Impedance matching discussion

[bumpy]

When you read reviews on assembling a great system there is always a great deal of emphasis placed on impedance matching between components. I believe this is so each component can successfully drive the next one in the chain. If it cant do that, there is a general softening of the sound quality and reduction of bass.

I am aware of a couple of facts

1. Cables etc may add a bit of impedance, but for short runs they are not a big contributor.
2. A high input impedance makes a component easier to drive than a low one. This shows itself in an obvious way where speakers of 16 ohm impedance are more easily handled than those of 4 ohm.

So my first question is "Does a fully passive volume control contribute significantly to impedance, or can it be viewed as 'invisible' ?"

[anthonyTD]

Its a bit of a soar subject this Chris, as there are many here that use passive preamps, which in my opinion; more often than not; hinder the signal more than help it, especialy at frequency extremes, its all about the Attenuator's ability; or should i say; its all about the source being able to see, and drive the input load of the amplifier correctly through the Attenuator, at diffrent volume positions, in order to maintain linearity of the original signal.
When you read reviews on assembling a great system there is always a great deal of emphasis placed on impedance matching between components. I believe this is so each component can successfully drive the next one in the chain. If it cant do that, there is a general softening of the sound quality and reduction of bass.

I am aware of a couple of facts

1. Cables etc may add a bit of impedance, but for short runs they are not a big contributor.
2. A high input impedance makes a component easier to drive than a low one. This shows itself in an obvious way where speakers of 16 ohm impedance are more easily handled than those of 4 ohm.

So my first question is "Does a fully passive volume control contribute significantly to impedance, or can it be viewed as 'invisible' ?"

[bumpy]

Thanks Anthony
There are a number of passives types available, for instance transformer based, ladder etc. Do you have any feeling which type present the biggest impedance issues?

[Barry]

With a transformer, the signal level depends on the turns ratio, whereas the impedance variation varies as the square of the turns ratio.

It is possible to design 'constant impedance' resistive attenuators; but they are much more complex than simple L-pads, as in conventional potentiometers, whether discrete step or continious.

[anthonyTD]

Where passive attenuators are concerned Chris, the ideal would be one that maintains its input, and output impedance at all useable volume positions.
Then you have to take into account the affect the cablings capacitance, reactance, and resistance has on the signal at diffrent frequencies, when combined with the attenuator, the input source, and power amps impedance.

Thanks Anthony
There are a number of passives types available, for instance transformer based, ladder etc. Do you have any feeling which type present the biggest impedance issues?

[Barry]

Where passive attenuators are concerned Chris, the ideal would be one that maintains its input, and output impedance at all useable volume positions.

Hence a 'constant impedance' design. The question to ask is what value of impedance should one choose? Should it be the geometrical mean of the output impedance of the source and the input impedance of the amplifier?

[anthonyTD]

:)
Hence a 'constant impedance' design. The question to ask is what value of impedance should one choose? Should it be the geometrical mean of the output impedance of the source and the input impedance of the amplifier?

[bumpy]

Thanks, I can understand the variables in the discussion but don't have a feel for the scale of the contribution from the passive.

Lets use an example.

A DAC is capable of driving an amp with an input capacitance down to say 10K. If the amp has a variable input capacitance down to near 10K, could the passive take it over the cliff into undesirable mismatching?

[Pharos]

This was alluded to on another thread recently.

The general rule is that for maximum voltage transfer*, which is the objective in one audio component feeding another, the source Z should be about 1/10th of that of the load Z.
*Power transfer not being a consideration in this scenario, but it is in transmission line situations.

Trying to design a passive to meet this, and with a constant O/P Z must be a real challenge, and it doesn't matter if the source Z goes very low, but that is unlikely.

With passives, (although I haven't tried to design one, but have designed many attenuators), the range of solutions may well be quite wide, the greatest problem AFAICS being trying to avoid a high O/P Z.

[Mr. C]

I love impedance matching, no seriously its is the holy grail of getting circuits correct imho

That is why we now own a 13.5Ghz vector network analyzer, virtual probing capability on the scope plus a TDR unit for more in-depth looking at back plane, transmission lines. We can also de-embedd the cabling, individual components and circuit board vias. With capability up to 40 Meters. Its great for truly very accurate impedance analysis of the full circuits. For example, for the amplifier output, through the speaker cabling, x/overs and back to the amp wonderful device.

[bumpy]

I keep reading of the 1:10 rule with impedance matching where 1 is the output impedance and 10 the subsequent input impedance. But is this the ideal, the maximum or the minimum requirement?

[Pharos]

Impedance matching has largely been regarded as inappropriate in the audio world; the concept is taken from radio frequency engineering where considerations of maximum power transfer have been important, particularly in line transmission.

In audio voltage is used as the parameter to connect equipment, disregarding the power transfer theory which is regarded as of little consequence.

Sometimes a source is connected to numerous I/Ps of devices, and when so, the advantages of high I/P Zs becomes pronounced; they load the source to a lesser extent. From memory of my time at the BBC, even with signal transfer around the buildings the I/P Z and I/p Zs were 600 ohms, but I do not know the rational for this, probably noise reduction on long lines.

Consider the O/P Z of a power amp, which we all want to be as near zero as possible to transfer as much voltage as possible to the load into which it feeds. Damping factor is an expression of maximising this ratio. This whole scenario is completely unbalanced.

Some attempt to feed speakers with a constant current source, with various supportive contentions.

[Firebottle]

I would say minimum.

[Barry]

Dennis,

Impedance matching is still used for some audio amplifiers. That is why an output transformer is used in valve designs, to match the anode loading of a few kilohms to the ~ 8 Ohm load of the speaker.

[Pharos]

My bad, yes of course Barry, my (internal context was one of the last47 years of SS equipment, I had forgotten the difficulties with matching with stages in valves. I built and used valve stuff from about '66 to '70.

[sailor]

The 1:10 ratio should be regarded as a minimum. That is the source should feed into an impedance at least 10 X higher.

I sent you an email yesterday.

[Pharos]

Yes Charles absolutely, so that the source amp doesn't 'know' that there is a load.