Components

This has been a highly debated subject for many years. The argument on whether one resistor sounds better than the next or this tube versus that tube or connector type A or connector type B, and so on. I have some theories concerning this subject and they follow:

• When comparing an electronic component (resistor, capacitor, integrated circuit or tube) how is this comparison done? It is a well known fact that to compare two pieces of electronics, the double blind A-B or A-B-X tests are the only way to compare. To conduct an A-B test a simple switchover between A and B is done. In an A-B-X test you listen to A, then to B and then to X which is either A or B. The listeners than note which they think X is, A or B. Much has been written on this subject and I shall not go into the where's and why's but suffice to say that the human brain whilst being very "smart" does not have a good "memory" when it comes to these tests. The accepted MAXIMUM time for switchover is 5 seconds and even this is considered too long by some experts. Let me qualify this. When we compare two pieces of equipment, we assume that they are comparable and that one is not grossly inadequate. The double blind test is elegantly simple. The equipment is set up with gains set to better than 0.1dB, the listeners have no idea which piece is playing at any given time and the switchover is done by a person who is not involved in the listening test. The switchover is instantaneous so no time delay exists between the listeners hearing one or the other piece of equipment. The listening panel's results are tabulated and the results are an average.
• So an interesting question arises. How do we compare say two resistors or capacitors or integrated circuits in a piece of equipment? To do this is almost impossible. One would have to wire a switch into the gear and then the main problem arises of the time for the break before make switch that there is NO component in circuit. Assume we could overcome this, how does one compare integrated circuits which have at least FIVE connections? In this case the op-amps (Integrated Circuit) can be both fed the same signal and then a switching system (either relay or mechanical switch) can select which op-amp is being listened to. I am very skeptical of those who proclaim that they heard the difference between one resistor and another. The time interval to change these parts is way longer than the brain can remember. I have done A-B and A-B-X tests with integrated circuits but I have used either 2 amplifiers, one with one IC and the other with another OR I have done as I describe a few lines above.
• The same argument is valid for capacitors and other components which some say sound different. Please read the short excerpt in the CAPACITOR link about a blind listening test comparing a piece of wire to a cheap electrolytic capacitor.
• Tubes fall under the same 'argument". Changing a hot tube, waiting for the replacement to heat up – the human brain simply cannot remember over this "long" period of time.

Cost of components

Why do some amplifiers or preamplifiers cost so much more. My answer is simple, "Because they can". It is all marketing hype, no more and no less. Of course we accept that a manufacturer can spend a lot of dough on the cosmetics, gold plate this, chrome that and so on. Twenty millimeter (3/4") thick aluminium panels, beveled edges, machined heat sinks do add massive cost to an item and these costs are most certainly reflected in the final price of the article. But they do not contribute one single thing to how it sounds. I do admit is it cool to look at an amplifier that we know weighs more than a grown man, looks better than a beauty queen and has rows of glowing power tubes like soldiers waiting to do battle.

If one compares two products of similar pedigree and one costs several times the other, the question begs, why? I refer to the car stereo market and to amplifiers in particular. Many amplifiers have come across our workbenches at Zed. I have repaired some and tested others. I have been asked why something is what it is and why there is a substantial difference in price. I have been asked about 105 deg C capacitors, 1% metal film resistors, film type capacitors, integrated circuits and cables. Unfortunately the media and others have hyped many of these issues with garbage and misinformation. Most media writers have little or no technical knowledge and DO NOT write from a position of knowledge! Also people out there tend to believe the media and normally do not have access to good information which has not been geared towards marketing.

One of the big buzz words now is 105 degree C electrolytic capacitors. I am asked continually if our amplifiers have these. My reply is always the same. We use them where appropriate. The issue of film capacitors arises all the time. What kind of film capacitors does Zed use? Are they from manufacturer A or B? Are they in the signal path? These questions which people ask me have been asked simply because the misinformation has been put out and people tend to believe what they read or hear – without question. I write from a purely technical and not from a sales point of view. The purpose of these technical pages is to educate, nothing more and nothing less.

Many so called purists out there harp on about the fact that the signal must never pass through a dreaded electrolytic capacitor.

My answer about these capacitor issues is as follows: Even if an amplifier (or preamplifier) is DC coupled from input to output, (that is no actual capacitors in the traditional signal path) the signal MUST pass through the powersupply, loop wise. Guess what 99.9999% of all power supplies have in them, great big ugly electrolytic capacitors. So the signal passes through these ugly large value electrolytics. Furthermore think what the signal passed through in the original recording process. It had to go through a torturous maze of transformers, cheap electrolytic capacitors, lots and lots of operational amplifiers (ICs), tape heads, equalizing circuits, fader controls, really low tech connecting cables…..should I go on?

And now all the techno-boffins are worried about the last 1% of the signal chain. This makes me laugh.

One of the other questions which arise is that of film capacitors (all types) versus ceramic capacitors. Please read the link CAPACITORS on the techtalk pages.

Oh I forgot to mention that another buzz is the use of oversized, both in dimension and value, of the capacitors in the secondary supplies in the amplifiers. Some people seem to think that the use of high value electrolytic capacitors in this part of the circuit helps the amplifier to perform better at low frequencies. This is just not true. I do not advocate the use of wimpy low value capacitors but there is a point of no return. Those large and heavy home amplifiers with the great big computer grade electrolytics (or lots of smaller ones in parallel) have to work from a 50 or 60Hz power line. With full wave rectification as standard, the capacitors are recharged 100 or 120 times a second. In large home amplifiers we need large value capacitors in the power supply. However in a mobile amplifier with the power supply frequency at say 50KHz, these capacitors are recharged 100,000 times per second. That is 2,000 odd times more often than the home amplifier!

The power which an amplifier can deliver is 100% dependent on the power transformer alone (assuming the power source can deliver – in a vehicle the battery + alternator, in the home the wall socket). Power supply capacitors contribute ZERO to the continuous power rating of an amplifier. It is synonymous to saying that wide tires on a car add horsepower!

Whilst we are on the subject of "buzz" items, the infamous TO-3 power transistor comes to mind. These look like they do simply because they were designed to fit into the same space as the octal tube socket for power tubes. Before the advent of plastic encased power transistors, the TO-3 was the only game in town. All one has to do is look at it and see that its physical construction is awkward at best. It requires no less than four mounting holes and to dissipate the heat correctly equal torque must be applied to both mounting screws. They require either insulating shoulder washers for each screw or an expensive socket which resides on the opposite side of the heatsink and has threaded inserts to accept the mounting screws. Wow, this was a mouthful just describing the TO-3's mounting criteria! TO-3 devices are typically rated from 50 watts to 250 watts of dissipation. Of course in the real world there is no TO-3 capable of dissipating 250 watts. All one has to do is refer to the SOA curves for a particular device and take any point on the curve and read off the volts and the amps. Multiply them together and look at the results. On the secondary breakdown part of the curve the V-I product never comes close to the rated dissipation number. Let's look at the popular TO-3 pair MJ15003/MJ15004. These are epi-base devices with an Ft of 2MHz and a gain of 25 with 5 amps of current. They are rated at 250 watts. At 100 volts they can pull 1 amp of current (100 watts), at 70 volts they can pull 2.2 amps (154 watts). On the thermal limitation part of their curve (single pulse signals only) the numbers are: At 50 volt they can pull 5 amps (250 watts) at 20v they can pull 12.5 amps (250 watts). So what can we learn from these numbers. First they are theoretical only and of no use in practice. The device certainly cannot dissipate 250 watts and if the temperature rises above 25 Deg C these current numbers MUST be derated (As all semiconductors are).

Plastic devices are no better than TO-3 and TO-3 are no better than plastic in terms of power dissipation if used within their limitations.

Plastic power devices are so much better, easier to mount, less expensive, better specifications, large selection of device types and more readily available. Why the big urge to use and see TO-3 devices? At the dawn of the car stereo age certain manufacturers used them, as higher power plastic devices were not available. The TO-3 devices were available in high voltage and high current types. The only plastic power transistors of consequence, was the TIP35/TIP36 pair. Whilst these have massive current capability they could not dissipate the heat due to their smaller base plate. About 20 years ago along came Toshiba with the TO-3P packaged 2SC3281/2SA1302 complementary pair which is similar in base plate dimensions to the TO-3 but it in a plastic housing. Either single hole mount or as most prefer bar or clip mount. Why was this device so good? That is easy. It had a high breakdown voltage (200v), it is rated at 150 watts and has a collector current of 15A. Where it absolutely outperforms the TO-3 devices is in its linearity, Ft (transition frequency) and gain versus current. These old TO-3 devices are slow hogs with Fts ranging from 800KHz to 4MHz where the Toshibas have Ft of 20MHz. Sanken makes plastics with FTs of 50MHz. Both Toshiba and Sanken now make many complementary power transistor pairs to satisfy all power requirements.

One supplier of amplifiers in the USA who brings in cheap Korean product made by Abyss Audio, uses TO-3 devices and then quadruples the price and more and claims that they are high end, boasts that they use TO-3 devices with a breakdown voltage of 250 volts. (These are the MJ15023/MJ15024 pair). Big deal! What does the 250 volt specification do for them in circuits that are typically limited to less than 120 volts? NOTHING AT ALL. This is just advertising hype by people who really know little about what they write. If the breakdown voltage of a semiconductor device is 10 or 20 volts above the voltage used in the amplifier that is safety margin enough. You can use a 10,000 volt transistor and you gain ZERO!

Sorry for moving sideways. But all the techno boffins who claim the superiority of TO-3 devices need an electronics lesson. They are more expensive than plastics and have ZERO advantage. My philosophy is "use enough power devices" and you will have a reliable amplifier, all other things being equal.

One BIG advantage of using lower power devices but more of them is the following. The current gain of a bipolar transistor goes down as the current through the device increases. So those using single TO-3 devices, where we at Zed may use two or three plastic types in parallel, are at a distinct disadvantage. The current gain of our output stage is significantly higher than the single TO-3 because the current through EACH of our multiple paralleled devices is lower and thus the current gain is higher. This means less current demand on the driver stages and lower distortion. The MJ15003/4 devices I mentioned earlier in this document have a current gain of 25 with 2 amps of current. In the same amplifier we would use TWO of our lower powered plastic devices from Sanken and their current gain at 1 amp is 180. So the gain of our output stage is 7.2 times greater, we have 20MHz devices instead of 2MHz and the total dissipation of our output stage is similar to the TO-3.

At Zed we choose the type of component which will do the job sonically and reliably. We use surface mount components wherever possible and the reason is reliability. Consider the following. Any component which has wires must be placed on the PCB and the wires must go through holes on the PCB. The holes on double sided material are through plated. This means that there is plating on the inside surface of the hole. The process of making the board is not perfect and so there is the chance of bad plating (We have seen this many times). With surface mount there are NO holes. The component is "shot" onto the board by high speed auto insertion machines. This method of assembly is many times more reliable than through hole insertion. When a leaded component is inserted onto the board, it's leads are "clinched" (bent) on the backside of the board to prevent it from falling off during handling prior to flow solder. These bent leads can, and do touch other leads if care is not taken. With surface mount there is no lead to bend.

Of course the larger components must be leaded types (Large capacitors, power devices, transformers, etc.) but these are all hand inserted and so little stress is placed on them.

Connectors, switches, potentiometers and other mechanical parts:

A lot of common sense prevails here. Power and speaker connectors can cost less than $1 or up to $3-4. Depending on the power requirements the sheer size will determine this. Most high quality amplifiers use either custom or production connectors which have set screws to clamp the cables and are usually made of brass which is normally gold plated.

At Zed we use high quality RCA sockets which do not break. They use Teflon inserts which are much better than the plastic used in most cheap RCA sockets we see on most car amplifiers. The supplier I mentioned above blabs on about their TO-3 devices and then puts penny RCA sockets on their amplifier (well Abyss in Korea do this).

The design of the chassis and all the other bits and pieces influence the cost of a car amplifier tremendously since the chassis is the single most expensive component of the amplifier.

