Amplifier Design Considerations- Output Stage

The output stage in a Power Amplifier, drives the speakers, and therefore is called upon to deliver heafty current and voltage swings, without distortion and noise.

CLASS A & B

The output stage is sometimes designed so that it is normally off, and starts conducting, only when the signal demands it.

This is refered to as a Class B stage.

The other extreme is that the output stage is designed to continiously conduct the Max current, irrespective of whether there is a signal or not. This continious current is called the bias.

When a signal arrives, part of the bias is converted to the signal and fed to the loudspeaker.

The balance continues to be wasted as heat.

This is Called Class A.

Typically, 90% or more of all commercial designs use a bias current in between the extremes of Class A & B. These are termed as Class AB.

Class A amps are usually better respected for their sound, than Class B amplifiers.

Class A amplifiers run HOT, and consume a HUGE amount of power ( specially the High Powered ones)

As an example, a 200 Watt Class A mono block amp will consume about 800 Watts, signal or no signal !

2 for stereo and left continiously on will consume 1.6 'units' of electricty every hour. Assuming Rs 6 per unit, it will cost the user over Rs 9,000 in electricity bills alone, every month ! And yes, Mark Lev recommended that their Monoblocks be left on continiously, for best results. ( They only had a standby switch, no power off from the Pre amp. ) That was about 10 years ago. Today's ML mono blocks do not indulge in such extravagance....

I was offered a pair of these monsters for about Rs 1 Lakh, last year, but the Electric bill scared me from even trying them out.....

QUASI COMPLIMENTARY & QUASI COMPLIMENTARY OUTPUT STAGES

When solid state power amps first emerged, only npn type of bipolar power transistors were available.

'Ideally' the output stage should deploy a symetrical, npn+pnp power transistor pair. Forced to use what was available, designers then used one of the npn power transistors driven at its input with a pnp driver transistor. The pnp + npn composite transistor behaved approximately as a pnp power ptrnsistor.

This was called "Quasi ( or false ) complimentary"

Most modern amp deploy Full Complimentary transistors, since today, matched pair npn + pnp transistors are available.

For the past decade or so, Qusi Complimentary output stages have been considered as 'compromise' designs.

However, about a year ago, they seem to have emerged as 'preferred' with the emergence of amps such as the Darzeel power amp.

So which is better .... Ah ! the latest take on that is " It all depends on the circuit design and how it is implemented'

Enough rambling from me on the topic.

Like I said initially, I don't even know if this is of interest on this forum.

If interested, do post ...

Cheers

hi amp nut,

if you could throw some, or a bushel full of light on how valve amps work it would be really nice. especially the bias part. what is bias and how does one bias a valve amp regularly?

thanks in advance

Enough rambling from me on the topic. Like I said initially, I don't even know if this is of interest on this forum. If interested, do post ...

Do continue. This is interesting. I knew a Class A amp consumed a lot of power. But I didn't think it would be this much.

Typically, 90% or more of all commercial designs use a bias current in between the extremes of Class A & B. These are termed as Class AB.

When you say "commercial designs" are you talking about solid state amps? O are tube amps included too?

on second thoughts perhaps it would be better to do the valve amp discussion under a separate heading? since this is just for the output stage? am sure there are a lot of valve amp fans here.

Shahrukh said When you say "commercial designs" are you talking about solid state amps? O are tube amps included too?

I guess that by a 'commercial amplifier' I would refer to a product that is sold as a finished product, and supported by some sort of a manufacturer's warranty....

Tube amps CERTAINLY included !

stevieboy said: hi amp nut, if you could throw ... a bushel full of light on how valve amps work it would be really nice.

I have worked all my time with Solid State, except some 30 years ago, when I took a course on Valve Radios, and worked with Valves.

Would be GREAT if some other forum memebers contributed to this topic too. Come on guys, there are many of you, who manufacture some Good Valve amps, so do share ....

A primer on Valve amps...... In a way, they are similar to npn transistors, ie they dont have a complementary pnp equivalent !

Also, Valves operate at relatively High Voltages and low current. On the other hand, speakers need relatively high current and low voltages.

Hence valave are interfaced with speakers using Transformers.

You can also parallel Many, Many Valves, and get them to drive speakers directly. These are OTL ( Output Transformer Less ) designs.

The reason for OTL.... the transformers are lossy and difficult to design. Their loss increases Rapidly with freq. Hence feedback needs to correct for Transformer losses. Transformers also add Phase shift. Phase shift is DIRTY. It converts the good negative feedback, into positive fedback, which makes an amp oscillate ( whistle) and blow itself and the speakers...

Ofcourse, when things are done right, Valve amps DO provide Blissful music ...

especially the bias part. what is bias and how does one bias a valve amp regularly?

VALVE BASICS:

A Valve consists of a Cathode which is usually heated by a filament, much like an incandescent bulb. When heated, the cathode emits elecetrons ( These are the thingies that carry electric current :-) )

The Electrons are negatively charged, and hence they will be attracted to any plate that is positively charged. Usually, a plate is located some distance away, and a high DC Voltage ( as low as 24 VDC but more typicaly 250 VDC to 1500 VDC ) ....Plate Voltage is applied to the 'Plate'. This plate is called the anode.

Now, a wire mesh can be introduced in-between the Cathode and the anode, but rather close to the cathode.

Since the mesh is close to the Cathode, a small voltage applied to the wire mesh, will repel some of the electrons just emmitted from the Cathode, and not let them travel towards the Cathode.

Hence the wire mesh is called the "Control Grid" since it can easily control the current flow between the anode and Cathode.


Small voltages at the control grid, can control a large current foow ( stream of electrons ) to the Anode.

Viola ! You have an amplifier!!

BIAS:
A small negative voltage ( typically -1.5 Volts ) is applied to the Control grid, and this is adjusted, till a predetermined current is established between the Cathode and Anode. This current flow is WITHOUT any input signal (music)

This is the BIAS voltage applied at the Control Grid, which sets the bias current between the Cathode and Anode.

