ESP Logo
 Elliott Sound Products Mains Power Quality 

By Rod Elliott (ESP)
Page Created 25 August 2014

Share |

IndexMain Index


We tend not to think too much about the power that we use for daily activities, and this includes sound systems. I doubt that anyone would be heard to complain that their morning coffee tasted odd because of mains interference or distortion, but there is an entire industry that will try to convince you that without their mains filter, sinewave reconstruction power supply, isolation transformer and/or $5,000 power cables your audio and video systems must be horrible (and your coffee will taste like cat pee as well!).

Mains distortion is commonly cited as something that will cause the soundstage to be contracted/ compacted/ eliminated, and that the distortion will cause a loss of clarity, soften dynamics and mangle the bass 'slam' from your subwoofer. Naturally, we can expect that micro-dynamics will be damaged beyond repair, and the 'air' between instruments will be cloudy and grey with a 35.7% chance of reproductional ineptitude.

Now, if you happen to be in (or near) a commercial or industrial area, there may indeed be various noises that pass down the mains distribution system and cause your system to generate clicks, pops, farts and other noises. If this is the case, you really might need some kind of filter, but if you never hear any of these things (or they are sufficiently infrequent they cause you no pain) then your power is perfectly fine just as it is.

Those who make and sell this equipment are often guilty of claims that are at best specious, and at worst downright lies. There is usually a grain of truth to the advantages they describe, but often it's what they leave out that's the most important. Don't expect them to tell you that their expensive kit will probably make no difference - expect instead to be told that the mains quality determines how good (or otherwise) your system sounds. In reality it usually makes no difference whatsoever.

One 'interesting' claim I saw ... "Everything we see and hear through our system is really the power from our home's wall socket manipulated to make music through our speakers by our electronics. The quality of that power is critical to the success of any high-end system." While this is superficially true, it ignores the details of the "manipulation" that happens in the electronics. There's nothing subtle about it - the power supply uses brute force to convert the incoming AC into DC, and if the conversion is good enough to remove ripple and noise, the DC will be exactly the same whether it comes from a generator, wall outlet or a very expensive AC power supply. The voltage needs to be the same, and the frequency needs to be close to the design value, plus or minus a few Hertz.

Mostly, the statement is marketing BS, and has nothing to do with reality.

Many of these 'mains reconstruction' devices are basically a high-power amplifier that outputs AC at the designated frequency and voltage, having first rectified and filtered the incoming AC from the mains outlet. The output has low distortion and is regulated, and the claimed benefits cover just about every area of reproduction. According to the makers, you can't afford not to have at least one of these wonders, even (apparently) if your system sounds just fine already.

Strangely, the incoming mains quality doesn't seem to affect their power amplifier, even though it will affect yours - after all, the 'regenerator' is powered from the mains.

You need to be aware that no mains reconstruction amplifier, filter or mains lead will have much effect with many of the common noise sources. If you hear a noise when your fridge switches on or off or when the vacuum cleaner is used, then the noise you hear is probably airborne and is not carried by the mains. None of the devices described here have any effect on airborne noise, which can only be fixed by suppressing the noise at the source. That means adding a filter to the device that causes the noise, rather than trying to get rid of it by expensive devices to power your hi-fi system

1.0  'Dirty' Electricity

There is a new breed of scumbag that's emerged from the primordial slime-pit. They will hold a meter close to your power leads or wall sockets and tell you how high the reading is and how it will ruin your health in ways that you never knew existed. Those brandishing the meters never actually tell you what it's measuring, and don't expect peer reviewed medical evidence to back up the claims. It's quite obvious that the meters detect frequencies above a few hundred Hz, but there is never the slightest word about how they are calibrated or the units being measured.

For all the good they do, they might as well tell you that your mains leads have 300 litres of horse feathers per furlong. Any meter reading is utterly useless without knowing the units, the accepted safe (or 'safe' if you prefer) exposure limits, and at least some idea of what is being measured and why. I've measured the mains at home, and I'm now down to only 27 litres of horse feathers per furlong, so that must be an improvement .

The mains can genuinely have significant high frequency noise along with the (more or less) sinewave that provides the power for your appliances. Some of this noise might be audible through your system, and if so (and if it bothers you) you will need to do something to (try to) fix the problem. Mostly, it's there whether you know about it or not, and it usually causes no noises, ill health or anything else - at least until someone measures it with a silly meter and gives you a scare.