A potentiometer is a typical "high" distortion component. Ideally if we were able to control the level or volume of an amplifier without the use of a traditional carbon film or plastic film potentiometer it would be better. At Zed Audio we use the potentiometer in our level stage in a shunt configuration. This means that all the actual potentiometer is doing is shunting signal to ground. The signal never passes through the control.

Below are some quotes from a well renowned engineer where he talks about cables. The same applies to some of the components used in amplifiers.

If one were to attempt to summarize the objectivist argument in the proverbial 100 words or less, it might read as follows: "High-priced cables are based on voodoo science, designed for gullible consumers who are so swayed by their cost, looks, and status symbol appeal that they delude themselves into believing they hear differences when such differences do not exist. The proof that the differences do not exist is that they are neither measurable nor provable in blind testing."

It is human nature to have an opinion based on such perceptions. When someone says, 'I hear a difference between this and that,' it is an opinion.

"Now, why is it that our senses fail us so? A big part of this, especially when we're talking about audio, has to do with the fact that what seems straightforward and reasonable is not. It seems straightforward and reasonable to listen to a piece of audio equipment and then develop an opinion on how it sounds. Unfortunately, due to the complex way that we humans take in sensory information and then combine it with prior knowledge and experience, the resulting perception may be incorrect. If we really want to know how something sounds, we must separate out prior knowledge and visual cues and force ourselves to only use our ears. This is why we do the test blind. Our eyes play a HUGE role in our perceptions of audio quality. Counter intuitive? Yes, but true.

"Now there are perfectly good reasons to buy expensive power cords. For one they look cool. 'The better it looks, the better it sounds.' (This has been shown to be the case in studies that compare sighted to blind tests using identical equipment.) Another reason to buy them is that they cost a lot. They are status symbols. Status is good in our culture. Fear is another reason. Fear of not doing everything you can to get 'the best sound.' Of course there is plain old faith. If you believe that they make your stereo system sound better, then they do! If having them makes you feel better, makes you more relaxed when you listen to music, gives you comfort in knowing that you have left no stone unturned in your pursuit of musical enjoyment, by all means, go for it!"

This specification is not often quoted in multi-channel equipment. I personally feel that it is somewhat of an over rated specification. Here are my reasons.

The ultimate signal source is still the vinyl LP. For those who use only CD as a playback source they are really missing out. The compact disc is convenient but it cannot compare to the best that vinyl has to offer. Of course the playback system for vinyl must be of top quality to take advantage of its performance. The cartridge, turntable and phono preamplifier must work in harmony. Now what does this have to do with channel separation? The moving magnet or moving coil (the two most popular types) limit the amount of channel separation possible. A very good cartridge may have 35dB of channel separation at 1KHz and this drops off at higher frequencies. What does 35dB mean? It means that the "leakage" between channels is one fifty sixth of the signal level. If we could magically remove the signal from one of the cartridge's output terminals, it would still output a signal which is coupled from the other channel but it would be at 1/56 of the level of that channel and again do not forget that this is at middle frequencies only. But for the sake of argument, let us accept this 35dB figure as being broadband.

For ease of arithmetic I shall pick some easy numbers from which we can see how this affects what we hear. Let us assume that the one channel of the cartridge has an output of 1mV (quite typical of magnetic cartridges and even some very high output moving coil types). The other channel into which the signal is "leaking" will output 1/56 of this 1mV (0.001v). This means that it shall have an output of 0.018mV (0.000018v).

The typical amplification signal chain is 40dB for the RIAA preamplifier, 20dB for the high level amplifier (Typical home preamplifier) and then another 26dB for the power amplifier. Let's see how much voltage we end up with at our loudspeaker. 40dB = 100x voltage gain, 20dB = 10x voltage gain and 26dB = 20x voltage gain. So the 1mV is x20,000 =20 volts and the other channel is 0.36 volts.

OK what do we have now? 20 volts across a 4 ohm speaker is 100 watts and 0.36 volts is 0.0324 watts (32.4mW). The separation is still 35dB but consider the power numbers! Thirty two thousandths of a watt, as compared to one hundred watts. I challenge anyone who has one speaker been driven with 100 watts of average power to even hear the other with 0.0324 watts of average power. So who needs more separation? I am not saying that we as electronic designers should not strive to have the channel separation as good as we can make it, but not go overboard with it. It is not that difficult to design equipment with channel separation of over 80dB at middle frequencies. The channel separation shall ultimately be determined by the phono cartridge. CD users do not have to contend with these issues.

Consider for one moment that our signal source is a CD player and typically they have excellent channel separation. If our amplifier "only" has an 80dB separation specification, this means that the un-driven channel shall have one ten thousandths of the signal from driven channel, injected into it. So using our 100 watt amplifier above, the un-driven channel shall have an output of 0.000001 watts (1 microwatt). Hardly worth worrying about I would think. A separation figure of a mere 50dB shall yield the un-driven channel to have an output of 0.001 watts (1 thousandth of a watt).

There seems to be a lot of hype and mystery concerning the sound of capacitors, the quality of capacitors and what capacitors actually do in a circuit. There are many types of capacitors including ceramic, polyester, polypropylene, polycarbonate silver mica, tantalum and electrolytic. Each has its own area of "expertise". One type of capacitor will perform well in a particular application and perform poorly in another. A capacitor is simply two metal plates separated by a dielectric. An electrical charge is stored between these two plates. The plates and dielectric can be made form several different types of material. The closer the plates are, the higher the capacitance and the larger the area of the plates, the larger the capacitance. The dielectric material affects the capacitance as well. Here is some brief information about the particular types.

Note: The unit of capacitance is the Farad. It is a very large unit and so we use the following to express capacitance

Microfarad is one millionth of a farad.

Picofarad is one millionth of a microfarad

So 0.001mfd is equal to 1000 picofarad (pF)

The above diagram shows a simple representation of a capacitor.

The capacitor is shown in RED.

In parallel with the plates of the capacitor is a resistance, Ins-Res of the insulation. Teflon has the highest resistance with polystyrene, polypropylene and polycarbonate coming in second. Third is polyester and then the ceramics COG, Z5U and XR7 with tantalum and aluminium last.

Dap< is the Dielectric Absorption. All capacitors when charged to a particular voltage and then the leads are shorted, will recover some of their charge after the short is removed. The ratio of the initial voltage to the recovered voltage is expressed as a percentage. In general, electrolytics are the worst and film types are the best.

The series inductance is shown simply in series with the capacitor itself. Electrolytics have the highest inductance mainly due to how the capacitor is made. The higher the frequency the higher this inductance becomes.

The ESR of the capacitor is the resistance that appears in series with the actual capacitor. The higher the voltage and value the lower the ESR. This resistance is made up of dielectric loss, lead termination resistance and the el ectrode resistance. This is nearly constant with frequency.

The dissipation factor or DF is the ratio of ESR divided by the capacitive reactance(Xc) which is given by this formula (Xc = 1/6.28 x F x C) where F is frequency and C is the value in Farads. Normally of course we have values in "Mfd" and so the the value of C can be in Mfd and then the numerator becomes 1,000,000 instead of just 1.

Ceramic capacitors are used in high frequency circuits such as RF. They are also the best choice for high frequency compensation in audio circuits. Now some may snub their noses when hearing this. We used to manufacture some high end amps for a Japanese company and they demanded that NO ceramics be used in the amplifiers. Pretty much all amplifiers have some sort of high frequency compensation to prevent them from oscillating and instability. The frequency at which these ceramics were doing their work was at 240KHz. Now I do not know about all of you out there but my hearing does not go out THAT high. Maybe these guys were distant relatives of the bat!

They come in values from a few picofarads to 1 microfarad. The voltage range is from a few volts up to many thousands of volts. Ceramics are inexpensive to manufacture and they come with several dielectric types. Types XR7 and Z5U are the least stable as far as temperature is concerned. They have a higher dielectric constant than the higher stability types like COG. The tolerance of ceramics is not great but for their intended role in life they work just fine.

Tantalum capacitors are made by depositing a film of oxide on tantalum. These are polarized types and are smaller than their aluminium counterparts. They are low voltage types only with a maximum rating of about 40 volts. We at zed do NOT use these as they are notoriously unreliable. They have a bad tendency to go leaky. I will NEVER EVER use a tantalum capacitor as they are so unreliable and my experience many years ago bears that out.

Aluminium Electrolytic capacitors are made by depositing a film of oxide on aluminium foil. The foil is formed for a specific voltage rating. These are polarized and of courser do not tolerate having reverse voltage applied to them. (Anyone been around when one of these larger value babies explodes – it is not a pretty sight?) They are also not happy campers if the rated voltage is exceeded (same thing, they will make a mess of the equipment and your face if you are too close) BUT higher quality types will tolerate about 5% over voltage. What happens in these capacitors that if one applies say 37 volts to a 35 volt capacitor it will actually reform its foil over time to the new applied voltage but its value will drop to keep CV a constant. Conversely if a lower voltage is applied it will reform to this new voltage and the capacitance will increase. Now do not get all excited and take your 10,000mfd 50v capacitor and use it at 25 volts and expect to get 20,000mfd out of it. There are limits to what these guys will do. It is also not a great idea to run electrolytics at voltages well below their rated voltage. A rule of thumb is about 25% lower voltage than rated is OK. They are normally made by winding the foils around in each other in a cylindrical way. High capacitance is easily obtained.

The ESR is the Equivalent Series Resistance and the higher the value and voltage the lower the ESR. The lower the ESR the less heat the capacitor will generate when current is drawn from it. Also closely related to the ESR is the available ripple current that a capacitor can tolerate. This is mainly of concern in power supplies. Most manufacturers offer many grades and sizes of electrolytic capacitors. There are of course both through hole and surface mount types. Within each category there are sub categories. Through hole types offer many more variations than surface mount. There are 85 deg C and 105 deg C versions leaded, snap in and screw type terminations. Electrolytics, especially those used in high current power supplies have a fixed lifespan and once a electrolytic decides it is tired of living, then it is off to "the pie in the sky". Typically their life span is from 1,000 to 3,000 hours depending on the quality.

Their tolerance is not good but then again a low tolerance component is not essential. Typically the value can vary from -50% to +100% of the nominal value.

There are non-polar electrolytics and these are mainly used in passive speaker crossovers.

Silver Mica Capacitors are one of the best types of capacitors. They have excellent stability and are available in low tolerance values down to less than 0.1%. They ARE sensitive to heat and are now used mainly in RF and tuned circuits. I like them in RIAA preamplifiers as I think they do sound better in that application.

Film capacitors encompass polyester, polypropylene, polycarbonate and others. Each has its own strengths and weakness. These are normally used in audio for filters, equalizers and power supply bypass duty. They are available in almost any value and voltages as high as 1,500 volts. They come in any tolerance from 10% to 0.01%.

Well now that you almost know all there is to know about capacitors (Only kidding) it is time to discuss how they sound and why we use certain types in particular applications.

Power supplies generally demand the use of electrolytic capacitors because they have high values in small packages. The value is determined by how much ripple can be tolerated and the voltage is determined by the voltage of the power supply (Duh!). Because this type of capacitor has inductance, it is normal practice to bypass them with film type capacitors in order to improve the high frequency characteristics. In 50/60Hz supplies I have never found these bypass film capacitors do anything to improve the sound.

Coupling capacitors are usually electrolytic (Yes this is not a typo error) and film types. The value of the coupling capacitor is usually determined by the load impedance which the capacitor "sees". If the value is too small for a given load impedance then the low frequencies will be attenuated at a rate of 6dB/octave.

Let us examine the circuit above. The capacitor has a reactance which forms a potential divided with the 10K ohm resistor. This circuit may represent the coupling of one circuit block delivering "Vin" at 1 volt to another circuit block shown as "Vout". We can consider the capacitor as a resistor whose value changes with frequency which is really its reactance (Xc).

