[James Romeyn]

For absolutely ideal audio performance, I stress no compromise, should speaker wire series resistance not exceed amplifier output impedance?   

[Davey]

I hope you're not expecting a yes or no answer to that. 

Amplifier output impedance might exceed speaker wire resistance in many cases.  But both are usually a small fraction of load impedance.  Either condition might be a condition where there is (essentially) no compromise.

Resistance is generally not the parameter of interest in speaker wire.  It's usually fairly low.  Inductance and capacitance are the parameters traded off with each other to create the sonic "character" of speaker wire.  And that "character" is accenuated with amplifier having higher output resistances.

Cheers,

Dave.

[James Romeyn]

Good points, thanks.

My bad: I should have mentioned the designer of the amp specifies an output impedance of .0015 Ohms, equivalent to a mere 3' of 4 AWG!

Post #3 and later are relevant here: http://www.audiocircle.com/index.php?topic=108971.new;topicseen#new.

Possibly the amp in question may be the rare case where series resistance overrides other considerations.  So far that's my experience anyway.  I prefer performance of zip cord speaker wire equivalent to 4 or 5 AWG over triple parallel Stan Warren recipe equivalent 6 AWG.  On other amps wire performance is the inverse of the Ncore amp. 

[Davey]

I don't understand the premise.  You need welding cable for speaker wire?    It's completely ridiculous.  Am I missing something?

Is this some kind of misunderstood Ncore recommendation that came from Hypex?

Even with nominal gauge speaker wire like say 16awg, losses across the speaker wire will be tenths of db's, worst case.

Cheers,

Dave.

[James Romeyn]

Ncore assembly instructions do not address external speaker wire at all. 

My prior understanding was that losses less than 1 dB are inaudible.  My personal Ncore experience is dramatically different, hence the post.

[James Romeyn]

I don't understand the premise.  You need welding cable for speaker wire?...

Well, this is nothing about need, because it's only audio!

I'm saying that in my experience, Ncore seems to need a lot more conductor than any other amp I've auditioned.  And it also seems like the old rule that losses less than 1 dB are inaudible does not apply.

With what amp have you experience having output impedance of .0015 Ohms?   

[Davey]

All other things being equal, losses much greater than 1db should be inaudible (when A/B comparing) because you should be matching levels.

The way I understand your experiment, you ran a pair of speakers in parallel and in series for a nominal/2 and nominal*2 (respectively) effective load resistance on your amplifier.  You've created quite a few extra variables in your evaluation, yet you conclude that speaker wire resistance is the parameter that caused the audible difference?
Do you see the problem there?

Most amplifiers function as (nearly) voltage sources.....the Ncore does as well.....so amplifier output impedance can generally be ignored...since it's so low.

Cheers,

Dave.

[Ethan Winer]

should speaker wire series resistance not exceed amplifier output impedance?

As long as you use reasonably heavy connecting wire, the main issue is the resistance of the voice coil, and possibly the passive crossover's inductor. A voice coil's resistance typically limits damping factor to around 50, no matter how low the amplifiers output impedance is.

--Ethan

[*Scotty*]

James, the Stan Warren recipe first posted on Harmonic Discord consisted of a star-quad configuration of 12 ga. conductors. This is equivalent to a 9ga conductor for both hot and ground legs of the speaker wire. A 9ga. wire has 0.7921 ohms per 1000ft..
When .7912 is divided by 1000ft. you have 0.0008 ohm/ft.. A 10ft. run of 9ga. wire will have 0.008 ohms. This comfortably under .0015ohms output impedance of the NCore amp. A shotgunned pair of 10ft. 9ga. speaker wire would have about .004 ohms resistance, or about 1/3rd. the output impedance of the NCore.
Scotty
Calculations of resistance were based on information from the Handbook of Electronic Tables and Formulas for American Wire Gauge.
found at this link.
http://www.powerstream.com/Wire_Size.htm

[Steve]

To further expand, if no crossover network is used, a SS amplifier with very low output impedance (Z), the DC resistance of the voice coil is zero, the speaker wire could lower the damping factor by up to half.

However, since the DC resistance of the voice coil is several ohms, the
damping factor change is very small. For instance,

1) an "8 ohm" speaker and
2) DC voice coil resistance of, say, 5.5 ohms (please specify if smaller resistance has been found), and
3) amplifier output impedance of zero ohms
4) the damping factor at 8 ohms is 1.454.
5) at 5.7 ohms minimum Z, the damping factor would be 1.036.

Even at this low damping factor, the wire would cause minimal change.

Depending upon the crossover frequency and say, a woofer's resonance peak of 40 ohms, the maximum damping factor would be approximately 7.27. Adding a typical  crossover would decrease the damping factor further.

Midrange and tweeter would also apply, with damping factor also varying with frequency. Imo, wire size would minimally affect damping.

With other types of speakers, the damping factor could be less.

Cheers.

ps. A piezo tweeter would be an exception as it normally has a high impedance and no crossover is often used.

[Davey]

Steve,

Your post is nonsense.

Damping factor is defined as Zload/Zsource.  If you assume Zsource is zero ohms then obviously you create a divide-by-zero error and can't compute an answer.

Here's a much better (typical) example:

Zsource = .1 ohms.
Zload = 8 ohms.

DF = 80.

Speaker wire resistance adds to Zsource and thus lowers the damping factor.
Say you used a 10 foot chunk of 12 awg wire.  You'd create a round trip resistance of approximately .032 ohms.

Zsource = .132 ohms.
Zload = 8 ohms.

DF = 60.