As a valve ages, its Cathode depletes, and cannot emit as many electrons as when new

Hence the voltage at the control grid needs to be adjusted to re-achieve the specified current flow between the Cathode and anode.

Thats the purpose of "adjusting the bias" and its required fairly frequently in some Valve amps.

Some new ones have an Auto bias circuit....

( Finally, like us humans, Valves too, age, splutter and die.. )

Hope I did not loose you guys in this looong monologue.

Amp_Nut schrieb:

VALVE BASICS: A Valve consists of a Cathode which is usually heated by a filament, much like an incandescent bulb. When heated, the cathode emits elecetrons ( These are the thingies that carry electric current :-) ) The Electrons are negatively charged, and hence they will be attracted to any plate that is positively charged. Usually, a plate is located some distance away, and a high DC Voltage ( as low as 24 VDC but more typicaly 250 VDC to 1500 VDC ) ....Plate Voltage is applied to the 'Plate'. This plate is called the anode. Now, a wire mesh can be introduced in-between the Cathode and the anode, but rather close to the cathode. Since the mesh is close to the Cathode, a small voltage applied to the wire mesh, will repel some of the electrons just emmitted from the Cathode, and not let them travel towards the Cathode. Hence the wire mesh is called the "Control Grid" since it can easily control the current flow between the anode and Cathode. Small voltages at the control grid, can control a large current foow ( stream of electrons ) to the Anode. Viola ! You have an amplifier!!

Just an additional note to compliment what Amp_nut already wrote:

there are 3 basic categories of tubes/valves & they are
(1) triode
(2) tetrode &
(3) pentode

what Amp_nut described above is the construction & brief working of a "triode" valve. It's called triode because it has three elements in it: a cathode (-ve), control grid & a plate or anode (+ve).
The music signal is fed into the control grid & it modulates the control grid voltage. Basically, the music signal "rides" on the DC bias. So it's very imp to get the bias correct; otherwise, if the music signal goes too low or too high, the control grid voltage will hit a floor/min or a ceiling/max based on the B+ & B- supplies in the power amp. This will clip the music signal & create distortion.
Examples of triodes are: 12AX7, 12AU7, 12AT7, 6922, 7308, 6N1P, 6N23. The 12Axxx, 6922 & 7308 tubes are what they call "dual triodes" because they have 2 cathodes, 2 control grids & 2 anodes in 1 bottle.

The tetrode valve has all the 3 elements as stated by Amp_nut + 1 additional element called a "screen grid". This is placed between the control grid & the plate/anode. About half the voltage applied to the plate/anode is applied to the screen grid to enhance the amount of electrons flowing thru the tube. Thus, a tetrode has more gain than a triode.
Examples of tetrodes are EL34 & EL84. However, EL34 tubes are almost always used as triodes (better linearity) hence their "lush" midrange.

the pentode valve has all the 4 elements of the tetrode + 1 additional element called the "suppressor grid". This grid is placed between the screen grid & the anode. It is negatively biased & usually internally connected to the cathode. It's job is to suppress the electrons that bounced of the anode & would have been absorbed by the screen grid. The reason for this is that one needs to ensure that the screen grid does not conduct too much current otherwise it might get damaged too soon thereby decreasing tube life.
Examples of pentodes are KT88, KT66 (the ones used in the original Quad II power amp).
Sometimes KT88 are also called beam tetrodes.

Amp_Nut schrieb:

BIAS: A small negative voltage ( typically -1.5 Volts ) is applied to the Control grid, and this is adjusted, till a predetermined current is established between the Cathode and Anode. This current flow is WITHOUT any input signal (music) This is the BIAS voltage applied at the Control Grid, which sets the bias current between the Cathode and Anode. As a valve ages, its Cathode depletes, and cannot emit as many electrons as when new Hence the voltage at the control grid needs to be adjusted to re-achieve the specified current flow between the Cathode and anode. Thats the purpose of "adjusting the bias" and its required fairly frequently in some Valve amps. Some new ones have an Auto bias circuit.... ( Finally, like us humans, Valves too, age, splutter and die.. ) Hope I did not loose you guys in this looong monologue. :L

awesome awesome awesome!!!! i have some more poring over this topic to do and have tons of new questions

for starters - whats voltage and current? current is the flow of electricity and voltage is how high/much/strength? ok this i guess is as basic as it gets

second what's npn and pnp???

The reason for OTL.... the transformers are lossy and difficult to design. Their loss increases Rapidly with freq. Hence feedback needs to correct for Transformer losses. Transformers also add Phase shift. Phase shift is DIRTY. It converts the good negative feedback, into positive fedback

thirdly, by 'loss increasing with freq' what exactly do you mean? feedback means feedback of the music signal? current? ok i'm totally clueless here so don't laugh what is good negative feedback? and positive feedback?

well that should keep you busy till the other valve guys get here

bombaywalla your post requires a leeetle more reading so ur spared the questions for now. but great stuff. thanks guys!

For starters - whats voltage and current? current is the flow of electricity and voltage is how high/much/strength? ok this i guess is as basic as it gets

Nice to see un-inhibited questions !

Think of a Pipe carrying water.

The Pipe is the equivalent of the wire or Cable.

The AMOUNT of water flowing is the current.

The Pressure difference between the 2 ends of the pipe is the Voltage Drop accross the Cable.

A long pipe will impede the flow of water through it. That is similar to resistance/ impedence of the Wire.

Cheers

viren here if i remember right does not use global feedback in his valve design. any more light on negative feedback, global feedback? what it is and pros and cons of incorporating it in a design?

I got this question as a PM. The sender felt it too basic.

I think it is a GOOD Question, so I am posting it here :

This is too basic a query. What's the difference between current and power? What do they mean by a high current amp? Does high current mean it is "closer" to a class A amp?? Am I all confused?

I have explained the difference between Voltage and current.