So the formula for Vout is [Vout/Vin = 10,000/10,000+Xc] and [Xc = 1/6.28xFxC]

We can now throw some numbers into our formulas and see what comes out. Let us pick a frequency of say 1KHz and a capacitor value of 0.22mfd as a first example. Solving for Xc we get Xc = 723 ohms. Solving for Vout we get 0.932 volt. This means that at a frequency of 1KHz a 0.22mfd capacitor will cause the output voltage to drop to 0.932 (a 0.61dB drop). The frequency at which Xc = 10K ohm is 72.37Hz (I simply solved for F in the formula above)

As a coupling capacitor in a full range circuit the 0.22mfd is clearly inadequate since it rolls off the response from over 1KHz and is 3dB down at 72.37Hz. What do we do to calculate the value of C which will allow decent low end response? We shall pick a frequency at which we want the signal level at Vout to drop to -3dB as low as we desire. We shall choose 2Hz. Solving again for C in the Xc formula above we get 7.96mfd. We would then use a 10mfd. This 10mfd can be a regular polarized electrolytic, two electrolytic capacitors wired back to back with common polarity terminals joined, a non-polar version or a film type. The film type will be large and expensive. In op-amp circuits where the DC supplies are less than 20 volts, a film type rated at 25 or 63v is not that large. In tube circuits however where there are hundreds of volts, a 10mfd 450v capacitor is a large specimen. At Zed we go one step further and increase the value fivefold so the -3dB break frequency goes down well below 1Hz. My opinion is that if the value of the capacitor is so large so as to make a break point well below 1Hz the electrolytic capacitor does not degrade the sound.

An amusing story: About 14 years ago Zed Audio was building subwoofer amplifiers for a large well known speaker company in Chatsworth California. One of the projects involved a high pass active preamplifier whose output was to be 6dB/octave high pass crossover (same as the circuit above) with some simple elegant electronics. The BIG question came up from management and sound gurus whether one could tell the difference between various high quality film capacitors. I said "no" and I got hammered for this comment. I set up a test where I said they could not even hear a cheap electrolytic, never mind a film type. "You are crazy, full of you know what" were some of the comments I received. So we set up a double blind test. We had a fancy turntable, Audio Research tube preamplifiers and power amplifiers, speaker cable as thick as your arm and all the other high end toys needed for a sound system. Between the preamplifier and power amplifier a two pole switch was inserted and the switch was to select either a dead short or this one penny electrolytic capacitor. I soldered the capacitor to the switch but unbeknownst to the audio boffins, I wired a short across the capacitor so in either switch position they were listening to the same thing – a dead short, a piece of wire! This was going to be fun – I knew that.

The gurus put their favourite album on the turntable and away we went. One of their technicians was flipping the switch at the listeners' command. Back and forth we went for over ten minutes with all saying "Yes that's the capacitor, no that's the wire". So we stopped and I called all these gurus over to the switch and showed them the dead short across the capacitor. Red faces, curses etc and I was a bad boy and they were fools. Most of this "component" sound is in one's head. You have to hear the difference after you have spent $55.00 on your new coupling capacitors!

Power supply bypass capacitors are sometimes required in low level circuits. These should be a pair of ceramic capacitors (Yes ceramic as they perform best at high frequencies) placed as close to the power supply pins of the integrated circuit or discrete circuit block and some electrolytic capacitors placed reasonably close. Film capacitors will work fine but as they are more expensive than their ceramic counterparts I see no reason to use them.

Switching power supplies require the use of either low ESR capacitors or a larger quantity of "regular" ESR capacitors. It makes absolutely no difference which method is employed as long as the final ESR and ripple values are arrived at.

High frequency compensation capacitors in audio circuits always operate at very high frequencies. In our opinion the best type to use are ceramic capacitors. Some disagree and claim that film types are better. Since these compensation capacitors are used to roll off extremely high frequencies I seriously doubt that one could hear the difference. Unless you are of course a relative of our beloved little bats.

Listening to different capacitors is a time consuming. In order to compare two different types of say coupling capacitors in an audio circuit requires that you install the two types and a double pole double throw switch (DPDT) so that a direct A-B comparison may be done. The switch should be a gold plated low contact resistance type. Ideally you must install the switch at the point in the circuit where the capacitors live. This may sometimes be difficult to do because it may not always be simple to solder in the parts. The "crook" method is to do the switching between two pieces of equipment like a head unit or preamplifier and the power amplifier. The simple circuit is shown below with only one channel shown for clarity.

The two capacitors marked C1 and C2 are to be compared. The double pole double throw switch (DPDT) is toggled between the two capacitors. The Resistor shown as "x" ohm is what the capacitors "see" as their load impedance and this resistor is normally the input resistor of the next circuit block or equipment.

This subject along with some others has to be at the top of my top ten peeve list. In the past years there has been so much hype written and sprouted by untrained mouths that it could fit in a 4 inch thick book. When profits on traditional stereo components started going south, the snake oil and magic suppliers had to find new avenues for revenue – and they surely found one in cables. Interconnect cables, speaker cables and power cables. As the saying goes. "There is always room for one more sucker" and boy there were and are lots of those suckers running around. But there is hope and another saying goes like this, "You can fool some of the people some of the time but you cannot fool all the people all of the time".

There is no doubt in my mind that there are bad cables out there. Bit it is like an amplifier designer deliberately designing a real bad sounding product. Why would a cable manufacturer go out and make bad cables. Yes they do exist but the average audiophile should never come into contact with these products.

The following discussions will be for audio frequencies and so any frequency above 100KHz will be ignored. At last check my hearing did not go near 100KHz!

Interconnect cables are the first cables which the signal passes through from signal sources to preamplifier (in the home stereo) and from head unit to processor or amplifier (in the car stereo). We shall only discuss cables with RCA plugs at each end as they are by far the predominant cable type in this category. As in any interconnection, there are THREE main issues to consider. First the source's output impedance, secondly the cable's electrical impedance and lastly the receiving gear's electrical input impedance.

Note:Please read my link on Balanced line as well

The source has a finite output impedance and in the case of CD players, FM tuners, etc. this is in the order of 10 ohms to hundreds of ohms. The lower this source impedance the better. Most of these sources use op-amps as their final output stage. I shall talk about op-amps often in the various tech talk subjects and these are typically integrated circuits. They are in essence small "power" amplifiers with microwatt or milliwatt capability. Most op-amps do not like capacitive loading. All cables exhibit capacitance between adjacent conductors and they present a capacitive load on the preamplifier output stage. The simplest way to cure this is to insert a low value series resistor in the output of the op-amp which isolates the output stage inside the op-amp from the capacitor. The value is anywhere from a few ohms to hundreds of ohms, the latter being more typical. So what we end up with is an output stage with a series resistor in the "hot" leg and hanging on this is a shielded cable. Let us examine the drawing below.

The op-amp of the source equipment is at the left side. It has an output impedance of Zo and in series we have added a 220 ohm resistor (typical) which isolates this op-amp from the capacitive loading effects of the cable AND the input capacitance of the receiving equipment. I have lumped these together as "Cin".

The shielded cable also has inductance. This is determined by the area between the inner and outer cores and the diameter of each core of the cable and is proportional to this. Typically inductance in interconnects is very low indeed and can be neglected in our audio discussions. Of course at very high frequencies it becomes a factor since inductance (L) is proportional to frequency (F). The inductive reactance is XL = 6.28 x F x L. (The 6.28 = 2 x Pi where Pi is 3.14)

The shorter the cable, the lower are its reactive components being inductance and capacitance. This stands to reason as cable has xx micro-farads of capacitance per unit length and yy Henries of inductance per unit length. If we could make our cables less than say 6.35mm (0.25") all would be well in audio land since then the cable would be essentially out of the picture. Of course this is impossible and so we must deal with practical cables.

Back to our op-amp and let us examine the effect which the cable has on the performance. The series resistor which we have manually inserted in the output of the op-amp (The Zo of the op-amp is added to the 220 ohm we added) forms a low pass filter with the lumped cable and destination equipment's input capacitance. Zo depends on the amount of negative feedback which is applied in the source. Let us assume it is 50 ohms which is typical for most op-amps. We now have a total series impedance of 270 ohms. Cable capacitance in either those high end "snake oil" types or the RCA interconnects which come with the average VCR are about 150-195pF per metre or 50-60pF per foot maximum and typically are far lower. I have some cheap RCA-RCA cable at the factory which has a capacitance value of only 8pF per foot.

Note:(0.001mfd = 1000pF)

In a home stereo environment interconnects are relatively short, mostly less than 1 metre in length and in automobiles the longest is maybe 5 metres. Let's do some calculations based on these two cases and also taking into account the input capacitance of the destination equipment. These examples only take the series output resistors of the source into account.

In the case of the 1 metre cable, C-lump is 50pF + Cin (say 220pF) = 270pF

C-lump forms a potential divider with the 270 ohm resistors. The capacitive reactance of C-lump (C) is Xc = 1/6.28 x F x C.

At 1KHz Xc = 589,761 ohms and at 20KHz it is 29,488 ohms and at 100KHz it is 5,897 ohms.

So now using our formula to calculate Vout we arrive at the following. (Vout = Xc/Xc+R)

1KHz : Vout = 589761/589761+270 = 0.999 of the input which is 0.999 volt, a drop of 0.0086dB hardly audible I would imagine.

20KHz: Vout = 29488/29488+270 = 0.9909 of the input which is 0.9909 volt, a drop of 0.079dB again not audible unless you are bat!

100KHz: Vout = 5897/5897+270 = 0.9562 of the input which is 0.388dB which is barely audible EXCEPT that our hearing does not go anywhere close to 100KHz! (Unless you are one of those cable manufacturers who claim that we can hear several octaves above 20KHz but then again they may be relatives of the bat)

In the case of the 5 metre cable, C-lump is 250pF + Cin (say 220pF) = 470pF

At 1KHz Xc = 338,799 ohms and at 20KHz it is 19,439 ohms and at 100KHz it is 3,388 ohms.

So now using our formula to calculate Vout we arrive at the following. (Vout = Xc/Xc+R)

1KHz : Vout = 338799/338799+270 = 0.999 of the input which is 0.999 volt, a drop of 0.0086dB hardly audible I would imagine.

20KHz: Vout = 19439/19439+270 = 0.986 of the input which is 0.986 volt, a drop of 0.112dB again not audible.

100KHz: Vout = 3388/3388+270 = 0.926 of the input which is 0.66dB which is audible EXCEPT that our hearing does not go anywhere close to 100KHz! (Unless you are one of those cable manufacturers who claim that we can hear several octaves above 20KHz but then again they may be relatives of the bat).

If we lower the value of the resistor we insert in the output lead of the preamplifier to say 50 ohms, you may redo all the above calculations and the differences are not worth discussing. Also reducing the value of the destination equipment's input capacitor ** will also not change the numbers by any significant amount. Do the mathematics for yourself. Even if cable manufacturers eliminated nearly all capacitance, the input capacitance of the destination equipment will be dominant. There are cables out there that have very low capacitance of around 26pF per metre (8pF per foot) and the results with these cables would be mathematically different but not audible.

So what have we learned from the above calculations. Yes we must use low capacitance cable. We must know what the input capacitance of the equipment is and we must also know the total output impedance of the source equipment. Knowing these parameters we can easily calculate the dB drop at any frequency. Of course at lower than say 15KHz the drop is of ABSOLUTELY no consequence.

Another point to consider is that the low output impedance of the source will tend to discharge the cable/equipment capacitance very quickly so this notion of the charge held by the cable is nonsense.

Almost all manufacturers DO NOT specify the input impedance of their equipment in a 20Hz-20KHz bandwidth. If they did it would be relatively simple to work out the capacitance at the input. The amplifiers which Zed Audio produces have low pass 6dB/octave filters at the RCA input. The 3dB point of these filters is set at 339KHz. The reason for these filters is to prevent very high frequencies from entering the input stage of the preamplifier.

The DC resistance of these cables is very low and even if it was 1 Ohm per metre this 1 Ohm is not even a factor compared to the series output impedance of the final stage of a preamplifier and then any added resistance completely swamps any cable resistance. This is easily confirmed by anyone with access to an Ohmmeter. Just measure the resistance from RCA tip to RCA tip of your favourite 5 metre cable.