If we look at it dispassionately, anything that affects the waveform of the AC mains can be classified as 'dirty electricity', since noise is simply extra signals added to the mains at various (and often random) frequencies. Distortion is caused when the mains has harmonic frequencies of the base 50/60Hz waveform. Most distortion will be odd harmonics (e.g. 150, 250, 350Hz etc. for 50Hz mains). Even harmonics mean that the waveform is asymmetrical and contains a DC component. This can and does happen, and there is an article on the ESP website that explains how it happens and how to remove any DC offset - see Blocking Mains DC Offset.

1.1  Distortion

Part and parcel of the mains these days is some degree of distortion. Connected to the grid is a vast number of switching and transformer based power supplies, and these only draw current at the peak of the AC waveform. With enough of them, it's inevitable that there will be distortion, and this typically shows up as a sinewave with the peaks flattened, as shown below. At this stage, other noises on the mains are not being considered - only distortion.

Figure 1 - Typical 'Flat-Topped' Mains Waveform

The question is ... does it matter? Quite obviously, if mains waveform distortion made a difference to how an amplifier or preamplifier sounds, it should be eliminated. If you enjoy listening to a pure tone from the mains at 50 or 60Hz, then 4-5% distortion would be a serious problem for you. We need to examine what happens in a power supply to allow us to decide whether distortion on the mains is a problem or not.

The vast majority of power supplies used for home audio equipment (regardless of price) use a traditional 'linear' transformer based power supply. Almost without exception, these draw current at the peak of the mains waveform, and help to create the waveform seen in Figure 1. By implication and in reality, that means the the voltage that appears across the transformer primary will have exactly the same distortion components as shown above, even when presented with a pure sinewave!

Yes, you did read that correctly. Even if you have paid $thousands for a pure sinewave mains 'regenerator' or similar, the voltage across the secondary winding (after the winding resistance has been taken into account) will look just like that shown above. This happens because the mains series resistance and that of the transformer allow the voltage to collapse when current is drawn. Since current is drawn only at the waveform peaks, the peaks of the sinewave are truncated.

The only exception is if your equipment uses a switchmode power supply with active power factor correction (PFC), but these are uncommon in hi-fi systems because they add considerable cost and complexity and aren't warranted (or legally required - yet) for normal home use. In many cases, the mere mention of a switchmode power supply is enough to send dedicated audiophiles running in the opposite direction, because many feel that a high frequency switching power supply can never sound any good. Never mind that a switchmode supply with PFC has extremely good regulation and the DC output is completely free of mains frequency ripple. However, it is true that there will be some high frequency components superimposed on the DC, and these can interfere with the audio unless care is taken during design.

The slightly distorted waveform actually results in a small improvement, with lower ripple and noise than if the rectifier and filter capacitors are fed with a pure sinewave. This happens because the filter capacitors have a tiny bit longer to charge while the mains is at its peak. To demonstrate, the FFT spectrum of the ripple current waveform is shown below. The DC voltage is nominally 35V, and the circuit is shown in Figure 3 with the ripple voltage shown in blue (but not to scale).

Figure 2 - Ripple Voltage Spectrum Of Sine And Flat-Top Mains Input

In reality, the above is a bit silly, because it's looking at signals down to 100nV which can only ever be resolved using a simulator. Anything below a few millivolts is not an issue, and measurement uncertainty makes it almost impossible to measure accurately. However, the trend is clear, and the ever-so-slightly lower noise with the flat-top waveform is obvious. To make it a little clearer, I chose not to include the transformer's series resistance, so the rectifier is supplied directly from the voltage waveform. With the typical transformer and mains resistance included there is so little between them that it's of no real consequence.

Figure 3 - Test Circuit For Sine And Flat-Top Mains Input

As shown in Figure 3, the simulations did include a token 100mΩ of series resistance. The ripple voltage for the two power supplies simulated was 249mV RMS with a pure sinewave input, and 230mV RMS with a flat-top waveform. DC output voltage and current were the same for both, with 32.4V DC at a current of 980mA. I used a 10,000uF filter cap with 10mΩ ESR, and the peak input current with a sinewave was 9.15A, reducing to 6.1A with the flat-topped waveform. This shows another benefit of not eliminating the distortion from the mains - it reduces the peak (and RMS) charging current, although as noted earlier the real differences are smaller than those simulated.