Many amplifiers have an output impedance lower than 0.1 ohms, and most speaker drivers (being electro-mechanical devices) have Z which varies with frequency.  So, obviously there are variables.  And, as Ethan alluded to.....the driver itself "self damps" based on its own VC resistance which effectively limits the "true" damping factor to values near 50.

Cheers,

Dave.

[Russell Dawkins]

To further expand, if no crossover network is used, a SS amplifier with very low output impedance (Z), the DC resistance of the voice coil is zero, the speaker wire could lower the damping factor by up to half.

However, since the DC resistance of the voice coil is several ohms, the
damping factor change is very small. For instance,

1) an "8 ohm" speaker and
2) DC voice coil resistance of, say, 5.5 ohms (please specify if smaller resistance has been found), and
3) amplifier output impedance of zero ohms
4) the damping factor at 8 ohms is 1.454.
5) at 5.7 ohms minimum Z, the damping factor would be 1.036.

Even at this low damping factor, the wire would cause minimal change.

Depending upon the crossover frequency and say, a woofer's resonance peak of 40 ohms, the maximum damping factor would be approximately 7.27. Adding a typical  crossover would decrease the damping factor further.

Midrange and tweeter would also apply, with damping factor also varying with frequency. Imo, wire size would minimally affect damping.

With other types of speakers, the damping factor could be less.

Cheers.

ps. A piezo tweeter would be an exception as it normally has a high impedance and no crossover is often used.

I find this surprising, coming from an amplifier designer/manufacturer - or else I'm missing something here. Could this be a fundamental misunderstanding of some type?

[Steve]

Both Davey and Russell,

I used a zero amplifier output Z as some approach zero ohms. At, say, .008 ohms, that is pretty close.

If one does a thevenin equivalent circuit analysis (see photo below), the DC resistance of the voice coil appears in series with the output Z of the amplifier (assuming no crossover is used that would add to the resistance/reactance in series).

As such the maximum damping factor possible is actually small. (See my previous post for figures.)



Cheers.

[Davey]

Steve,

Your assertion is that transducer Rvc is part of the source and not the load?  (At least with regard to damping factor computation?)

Well, that's a novel way of looking at it. 

Cheers,

Dave.

[Steve]

Steve,

Your assertion is that transducer Rvc is part of the source and not the load?  (At least with regard to damping factor computation?)

Well, that's a novel way of looking at it. 

Cheers,

Dave.

I think you misundersand Davey. It is called a "thevenin circuit"; check Linkwitz and others. A "thevenin circuit", Kirchoff voltage and current laws need to be understood and are necessary for nearly all designing. For instance your computer programs for designing electronic circuits would not work without an understanding of thevenin equivalent circuits and Kirchoff's laws. (The picture posted is not mine but from another electronics site.)

Anyway, here is what I previously stated.

If one does a thevenin equivalent circuit analysis (see photo below), the DC resistance of the voice coil appears in series with the output Z of the amplifier (assuming no crossover is used that would add to the resistance/reactance in series).

As one can see, the thevenin equivent circuit reveals that Revc is the voice coil DC resistance of the speaker driver and is in series with the amplifier output, not "part of the source". Would one conclude that the crossover is also part of the amplifier/" because it is connected to the amplifier?  I do not think so.

As such, a thevenin equivalent circuit is factual, not assertion. Anyway, the speaker wire resistance is usually very small vs the voice coil resistance.

Cheers.

ps. We could add the amplifier output Z, speaker wire resistance and capacitance etc to the thevenin equivalent circuit, but they are extremely small values. Let's keep it simple.

[Russell Dawkins]

My understanding is that damping factor in this context is the ratio between the load impedance or resistance and the source resistance. The load resistance is, with a passive speaker, the combination of the the voice coil resistances plus the influence of the crossover components. All of these combine to make the load impedance which, of course, varies somewhat with frequency. As a result of this variance, damping factor is usually quoted as a function of an assumed resistive load of 8 ohms, for convenience.

[*Scotty*]

In as much as most amplifiers incorporate negative feedback, how does one then account for the virtual negative source impedance that its use engenders when attempting to calculate an amplifiers damping factor?
Scotty

[Russell Dawkins]

At moments of negative source impedance the damping factor is beyond infinity?

[Davey]

Steve,

I understand Thevenin equivalents.  We're talking about damping factor here.

In your previous post you made a simple damping factor computation of 1.454 (8 ohms divided by 5.5 ohms.)  That's where your explanation was questionable.

The standard definition of damping factor is Zload/Zsource.........not Znominal/Revc.

A pretty good reference here:  http://en.wikipedia.org/wiki/Damping_factor

Cheers,

Dave.

[Steve]

Steve,

I understand Thevenin equivalents.  We're talking about damping factor here.

In your previous post you made a simple damping factor computation of 1.454 (8 ohms divided by 5.5 ohms.)  That's where your error was made and your explanation needs to be directed to.

Cheers,

Dave.

Unfortunately, many misunderstand thevenin equivalent circuits as demonstrated by the link. If the voice coil had zero ohms dc resistance and 8 ohms reactance, then 8 ohms divided by 0,1 ohms amp output Z would give a high damping factor, 80. One just cannot get around the voice coil DC resistance.

As the thevenin circuit clearly demonstrates, the 8 ohms reactance not only sees the 0,1 amp output Z, but also the 5.5 ohms DC resistance. As such the damping factor is approx 1.45. Actually close to 1.1 at minimum speaker impedance (Z) of 5.7 ohms.

As mentioned earlier, the link is incomplete, and not accurate. A resistor load is different than a speaker load when making electrical measurements. As I suggested in a previous post, it would be beneficial to check out other, more reliable sources.

Cheers.