POWER is Voltage x Current.

also

Power = Voltage x Voltage divide By Resistance

Hence if an amplifier delivers 10 Volts ( RMS... and that's another matter ) with 3 Amps of current, it delivers 30 Watts.

What do they mean by a high current amp? Does high current mean it is "closer" to a class A amp??

A touch of Basics:

OHMS LAW: Voltage = Current x Resistance ( or Impedence )

Typically Speakers are 4 Ohms or 8 Ohms.

Lets consider 8 Ohms first.

If you have an Amplifier rated at 30 Watts into 8 Ohms, we can use the formula :

Power = Voltage x Voltage divide By Resistance

30 = Voltage x Voltage divide By 8

Voltage = 1.94 Amperes of current, say 2 Amperes of Current.

However, if the same amplifier is to deliver 30 Watts into a 4 Ohm speaker, using the same formula, it needs to pump out 3.87 Amperes ( say 4 Amperes ) of current !


So an Honest 30 Watt amp should be able to pump out Atlest 4 Amperes of current.

NOW THE REAL WORLD:

1. A 4 Ohm speaker is actually 4 Ohms only at some frequencies. Its actual impedence actually varies significantly .... often from 3 Ohms ( or even lower ) to about 6 Ohms.

Hence an "Honest" 30 Watt Amp may not be able to deliver 30 Watts into that speaker at all frequencies !

2. Speakers have an IMPEDENCE, not a Resistance. ( An someone on the group address this difference ? )

Impedence effectively 'wastes' power, and more Voltage and current is required to drive power into an Impedence .

I'll leave the Impedence matter, for someone else to address, please ? )

In view of 1 & 2 above, Serious amplifier designers, ensure thgat the amplifier can deliver Much ore current than simple maths indicates.

Hence they may design a 30 Watt Amp capable of delivering current ( maybe for a short durations ) of 5 Ampere or more.

These are 'High Current Apmlifiers'.

A Class A Output stage does not automatically deliver high current.

A 30 Watt Class A amp for example will have to run a current in its output transistors of well over 3 Amps ( 3 x 1.414 actually ... that 1.414 is a conversion factor for DC to AC RMs ... ). Lets Say it has a continious DC current of 5 Amps in the output Transistors.

Using the Formula:

Power = Voltage x Voltage divide By Resistance

we see that 30 Watts into 8 Ohms implies an output voltage of 15.5 Volts AC RMS or 15.5 x 2 x 1.4 = 43.7 Volts DC.

Keeping a few volts extra for the Output Transistors, they will have atleast 48 Volts DC accross them.

So lets see how much power is required, for a 30 Watt 'Honest' ( Not Very High Current ! ) Class A Amplifier :

50 Volts x 5 amperes = 150 Watts.

AND THAT IS FOR 1 CHANNEL.

2 Channels: 300 Watts being 'wasted' as heat continiously, whether the Class A amp is playing music or not.

Want a High Current Class A ?? Phew !

viren here if i remember right does not use global feedback in his valve design. any more light on negative feedback, global feedback? what it is and pros and cons of incorporating it in a design?

Viren, Bombaywalla and others.... PLEASE contribute to this and the other queries...

what's npn and pnp???

A transistor is made like a sandwitch... 2 slices of Bread on the outsides and different stuff in the middle.

For an npn Transistor, the outside layers are are made of a material with excess Negative particles ... This is Called 'n type' material.

The centre is made of material with excess Positive particles ... This is Called 'p type' material.

Hence the npn transistor has: n-type + p-type + n-type

Its cpmliment

pnp transistor has: p-type + n-type + p-type.

The 3 terminals of a Transistor (applicable for both, npn or pnp transistors ) are :

Emitter Base & Collector

These correspond to

Cathode Control Grid & Plate in a Valve (Vaccuum Tube)

Yes, get a 13.5 Volts, 5 Ampere Regulated PSU.

this is from the t amp post...

so the t amp would put out roughly 70 watts of power?

me just doing a practical lesson here

A transistor is made like a sandwitch... 2 slices of Bread on the outsides and different stuff in the middle.

unknowingly, you hit on the right motivational example.. food!!!

Nah.....I WISH !

The amount of current that will pass thru the resistor will depend on the Volage accross the resistor ( or speaker)

Hence for the T-Amp, powered by 13.5 VDC.

First convert the DC to Volts AC RMS by dividing by 2x1.4

13.5 / 2.8 = 4.8 Volts RMS ( in practice some Voltage will be lost inside the T-Map's transistors, and the real world value will be less.

Now Power = Voltage x Voltage divide by Resistance

Assume a 4 Ohm speaker :

Power = 4.8 x 4.8 / 4 = 5.8 Watts.

For an 8 Ohm speaker, you will get Half that power...

voltage = current x resistance

therefore 4.8 = 1.2 x 4 for a 4 ohm speaker and what's the equation when it's an 8 ohm speaker? 4.8 = 0.6 x 8 so the current requirement eases off and becomes a more gentle load?

thanks for the patience. math was never my forte.

when it's an 8 ohm speaker ...... the current requirement eases off and becomes a more gentle load ?

yessss !

looks like you have got a good feel now...

i always knew that as the speaker resistance reduced the current requirement went up. just didn't know how exactly it all worked. now i do

Interesting! But way too much math. And please don't say "it's elementary stuff" !

And please don't say "it's elementary stuff" !

yeah too much math i agree but wot to do? physics will be physics wonly

Hi Amp_nut..this is something which makes other forums envious..I'm sorry I initiated this and was away for a while... buurrpp

back to talk..whats the role of an Op amplifier.or maybe even give me a brief explaination as to what vital components go into an amp..step by step..

whats the role of an Op amplifier ?

Put briefly, and rather crudely, an Op Amp is one Huge Blop of gain.

Its a circuit block with a gain for approx. 1 Million !

Not much that you can do with it on its own.... but you can apply feedback, and make it do almost anything you want it to, as an amplifier.