As I stated above, the inductance of interconnect cables is very low. A reasonable quality interconnect should have inductance of about 0.48 micro Henry per metre (0.15 micro Henry per foot) What does this inductance do to our sound? Let's examine the following diagram.

What we have is the lumped inductance of the 5 metre cable shown as "Lx" and it's value is 2.4 micro Henry. Typical cable inductance is about 0.48 micro Henry per metre (0.15 micro Henry per foot). The inductive reactance of a 2.4 micro Henry inductor is XL = 6.28 x F x L

As we see from the formula, the higher the frequency, the higher the value of XL. This value of XL will add to the value of the two resistances, Zo and Rx. Three of these will form a low pass filter with C-lump. So the total series impedance is Zo +Rx + Lx. Lets us calculate this at three frequencies again.

1KHz: XL = 6.28 x 1000 x 2.4/1,000,000 = 0.015 ohms. Not worth considering since Zo and Rx are 270 ohms!

20KHz: XL = 6.28 x 20000 x 2.4/1,000,000 = 0.3 ohms. Not worth considering since Zo and Rx are 270 ohms!

100KHz: XL = 6.28 x 100000 x 2.4/1,000,000 = 1.5 ohms. Not worth considering since Zo and Rx are 270 ohms!

So the inductance of a 5 metre cable is totally negligible and certainly on a 1 metre cable it is even less of a concern.

The material of the cable is some what important. The cable should be flexible, so multi-strand cable is generally used. The insulation should be of good quality and the RCA jacks at each end should be durable and not clinch the female RCA sockets on the equipment. All the mumbo jumbo about OFC wire, silver wire, ten thousand stranded wire is nonsense. The reactive properties of these cables are affected to such a small degree by the wire material used for their construction that it is hard to believe that they contribute anything. I am not saying for one moment that one should use interconnects made from slivers of iron but simply that if reasonable care is taken in the manufacture of cables they will not affect the sound. There has never been any proof that the dielectric properties of the insulating materials affect the sound in any way.

Note:A very easy way to compare two RCA cables is to use a high quality signal source such as a CD player (A high quality turntable with an outboard phono preamplifier is even better). Use a Y-adaptor and connect the two sample cables to the Y-adaptor and then one cable to any high level input on your home preamplifier and the other to the tape monitor loop. The Y-adaptor cannot impart any change to the sound – I do not believe it. Also the tape monitor switch contacts are in circuit whether the switch is in "source" or "monitor" positions. Now sit in your favourite armchair with your favourite music playing and ask someone to alternately flip the tape monitor switch to either position at your command. You of course cannot know which position it is in at any given time. There will be NO difference in the sound with any pair of test cables you choose. I have tried this with cheap VCR cable and $900/foot cable – no difference whatsoever. Rather use acceptably good cables and use your hard earned money for something else. Nearly all who buy useless expensive additions to their sound systems (be it car or home) have to "hear" a difference in order to justify their expensive toy.

Skin effect is often talked about. The electrons tend to flow on the outside surface of the wire as the frequency increases. The following equation approximates the difference between the DC resistance and the AC resistance of a strand of cable.

Res AC = Res DC x n x square root of freq (MHz) and n is the wire gauge factor. For typical interconnects it is about 7.

So at a frequency of 1MHz, Resistance at AC = 10 x 7 x 1 = 70 ohms. Assume that we have a high DC resistance cable of 10 ohms. The 70 ohms added to all our above equations does not change the answers significantly especially that I have just done this at 1MHz. This equation is not 100% accurate for lower frequencies but is a good enough indicator. Plug in say 50KHz to the above equation and we get 3.5 ohms of AC resistance. This is of course lower than the DC resistance but what is clear is that at all audio frequencies, skin effect has NO effect on the sound.

One cable manufacturer says the following and I quote

· it has to be high quality, solid core silver

· diameter must not exceed 0.5mm (skin effect and its implications)

· insulation has to be as thin as possible (dielectric absorption, electrostatic micro-discharges)

· insulation has to be as natural as possible (unbleached cotton)

· connectors have to use as small as possible quantities of metal

· sensitivity to vibrations

What a lot of rubbish if ever I have heard.

Another respected home amplifier manufacturer states the following and I concur 100%

The main implications are that the cable used should have reasonably low values for its capacitance and d.c. resistance per metre. From the above, a capacitance of around 100 pF/metre or less seems likely to be adequately low for interconnects that are no more than 2·5 metres long unless the source impedance is significantly higher than 600 Ohms. In practice, most good quality domestic audio sources are likely to have a source impedance below 600 Ohms, and the interconnects employed may often be only 1 metre in length. Thus even keeping to no more than 100 pF/m seems to be erring on the side of caution.

When the load impedance is much higher than the impedance of the source and the characteristic impedance of the cable the signal current is likely to be relatively small. Since this is true in most domestic systems it seems reasonable to expect that effects due to interconnect co-ax inductance, series resistance, and internal impedance should be very small, and it is questionable whether they are audible. Given its shielding properties co-axial cable seems a good choice for interconnect provided that we follow the general implications drawn above.

Speaker Cables are the next on our list. They are different from interconnects because of the source impedance and the load impedance. The same situation exists as far as cable resistance, capacitance and inductance is concerned. In the case of speaker cables driving conventional dynamic speakers the capacitance of the cable is of very little concern** but the DC resistance and inductance are.

**Some power amplifiers behave adversely when loaded with significant capacitance.

The DC resistance of the cable should be an order of magnitude lower than the load impedance simply because we do not want any volt drop across the cable. How do we obtain low DCR? Simply use heavy gauge wire. The table below shows the resistance of #10 wire using various metals.

The table shows that silver and copper are so close to each other that the difference is negligible but the cost of silver is a lot higher than copper. In a car system the longest speaker cable length is maybe 5 metres. So if we use #10 copper wire the resistance is 0.03246 ohms per leg for a total resistance of 0.0649 ohms. Let as assume we are using a 500 watt amplifier to drive a 4 ohm speaker. The RMS current is 11.18 amps and the peak is 15.8 amps. Using Ohm's Law the volt drop over the length of this cable is 0.0649 x 11.18 = 0.71552 volts RMS or 1.02 volts peak. Assuming the amplifier was playing at maximum power (an impossibility of course since we cannot average 500 watts from a 500 watt amplifier but let us be ridiculous and say we can) then the volt drop of about 1 volt as compared to 44.72v (500w with 4 ohms) is a drop of 13 watts or 1.22dB. Of course this does not occur in practice and a realistic number due to a crest factor of 10dB is an average of 50 watts per channel.

The RMS current is 3.5 amps and peak is 5 amps. So the volt drop is 5 x 0.0649 = 0.32 volts. Not a number to get excited about!

The capacitance of speaker cable is normally low and can be ignored BUT there are some of these fancy cables that have high capacitance and this can cause problems with amplifiers which are on the borderline of stability. Typically using say #8 wire for speaker wire lengths of less than 6 metres (20 feet) will not cause any problems for 99% of amplifiers.

The inductance of speaker cable can play a part in affecting the sound. Again common sense should rule. Keeping the inductance low is not a problem in car systems as the speaker runs are very short. Home systems are a little more complex. Let's look at a few examples.

The inductance of twin speaker cable is given by this formula

L= 0.913 x log (w/r) micro Henries per metre or L= 0.281 x log (w/r) micro Henries per foot where "w" is spacing between cable centers and "r" is the radius of each conductor.

Typically the ratio "w/r" is about 3 with average #10 or #12 wire. Solving for "L" in the above formula we get 0.435 micro Henries per metre or 0.134 micro Henries per foot. OK this looks good and now we can apply this to say a 5 metre (15') run of cable. The answer is 2.175 micro Henry. What does this mean to the sound and how can we measure it? The circuit below shows what happens.

The inductance is in series with the speaker, and let us assume it is a 4 ohm speaker which remains at 4 ohms from 20-20KHz (not true as the impedance will rise near 20KHz). The inductive reactance is XL = 6.28 x F x L where F is frequency and L is in Henries so 6.28 x 20000 x 0.000002175 = 0.273 ohms. So what we have at 20KHz is a series impedance of 0.273 ohms with our 4 ohm speaker and the attenuation is -0.573dB. Vout is 9.36 volts. This 0.573dB loss is not typical of the real world since the impedance of the tweeter will rise to almost 10 Ohms at 20KHz and so the attenuation falls to -0.23dB. At 20KHz we would be hard pressed to hear a 0.5dB drop. If we move our head off axis from the tweeter by a say 20-30 degrees the dB drop is far greater. I challenge anyone to drive their car and keep their head still and on axis with the front tweeters. (Do not attempt this as it is dangerous)

The ratio of "w/r" should be kept as small as possible to keep the inductance of the cable short. At lower frequencies the inductance has less effect since the inductive reactance reduces with lower frequencies. At 1KHz it would be a twentieth of the above at 0.013 ohms.

Dielectric absorbtion is also not worth worrying about because at ALL audio frequencies the worst dielectric which is PVC has shunt impedances of many millions of ohms. This parallel impedance cannot affect a 4-8 ohm speaker.

Skin Effect is caused by the self inductance of the wire. This causes the inductive reactance to rise at higher frequencies and electrons are forced to the surface. The circumference of the wire is therefore preferred at these higher frequencies and so the net resistance of the conductor is increased. The center core of the wire is not used. At 20KHz the losses with a 4 ohm speaker are less than 0.01dB with a 3 metre (10') cable. I hardly think that this is worth worrying about. I quote from a renowned source.

EMI and RFI are not a problem with speaker cables since the impedances are so low that it is very difficult for these types of signals to enter the cable. I of course assume that your speaker run is not 3Km (1.8 miles) long which may act as an antenna.

One of the cutest products I have seen are cable lifters. Yes you read this correctly – cable lifters. These are little stands which are used at maybe 1 metre intervals to lift the speaker cables a few cm off the floor. So, several must be used on average length speaker runs. They claim that by lifting the cable off the floor affects the sound. They are about $20-$30 each! Of course they do not affect the sound BUT after spending a few hundred bucks on this junk you better hear some difference @#$%. You could of course just use small pieces of 2x4 wood blocks and save your money.

Power cables both for 12 volt and 120/230v 50/60Hz. Don't you love those 12 volt #2 gauge power cables with the little arrows printed on the insulation which indicate that the cable must be connected with a certain "polarity". What is amazing that there are some out these who ACTUALLY BELIEVE THIS JUNK! The cable which delivers current to your amplifiers has one mission in life, to deliver the current without losses. It has NO knowledge of which way it is connected (Electrons are dumb and do not care either) This means one thing and one thing only. Use the largest gauge wire possible. This cable carries DC and so there is no skin effect, dielectric absorption, inductance issues, capacitance issues. In fact if the cable is inductive (which it is in fact because every piece of wire no matter how short is inductive) it is advantageous since the inductance helps to reject alternator noise.

The connections at each end should be of high quality and the connectors should be of robust construction. Contact resistance is important to avoid losses and again common sense should prevail here. Zed Audio does not recommend the use of power distribution blocks when using multiple amplifiers. The reason is that when an amplifier draws current it causes voltage spikes to be induced in the power lead. These can interfere with other amplifiers. The capacitors used in every amplifier across the power input are not large enough to dampen these spikes. The battery being an enormous battery has the capacitance to do this. We advocate the use of separate power cable for each amplifier.

AC power cables for the home stereo have to be the biggest con which cable manufacturers and suppliers have pulled on the unsuspecting public. The claims of how "the noise floor of your home system is reduced" and "the sound seems to come out of total blackness" is their favourite. "The mids and highs seem to be so much more transparent than before once I substituted my power cables with cable XXX". Only a fool can believe such silly comments.