So, ensuring that you have a perfect mains sinewave makes little or no difference, but the pure sinewave is actually slightly worse than the normally distorted mains in all significant respects. From this we can conclude that mains distortion is not a problem, and will not result in more ripple or noise. In fact, both ripple and noise are reduced very slightly if the mains is distorted, as is the current drawn from the mains (peak and RMS). I can guarantee that you didn't see that coming, and nor did I until I ran some representative simulations.

1.2  Impedance & Regulation

The impedance of the mains is normally quite low, and as a direct result, load regulation is at least fair. At a power outlet at home I measured the impedance at 0.8 ohm. This means that a 2,300W load (10A, the maximum for a standard outlet in Australia) will cause a voltage drop of 8V RMS, and represents a regulation of about 3.5%. While this isn't wonderful, it's generally considered perfectly acceptable and never causes any problems with sensibly designed equipment. However, that's by no means the end of the story on regulation.

The mains voltage can be expected to vary by up to ±10% from the nominal supply voltage (see Note below). So, if the mains is nominally 230V, expect it to vary between 207V and 257V. 120V mains can vary between 108V and 132V. The limits are rarely reached, but your electricity supplier cannot guarantee that you'll always get the exact voltage specified. People who design equipment know all of this, and will nearly always ensure that the equipment they make will function normally across the full voltage range.

Note: the regulations vary from one country to another, so you might find that your supplier 'guarantees' that the voltage may vary by perhaps +10% or -6% (or other similar numbers), so the above may be somewhat pessimistic. There will be exclusions though, and 'brown-out' conditions can happen any time due to network faults. A brown-out is a condition where the voltage falls (well) below the nominal for an extended period. The voltage may fall by 20% or more.

There will also be losses within your house wiring, but for typical home hi-fi systems these can generally be ignored - especially if you have a dedicated power circuit for your audio-visual equipment (which is a very common approach for 'high-end' systems). If you do use an existing circuit, use one that's not connected to the fridge or anything else that may create electrical interference.

Preamplifiers almost always use regulated supply voltages, and the regulator ICs will usually maintain the voltage within a few millivolts of the design voltage over the full voltage range. Power amplifiers rarely use regulated supplies because they aren't necessary and just add cost and heat to the product (heat because regulators are not very efficient and need substantial heatsinks). Even over the full voltage range (e.g. 207 - 252V RMS), the power difference is only 1.7dB, and that assumes that the amplifier's power supply has perfect regulation!

Needless to say, this isn't the case, but the final error we get with this simplistic approach is quite small, so the figure of 1.7dB is quite reasonable. If your system is operated so close to the limit during critical listening sessions that 1.7dB will be the difference between clean and clipping, then it's well past time that you upgrade to a more powerful amplifier. Remember that the figure of 1.7dB is the total variation, from the full ±10% mains voltage change. A more realistic ±5% variation means that the voltage will change from 219V to 241.5V, a change of only 0.85dB. This is negligible.

Not that there is anything wrong with a regulated mains supply of course. However, it has to be considered on the basis of cost vs. benefit, and for most people the cost will be far too high for the very small benefit you receive. Remember that any mains regulator device will have losses itself, so not only is the device itself expensive, but it may be very costly to run and may also add a significant heat load to your listening room. In hot weather that means air-conditioning systems will be working harder too, leading to comparatively high operating cost for a generally completely inaudible end result. Of course the heat isn't wasted during colder months, but a mains reconstruction amplifier is a very expensive room heater!

Where the published 'benefits' step over the line is when they try to convince you that a 'BrandX' mains reconstituting unit (or other fancy device) will "increase the audible detail, bass 'slam', micro and macro dynamics*, etc.". This is clearly nonsense, and is right up there with $5,000 mains cables in terms of fraudulent claims. The simple reality is that regulated mains will do none of these things, because your amplifier already has a power supply that was determined to be somewhere between 'perfectly adequate' and 'way over the top', depending on the manufacturer. Adding an external regulator is simply a waste of money if you assume that any of the alleged improvements will make your system sound better.