( I do hope the feedback question posed earlier is taken up by other forum contibutors.... Come on, lurkers... )

Usually Op Amps cannot deliver much current... about 50 mA to 100mA. ( 1000 mA = 1 Ampere )

Hence Op Amps are typically used in low signal level stages, such as pre amplifiers...

Not all op amps are the same. There are Hundreds of Op Amp ICs available commercially, each optimised for a different purpose, eg Low Noise, High Bandwidth ( ie freq response ) etc..

stevieboy schrieb:

viren here if i remember right does not use global feedback in his valve design. any more light on negative feedback, global feedback? what it is and pros and cons of incorporating it in a design?

Loaded questions, stevieboy!!

Feedback, in general, is a concept used in electrical design wherein part of the output signal is inverted in phase & fed back to the input. Inversion of phase means that the fedback signal is 180 out-of-phase w.r.t. the output signal. The ampl of the signal being fedback is dictated by the 'feedback gain' that the designer puts in the feedback path.
The reason for employing negative feedback is that we want to keep under control certain design criteria that are important to the performance of the circuit such as output noise, harmonic distortion, intermodulation distortion, output impedance to mention a few.
At the input when we subtract a portion of the output signal from the input (remember, it IS a subtraction because the fedback signal is 180 out-of-phase with the input) we create an 'error' signal. This error signal informs us by how much the output signal differs from the input. If the error signal is large, the circuit will work to reduce the output amplitude so that the error signal decreases. And vice-versa. The basis reasoning behind this is that if the circuit is operating with a large(r) output, it is probably operating beyond is linear (biased) region thereby creating distortion at the output beyond what the designer is willing to accept.
The negative feedback employed is real-time in that it happens as the circuit is operating BUT...............
there is a delay in propagating the signal from the input to the output (we all know that if a signal is fed to the input of a circuit, it takes a finite amount of time before it appears at the output. This time can be very small, say microseconds or even nanoseconds, but it is NOT zero). Hence, the negative fedback signal is time-delayed w.r.t. the input when it is used to create the error signal. So, the present/right-here-right-now input signal is being compared to a delayed & amplified version of itself (the fedback signal) & the circuit is making a decision to inc or dec its output.
In AUDIO, where the human ears are very sensitive to distortion in the 300Hz-3KHz region, this concept of negative feedback is a fundamental problem. If you have ever watched a music or speech signal on an instrument called an oscilloscope, you would have noticed that music/speech is a rapidly changing signal - it can change dramatically in amplitude & phase depending on the music. Think Beethoven's 5th symphony - how it cresendos & how is becomes barely listenable. Think Chopin's piano concertos where he's banging the piano keys & then, the next instant, he's barely tapping them. The music signal is going thru a collossal change.
So, in negative feedback, one is using "yesterday's" fedback signal & comparing it to "today's" input signal & making a decision. The issue is that "yesterday's" signal might have no resemblance whatsoever to "today's" signal. So, how can the decision made by ckt be relevant?
This is why negative feedback has been the bane of audio for all these years.
Designers used this concept in the 1960s, 1970s, 1980s to keep o/p impedance & distortion low but the music thru these power amps (used esp in power amps because they are large signal circuits as compared to preamps, which are small(er) signal components) sounded like s***! The sound was closed in & 2-D. However, the bass was fantastic ('cuz the amp o/p impedance was low hence the amp's damping factor was relatively high & it could damp that woofer, by God!). Also, these amps measured superbly on the testbench in the lab. So, if Stereophile ever got hold of these amps & John Atkinson measured them in his lab, he would generally wet his pants when he saw the measured performance! However, their sound was anything but good.

Now, the above negative feedback that I just described is called "global negative feedback" 'cuz the output signal is fedback all the way back to the input. This feedback circumnavigates the entire circuit. It is the worst possible feedback that can be employed in audio & the amps using this feedback sound the most "hi-fi".

The other type of feedback that can be used is called "local negative feedback". As the name implies, it is very localized within the circuit. As a simile: if you have a toothache, you apply Zylocaine to your gums where you have the toothache, correct? This is like local negative feedback: you have a problem in a particular spot, you take care of that spot only. You do not go to the doctor & ask him for general anesthesia for your toothache. This would be global negative feedback: take care of the whole body. Local negative feedback is superb & can allow the designer to achieve his/her design goals & make the power amp sound absolutely incredible. There are a few well-known companies that employ local negative feedback: Pass Labs & his 2 older companies- Threshold & Forte. The 2nd manuf is Ayre. There are others, I'm sure, but the list is too large to mention here.

This is an important statement: Local negative feedback ALWAYS exists in EACH & EVERY electronic device/transistor & in each & every circuit. It might be hidden to the casual observer but the trained eye will find it.

So, local negative feedback simply cannot be shrugged off. In fact, no electronic device would operate if there wasn't local negative feedback.

Usually, local negative feedback within a transistor is not enough to linearize the circuit & so the designer uses local negative feedback around an amplifier stage - input stage, gain stage, output stage or bias stage. The feedback is local to these stages & its influence does NOT transcend that particular stage. Local negative feedback might be employed in more than 1 stage at the same time but they do not influence each other; their overall effect is felt in the entire circuit.

So, in audio, global negative feedback should be avoided if you want your audio gear to sound like music. "Sound like music": this is a loaded statement 'cuz diff people are at diff stages in this audio hobby implying that their ears are at diff levels of being trained to hear the output sound. A relatively newbie will not perceive the ill-effects of global negative feedback while a more experienced person will pick it up immediately. The more you expose yourself to live music & the more you understand what the timbre & tonality of instruments & voice are, the better you'll become as an audiophile & the better your ears will be trained.
Local negative feedback is preferred & these days the audio manuf have realised it themselves. Global negative feedback is a "badword" these days BUT............
some manuf STILL use it. It comes under several guises: in valve/tube circuits, it is called "ultra-linear" operation. Take a look at the schematic for such a power amp!!