I have a catalog in front of me and I see real hot specials. How does $175 for a power cord sound? Or even $649? No way those are cheap… let us look at one for $1,995.00 yeah nearly two grand for a dumb power cable that this company claims will almost cure the common cold if asked to do so! If $1,995 is too much they have a slightly less expensive model at $998.00 – wow what a bargain. There is another company offering what they call the "Magic Power Cord" at a steal of a price of $1,499.99 for 1.8 metres (5.85 feet). The reporter who "tested" the cable said "The first thing I notice is the presentation becomes BIG, stage width, depth and the space between instruments all seem larger, images are more solidly located in space and don't wonder about" I think this dude has been wondering about himself. It performs magic on your wallet and that is about all it does. It lifts the Dollar bills right out of your back pocket without you being aware.

A long with the power cable rip off artists are the guys who sell AC outlets. They paint them a new colour and claim all wild and wonderful things for "their" scientifically designed AC outlet. Or what about those hospital grade types which cost almost $50 per pair! The same company sells a power strip for $199 for 6 outlets, that is $33 per outlet!

I think that maybe these power cables at two grand and the hospital grade outlets with nice colours and silkscreened logos would do better if we could rewire our home with solid 99% pure silver wire, replace the circuit breaker board with gold plated contactors and fuses and then the ultimate demand of your city that they run the same silver wire from YOUR home to the substation and then from the substation to the main distribution grid. Wow what a deal and then the music will jump right out at you. PS I just woke up from my dream.

One simple question one should ask. The power to your home comes from a transformer some many kilometers away in some cases. It enters your home and a distribution board with many circuit breakers distributes the power throughout your home. NOW HOW CAN THE ADDITION OF MAYBE 2 METERES (6') OF CABLE AFFECT THE SOUND? The electricity has traveled a torturous route to get to you and now it encounters a magical piece of cable and all of a sudden all the ills of your stereo are cured. I say, you must believe this if you have spent a stupid amount of money on your new power cable because how would you justify this expense to your family and friends! Reading the hype which these snake oil merchants put out makes me sick, but if these kinds of things ring the bells of some people, be my guest.

I am a firm believer in direct A-B blind listening tests. The brain cannot remember the differences (if any) between two pieces of gear when the time between the listening experiences is more than a few seconds and even then this may be too long. Immediate comparison is the ONLY way to find out if one can detect any difference. Those who say "I took out the standard power cord and replaced it with brand XXX and wow the sound was transformed" are simply lying. The time interval to substitute one power cord, interconnect or speaker cable with another is just too long to allow the test to have any meaning whatsoever. To compare cables is a tedious process. A multi-pole changeover switch is required or a switch with multi-pole relays. I did this once with the power cable to my preamplifier. I borrowed a "snake oil" type IEC power cord and compared it to a standard off the shelf #18 version. I had a friend do the switching and I could not detect any difference. He listened and he could tell no difference.

The brains of those who spend money on these magical wires, cable, etc. have to "hear" a difference because the shame attached to not hearing any would be too much too bare.

It is sort of similar to those claims made by the companies who advocate a 600 pound block of exotic rock taken from the frozen wastes of Antarctica and place it carefully on the top of your CD player. This will enhance the upper midrange of your CDs whilst also improving your stamina to run the mile in under 4 minutes. Also it shall allow the resident bats in your home to hear the music with more air and prescence. In addition use those $290 rubber feet under you speakers to improve the bass. I read some of these audiophile magazines (there are nice pictures inside I must admit) and the claims made by both advertisers and reviewers alike astound me. They really do believe the rubbish they print. The one magazine I read (it shall remain anonymous) has reviewers who review some gear and then 9 months later review an updated version or a similar piece from another supplier and start saying how the subtle differences between A and B are bla bla bla. Give me a break, how can they remember what A sounded like 9 months ago? You know what it is about – ADVERTISING. If they did not write this junk the companies would not advertise and I would have no magazine to look at some nice pictures.

This is probably one of the most controversial topics in car stereo. Most are mystified by both the specifications and the way in which amplifiers are bridged. The term "Bridging" is really not a correct description of what we are trying to achieve when connecting amplifiers in this mode. The old professional term was "Strapping". Anyhow we shall stick with "Bridging" since we are all used to it.

When we bridge an amplifier we actually stack the two channels one on top of each other. Not physically but electronically. What we do is to ask one channel to handle the positive side of a waveform and then the other channel to handle the negative side. This sounds rather confusing so let's show an example. We have a 100w/ch amplifier at 4 ohms which doubles power into 2 ohms. The power supply rails are +/-33v (33 peak x 0.7071 = 23.33 volts RMS). We require that we deliver 20 volts into our 4 or 2 ohm load. So the 23.33 is fine as we have some losses in the output stages due to output device saturation volt drops. The specification of this amplifier in bridge mode would read as follows: 200 watt mono into 8 ohms and 400 watt mono into 4 ohms. Look at these numbers carefully and they are not magic. The 200 watt 8 ohm is derived from the 100+100 at 4 ohm and the 400 watt 4 ohm from the 200+200 at 2 ohm specifications.

What has happened is that the load impedance "seen" by each of the two channels in bridged mode is 50% of the total load impedance. So an 8 ohm load in bridge equals a 4 ohm load per channel in two channel mode and a 4 ohm load in bridge equals a 2 ohms load per channel in two channel mode. Since our amplifier is only rated down to 2 ohms per channel, it CAN ONLY drive a 4 ohm load in bridge mode.

Let's return to those supply rails. Assume we wanted to build a single channel amplifier which is equal in power rating to out stereo amplifier when bridged our mono block specification would be: 200 watts at 8 ohm and 400 watts at 4 ohm. So as a designer I must choose the power supply rails which will allow me to deliver this power. The exercise is simple. Here goes. 200w/8 ohm or 400w/4 ohm means I must deliver 40 volts RMS into my load. (40x40/8 = 200 and 40x40/4 = 400) Take 40v RMS and we add in some extra for output stage losses (say 3 volts). Now we have 43 volts RMS, multiply it by 1.414 (To arrive at the peak for the supply rails) and we get 60.8v. This is very close, but not equal to our stereo 100w/ch amp which used 33 volt rails for a total of 66v – when in bridge mode. Here we only have 60.8v rails in the mono block. Why the difference?

The bridged amplifier has the output stages of each channel in SERIES, and because of this there are DOUBLE the volt drop losses in the output stage. In the mono block however we just have a single push-pull output stage with ONLY ONE volt drop loss.

How do we bridge channels of amplifiers? This is done by simply driving one of the channels with a 180 degree phase inverted copy of the signal OR making the second channel of the bridged pair an INVERTING amplifier. We use the latter method at Zed Audio. It is important that each channel has the same gain structure so that the final signal from the bridged pair has equal positive and negative halves of the waveform. Typically in practice this is never the case and the result is that the channel with the lower gain, technically clips its portion of the wave a little early. We have not found this to be a serious problem because the error is so small.

Referring to the above diagram I have depicted the channels using the op-amp symbol of a triangle with a + and – inputs and an output. Power supply rails are left off for clarity. The signal is fed to both channels (The Y-Adaptor which you use to feed the mono signal to both RCA inputs) and the signal is phase inverted on the bottom channel. The gain structure of the NON-INVERT channel is R1+R2/R2 and the gain of the INVERT channel is R4/R3. This diagram shows one way of inverting the signal. As I said above one may use two NON-INVERTING amplifiers and simply have a phase inverter in front of one of the channels.

Let us examine how the amplifier delivers the power. Let us consider the top NON INVERTING section. It runs of +/-33v rails. It can therefore develop 20v RMS from its output with reference to ground. (Do not forget that this is our same 100w/ch 4 ohm amplifier example) Let us say a positive waveform is injected at the input at an amplitude of 1 volt RMS. Let us also assume that each pair of the bridge has gain structure of 20x (It will amplify an input signal twenty times). The output from the top amplifier is 20v RMS in a positive direction. Meanwhile the bottom INVERTING channel has not been sleeping! It has taken the 1v input and inverted its phase and therefore will develop a 20v RMS signal BUT in the negative direction. Since the load (speaker) is connected between the two "hot" outputs and the top channel swings the signal 20v positive and the bottom 20 volts negative, a total of 40 volts is impressed across the load. This is how the 40 volts is developed. Remember 40 volts into 8 ohm equals 200w and into 4 ohm equals 400w.

The two channels being bridged do not have to be a pair in a single chassis. Zed's mono blocks are bridgeable by simply feeding the second mono block from the first with a phase inverted signal.

An amplifier pair which is in bridge mode CANNOT be re-bridged with another amplifier which itself is in bridge mode or another non bridged amplifier. This is because of the grounding system we at Zed Audio use. This would only be possible to bridge two bridged amplifiers if their grounding systems are 100% isolated and the centre tap of the transformer secondary swings with the signal. I know of no mobile amplifier which does this and there are very few professional amplifiers which have their transformers configured to accommodate this.

Note: When bridging any amplifier, please keep in mind the minimum load impedance which the manufacturer states. If it is 2 ohm per channel then minimum bridge impedance is 4 ohms. So minimum bridged impedance is twice the minimum per channel.

There seems to be some confusion about balanced line systems. They have been used for ever in professional installations and in radio frequency systems. We shall confine ourselves to the audio side of this discussion.

Note:In the following discussion I only talk about a single channel of the traditional stereo pair. This is for clarity purposes only.

An unbalanced line has only TWO conductors, a hot and a ground. When linking two pieces of equipment with an unbalanced connection, the ground between the two becomes common by virtue of the single ground return circuit. The hots are of course connected together. This system has both advantages and disadvantages. The advantage is it is only a two wire circuit, typically RCA connectors are used and it is less expensive than balanced line circuits. The equipment transmitting the signal needs only a single ended output stage using the hot and ground. The receiving equipment also only needs a single input stage that uses only hot and ground. Disadvantages are that long cable runs cannot be used without high frequency loss UNLESS the transmitting equipment has very low output impedance and can deliver some reasonable current. Never the less it is not advisable in a high quality system to have unbalanced runs of more than 10 metres (30 feet). An unbalanced system is prone to noise pick up unless extreme precautions are taken. Cables should be well shielded (preferably with a second outer braided shield connected to the chassis of the vehicle). Ground loops can be formed unless precautions are taken with grounding techniques.

Note:See the section on ground loops under Techtalk.

Balanced lines are kind of the opposite of their unbalanced cousins. More expensive due to the 5 wire system and the connectors are more expensive. XLR plugs/sockets are more costly than RCAs (Unless you are stupid enough to spend $3,000.00 on your new 1 metre RCA-RCA "high end" cable). The transmitting equipment must have anti-phase outputs to drive the balanced line hot legs and the receiving equipment must have a differential input circuit. A balanced line circuit with a good CMRR will reject noise if it is picked up on both "hot" legs. Ground loops are easily avoided but still care must be taken with grounding of the various pieces of equipment.

In Fig 1 showing the unbalanced connections, the signal exists between the HOT and the GROUND leads. As shown the green ground connections MUST be made at each end to complete the circuit. I have shown the two pieces of equipment in a schematic format so one may understand more fully the various impedances involved. The resistor labeled as "Zout" is typically less than 500 ohms from a line level source. The resistor on the receiving equipment labeled as "Zin" is typically many thousands of ohms whether this piece is another line level processor or an amplifier. A simple potential divided is made from Zout and Zin so that the effective voltage which the receiving piece actually gets is Zin/Zin+Zout. A simple example if Zout is 500 ohms and Zin is 10,000 ohms the attenuation caused by Zout is 10,000/10,000+500 = 0.952. So if the output voltage of the sending equipment is 1 volt then the receiving equipment only gets 0.952v (-0.427dB). This fine when there is no capacitance involved, but this is not the rule.

