple, finding suitable replacements may be more difficult.

A shorted rectifier diode could cause potentially massive amounts of damage as then any components after it would be subjected to AC, something which neither electrolytic capacitors nor transistors would take kindly to - if you had a piece of audio kit which had suffered this kind of catastrophic fault, my advice would be to not even attempt repairing it; the potential for components further down the line to be damaged but remain 'walking wounded', only to fail later would be too great - my main concern would be large reservoir capacitors which could start to pass DC current if they had been reverse biased.

That said, semiconductors can also fail open circuit, although I'd say this is a far more uncommon failure mode at the low voltages which solid state audio equipment uses.

A favourite for transistor faults are the power regulators in vintage Pioneers. For reasons I never understood, the regulator transistors in many models of receiver and amplifier are attached to tiny heatsinks and are run far too hot, in some cases eventually destroying the solder joints which connect the component to the PCB (SX-880 comes to mind here, the transistor which powers the low level circuits can hit 140°C). Usually this will also cause heat damage to the components around the heatsink, replacing everything and modifying the PCB for a larger finned heatsink should solve the problem permanently.

You can do basic 'go/no go' tests for dead shorts with a normal ohmmeter but if you're repairing a lot of electronic gear then a more expensive multimeter with a diode tester function which gives you a junction voltage reading and ideally a beep confirming a good or bad junction is better.

One more thing to mention is that when a semiconductor goes short circuit, it may be in a position where it has a lot of current available and can sometimes pull enough current to destroy tracks on the PCB - if you're repairing a fault caused by a shorted component, examine the PCB carefully for signs of damage, it might prevent headaches later. PCB tracks can be repaired but it can get tricky.

 

Intermittent faults

Whereas the largest cause of point blank failure in an amplifier is usually through excess power/heat, in my experience intermittent faults in any type of audio gear are usually mechanical in form, one way or another. This could be broadly classed as anything from switches to dry solder joints.

I should know better than to endorse this, but the classic advice to 'give something a whack if it stops working' does actually show some wisdom, even if it doesn't actually 'fix' anything. Take an example of a typical fault caused by a dry solder joint or loose connector - the gap between the two parts might only be in the order of microns thick, but unless the potential difference (voltage) between those parts is very high, the voltage won't be able to track across the gap and make a circuit - what you end up with is effectively a very high value resistor which behaves in a very non-linear way.

Usually, once a current can flow across a bad connection (for example, if you give that connection a whack and those parts momentarily touch each other), it will 'wet' and will then make a reasonable connection - as long as enough current keeps flowing and the gap doesn't become too large (again, we're talking microns here).

That's another reason why audio equipment is so prone to faults in tone and function switches and loudspeaker relays - because low level audio signals in those areas often have no DC bias and so are inherently balanced around 0V, the signal is constantly changing in magnitude (and hence in strength); as the signal reaches its peaks it becomes strong enough to track across the bad solder joint or oxide covering on the switch but then gets cut off again as it reduces in magnitude and can no longer bridge that gap - and remember, with music signals this process is probably happening hundreds of times a second or more. That's why bad connections in audio circuits usually manifest as heavy distortion with pops and clicks.

 

Switches and carbon pots

The switches on audio equipment are usually precious metal of one kind or another, precious metals conduct electricity very well making them good for fragile low level audio signals but some (like silver) will oxidise over time, made far worse when the equipment hasn't been used for a time - working the switches in normal use scrapes off any oxides as they build. The switches need to be cleaned with a good quality contact cleaner which will loosen the oxide and leave a film on the cleaned contacts to protect them for a while.

On a rare occasion there are still problems, at this point I'll usually just remove them and take them apart to clean manually rather than dousing them in cans of cleaner and hoping for the best. Bear in mind that most audio switches weren't ever designed to be taken apart, it takes some skill along with patience and some fine tools to get them opened and assembled again in usable condition. 

Technics rocker switch, you can see the oxides on the contacts

In some types of switches such as the rockers used in many vintage Technics, the contacts are covered by a plastic carrier, making it difficult to get enough cleaner to the contacts. The contacts are cleaned using Brasso and cartridge paper of all things (Brasso works very well as long as you remove all traces afterwards). Everything is then flushed clean with contact cleaner and coated with protector, any moving parts like rockers coated in silicone grease and everything fitted back together.

On Pioneers at least, tone and muting switches are favourite for causing faults because the signal levels are usually quite low at this point in the circuit  Other apparent faults range from bad imaging (more bass/treble/volume in one channel than the other making the stereo image unbalanced), volume gradually fading in and out at random (especially on music with solo instruments), pops and clicks, white noise, whistles and howling sounds (these last two especially in cassette decks) - the list is endless.