*   The terms 'micro' and 'macro' dynamics are pretty much the exclusive domain of hi-fi writers, and the terms have no significant meaning. The resolution (micro-dynamics if you like) of a hi-fi system is not affected by mains distortion because it runs from DC! The mains distortion is not magically transferred to the DC. Bass performance mainly depends on the amp and its power supply, not on the mains which will always have better regulation than the power supply.

As for using an external regulator and/or mains reconstruction amplifier (and that's what they are - an amplifier) for TV sets and the like, bear in mind that the vast majority of TV and other video gear use switchmode power supplies, which don't give a rodent's rectum about the incoming mains waveform. As long as the peak voltage is high enough for them to operate, a switchmode supply doesn't care if the input waveform is a sinewave or a square wave.

None of this means that a regulated mains supply isn't desirable. In an ideal world, the power to our houses would be the exact voltage intended, but this will (and can) never be the case in the real world. However, the vast majority of equipment won't care if the voltage changes within normal limits, and the result will normally be completely inaudible. Remember that the equipment manufacturer has already designed the power supply to accommodate normal variations and to minimise noise. A stabilised supply may be a good investment if you normally experience large variations, or if the voltage regularly rises to more than 10% above the nominal value.

The general principles of voltage stabilisers are described below. There are many different types, with many having fairly large steps (perhaps 5V RMS or so, sometimes more). These are probably alright for the odd industrial process, but are best avoided for h-fi. For the intended purpose, some of the commercial units may be acceptable. but you need to verify that your equipment and the stabiliser will play nicely together. As noted above, there is rarely any need.

1.3  Noise

With most equipment using a mains transformer, there is already a pretty good filter - the transformer itself. Because of its inductance (primary and leakage), high frequency noise is attenuated automatically, and common mode noise (applied equally to both active and neutral) is largely rejected. Unfortunately, most transformers don't have an electrostatic shield between primary and secondary. When fitted, this will afford excellent protection against noise coupled between the windings via the inter-winding capacitance. This notwithstanding, not very much noise can get past a 10,000uF filter capacitor!

Despite glowing recommendations from deluded users, don't expect any noise filter to make a substantial difference to your system (positive or negative). Unless you have audible noise that is determined to be due to noise on the mains, a filter will not make the system sound 'better'. Internally, your amplifier, preamp, etc. converts the AC to DC, and DC has no 'sound' of its own. The worst that can happen is that a certain amount of noise might contaminate the DC so it becomes noisy. This is actually surprisingly uncommon.

Noise on the mains covers a very wide range of possibilities. Audio frequency noise comes in many forms and has many causes. Some are even deliberate, such as the use of 'ripple control', where a medium-frequency (typically from a few hundred Hz up to 2kHz or so) is superimposed on the mains to turn off-peak and other systems on and off remotely. In addition, there are many other causes, ranging from momentary shorts (small animals and tree branches causing wires to touch), lightning, and a myriad of industrial processes.

Because transmission wires are often very long, they also make good antennas and pick up radio frequency signals. In reality, not very much of this ever gets through to your equipment, and it's not usually a problem. Clicks, pops and other noises can be created by switches, small 'universal' motors (as used in vacuum cleaners for example) and refrigerators, the latter being a common source of transient noise, especially for older models. There are countless others of course, and some will be troublesome, others not.

You don't need to regenerate the mains to get rid of noise. There are many filters that will help to clean up the mains, but some noises will defeat all your attempts to get rid of them. This may mean that either the filter doesn't live up to expectations (so return it and get a refund), or in some isolated cases the noise might be coming in via the protective earth lead. Airborne noise from nearby switching (especially motors and inductive loads) will not be reduced by a mains filter.

So-called 'surges' include large spikes created by lightning strikes and much smaller short duration spikes from inductive loads as they disconnect from the mains. It's also possible to get mains voltages that are much higher than the normal range would indicate, and these are invariably the result of a fault condition within the mains distribution system. Very high voltages (greater than nominal voltage +10%) can also be fairly common in rural areas, and may warrant a stabiliser or regulator in some cases.

Lightning is the worst thing that can happen. If it occurs nearby, it will usually cause a great deal of damage. In severe cases, nothing will survive, including the protective devices intended to prevent damage to equipment. The energy in a lightning strike is truly scary. There's an old saying that lightning never strikes the same place twice (not actually true), and I've always maintained that's because the same place isn't there any more .

