Input impedance-Balanced operation

[Ianmac]

I'm driving a power amplifier in balanced mode from a DAC source with balanced output. ( Benchmark and Lavry dacs)
I'm getting reports from a friend of different sonics when using the Lavry with different amps and cables
The nominal input impedance of the power amps is 20k to 100k ohm.

Should the input impedance of the power amplifier be made to be 600 ohm to match the balanced outputs of the DAC and minimise the effect of different cables
Many years ago I came across amps with an input impedance switch allowing a hi Z or 600 Ohm operation.
The Hi fi amps I have come across here in the UK all seem to be hi Z.

Any advice ?

[DF96]

Should the input impedance of the power amplifier be made to be 600 ohm to match the balanced outputs of the DAC and minimise the effect of different cables

Only if the cable is very long and definitely 600 ohm impedance. 'Very long' is unlikely - you would need at least 10's of metres. '600 ohm' is very unlikely for an audio cable, as resistance dominates so the cable characteristic impedance over much of the audio range will not be resistive.

[Ianmac]

Thanks for your reply.
I have fired the question to both Lavry and Benchmark. Benchmark in their handbook give one details of the output attenuator and cable length compensation. I'm a little mystified how they can compensate for a cable when they dont specify the terminating impedance.

There is a suggestion that the input impedance of the amp should be set lower than than 20k- 100k by a resistor across the input of the amp in the region of 3K.( but not as low as my suggested 600 Ohm) This is well within the drive capacity of the DAC output stages and would decrease the effect of cable parameters on the sonics.

I use Hypex amps so I will fit a 3k resistor across the input and see (hear ) the result

[Speedskater]

Almost no impedance matched (output impedance to impedance) audio equipment has been designed in the last 50 years. Almost all equipment now operates with a low output impedance into a high input impedance.
In any case analog interconnect cables do not have a (single) characteristic impedance at audio frequencies (no mater what their length may be)(with a very, very few exceptions).

[Ianmac]

Thanks Kevin,
I agree with both your points.
I suppose I'm really saying how High an input impedance should I have. ( My Hypex is 100K,Cyrus 20K) My thinking is that if I get this as low as possible allowing for the drive capability of the DAC I will reduce the effect of the interconnect cable capacitance and inductance.
I run 14M between DAC and Amp.

[DF96]

14m is still quite short. I suppose you could try to equalise the cable either by adding series inductance (as originally suggested by Oliver Heaviside for telephone cables) or shunt resistance. The idea is to get the ratio of (series resistance+inductance) to (shunt capacitance+conductance) to be constant across the audio frequency range. I doubt if it will make any difference, except for a small increase in distortion as the source has to work a bit harder.

[AndrewT]

Input impedance of balanced connection is discussed by Jung and Whitlock.
Whitlock mentions the >10M equivalent impedance of transformers and talks of electronic balancing never being able to match that excellence.
Jung shows a circuit with 1M5 as the input impedance and from his report seems to me to suggest that the higher the better.

From both we hear the same message, very low Rs.

As far as my understanding goes the ratio of input impedance to Rs is a goodness factor in balanced impedance connections.
It is also my understanding that low Rs gives better immunity from interference.

[Ianmac]

Thanks Andrew, I'll search out the Jung and Whitlock papers.