Sorry, loooooong answer to what seemed an innocent question.
Hope that this sheds some light on the matter. It's a very complex question that actually requires knowledge of control theory (taught in engineering school).

GOOD Post !

You have covered all the bases !

Cheers

Amp_Nut schrieb:

Hence if an amplifier delivers 10 Volts ( RMS... and that's another matter ) with 3 Amps of current, it delivers 30 Watts.

OK. Amp_nut, it appears that you are talking about the amplifier output. Am I correct?
If so, 30 Watts is DC power & it is DC RMS.
I think you & I agree here. Correct?

Amp_Nut schrieb:

Power = Voltage x Voltage divide By Resistance 30 = Voltage x Voltage divide By 8 Voltage = 1.94 Amperes of current, say 2 Amperes of Current. However, if the same amplifier is to deliver 30 Watts into a 4 Ohm speaker, using the same formula, it needs to pump out 3.87 Amperes ( say 4 Amperes ) of current !

OK. once again the 1.94A of o/p current is DC RMS. Agree? Same deal w/ the 3.87A - it's DC RMS.

Amp_Nut schrieb:

NOW THE REAL WORLD: 1. A 4 Ohm speaker is actually 4 Ohms only at some frequencies. Its actual impedence actually varies significantly .... often from 3 Ohms ( or even lower ) to about 6 Ohms. Hence an "Honest" 30 Watt Amp may not be able to deliver 30 Watts into that speaker at all frequencies ! 2. Speakers have an IMPEDENCE, not a Resistance. ( An someone on the group address this difference ? ) Impedence effectively 'wastes' power, and more Voltage and current is required to drive power into an Impedence . I'll leave the Impedence matter, for someone else to address, please ? )

OK. I'll bite!
The push back from an electrical ckt when there is current flow is called "impedance" in general. This is so because, unless we know for a fact, electrical circuits, in general, do not a constant over frequency push back (for a lack of a better word) to current flow.
Thus, impedance is usually written as a complex number (what you were taught in 10th standard, unless you decided that sleeping in class was a better alternative!). So, complex numbers are written as i + kj or a + jb depending upon which part of the world you live in. Same difference! the "a" denotes the pure resistive element in the impedance - the one that burns current & gets hot to the touch. The 'kj' or 'jb' part describes the reactive portion of the impedance - the part that us made from capacitors &/or inductors. These 2 elements do not dissipate heat as a resistor; rather, hold charge in an electrical field (capacitor, C) or in a magnetic field (inductor, L). Charge is put onto the C or L in one part of the music signal or power signal & is removed during the other part of the music signal. If one looks at a power or music signal & takes a long-term average of it, one will find that the long term average is zero! Thus, capacitors & inductors play this game of taking & giving charge & over the long-term have no net charge in them.
Another thing to note: both C & L components have an impedance that is frequency dependent. For a C, the impedance at DC/0Hz is infinite - practically some very large number. As the freq increases, the impedance of the C decreases. The L component is exactly opposite - it's a dead short at DC/0Hz & as the freq goes up, its impedance increases.
R, L & C components are the only components that can be used to make a passive filter in a speaker to construct its cross-over network (fancy name for an audio freq filter). So, you can now see why the speaker impedance is not constant over freq. It's those x-over L & C components!
Since the L & C components are frequency selective by their very nature i.e. allow some freq to pass thru unimpeded, allow some others to pass thru with some attenuation & allow still others with a lot of attenuation, you can see that certain music signal frequencies get treated differently as they pass thru a speaker x-over network.
If we assume (& this is a bad assumption but I need to make it so that we have just 1 variable while holding the other constant) that the amplifier output is constant over frequency, then we can assume that the amplifier is putting out 30W (using Amp_nut's example) 20Hz-20KHz. However, if the speaker x-over network selectively treats the freq differently, the power in certain freq will be attenuated before it actually reaches the speaker driver/cone. Needless to say, what you will hear at that frequency(ies) will be inadequate SPL. What will you do? You will turn up the volume!
That is why Amp_nut said "Impedence effectively 'wastes' power, and more Voltage and current is required to drive power into an Impedence".

Amp_Nut schrieb:

A 30 Watt Class A amp for example will have to run a current in its output transistors of well over 3 Amps ( 3 x 1.414 actually ... that 1.414 is a conversion factor for DC to AC RMs ... ). Lets Say it has a continious DC current of 5 Amps in the output Transistors.

Amp_nut: why are you converting from DC to AC? At the amp output we are DC all the way. Power is spec'd at DC RMS. Are you trying to calculate the AC power consumed from the mains?

Amp_Nut schrieb:

Using the Formula: Power = Voltage x Voltage divide By Resistance we see that 30 Watts into 8 Ohms implies an output voltage of 15.5 Volts AC RMS or 15.5 x 2 x 1.4 = 43.7 Volts DC.

Amp_nut: I believe that this is not correct because the amp output voltage is DC. The amp's o/p voltage should be 15.5V DC. Am I missing something here?

(we all know that if a signal is fed to the input of a circuit, it takes a finite amount of time before it appears at the output. This time can be very small, say microseconds or even nanoseconds, but it is NOT zero).

How right you are bombaywalla, in assuming that all of us here have some rudimentary understanding of electronics.

Cheers
K

P.S - Keep it up.This is facinating, even as I struggle to understand the mathematics of it.

Bombaywalla said: Amp_nut: why are you converting from DC to AC? At the amp output we are DC all the way. Power is spec'd at DC RMS. Are you trying to calculate the AC power consumed from the mains?

I am not sure I have understood your comment....

DC is ONLY present INSIDE the amplifier. DC is required Internally by the amplifier, to make it work.

The outside world of signals and Music are all AC ( Alternating Current)

There should be No DC at the input or speaker terminals.

Output Power at the speaker terminals is AC Power.

RMS is a mathematical manipulation / calculation used for Varying ( AC ) signals.