Lightning notwithstanding, there is sometimes the need for a mains filter and it should have inbuilt protection. It won't save your gear from a catastrophic event (nothing will), but it will eliminate most noise and provide a measure of safety to ensure that most spikes and other disturbances will be absorbed by the filter board rather than your equipment. Having said that, I've run my system for close to 30 years in my current location without any 'protection' other than that included in the equipment itself.

Figure 4 - Typical Mains Filter

An example of a suitable filter is shown above. It includes MOVs (Metal Oxide Varistors) to help protect against transients, a common-mode and two additional chokes (inductors) to filter noise. All capacitors marked 'CX' are X2-Class, 275V AC types, and those marked 'CY' are Y2-Class electrically safe (and certified as such) types. Filters that include significant capacitance to earth are not legal in most countries, and may cause electrical safety switches (aka RCD, ELCB, GFI, etc.) to trip because of earth current.

The first inductor is a common-mode type. These offer minimal impedance to differential signals (the mains itself), but high impedance to common mode noise. L2 and L3 are normal filter chokes and these provide protection against differential noise on the mains. In extreme cases, fitting Y-Class caps in parallel with the X-Class types will help to reduce RF noise because they are ceramic types and have very good performance at high frequencies.

The 1Meg resistor may appear to have no purpose, but it's there to ensure that the X-Class capacitors can discharge when the unit is disconnected. Without it, the caps can retain a significant charge for many hours, and they represent a potential shock hazard. The resistor will discharge them to a safe voltage in under 1 second.

Figure 5 - Internal Photo Of Mains Filter/ Power Board

The power distribution board in Figure 5 is fairly typical of these products. There is some fairly basic filtering - nothing as elaborate as shown above. There's quite a number of oversized MOVs which I quite like, and there are two thermal fuses included in case the MOVs get hot due to excessive dissipation. In common with most similar units, it has protected pass-through connections for a phone line and TV antenna, and of course it comes with all the outrageous claims and guarantees that are so common with these products.

1.3.1  Ferrites

You can get some additional basic filtering by using a split ferrite block in a plastic housing, clamped around the mains lead, similar to that shown below. These can be surprisingly effective, and are often found moulded onto the leads for LCD computer monitors (mains and/or video leads). When used, it's most often because the product would not pass EMI tests (e.g. CE, C-Tick, VDE, UL, etc.) without it. This alone tells you that they are effective - both for keeping equipment noise out of the connecting cable, or preventing external noise from getting into the gear itself (or both).

Remember that many noises are airborne, and adding a mains filter will have no effect. Airborne noise (which includes RF) can enter the system via a multiplicity of methods, including speaker leads, interconnects (especially non-shielded 'audiophile' types), or even via the mains earth. Sometimes, simply passing speaker leads through a clamp-on ferrite block can help, but elimination can be very difficult and is often not intuitive.

These split ferrite cores can be particularly effective, and where noise or RF is a problem they should be fitted to all speaker leads, as close to the amplifier as possible. In severe cases, you might need to include them on signal leads as well, and you may need to use them at both ends if the RF still manages to get through. They are not always a complete cure of course, but they are cheap and generally work very well. They are usually available in a variety of sizes.

1.4  Frequency

The mains frequency is remarkably accurate in most countries, and will never vary by enough to cause any problems. Short term variations are extremely small, as they must be to ensure that the distribution grid doesn't fail completely. Your home might be supplied from several power stations at once, and the outputs from each can't even drift by a few degrees in phase, let alone by a few Hertz. It's outside the scope of this article to discuss this in detail, but feel free to look it up, or even measure the frequency with an accurate frequency counter to verify it for yourself.

It's very unusual for any mains powered device to care about the frequency - provided it is never lower than the minimum design value for anything that uses a transformer based power supply. For example, transformers designed for 60Hz may overheat and fail if used at 50Hz. See Voltage & Frequency for more info on this topic.

However, the reverse is not true, so equipment designed for 50Hz will work just fine at 60Hz, provided that the voltage can be set to suit the 120V mains. This might be via a voltage selector switch, internal jumper or perhaps an external transformer. Any AC source that claims to make the mains frequency 'more accurate' is a scam, because it's already perfectly acceptable. In addition, even if it were to drift by (say) 0.5Hz, your equipment will still function exactly the same - it makes no difference.

