Impedence matching in Audio

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
Today I read something tht contradicts with something tht was taught to me decades ago.

I am not trying to prove anyone or any theory wrong.

What I want is to clear this misunderstanding, whether it is me or any other.

This topic is purely based on Audio and Audio only. Any other source is out of the question and outside this topic mostly.

From what I have learned Audio circuit deals with something known as Impedance. Impedance reflects to AC signals and Audio consists of purely AC signals, so any circuit built to manipulate it has some resistance to it, whether it is an input or an output.

Mainly Impedance is concerned in Amplifiers. It does not matter a balanced XLR preamp or High Power amp driving 8Ω Transducers.

An amp can have a high input impedence in order of KΩ and an output impedance which is much much lower.

Today one member of AAC wrote that impedance matching is never done in Audio amps.
My question is why?

Taking a commercial amp, say for eg: a 100W amp.
It specs says, Input level is 700mV, has input impedance of 47KΩ.
Delivers 100W to 8Ω at distortion level of 0.04%.

If Impedence matching is not a concern, can anyone explain how a 100W, 16Ω speaker that is driven by the above amp at it's full capacity will deliver 100W of power.

And if the input is connected to a source of 700mV at 20KΩ could deliver 100W to a 8Ω speaker.

I know I have proven that impedance shud be matched when taught electronics decades ago. Now I am a bit rusty at the theory part but I know I can prove that impedance shud be matched for maximum power transfer, whether it is to an input or to a speaker

I will need to go back to the books and dig up my tutorials to show it here.
This might take time.
In the meantime I would like to know what every respectable member here knows about this theorem

Thank you
 

thatoneguy

Joined Feb 19, 2009
6,359
With Audio Power Amplifiers, the output impedance is put as close to zero as possible, some are stable down to 0.5Ω and lower. This allows the amp to source more current across a wide range of frequencies and speakers.

Trying to match an amplifier to a speaker is essentially not possible, as the voice coil is an inductor, the impedance of which changes drastically depending on the input frequency and position of the speaker cone.

As for low level, most of the higher end pre-amps I've seen have been DC Coupled rather than AC coupled, generally following the same rule of high input impedance, low output impedance.

To do otherwise would be tuning for a specific range. Audio amplifiers have a very wide/low Q, resulting in flat frequency response in the audio range.

--ETA: The perfect "Audiophile Amp" has a "flat" bandwidth from DC to Light, Infinite input impedance, and zero output impedance.
 
Last edited:

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
With Audio Power Amplifiers, the output impedance is put as close to zero as possible, some are stable down to 0.5Ω and lower. This allows the amp to source more current across a wide range of frequencies and speakers.

Trying to match an amplifier to a speaker is essentially not possible, as the voice coil is an inductor, the impedance of which changes drastically depending on the input frequency and position of the speaker cone.

As for low level, most of the higher end pre-amps I've seen have been DC Coupled rather than AC coupled, generally following the same rule of high input impedance, low output impedance.

To do otherwise would be tuning for a specific range. Audio amplifiers have a very wide/low Q, resulting in flat frequency response in the audio range.

--ETA: The perfect "Audiophile Amp" has a "flat" bandwidth from DC to Light, Infinite input impedance, and zero output impedance.
Can u prove ur theory?
 

Audioguru

Joined Dec 20, 2007
11,248
Your example of 100W into 8 ohms is from an amplifier with an output of 28.29V RMS. 28.29V RMS is 80.0V peak-to-peak so the amplifier has a 90V power supply.
With a 16 ohm speaker the current is less so the output swing will be a little higher, maybe 82V p-p then the RMS voltage is 29V RMS and the power into 16 ohms is 52.6W.
You must increase the power supply voltage to get 100W into 16 ohms.

Audio amplifiers have a spec called "damping factor" which is determined by their output impedance. Most have a damping factor from 200 to about 800 so their output impedance is from 0.01 ohms to 0.04 ohms.
But old vacuum tube amplifiers had the same output impedance as their load.
 

marshallf3

Joined Jul 26, 2010
2,358
But old vacuum tube amplifiers had the same output impedance as their load.

True, had many an old amp that had tapped output transformers selectable to match 4, 8 or 16 ohm speakers.
 

nomurphy

Joined Aug 8, 2005
567
It seems to me that you may be misapplying the term "impedance matching". In audio, it does not necessarily mean driving a 47K output into a 47K input and therefore matching the two up. It can mean "matching" an output impedance of one value or range, to an input impedance of a completely different value or range (this is also done in telephone circuits with a line-matching transformer).

There are times and reasons to match I/O impedances in audio, and other times it is not desired.

With tube amps (be they old or new), the output uses an impedance "matching" transformer, because the relatively high-impedance output tube cannot directly drive the relatively low-impedance speaker (way too much current required of the output tubes). Therefore, the xfmr "matches" the tube impedance on the primary, and the speaker impedance on the secondary. The xfmr essentially converts a high-voltage, low-current, output into a lower-voltage, higher-current output. With xstr amps, the xfmr is no longer necessary, because the xstr output section has a much lower impedance than the speaker and can provide the neccesary volts/amps directly -- per "Guru's" explanation (dampening factor).

However, the use of an output xfmr can be used in xstr amps, and were common in older amps when designs were transitioning between tubes and transistors, because the xfmr is very handy at converting voltage and current.