Given time, I will post seperately on " What Is A Watt ? "

Cheers for now....

RESISTANCE & IMPEDANCE

Let me take a shot at explaining the difference between resistance and impedance.

Consider a straight bar of metal. If you push it from one of its ends, the push is conveyed to the other end of the bar immediately.

On the other hand consider a soft metal spring. If you push it at one end, the spring compresses and takes time before the push travels to the other end of the spring. Hence, there is a time lag between the action and reaction in a spring.

The metal bar is similar to a pure resistance. As soon as a voltage is applied, a current flows through the resistance. The current and voltage occur at the same instance.

However, the spring is similar to an inductor. When a electrical voltage is applied to the inductor there is a time lag before the current starts flowing in the inductor. The same AMOUNT of current will flow in the resistor or inductor if their resistance or impedance is the same. However in the inductor the current flow takes place after some time.

However, the watts of power require the product of voltage and current AT THE SAME INSTANCE.

Hence, let us say a voltage varying between 0 to 2 volts is applied to an inductor. The voltage may reach 1 volt before any current starts flowing in the inductor.

By the time the voltage at across the inductor is 2 volts, only 1 ampere of current may be flowing through the inductor. Hence the power delivered to the inductor is only :

2 volts x 1 ampere = 2 watts

Compared to this, if 2 volts is applied across a resistor, an instantaneous current of 2 ampere will flow through it delivering a total power of 4 volts into the resistor.

Hence it is clear to see that an impedance does not absorb as much power as a resistor.

Loud speakers have a major component of impedance rather than pure resistance. Hence to pump the intended power, often a larger current delivering capacity is called for, than what simple Ohm's Law indicates for a pure resistance.

That is why Amp_nut said "Impedence effectively 'wastes' power, and more Voltage and current is required to drive power into an Impedence".

Amp_Nut wrote: Power = Voltage x Voltage divide By Resistance we see that 30 Watts into 8 Ohms implies an output voltage of 15.5 Volts AC RMS or 15.5 x 2 x 1.4 = 43.7 Volts DC. Bombaywalla said: Amp_nut: I believe that this is not correct because the amp output voltage is DC. The amp's o/p voltage should be 15.5V DC. Am I missing something here?

As I clarified in my earlier post today, the DC is ONLY inside the amplifier, and is required to power the circuits.

The circuits have a certain max DC voltage that they are supplied. This DC voltage can be converted to an AC voltage and the AC Voltage Peak to peak can be at most ( usually a touch less) tthan the DC voltage power to the circuits.

Hence the max AC (RMS) voltage that can be outputted by the amplifier is the DC Voltage divide by 2.83 ( ie 2 x square root of 2 )

Hence a 60 V DC power supply can generate only 60 / 2.83 = 21.2 Volts AC RMS

Into an 8 Ohm resistor, it will deliver 56 Watts.

Into a 4 Ohm resistor it will deviler 112 Watts, IF THE AMPLIFIER OUTPUT STAGE CAN DELIVER TWICE AS MUCH CURRENT.

A LARGE number of amplifier cannot achieve this feat :-(

What's A Watt ?

There have been lots of interpretation on the 1 watt of electrical power. Almost all these interpretations are intentionally corrupted for marketing purposes to provide consumers with highly inflated and totally irrelevant "specifications."

Let us take a proper look at "What's A Watt ?"

1 watt of electrical power is simply the energy produced when 1 ampere of electrical current close through a resister produccing a votage drop of 1 volt across the resistor.

Hence : Watt = Voltage x Current that occur at any instant of time.

DC WATTS :

In the case of DC power that is where the voltage and current do not change with time, 1 watt is simply voltage x current.

AC WATTS :

All music signals vary with time.

Equipment is often tested using Sine waves. A Sine wave is gently varying signal. How fast that signal varies every second is its frequency. Hence a 500 Hz signal varies 500 times every second.

Hence it is necessary to define what is a watt for an AC signal.

Mathematicians have arrived at a quantity called RMS (Root Mean Square). The RMS value of a varying signal gives the exact equivalent heat or watts of energy contained in that signal.

RMS is like a simple "average" but more accurately indicates the power or energy contained.

Let us say a voltage at different instances is 2, 4, 6 and 10.

Its RMS value is :

2x2 + 4x4 + 6x6 + 10x10 divide by 3. Now take the square root of the answer = 7.2

Hence the RMS value of 4 numbers viz : 2, 4, 6 & 10 is 7.2

( Its arithmetic average is 5.5 ! )

The peak of the 4 numbers is 10. But you cannot use that number to calculate "PEAK Watts' by taking Peak Voltage and multiplying it with Peak Current !

That does not give Peak Watts, only PEAK Bull Shit !

Krish schrieb:

How right you are bombaywalla, in assuming that all of us here have some rudimentary understanding of electronics. Cheers K

The posts were directed mainly at Stevieboy & Shahrukh, who had posed the questions in the 1st place & who professed their ignorance in these electrical principals. I'm *not* trying to be condescending. FWIW.

Hi,

Some more thoughts on feedback:

Feedback is basically a corrective measure to reduce non-linearities in a circuit. It is also used to adjust gain in a high gain circuit.

The best way to design a circuit is to make it absolutely linear, so the output follows the input exactly. If everything were linear, no feedback would be required. Obviously, there are no perfect amplifiers.

There are three basic stages in an amplifier - the input, the driver, and the output stage. The attempt is to make each stage as linear (perfect) as possible. The input and driver stages usually run in class A mode, handle the full signal, and so are quite linear. A small amount of local feedback within the stage is used sometimes to smooth the signal out some more.

It's the output stage, handling high power, where nonlinearities come up. The signal is split in a push-pull stage, goes through a crossover region; devices heat up and change characteristics; the speaker being a reactive load feeds back into the amp; and so on. In a valve amplifier, the output transformer, with its leakage inductance, and stray capacticances, causes all sorts of interactions. Here, if anywhere, corrective action is required.