These kinds of faults are often highly dependent on factors like temperature. Switches are probably the favourite cause of oddball faults in vintage audio - high pitched squeals in cassette decks can often be traced to a dirty/jammed main record switch, faults on amplifiers and pre-amps which occasionally lose volume and/or high/low frequencies on one or even both channels can often be traced to bad switch contacts.

Loudspeaker relay

I've had graphic equalizers and receivers play completely dead on both channels all because the main tape monitor switches hadn't been used for years and had built up enough oxides to open the switch contacts on both channels, the oxides build up to a point where the force pushing the contacts together is no longer strong enough to scrape through the oxide layer and hence the two contacts can't make a proper connection.

On amplifiers which switch the loudspeakers via relays, the relays can cause problems - especially if the amplifier has had a fault in the past which has put DC across the speakers and continually tripped the relay, arcing the contacts where they were switching heavy currents.

Relays are like any other mechanical component, they do gradually wear out; the coil will last virtually forever as long as it's not abused, however the contacts will cause problems - most relays use pad contacts and hence there's no scraping action to clean them. Over time, oxides form on the contacts and the connection is gradually broken - again, this will often cause issues with sound cutting out and heavy distortion, especially at low volumes. If you're restoring an amplifier, it's a good idea to either disassemble and clean the speaker relays, or just replace them.

 

Solder

Most vintage audio gear was wave soldered (much faster and much cheaper than regular soldering) and this can cause problems over time where joints start to go bad.

Speaking for Pioneers at least, solder quality varies widely from bad to very good, usually the later models are better (x500 series and earlier amps can sometimes be a nightmare). Another problem is that as the joints go bad, they sometimes collect impurities and can't simply be reflowed - usually the only way to get a good joint in this case is to remove every trace of old solder and replace it with new, even then the component leads are sometimes badly oxidised and need scraping clean.

In my opinion, the only 'proper' way to reflow old solder joints is to remove all the old solder and to re-make the joint with new solder alone - some people say to just touch each joint with an iron to reflow it, but I don't agree with this - as I've mentioned before, the old solder contains too many impurities and the reflowed joints you'll make will look ugly and probably go bad in a very short space of time. Besides, a well done, hand soldered board will always be of higher quality than a mass-produced, wave soldered one.

People talk a lot about solder - my own, personal opinion (and again, this is my own, PERSONAL opinion) - is that modern lead-free solders just don't work well inside old audio equipment - I've had many issues with it not 'wetting' with the remaining traces of old lead solder left on the PCBs and component leads, I also don't like the way that lead-free always seems to give a visibly 'dry' looking joint, even when you know that joint should be mechanically perfect. Even at work, on the shop floor they use lead-free (for RoHS compliance), yet in the labs we use old style 60:40 tin/lead, because it flows much better, gives a better finish and is more durable over time.

My own choice is Multicore brand 60:40 tin/lead solder - or, if you can justify it, the version with a few percent of silver added - to lower contact resistance.

 

Wiring and connectors

Even fractures in wiring looms can sometimes cause strange faults, I remember once trying to trace the source of a fault in a vintage Pioneer receiver which would kill the display lamps after it had warmed up for 15 minutes - turned out to be a break in one of the wires which expanded with heat. Single core wire (such as the type which Pioneer used) is more vulnerable to this kind of fault, although stranded wire can too, if it's been stretched or creased too much.

Connectors can also be a problem waiting to happen. Any place where a wire joins a PCB or other type of component, it needs to be connected some way. The most reliable way is to solder it, but this isn't viable for mass produced electrical gear and makes repair or servicing much more difficult if that part needs to be removed at any point. Pioneer used wire wrap pins, which are supposed to give a gas-tight connection; to be fair, these don't cause anywhere near the amount of problems that they get blamed for, but I usually clean and solder them.

Brands such as Technics and Quad relied mostly on connectors - this is fine, but the actual connection between the contacts isn't necessarily gas-tight and after a few years the contacts will begin to oxidise.

 

In summary

In my experience when dealing with audio gear (especially amplifiers), the order of faults from the most to the least common are:

 

- Mechanical faults or damage (this covers anything from dry solder through to wiring/connector faults, dirty switches, worn relays, bad panel fuseholder connections)

- Blown main fuses (usually caused by something going short circuit, fault actually lies elsewhere, probably shorted semiconductors)

- Shorted semiconductors (especially output transistors in power amps, followed by voltage regulator power transistors)

- Overheated mains transformers (more common on modern 'mid-fi' audio gear)

- Damaged resistors, usually gone high in value through heat but possibly open circuit if wirewound or fusible

- Degraded electrolytics, usually a source of noise problems in power circuits; I don't see THAT many faults with these though, nothing like the problems they cause in the power and control circuits in LCD TVs etc.

- Other capacitors; highly unlikely to cause problems unless they've been abused through heat or mechanical damage or damaged by a fault somewhere else (for example having way too much voltage placed across them - again, this is uncommon in well designed audio gear)

 

Back to amplifier rebuild main page