In the example you provided of a 47K input, it wouldn't be desirable to match this with a 47K output from the driver; the driver should have a much lower impedance to avoid noise contributions and amplitude/frequency roll-off. One of the reasons RIAA established a known input impedance for audio amps of 47K + capacitance range, was to create a standard. So, if the input impedance doesn't change significantly from one audio device to another, then the designer's/manufacturer's spec will not vary significantly when mixed/matched with other brand equipment -- as long as their driver is a much lower impedance relative to the known standard and provides adequate drive capacity across their specified audio spectrum.

With RF frequencies, it is important to match impedances, say 50 ohms to 50 ohms, to avoid undesirable attenuation and reflections that can occur at these higher frequencies (among other issues). Some of this may come into play in audio if one is trying to accurately reproduce or listen to square-waves whose fast edges consist of high (lots of) harmonic content, but usually this is not of great importance in audio design, which is one of the reasons the edges roll-off in audio (there are other reasons as well, such as tone controls and various filtering).

Impedance "matching" is usually used in audio when the output impedance can be greater or in the same range as the input, such as with some types of phonograph cartridges. A "matching" transformer is also used in differential mic inputs, typically using XLR connectors, to convert a differential low-noise signal (high "common-mode" noise rejection) into a "single-ended" ground-referenced signal -- although, this can be accomplished in modern designs with high-quality op-amps without using a xfmr.
 
Last edited:

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
Hmm.
Everyone opinion are well established and are true.

So it can be said the impedance matching are done in audio but not always.

What my primary concern is tht to have the maximum power transferred, impedance shud be matched.

Wht I got confused is tht why are these not taken into account when matching speakers to amps.
With the damping factor and all an amp has a very low output impedance to drive a speaker.

But what about the fact that how much power is really transferred to a speaker with 8Ω froman amp with an output Z of 0.1Ω.

Any one can clarify on this problem
 

Audioguru

Joined Dec 20, 2007
11,248
A modern amplifier has an output impedance of at most 0.04 ohms, not 0.1 ohms.
It is easy to calculate how much power is lost in the 0.04 ohms because it is in series with the speaker and therefore has exactly the same current. The power wasted in the 0.04 ohms is almost zero.

Then calculate how much power is wasted when the output impedance of an amplifier is the same as its load impedance. Half the power is wasted.
 

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
I know what u are saying, but the problem is this is contradicting with the maximum power transfer theory.

Or am I misunderstanding the concept completely
 

Audioguru

Joined Dec 20, 2007
11,248
Maximum power transfer occurs with an amplifier that has a tuned output like an RF amplifier. It is resonant so its output voltage is higher than the power supply voltage with a high impedance load (then its output power is low), and its output voltage is nothing when its load is a low impedance (then its output power is low). Halfway, its output impedance equals the load impedance then there is maximum power transferred.

But a modern audio amplifier has an extremely low output impedance so its voltage does not change when it has no load or has its rated load. Almost all of its output voltage goes to the load.
 

Thread Starter

R!f@@

Joined Apr 2, 2009
9,918
What U say makes sense.
But Something is still missing, I still can't get my head around it.

Bythe way Audioguru, Sorry I came in a bit hard, I don't have anything against u. Ur input is most welcome. As any other.

Still. I need to go thru this concept. I guess I have to make a practical approach on this..

One more thing. How can one make an amps output Z high
 

Tesla23

Joined May 10, 2009
542
The theorem simply says that impedance matching results in maximum power transfer, so you only use it when you want maximum power transfer. This is actually quite rare, so any course you took that implied that you should always match impedances is wrong.

When you are trying to extract all the power from a source, impedance matching is a good idea. For example, if you have a radio antenna to collect signal power, you don't want to waste any, matching is probably a good idea. If it is connected to a low noise amplifier, there are times when you may get better peformance by mis-matching the antenna to improve the noise performance of the LNA.

Once you have achieved a reasonable S/N ratio, then we can often ignore power and worry about other things like noise, distortion, efficiency, maximum power, immunity to interference and simply convenience, and these generally drive interface design in audio circuits. Low output impedance sources driving high impedance inputs is simply convenient, for example you can parallel up inputs.

With audio power amplifiers it turns out that a reasonably efficient way to get a large amount of low distortion power to the speakers is to drive them from a very low impedance source. All power amplifiers are non-linear, and the load impedance that achieves maximum power output is almost never equal to the output impedance. I am surprised by the comments written about valve amplifiers, I'm no expert but I would think that the output impedance of most valve amplifiers would be very high, and the use of the output transformer for different speaker impedances is simply to present something close to the maximum power load impedance to the plate. Roughly you want to have voltage and current clipping occurring at the same time.

The only other major issue ar RF is that once the cables become appreciable fractions of a wavelength, then mismatches become a major headache. If you did the audio thing and had low impedance sources driving high impedance loads, the problem is that strange things happen at various cable lengths. If your cable is a quarter of a wavelength long then the high impedance load suddenly had become a low impedance load and your source won't be happy. If your load is 50ohms and you use 50 ohms cable, the load is still 50ohms at the other end of the cable.

If you are still confused, try to understand why you should or shouldn't match your toaster to the power outlet (thought experiment only!).
 
Top