If the corrective action, through feedback, is done within the output stage only, it is considered local feedback. In valve amps, this is through the ultra-linear transformer connections, and through cathode feedback (again, as a separate winding in the output transformer). With good design, this can produce excellent sound.

The other alternative is to include the entire circuit within the feedback loop, from output to input. This is global feedback. It does TRY to correct nonlinearities throughout the amplifier. But, introduces problems of its own. As explained by "bombaywalla", it's like the dog chasing its tail. Correction always happens after the fact!
Traditional measurements, Total Harmonic Distortion,THD, do show a reduction using global feedback. But subjective listening shows detrimental sound quality. What global feedback does is reduce second and third harmonics, but add small amounts of higher order harmonics. It is these higher order harmonics that create that harshness to sound.

So, good design suggests highly linear individual stages, with judicious amounts of local feedback!

Viren.

Nice and very well balanced post, Viren.

Also explained well, in simple language....

thanks bombaywalla, amp nut and viren. it took a few days but i finally got through this crash course whew. very very interesting i must say even though math and physics was involved

Loud speakers have a major component of impedance rather than pure resistance. Hence to pump the intended power, often a larger current delivering capacity is called for, than what simple Ohm's Law indicates for a pure resistance

so why are speakers built with more impedance than resistance if resistance is better in providing a gentler load to the amp? am sure there's some inherent properties that make using more capacitors than resistors?

this is certainly a nice change from simply endlessly discussing which speaker is better and which is lousier

I must admit, I'm a little lost over here. Nevertheless, I'll be taking printouts of all these threads and poring over them this weekend.

steveieboy, this is the time for some research of your own

try this for a starter

http://www.epanorama.net/documents/audio/speaker_impedance.html
http://www.whathifi....D=34&newssectionID=3

so why are speakers built with more impedance than resistance if resistance is better in providing a gentler load to the amp?

Ideally an inductor (speaker voice coil) should have no resistance, only inductance. But there is always a finite amount of resistance of the voice coil, as a rule of thumb half its impedance. Secondly inductive reactance or impedance increases with frequency, usually specified @ 1khz. Therefore 8 ohms at 1 khz would be lets say 20 ohms at 10 khz. In contrast to this resistance remains the same over the entire frequence spectrum.

Resistance of the voice coil dissipates electrical energy as heat wheras the inductance is what moves the coil in its former inside a magnetic field. The effeciancy of a driver is thus dependant on the ratio of impedance to DC resistance, the lower the resistance the higher the efficiency. So its always good to have more impedance than resistance. Crossover networks can be designed to have an almost constant impedance over the entire audible frequency range. To learn more about passive crossover designs refer to Rod Elliot Design of Passive Crossover Networks.

hey thanks for the links arj,

the what hi fi one was a little basic. covered stuff i already know. the other one went largely into territory i don't particularly want to venture. though i did get some useful stuff. i am doing parallel research of my own (in fact a lot of spin off points cropped up from amp nut and bombaywalla's posts) but nothing matches someone actually taking time out to answer your question. cos that way unwanted fundas are not introduced, keeping it simple and it's a learning ground for lots of other guys here who am sure are following this keenly but are too shy to ask a simple question.

more links from anyone else too most welcome.

me i'm going to have a today evening. i need a break!

The posts were directed mainly at Stevieboy & Shahrukh, who had posed the questions in the 1st place & who professed their ignorance in these electrical principals. I'm *not* trying to be condescending. FWIW.

Oops, sorry about that Bombaywalla.I was just kiddin',tongue firmly in cheek.

No offence to you.

Cheers

so why are speakers built with more impedance than resistance if resistance is better in providing a gentler load to the amp? am sure there's some inherent properties that make using more capacitors than resistors?

A speaker voice coil is basically several turns of wire, placed between a strong magnetic field.

The turns of wire, is Exactly how an inductor is made ! Hence a speaker inherently has a fairly lage inductance.

The coil of wire, when it moves rapidly in a magnetic field, has a voltage induced in it. Depending on the behaviour of the speaker in it enclosure ( and the room), at a specific freq ( the speaker behaves differently at different frequencies ), the current that folws may lead the voltage, creating a capacitive result, causing the speaker to behave as a capacitor.

Add to all this, a crossover network, which has several inductors and capacitors, and you have a fairly complex behaviour, from the speaker in terms of inductive and capacitive loading....

The capacitors and inductors in a crossover are necessary to route different frequency bands to different drivers, eg the bass to the Woofer, etc...

stevieboy schrieb:

thanks bombaywalla, amp nut and viren. it took a few days but i finally got through this crash course whew. very very interesting i must say even though math and physics was involved :pros

Welcome! sorry to have "lost" you in the material posted but there doesn't seem to be any other way to explain the concepts you asked. Hopefully re-reading + your parallel research will ease the situation for you.

stevieboy schrieb:

so why are speakers built with more impedance than resistance if resistance is better in providing a gentler load to the amp? am sure there's some inherent properties that make using more capacitors than resistors?