2.0  Voltage Stabilisers

There are several approaches to providing a stabilised mains voltage. Some may appear quite strange, such as a motorised Variac™ (variable voltage auto-transformer) that uses a servo system to physically rotate the Variac's moving contact to adjust the voltage. These can be fairly slow, but can provide almost perfect stability and regulation over the long term. This approach is uncommon, partly because it's not well known in DIY or hi-fi circles, and partly because few people need that degree of stability. See Transformers - The Variac for more information on Variacs in general. They are also expensive.

Figure 6 - Servo Controlled Variac Voltage Regulator

If the incoming mains is low (less than 230V), the Variac moving contact will be above the centre tap, and the voltage is boosted. If the mains is higher than it should be, the wiper will be below the centre tap, reversing the phase to the buck/boost transformer and reducing the voltage to the preset value. A very wide control range is available, but very fast correction isn't possible because of the motor drive. Efficiency is very good, and there's no waveform distortion.

There is another system that's very similar to a motorised Variac, and that uses tap-changing on a transformer that is designed to have a number of taps that are connected either with relays (mechanical or solid state), or again using a motor to operate a sliding contact that changes the tap in use as needed. Voltage taps may be as coarse as 5V steps or finer than 1V, depending on the application and design. The step response of these can be a problem with some equipment, as the voltage may fall to (say) 225V and will suddenly be increased to 230V in one step. Likewise, the voltage may rise to 235V before it's reduced back to 230V. It's unlikely that anyone would consider that to be a positive change in a hi-fi system. Smaller steps mean far more relays, although it's theoretically possible to use a weighted sequence such as 1–2–4–8–16 so that the control has a good range without excessive switching devices.

Another technique is called a ferro-resonant transformer. These literally use a mains frequency resonant circuit to saturate the core to a greater or lesser degree and provide very stable voltage and an almost complete rejection of noise. There's a fair bit of information on the Net, and some sites even manage to mention (often only in passing) that the output is commonly a squarewave (more-or-less), and it's not a good idea to use one to supply other transformers because the secondary voltage will be considerably lower than expected or the core may saturate. Sinewave ferro-resonant transformers are also available.

It's also possible to use a magnetic amplifier. Mag-amps (as they are often known) are a rather ancient technology, but they are still used in quite a few areas. I've seen several references to them being used for voltage stabilisation, and they show excellent stability, reasonably fast response (within a few cycles) and very high efficiency. While there is some electronic circuitry involved, it mostly operates at fairly low power levels and should be very reliable. It's probable that a mag-amp based stabiliser will beat almost any other technology for efficiency and low losses in general, but it's inevitable that some distortion will be created in the process. I don't know if that would create a problem or not because my experience with mag-amps is limited (although it's on the agenda to do some tests).

Then there are the electronic versions. These can use a rectifier and filter to produce DC, then a high-power amplifier to reconstruct the AC sinewave. Efficiency is generally rather low, and in some cases a smaller power amp will be used that is designed to only add (or subtract) the amount of AC needed to maintain the preset output voltage. This type of circuit can be extremely accurate and fast acting, and may also reduce mains distortion. However, as noted above, this is not necessarily worthwhile.

The general scheme is shown in highly simplified form below. To reduce power dissipation in the output transistors, the AC is rectified but not smoothed. While this does help, dissipation will be at the maximum when the input voltage is ~24V higher than the desired output voltage (buck mode), when the output devices are carrying the maximum possible current with close to the full rectified AC voltage across them. Dissipation is greatly reduced in boost mode, where the output voltage is higher than the input.

Figure 7 - Simplified Circuit For Electronic Voltage Regulator

The amplifier might operate from around 25V RMS via TR1, and will be able to adjust the mains voltage over the range of at least ±20V using a 1:1 transformer for TR2. If the mains voltage is low, the output from TR2 is simply added to the mains to increase the output voltage. To reduce the voltage when the mains is high, the amplifier inverts the output waveform so it's subtracted from the mains voltage. Worst case dissipation in the amplifier occurs when the incoming mains is either equal to the preset regulated voltage or above it, where the circuit has to reduce the voltage. This arrangement will work for an output of up to 2kVA or so with readily available power MOSFETs or bipolar transistors.