In Fig 1 I have placed a capacitor in parallel with Zin. This capacitor is made up of the actual capacitor at the input and the lumped capacitance of the shielded cable itself (Normally low). So now Zin is frequency dependent because a capacitor's reactance varies with frequency. The capacitive reactance is inversely proportional to frequency. Let us say Cin = 300pF (0.0003 mfd) and Zout and Zin as the example above. So the potential divider made from Zout and Zin has this new component Cin. We have two scenarios here. The total Zin at 20Hz is equal to the value of the resistor labeled "Zin". Why? Because Cin at 20Hz = 26.5 million ohms – hardly a factor. But at 20KHz the story changes. Cin has a reactance of 26.5K ohm at 20KHz which is in parallel with the resistive part of Zin. So now the calculation for final Zin is done from this formula. 10,000 x 26,500/36,500 = 7,260 ohms. So our input impedance has changed from 10K ohm at 20Hz to 7.26K ohm at 20KHz. What have we now? The signal will be attenuated at 20KHz more than at 20Hz due to this capacitor.

The original formula where the final voltage received was 0.952% of the transmitted voltage NOW VARIES WITH FREQUENCY. At 20Hz it remains 0.952 but at 20KHz it is calculated 7,260/7,260+500 = 0.935 (-0.583dB). So the response is down by 0.156dB at 20KHz with respect to 20Hz. What do we learn from this? If we make Zout as low as we can, it has less effect on the final voltage the receiving gear will receive from 20Hz-20KHz.

Looking at Fig 2 above the transmitting gear has anti phased outputs. The bottom opamp is simply a unity gain inverter. Now the signal is ground free and ONLY exists between the + and –phase outputs as shown. The connector is typically a 5 pin for stereo with 4 hots and a shiled ground. (I show only three for one channel). The cable is of course a 4 conductor and an outer shield type (I show again only 2 plus ground for clarity). The outer shield serves only one purpose and that is too shield the inner cores from noise pick up and must only have ground connection at one end. (I show it at the transmitting side) The inner cores should have each pair in a twisted assembly for additional protection against noise pick up. The receiving gear has a balanced input a shown. This is the simplest form of balanced line input (without the use of a balancing transformer). However the input impedance on each leg is different. On the + phase input it is R1+R3 and on the – phase it is R2. So if all four resistors are of equal value (typical if no gain is required) then the input impedance on the + leg is twice that of the bottom leg. What does this do? Well a similar calculation can be done as in the unbalanced system and here is what happens. On the + phase the attenuation is 20,000/20,000+500 = 0.975. (remember the Zout of EACH balanced drive leg is 500 ohms as an assumption). Now on the –phase it differs because the numbers are 10,000/10,000+500 = 0.952. So as we see there is a difference on the two input phase legs of the receiving gear of 0.2dB (A 2% inbalance). This will affect the CMRR (The ability of the balanced line system to reject noise) of the system and there are more complex circuits which allow the input impedances of both phases to be identical.

So what have we learned here? In professional systems where there are many metres of cable and many pieces of equipment the use of balanced line is mandatory. In home and car systems it serves ABSOLUTELY NO PURPOSE if the unbalanced system is correctly designed, the equipment is well designed and the installation is good. Keeping signal cables well away from high current power cables is a must and using double shielded audio line level cable will keep interference out. The line level drive equipment must have low output impedance to keep the hot signal legs at the lowest impedance with respect to ground.

Balanced line in the car is like chicken soup – it cannot hurt. It's major disadvantage is the connectors. Miniature 5 pin DIN plugs and sockets are typically used as they are small. They are difficult to work on and typically do not come with gold plating on either plugs or sockets.

More reading should be done on the link for Ground Loops.

The following circuit is a guide to calculating voltages with a simple two resistor divider which applies to Figs 1 and 2 above

Amplifier protection is not one of the items which an installer or consumer think about. It is another thing for us as designers as we must attempt to anticipate what may happen to the electronics in the field. The automobile is a hostile environment for electronics and so we must incorporate several protection mechanisms in the amplifiers to keep them operating. At Zed Audio we have always paid attention to this fact and incorporate several protection circuits in our amplifiers.

Thermal protection

is the first line of defense against heat – the arch enemy of electronic components. Transistors, Mosfets and diodes will operate quite happily at 100 deg C (Yes they must be temperature de-rated at these elevated temperatures), Integrated circuits, capacitors, resistors, connectors are not to keen to have their operating temperature exceed about 80 deg C. The reliability of ALL components is reduced when exposed to high temperatures so it is in our interest as designers to keep them low.

Large heatsinks are a must if forced air cooling is not employed. Class A-B amplifiers are not that efficient and they will generate lots of heat when driven hard especially into low impedances. Multi-rail (Class-G) can be employed and this allows the use of smaller heatsinks but at the expense of more complex electronics.

I would imagine that almost all mobile amplifiers have some sort of thermal protection. A thermal sensor is mounted to the heatsink and it is either in the form of an electro-mechanical switch or a thermistor. The switch is simple in that the remote turn on line can be simply switched off when the sensor reaches its cut out temperature. Some have hysteresis built in, others do not. A thermisitor is a device whose resistance will change with change in temperature. The type Zed Audio use goes lower in resistance as temperature rises. We incorporate it in a bridge circuit with hysteresis. What the hysteresis does is once the amplifier has shut down (we do this at about 75 deg C) it shall only turn on when the heatsink temperature has dropped to 65 deg C.

DC protection

is a form of protection which monitors the signal on the speaker outputs, removes the AC component (The music) and checks if the residual DC is less than a specified value which the designer chooses. Typically a DC level of about 4 to 7 volts is considered safe for speakers. So we have a sensing circuit which monitors the DC component at the speaker, and if this DC component is greater than 5.2v (either positive or negative), the power supply is shut down and latched off. The only way to reset the system is to turn the amplifier off, wait a few seconds and turn it on again. If the amplifier clips hard with non symmetrical waveforms and the net DC component due to this clipping exceeds the 5.2v the amplifier is shut down.

Reverse polarity

protection is to protect the power supply if the power leads are connected the wrong way to the battery. The old method was a great big fat diode connected across the power supply terminals in a reverse direction. If the leads were reversed, the diode would be in the forward conducting mode and would instantly blow the fuse (and hopefully not itself). The net of about 1v of reverse polarity (the conducting diode would have about a 1 volt forward drop) was insufficient to damage the power supply. Fortunately Mosfets have a built in diode which can carry the same current as the Mosfet. The Drain – Source junction is essentially across the incoming 12 volt and so the body diode does the same job as an external one would.

The remote turn on line suffers the same fate as the main +12v would if polarity is reversed. A simple series diode (0.6v forward drop) does the job. Zed uses a three transistor circuit in this application as we use it to control other functions in the power supply.

Short circuit protection

is probably the most difficult to implement. We must attempt to protect the output devices in the amplifier as well as the power supply. In addition the output devices must be protected against very low impedance loads. The way this has been done for the past 30 years is to use what is known as V-I limiters. Their name implies what they do for a living. The circuit shall monitor the volt-amp relationship in the output stage, and if the safety limit is exceed, the circuit would remove drive from the output stage. There is a problem however. When the circuit is activated and the drive to the output stage is removed, the shorted or mismatched load is not being driven with signal (remember it has been removed). The V-I limiter then says to itself "hey no more work to do, stop removing drive from the output transistors"). So the drive is instantly restored. Well the V-I limiter immediately senses the shorted/mismatched load and does its thing again. This on-off cycle continues and what it does is causes high frequency artifacts to be superimposed on the output waveform. BAD FOR TWEETERS is the result. How can this be prevented? In a mobile amplifier we have a switching power supply. Zed does not limit the drive to the output devices but we take the V-I limiter's error signal and inform the power supply politely that there is some sort of problem on the speaker line and that we will shut the power supply down – which we do. Once the power supply is shut off, and latched off, NO damage can occur to either section of the amplifier. In Zed amplifiers we also employ a static and dynamic V-I limiter. We allow the limiter to let the amplifier continue to operate into sub 1 ohm loads for a "few milliseconds" with a music signal but with a sinewave test signal the static threshold is what causes the V-I circuits to activate.

Radio Frequency (RF) protection

is done in several places within the signal path of our amplifiers. The first line of defense is at the inputs where we use a low pass filter set at 338KHz and again at the level control amplifier another low pass filter at 338KHz.

There have been many published facts about this topic and much more unsubstantiated things written and spoken about this subject. I hope to put forward some facts that will lay those unsubstantiated theories to rest.

A 100% efficient amplifier is just that, power in = power out, no losses, no heat and of course NOT POSSIBLE. There is no such thing as a 100% efficient amplifier.

There are several factors which affect how much heat an amplifier shall dissipate. We shall assume a perfect power supply and only concentrate on the audio amplifier at first. I shall come back to the efficiency of the power supply. The class of the amplifier determines how efficient the circuit is.

Class A the output stage conducts all the time that is through the full 360 degrees of the waveform.
Class B each half of the output stage conducts for 50% of the time that is through 180 degrees of the waveform
Class A/B is just a class B amplifier with the output stage idling current set to some tens or maybe hundreds of milliamps.
Class D are PWM amplifiers and have no relationship with analog designs.

Let us begin with class A amplifiers.

Class A amplifiers fall into two categories, single ended and push pull. Single ended types are less efficient than their push pull counterparts. Typical efficiency for single ended is from about 0% (No this is not a misprint) to 25% and push pull up to 35%. Single ended class A amplifiers shall be discussed since push pull versions are too a large degree high bias class B designs. So the following discussion will pertain to a single ended design.

The current in the output stage should be EQUAL or slightly higher than the load (speaker) current. This shall assure us that at no time will the output stage switch into class B. The following is a simple example of a pure class A amplifier rated at 50 watts into 4 ohms. Output voltage at speaker = 14.14v RMS or 20v peak

Output current through speaker is 14.14/4 = 3.53A RMS or 5A peak

The power supply must be +/- 20v constant but we must include the inevitable losses in the output transistors as they are NOT perfect switches so +/- 24v will be used.

Since we must have a constant current in the output transistors of 3.53A RMS or 5A peak and the power supply is perfectly regulated to maintain +/- 24v the dissipation in the output stage UNDER IDLE conditions is 48 x 5 = 240 watts (we must use the whole value of the power supply as both devices are conducting all of the time)– and this is ONLY ONE channel. A stereo amplifier shall dissipate 480 watts! The problem becomes worse if we design for a loudspeaker which is nominally 4 ohms but dips to say 2 ohms (not unusual)……….well be my guest and double the above dissipation numbers only because into 2 ohms the peak current is 10 amperes.

So any company who claims to have a pure class A amplifier for mobile use of more than a few watts per channel (and I have never heard of any company offering a 2 watt/channel car amplifier) is telling a tall story. The idling current of this 50w/ch amplifier optimized just for 4 ohm loads would be 480/12 = 40 amperes and this does not include any power supply efficiency calculations. Typically one can add 10-15% for power supply inefficiency. So the package efficiency is 100/552 = 18.1% not exactly conducive to long battery life!

There are amplifiers where the idling current is reduced and so at higher power levels the amplifier does switch to a class B type.

Another problem with Class A amplifiers is that their CMRR (Common Mode Rejection Ratio) is poor. The CMRR is a measurement of how effectively an amplifier rejects noise or ripple on the power supply rail(s). A typical class B amplifier has a CMRR of over 80dB whilst a class A amplifier is 30-40dB worse. Due to the very high idling current, a class A amplifier's power supply has a few volts of ripple, whilst a class B amplifier which has very low idling current has a power supply with millivolts of ripple. The class A amplifiers noise can be improved by using an electronic regulator which filters out most of the power supply noise BUT to use these the power supply voltage pre-regulator must be higher. So in our example above the power supply could be as high as +/- 30v. Dissipation (including the regulators) is now 60x 5 =300 watts and make two channels and this is 600 watts and then add in the power supply inefficiency and we have 690 watts. Efficiency is now 50+50/690 = 14.5%.....wow we now have a 50w/ch amplifier idling at 57.5 amperes. One more bombshell, at idling the amplifier's efficiency is 0%, a big fat zip. Why well the output is zero and 0/690 = 0. As the power increases, efficiency will rise. At 3 watts per channel efficiency is 0.87%!