To compliment what Behram & Amp_nut already posted: a speaker manuf doesn't start by saying 'let me build a speaker with more resistance than capacitance' or vice-versa.
since a human ear cannot "listen" to electrical signal as they exist in an electrical ckt, we need a TRANSDUCER that will convert electrical signals to mechanical signals. Additionally, the human ear listens to sound when air around it compressed or rarefied.
So, what device should I "invent" to convert electrical to mechanical.
Faraday's Law comes to mind & from there if a paper/plastic/carbon fibre cone is attached to a winding of electrical wire in which the electrical signal (that I want to convert) flows & I immerse this winding of electrical wire in a magnetic field, I find that I can compress or rarefy the surrounding air to create sound!
However, what is the name of this winding of electrical wire? An inductor (L) - a reactive component of the impedance. There doesn't seem to be a way out of this unless we begin to hear differently.
So, we have to deal w/ reactive components of the impedance in a speaker system.
As Amp_nut pointed out, we also have a filter inside the speaker box right behind the speaker terminals called the cross-over network whose job it is to divvy up the audio spectrum & pass along the appropriate sections to the appropriate driver - low freq section to woofer, high freq section to tweeter & middle section to midrange.
The only way we know how to divvy up the audio spectrum is to use R, L & C components. Once again there doesn't seem to be a way out (so far). Hence, once again, we need to deal with reactive components & speaker impedance varying with frequency (because the reactive component impedance varies with frequency).
In fact, the amp doesn't "see" the voice coil (that winding of wire immersed in a magnetic field which is attached to the cone) because of the speaker cross-over network. The amp basically pours its o/p power into the x-over network & loss in this network strain the amp.
There is an electrical network called a ZOBEL network that can be used to trick the power amp into thinking that the impedance is flat over 20Hz-20KHz.
Speaker manuf Merlin Music uses such a network if you have looked @ the its speaker binding posts in the rear.
Additionally, several speaker manuf use a Zobel network across the woofer terminals (inside the speaker box invisible to the user) tricking the x-over network into thinking that the woofer has a constant impedance over freq.
Both locations of Zobel networks ease the strain on the power amp.

Krish schrieb:

Oops, sorry about that Bombaywalla.I was just kiddin',tongue firmly in cheek. No offence to you. Cheers

Hi Krish,
Thanks! No offense taken!

The only way we know how to divvy up the audio spectrum is to use R, L & C components. Once again there doesn't seem to be a way out (so far).

that's exactly what was going round in my head. surely someone somewhere can invent a more efficient way? hmmmm...

planar speakers dont use voice coils right? apart from quad why aren't there more mainstream speakers like that? and why is it so exp? is it cos of the plates/membranes? discussion anyone? havent had the good fortune of hearing a speaker like that so far. apparently they present a really 'togther' sound

stevieboy schrieb:

planar speakers dont use voice coils right?

Correct! planar speaker examples are Magnepan (Maggies), the older & highly revered among its fans - Apogee speakers. These types of speakers use a long strips of metal ribbons that are made somewhat electrically conductive.
Ribbon speakers examples are Newform Research (Canadian), VMPS, Piega & many others. These use long strips of mylar made mildly conductive. These ribbon tweeters come is many sizes from 6" or less to 48".
Electrostat speaker examples are Martin-Logan, Cadence (India), Soundlabs, Innersound to name a few.
All of them reply on a light-weight transducer to vibrate (move back & forth is a local magnetic field) to create sound. Since the transducer is so very light, it has a very quick response time. Hence the light, airy, holographic sound from them - there is hardly any mass of the transducer to be driven so it can respond to the vagaries of the music signal like no other transducer.

stevieboy schrieb:

apart from quad why aren't there more mainstream speakers like that?

IMHO, Quads are not mainstream. It's a name that that has become synonymous w/ planar speakers. Peter Walker of Quad invented this sort of driver & his speakers have found zealous fans worldwide. There seem to be equal detractors of Quad speakers as there are fans. I've never heard a Quad but I have a Magnepan. The Maggies are indeed holographic in the mids nearly like an electrostat. However, they do not play a broad spectrum of music genres with with equal panache.

stevieboy schrieb:

and why is it so exp? is it cos of the plates/membranes? discussion anyone? havent had the good fortune of hearing a speaker like that so far. apparently they present a really 'togther' sound

The manuf of planars, ribbons & electrostat panels is a very specialized process. It is also an unconventional driver. One needs to make these drivers very large if one wants to get the entire sound spectrum from them. Otherwise, the manuf compromises & puts a ribbon tweeter with regular cone drivers like VMPS & Piega.
Not many people can have accomodate a speaker that is 5-7 feet tall & about 4-5 feet wide in their houses! It's an intimidating size for a speaker & the wife is likely to reject it in her living room.
Thus, planar, ribbon & ES speakers are quite low volume compared to regular cone driver speakers. Hence the higher cost.

[quote]

I've never heard a Quad but I have a Magnepan. The Maggies are indeed holographic in the mids nearly like an electrostat. [/quote]

Which of the maggies do you have ? do you need a sub to flesh out the Bass ?

Ive heard folks say that with a powerful enough amp, the bass can be quite ok. id that true ?

I am sure more folks would be interested to learn about your setup

Arj schrieb:

Which of the maggies do you have ? do you need a sub to flesh out the Bass ?

No, I do not own Maggies. I heard a pair at a friend's house. I cannot remember the model number, sorry! They were not the biggest size that Magnepan makes.
I would say that these speakers produced sufficient bass for the small room that they were in. If I had to guess, I'd guess that the room would have been approx 300 sq-ft. These speakers had bass quality rather than bass quantity (ported speakers).
However, this friend listens to a lot of Jazz trios, quartets, vocal, country folk.
I listen to a wider variety & so when I played some of my stuff on his system, the Maggies were too polite. I wanted some more growl in the music (which I think it demanded). Anyway, that friend realized this as well later & those Maggies are now sold!

Arj schrieb:

Ive heard folks say that with a powerful enough amp, the bass can be quite ok. id that true ?

No idea, Arj!
You throw a large enough amp at almost anything & you could solve the issue 95% of the time. However, at what financial cost? & at what sonic cost?
Most of these kilowatt amps have lots of muscle but no soul! So, they end up producing great sound but no music. My opinion, of course.

Arj schrieb:

I am sure more folks would be interested to learn about your setup

Well, you asked so here it is. However, I believe that it's not relevant to this discussion.
My setup consists of a Wadia 861SE CDP, a Bluenote Bellavista Sig turntable w/ a Benz L2 MC cartridge, a Convergent Audio SL1 Sig Mk3 preamp, Symphonic Line RG4 Mk3 monoblocks, Green Mountain Audio c1.5i 1st-order x-over speakers w/ regular cone drivers.