The same thing is done by several manufacturers, but using a Class-D amplifier which improves efficiency (up to 96% claimed), but at the expense of complexity. As shown, maximum efficiency will be around 80%, but the normal operating efficiency will be somewhat less depending on the incoming mains voltage. The worst case average dissipation in the output devices can be as high as 500W (simulated with a 1.8kVA output), which is a lot of heat to dispose of.

3.0  Isolating/ Balancing Transformers

Some users have added a mains balancing transformer, and again users will maintain that their system sounds 'better' as a result. The idea behind this is that the mains is inherently unbalanced, with the neutral conductor connected to protective earth, often at the switchboard of each dwelling serviced by a distribution transformer. For unknown reasons, many people seem to think (or even think they know) that balanced connections sound 'better'. This is nonsense, unless using a balanced connection also results in less external interference that produces audible noise.

There may be situations where the use of an isolating transformer set up to provide a floating or balanced mains supply may help to reduce noise. There is also the probability that the transformer will also reduce the regulation you expect from the mains. If some of your equipment uses a switchmode power supply, the overall noise and distortion experienced by other equipment connected to the same supply may increase. Isolating/ balancing transformers aren't a magic bullet that will make your system immune from noise. Any substantial noise reduction is likely to be the result of additional filtering that may be included with the transformer (or indeed by the transformer itself), rather than the result of balancing the mains wiring.


Ultimately, the decision to use a voltage stabiliser, balancing transformer or just a filter is up to the individual. However, claims that using an all-singing all-dancing mains reconstruction device will make your system sound better are either gross exaggerations or completely false. A sinewave input does not make your audio sound better, but anything that reduces or eliminates audible mains noise is a worthwhile improvement.

There are many myths around the mains - especially including those that involve very expensive mains cables. Most hi-fi equipment has significant filtering (mostly provided by the transformer itself and the smoothing capacitors), and that usually removes most of the noise that is carried by the mains. Once the AC mains is converted to DC, it is nonsensical to assume that there is any audible difference between DC from highly filtered and regulated power supplies and that from the same supply when it's powered via an expensive mains lead, a complex filter or a mains 'reconstruction' device. The one exception to this is where adding the device reduces audible noise.

All that any of these devices can do to change anything is remove impulsive noises or other audio frequency interference from the mains. If your system doesn't have any noises that come from the mains, then adding expensive and/or complex external systems will do exactly nothing. Sinusoidal mains don't make 'cleaner' DC, and if noise happens to get into the audio path from the safety earth then none of the options will help much - if at all.

Depending on where you are (near an industrial area for example), the safety earth might have some noise. In this case you need an electrician to install a dedicated earth stake that complies with all regulations, rather than pay for costly external gizmos that probably won't help anyway.

In the vast majority of cases, no double-blind testing is ever done by people who claim huge differences, and anyone who insists that the system's sound stage (imaging), midrange clarity or high frequency reproduction is 'better' is probably deluding themselves.

Finally, it's worth stating again that no mains reconstruction amplifier, filter or mains lead will have much effect with many of the common noise sources. If you hear a noise when your fridge switches on or off or when SWMBO (she who must be obeyed) uses the vacuum cleaner, then it's quite likely that the noise is airborne. Filters, regenerators and stabilisers will have no effect on airborne noise, and the problem can only be fixed by suppressing the noise at the source.

1.     Ferroresonant Transformers
2.     Voltage Stabilisation Techniques
3.     Magnetic Amplifier Voltage Regulator System, US Patent 3323039 A (1967)
4.     Magnetic Amplifiers, another lost technology (US Navy, 1951)
5.     AC Voltage Stabilizers & Power Conditioners


IndexMain Index
ArticlesArticles Index

Copyright Notice. This article, including but not limited to all text and diagrams, is the intellectual property of Rod Elliott (Elliott Sound Products), and is Copyright © 2014 - all rights reserved. Reproduction or re-publication by any means whatsoever, whether electronic, mechanical or electro- mechanical, is strictly prohibited under International Copyright laws. The author (Rod Elliott) grants the reader the right to use this information for personal use only, and further allows that one (1) copy may be made for reference. Commercial use is prohibited without express written authorisation from Rod Elliott.
Page created and Copyright © 25 August 2014 Rod Elliott.