Now for class B amplifiers

Class B amplifiers by definition have zero dissipation in the output stages at idle BUT all amplifiers for audio are designated A-B. The reason we introduce a small amount of idling current in the output transistors in order to get rid of crossover distortion. This current in an amplifier of say 100 watts is typically 30-70mA. Let us use the same numbers as in the class A example.
Power supply is +/-24v.
Load is 4 ohms.
Output voltage is 14.14v
RMS current in load is 3.53A or 5A peak or 3.18A average current.

Since the amplifier is a push-pull design (all class B amps are) we only consider half the power supply voltage. So 3.13 x 24 = 76.32 watts and we get 50 watts out of it. Efficiency is 65.5%. If the output transistors were perfect switches the efficiency would be 78%.

The efficiency of a class B amplifier changes with output power. Let's examine a simple example. Let's say we have a power supply of +/- 50v. We also have a 10 ohm load (easy for calculation). Let's assume the output moves 10 volts positive. Then 20 volts until it reaches the rail of 50 volts. The output transistors are perfect for this example, NO LOSSES.

So as you can see the dissipation in the output transistors increase to a peak and then decrease. If we did this volt by volt maximum dissipation in the output transistors would be at 44% of absolute unclipped power.

The class A-B amplifiers we use and talk about are operating in class A mode only to extremely low power levels. Let's see what's happening. The 50w/ch amplifier is set to idle at say 50mA (0.05 amperes) and we have a 4 ohm load. Remember Ohm's Law. IxIxR= Power.

0.05x0.05x4=0.01 watts. Yes 0.01 watts or 10mW. A typical 50 watt amplifier runs in class A up to TEN THOUSANDTHS OF A WATT, NO MORE NO LESS.

Lastly Class D (PWM) amplifiers

This type of amplifier uses MOSFETS as switches. A high frequency carrier is mixed with the audio signal and the output Mosfets are on or off depending on the average level of the audio signal. Simply put when a positive pulse of audio exceeds the absolute value of the carrier, then the positive Mosfet turns on. This action happens at the frequency of the carrier (Typically > 100KHz). A low pass filter removes the carrier from the signal to be applied to the speaker and what is left is amplifies audio. There are numerous ways of achieving this result but at the end of the day the LowPass filter must be used to remove the high frequency carrier. Class D amplifiers for low frequencies are fine but in our opinion they kind of suck for full range. Due to the fact that the output Mosfets are either on or off, there are much smaller losses than their analog counterparts. Efficiencies as high as 95% are attainable but typically 80-90% is practical and this varies with output power and load. The higher the power, the higher the efficiency, the lower the load, the lower the efficiency. The efficiency numbers manufacturers quote are those at maximum output into the highest impedance (4 ohms?) but this is misleading since who can play their amplifier at maximum power?

What is it? How is it caused and how can we prevent it?What does it do? We shall deal with each of these topics one at a time.

What is it: Almost everyone assumes that amplifier clipping is the sole domain of power amplifiers? This is not true. A preamplifier is just as prone to clipping as power amplifiers. If the level of signal is high enough to cause the preamplifier to clip, the power amplifier being a faithful servant will just amplify the clipped signal it receives. For the purposes of this discussion we will assume that our amplifier/preamplifier models use a bi-polar power supply (as almost all audio electronics does today) and therefore the signal swings from a level of zero to either the positive or negative supply rails ("rail" is a commonly used term which we use to describe a power supply output). We shall also assume that the electronic building blocks are in the form of operational amplifiers WITH negative feedback. Op-amps as they are called, are TWO input ONE output building blocks. Input is to either positive or negative ports and feedback is taken from the output and returned to the (-) input. I shall show the op-amps as in the diagram below even though it is not the standard schematic symbol for an op-amp.

Note: A power amplifier (that is one which can drive a loudspeaker) is nothing less than just a high current preamplifier. Preamplifiers can and do run off very high rail (there we go again with the term "rail") voltages – they just do not have the capability to source lots of current.

The gain of this amplifier is defined as the ratio of R1+R2/R2. So if R1=10K ohm and R2 = 1K ohm the gain is 10,000+1,000/1000 = 11 times. So whatever signal is applied to the input, it shall be amplified 11 times PROVIDED the output signal does not exceed either rail voltage. Let's assume the + and – rails are each 30v. This means that the output rail can move towards either the +30 or the -30 volt rails depending on whether Q1 or Q2 is turned on. They CANNOT both be conducting at the same time. The diode D1 is shown as the source of bias which sets up the idling current for the class A-B output stage. I have just shown it to represent the fact that this building block of ours is a class A-B output stage.

Note: We can use the peak value of the input signal versus the peak value of the output signal but this does not equate to the real world so I shall use the RMS volts in versus RMS volts out. 1 volt peak = 0.7071 volt RMS. We shall assume perfect transistors for Q1 and Q2 so they have no losses. This means that with the 30 volt rails, the maximum RMS output voltage is 30 x 0.7071 = 21.2 volts.

Now the above schematic could represent a preamplifier stage or a power amplifier. The difference is simply the value of the LOAD IMPEDANCE we place at the output. If this was a preamplifier the load impedance would be in the order of thousands of ohms and if it was a power amplifier it would be say 4 ohms.

Note: We are also assuming that the rails have infinite current capability (the 30 volts remains constant) and that the transistors Q1 and Q2 can pass any current we require.

If we keep the gain at 11 times this means that the maximum input voltage allowed in order to prevent clipping is 21.2/11 = 1.93 volts. Applying more than 1.93 volts to the input will cause the transistors Q1 and Q2 to "bump" their heads at their respective rails. This is known as clipping. The waveform at the output of our amplifier looks like the diagram as shown below

Amplifier clipping

As we see the top and bottom of the sine wave has been flattened. When the waveform rises (or falls) to it's respective rail and it reaches a value of 21.2v RMS (30v peak) it cannot increase/decrease any further.

If we change the gain of the amplifier to say 20 times, (change R1 or R2 in the abive circuit) the only difference is that we shall require less input voltage to drive the amplifier to a given output voltage. Since we know that the maximum output before clipping is 21.2 volts, dividing this by 20 yields a value of 1.06 volts at the input. Conclusion, clipping is not affected by the initial design of the gain structure (since we can apply higher input voltages if the gain is low and visa versa) but only by the value of the power supply voltages.

How is clipping caused.

Simple by allowing the output stage of a preamplifier or power amplifier's output voltage to reach the value of the power supply voltage. A typical mobile audio system consists of a head unit, an equalizer and then an amplifier. It looks like the diagram below.

We shall begin with the head unit. Working from a 12v nominal power source (The battery) it is capable of delivering a theoretical output of 12/2 = 6v peak now x0.7071 =4.24v RMS

Typically this output voltage is limited to slightly less than 4 volts RMS. So the first stage of protection is to ensure that the output of the head does not exceed this nominal 4 volts. How do we do this? The only sure way is to use your favourite CD which has the loudest passages and connect an oscilloscope to the RCA outputs and confirm that with the volume control set at maximum, there is no clipping of the complex music waveform. By setting the timebase control on the 'scope it is very easy to see clipping. We shall assume for the moment that the head unit manufacturer has designed the product correctly and that the highest modulated CD will not allow any form of clipping at the outputs.

Now for the tricky part. We have a head whose output varies from 0 to 4 volts depending on the volume setting. At 0 volts output we will not induce clipping anywhere!!!

The processor which works off a typical split rail of +/-15 volts can output in practice 9v RMS. Normally these units have some sort of gain controls and for this example let us assume it only has an input level control (See also the links on Level Control and Level Matching). The max. of 4v being the worst case from the head now enters the processor. Hopefully the input level control is in the correct place in the processor's signal path and we shall assume this to be true. This level control obviously changes the gain structure (how many times the processor will amplify the in-out signal level). We shall assume that it can change the gain from -10dB to +10dB.

Note: Decibels (dB) is a logarithmic way of denoting a ratio. 10dB in voltage terms is calculated from this formula dB = 20 x log to the base 10 of the ratio. So +10dB in voltage terms is 3.16 times. -10dB is 0.316 times.

Back to our processor. All crossover controls and equalizer settings are flat. So the equalizer has no effect now and the crossover settings are such that we are measuring voltages within a particular passband. (Let's assume it is a 2 way crossover set at 100Hz and we are testing a signal at 1KHz which is far away from the 100Hz at nearly 4 octaves).

Setting the gain at 0dB we would get 4 volts out of the processor which is well below its clip point of 9v. If we set the processor's level control to any position less than 0dB we are still in a non clipped mode. Moving this level control to the +dB side we start moving towards the "red" zone. Remember that we have 4v from the head. The processor can output 9v, this means a MAXIMUM gain structure of 9/4 = 2.25 times! (2.25x = 7dB).

So we can only set our processor's level control to +7dB before the onset of clipping.

What we have in effect done is put a safety net up by setting the processor's gain to +7dB we guarantee ourselves that NO clipping can occur no matter where the level control of the head is set.

Note: What we have to know is the absolute max output of the head at max volume setting WITHOUT clipping. This is our start point and if it is incorrect the rest of the exercise is pointless!

So now we have a processor which can output up to 9v with NO clipping. Our 150w/ch amplifier is the next victim. Let us assume (yes we must assume things for examples) that our amplifier has a level control whose range is 200mV to 9v for rated output of 24.7v.

Note: Please do not be confused with my references to an amplifier's output voltage and not its power. The amplifier is just a great big old preamplifier which can source lots of current but at the end of the day it still puts out VOLTS and sinks AMPS. It is easier when referring to gain structures and voltage gains to use volts at each end of a piece of gear. There is NO reason in the world why we cannot specify a preamplifier's output as power. It may be very small but it still is power. You will notice that Zed Audio specifies the output of the amplifier in volts, amps and then watts. The volts/amps is a more accurate way of describing what the amplifier is capable of.

If we set our amplifier's level control to 9v (We assume the amplifier puts out a wee bit more than 24.7v before clipping) then we still have a system which does not clip. (Do not forget that we set the processor level to +7dB to have a 9v output)

So we now can repeat the same thing with the amplifier. Setting its level control to say 2 volts will 100% guarantee that we can clip the amplifier with ease. How do we prevent this from happening. First do not power up the system – no only joking! First make sure that the level control on the processor is then set back so that with 4v in from the head it can then only output 2 volts. Therefore we must set the level control of the processor to 2/4=0.5 times = -6dB.

Note: When boosting ANY equalizer control by "n"dB this is equal to lifting the gain by "n"dB. Even though the equalizer boosts the level in a relatively narrow band of frequencies the amplifier does NOT differentiate. So this "n"dB must be taken into account and the level should be reduced by "n"dB at either the processor or amplifier. Processors with both input and output level controls are far better at level matching than those with only one level control.

For more on this subject see the link on Level Matching.

What does it do?.

Clipping is the arch enemy of speakers, especially higher frequency drivers. It is probably the biggest cause of speaker failure. Looking at the diagram below which shows a clipped sinewave we see from the time axis that the waveform remains at a high amplitude (either positive or negative) for a period of time which is longer than the time it spends when the sinewave is not clipped.

Amplitude Time

The result of the speaker cone "spending" too much time at one end of its travel will cause voice coil overheating, deformity of the cone/spider assembly. Another effect of amplifier clipping is that harmonics are generated from the fundamental. Assume a 100Hz wave is being clipped. Harmonics at 200Hz, 300Hz, 400Hz, etc are generated. As the harmonic number increases, its amplitude decreases. The amplitude of these higher frequency harmonics is determined by how hard the amplifier clips at the fundamental frequency.

Because high frequency drivers are fragile as compared to high power low frequency and midrange drivers, they are more susceptible to damage. These high frequency harmonics do not generally damage low frequency drivers but this is not always 100% true.

Let us use a 200 watt amplifier as our example and let it be clipping at say +6dB worth of overdrive. +6dB of overdrive in power terms is calculated from the formula [dB=10 x log to the base 10 x power ratio]. Putting the numbers in the formula yields an answer of 4 times power. So the 200 watt amplifier will "attempt" to put out 800 watts. When an amplifier is hard clipped it puts out essentially a square wave which looks like this:

The area under the squarewave represents power and if one compares this with a sinewave at the same frequency, then it is obvious that the area under a sinewave is much less than the square wave.

Music is not constant in its peak amplitude. The ratio of average power to peak power is in the order of 10-20dB. (10dB = 10 times power and 20dB = 100 times power). I would imagine that modern rock and roll/rap music the value is closer to 10dB. This means that with typical music the average power when using a 200w/ch amplifier is in the order of 20 watts per channel with the peaks rising to 200 watts. Anything higher than the 20 watt average will most certainly push the amplifier into clipping. With this scenario the tweeter in a typical bi-amplified system or one with passive crossovers will receive about 10-15% of the power. So the tweeter's power is about 20-35 watts with our 200 watt amplifier. This is a lot of power for any single tweeter. But let us assume it is OK with this.

When the amplifier clips the energy into the tweeter is many times greater than with unclipped signals. (Of course the amount depends on the degree of clipping but it has been found that people will listen up to 10dB of clipping ). When this happens the compressed wave (now very close to a square wave) is absorbed by the tweeter (and do not forget about all the harmonics) and at this stage the tweeter goes to "the pie in the sky".

Low frequency drivers are more tolerant of clipping simply because of their more robust construction. I have however seen many a woofer damaged through been overloaded on a continuous basis.

The above discussions have assumed that the waveform is symmetrical about the zero line. Unfortunately music is not like this. The positive half of the wave may not be the same as the negative half. As an example let us assume that this is so and that the positive part of the wave at time zero is larger in amplitude than the negative half. When the amplifier clips, the area under the positive half is more than the negative half and because square waves are being generated by the amplifier the DC component on the speaker rail will not be zero – as it should be.

Remember one fact. DC is a constant voltage. 10 volts positive DC (ref zero) is just that. If our amplifier was flat to DC and we put in a DC signal the amplifier would simply do it's job – amplify and the output at the speaker rail would be a larger replica of the input. AC on the other hand is just varying DC. A sinewave begins at 0 volts. It rises at a particular rate (determined by the frequency) to its peak value and then declines to zero and repeats the same thing below the zero ref line. BUT at any given time during the single cycle of the sinewave it has an absolute value. The average is zero. A square wave (clipping!!!!) is similar but not the same. The square wave starts at zero, rises very rapidly to it's peak value, stays there for a time (determined by the frequency) and then returns to zero and the other half of the cycle is below the zero line. The average of course is zero ONLY if the positive half of the square wave is equal to the negative half.

With music and clipped amplifier the average is not zero and in our example above the speaker rail will tend to move positive DC for the period of that non symmetrical clipped wave. DC on a speaker for a sustained period of time (Constant amplifier clipping) will sustain damage.

Audio amplifiers have been put into different "classes" The class is dictated by the way the output stages operate. For audio we have five basic classes but one of them pertains to how the power supply operates.

Class A

Amplifiers are probably the best sounding of all the classes. The output stages operate at a constant current equal to or greater then the current which the load requires. This means that the output devices (Bipolar, Tubes, Mosfets, IGBTs) are never driven into cut off. They conduct through 360 degrees of the output waveform. All analog amplifiers have input and driver stages which pretty much all operate in pure class A mode. They can because their heat dissipation is relatively low. What are the disadvantages to pure class A amplifiers. Heat! A reasonable powered amplifier dissipates enormous amounts of heat. Please refer to the discussion on Amplifier Efficiency to see some simple calculations of how efficient a class A amplifier is. For this discussion let's use an average figure of say 15% which is a conservative number. This means that for every 100 watts of power into our amplifier, 85 watts go out as heat and 15 watts go to the loudspeakers! Not a very good situation if you must pay the electric bill.

Class A amplifiers are configured in either single ended or push pull. Single ended means that the output stage consists of a single amplifying device (Transistor, tube etc.) and it is normally driven from a constant current source but in tube designs it is a transformer. The amplifying device has no ability to sink current, only source current to the load. A push pull stage has two devices, each one delivering current to the load on each half cycle of the waveform. Some consider this type a "hard biased class B" output stage.

For an amplifier to be classified as "Class A" it is required that the standing current in the output stage be equal to or greater than the maximum load current. This means that if we are using a typical 4 ohm speaker, its impedance may drop to say 1.5 ohms at some frequency. If we have a 50 watt per channel amplifier this requires 14.14 volts to be developed across the load (speaker). So with a 4 ohm load, the current is 14.14/4 = 3.53 amps RMS or 5A peak. With a 1.5 ohm load it is 9.42 amps RMS or 13.32A peak. So in order for our amplifier to remain in pure class A (assuming the typical 4 ohm speaker goes to 1.5 ohms) it must idle at 9.42 amps RMS PER CHANNEL.

Let's see how the numbers turn out. The 50 watt amplifier runs off supply rails of about +/-25 volts (let's also assume that this is regulated). The wattage at idle is 13.32 x 50 = 666 watts per channel. This is a total of 1332 watts of heat when the amplifier is just sitting around doing nothing! Let us compare this to if the amplifier ONLY had to drive a 4 ohm load the dissipation would be 5 x 50 = 250 watts per channel and 500 watts total, still not an insignificant amount of heat.

Any company who claims to have a pure class A amplifier (of reasonable power output and I do not mean 3 watts per channel) for the automobile is simply not telling the consumer the truth.

There are those who will say that "our amplifier model XX operates in class A up to YY watts and then it switches into class B. This is none sense, a class A amplifier by definition NEVER operates in class B - period.

Class A amplifiers have some disadvantages as far as the power supply is concerned. Due to the high idling current, the power supply must be well filtered to avoid hum and noise. In an amplifier running off 60Hz AC this hum can be significant and the best way to eliminate it is by using fully regulated power supplies. In a car amplifier there is no 60Hz but in order to keep he signal free of noise, regulators in the power supply rails should be used. The common mode rejection ratio is very poor in class A amplifiers and the regulators help to allow the rejection of power supply noise and ripple.

Class B

amplifiers are those that only conduct through 180 degrees of the output wave form and are all push pull by design. The output stage has device(s) for both halves of the waveform. This means that when the positive device is conducting and sourcing current to the load, the opposite device is cut off (not conducting). This is kind of analogous to two people on opposite ends of a rope. One pulls and the other can relax and visa versa. During the time at which one person ceases to pull and the other starts to pull, the rope is limp (in a cut off stage). This is the region in which one half of the output stage stops conducting and there is a short period of time before the other half begins conducting. This time interval is what we call "crossover distortion". It manifests itself as a small time step on the sinewave on both the positive and negative going halves of the sinewave.

In order to reduce this crossover distortion to a small value we introduce a small idling current in the output devices. This current is normally in the order of milliamps. What this does is causes the output devices to conduct for slightly more than the 180 degrees. Depending on the amount of idle current introduced this may be up to 200 degrees and more. Technically the output stage runs in class A up to a very small power output. Consider a 100 watt amplifier which has the idling current set at 50mA which is not untypical. The load is 4 ohms(R). Power is I x I x R. So 0.05 x 0.05 x 4 = 0.01 watts. This amp runs in class A up to a power level of 10 milliwatts – hardly class A but theoretically it is. If we idle this 100 watt amplifier hard at say 500mA, it will run in class A up to a power level of 0.5 x 0.5 x 4 = 1 watt!

I assure you that a 100 watt class B amplifier idling at 0.5 amp (500mA) will run hot. The typical rail voltage is +/-35v and the dissipation at idle is 35 x 0.5 x2 = 35 watts for each channel.

Class A-B

amplifiers are just a variation of a class B amplifier as described above. Most amplifiers made today are A-B simply because a pure class B would be unacceptable for audio owing to its high distortion at crossover. Otherwise the two are identical.

Class D

amplifiers are of the switching variety. Technically they are Pulse Width Modulated switching power supplies where the modulation is the audio signal. Typically a high frequency carrier (50KHz-500KHz) is converted to a triangle waveform. This triangle waveform is fed into a comparator together with the incoming audio signal. The resultant PWM waveform is fed into an output stage which alternately switches either the positive switches on or negative switches on depending on the polarity of the incoming waveform. Since the Mosfet switches are either on or off, their efficiency is close to 100% but not quite there! Losses in the Mosfets are due to their finite on resistance and the losses which occur during their transition from off to on and back to off states. The high frequency pulse train must then be demodulated back to an analog form in order that the loudspeaker can reproduce it. This is done with a passive L-C filter whose cut off frequency is normally higher than the highest audio frequency the amplifier is being asked to reproduce. So in a 20Hz-20KHz amplifier a 25-30KHz cut off filter would be used. Feedback is nearly always implemented to get the distortion low, the output impedance low and the noise low.

Class G

amplifiers are really not a separate class but rather a variation of a class A-B amplifier. Their difference is almost entirely in the power supply. Referring to our notes under amplifier efficiency, it can be seen that if the amplifier is driving lower levels of signal, there is a large amount of net voltage impressed across the output devices. Multiply this net voltage by the current through the devices and we get HEAT! How can we reduce this heat? Well the current is "fixed" in that we need it to drive the speaker. So if we can reduce the value of the supply voltage we shall reduce the V-I product = HEAT.

Let us examine the above diagrams. A single transistor shown for simplicity but the bottom half of the output stage is a symmetrical PNP transistor

In the first row we see that the power supply is fixed at 50 volts. The load is 10 ohm (easy arithmetic). When the output is 0, the current through the transistor is 0 amps and VxI = 0 watts, when the output is 10 volts, the current through the transistor is 1 amp and the VxI = 40x1 = 40 watts (there is a net of 40 volts across the transistor). As we move to the right the calculations are similar just the V-I product changes as the output goes higher and the net voltage across the output transistor becomes lower.

As we can see the V-I product of the transistor generates the heat. If we could magically make the supply voltage track at say 0.001 volts ABOVE the required signal then the V-I product would always be 0.001 x [the current through the transistor]. If the current was 1 amp then V-I = 0.001 x 1 = 0.001 watts. Very nice indeed but a dream. If the current was 10 amps the V-I would be 0.001 x 10 = 0.01 watts, both insignificant dissipation figures in the transistor. This scenario is of course is not possible so we get to a practical situation.

In the second row of the above diagram, I have kept the supply voltage 5 volts ABOVE the required output voltage. So in each case the V-I product is always 5 x [the output current] since the voltage across the output transistor is always 5 volts.

The table below pertains to the first row of the main diagram. (Fixed Power Supply)

The table below pertains to the second row of the main diagram. (Variable Power Supply)

* Not 100% true as we have idling current but is still very small.

** Not true in practice as the transistor is NOT a perfect switch so its saturation voltage may be 0.5 to 3 volts depending on the max current flowing through it.

As we can see from the right columns of each table, the dissipation in the output transistors is greatly reduced in the second example.

The third part of the main diagram is a practical implementation of a variable power supply or Class G. The output stage has an initial supply voltage of 15 volts. As the peak value of the output increases, a control circuit monitors this. Just before the output signal is clipped, the second transistor which is supplied from a 30 volt rail is turned on, the 30 volts is applied to the main amplifier's transistor through one of the commutating diodes. Thus the transistor now has a "new" supply of 30 volts AS LONG AS THE SIGNAL IS ABOVE THE INITIAL 15 volts. The procedure is repeated when the signal reaches the 30 volt threshold then the supply is switched to 50 volts. Not as good as a continuously variable tracking supply but far better than a single fixed supply. With typical music the supplies will vary between the 15 and 30 volt rails 90% of the time and only peaks will require that the supply be switched to the 50 volt level. The table below compares the practical circuit to the first two rows, the first which is a fixed supply of 50 volts and the second which bumps the supply in 5 volt increments. This second example is practical but would require FIVE power supplies at 5v, 15v, 25, 35v and 50v and this would be a high parts count circuit.

The table below pertains to the last circuit of the main diagram.

As we see the dissipation in the output stage is lower than the single supply method but not as low as the 5 stage power supply. In our experience a 3-4 stage is a practical maximum.

Class H

is just a subtle variation of class G and was really only used by Soundcraftsman in their older amplifiers.