[art]

The capacitance of a cable is not always constant. The measurement depends on the frequency at which it is measured.

I see we now have 2 people who think they know T-line theory. But what the hell do I know? At my (old) day job, I was only the in-house "expert" on transmission lines.

You know what.........believe what you guys like. Try designing a high-power amp that is unconditionally stable, into any load, no matter how bizarre it might be. Or unlikely to be encountered in the normal course of business. Then lecture us on how poor of a job guys like us do.

In the meantime, stop buying and using high capacitance cables.

I am through with this thread. It is pointless for me to devote any more time to it.

Pat

[Niteshade]

"Try designing a high-power amp that is unconditionally stable, into any load, no matter how bizarre it might be."


I never said, "any load".  I stated amplifiers should be designed to operate into reasonable load variations.  Nobody expects an amplifier to operate under ALL conditions! That would be unreasonable.

[opaqueice]

I see we now have 2 people who think they know T-line theory. But what the hell do I know? At my (old) day job, I was only the in-house "expert" on transmission lines.

What the heck do transmission lines have to do with this, anyway?  I thought we were talking about speaker cables. 

Last time I checked, wavelengths at audio frequencies were in the tens or hundreds of kilometers.  That's a pretty long cable. 

[Niteshade]

Transmission lines cover a broad spectrum of 'interconnects'. Even telephone wire is a form transmission line. High Tension lines are transmission lines as well. It's all about getting a signal from point A to B through a solid medium and the complications in doing so. Speaker cables are indeed transmission lines. There are even balanced and unbalanced lines. I.e. Balanced microphones that use XLR connectors. In the RF world, ladder line is a balanced line technology and coax is unbalanced. All very neat stuff!

[opaqueice]

Transmission lines cover a broad spectrum of 'interconnects'. Even telephone wire is a form transmission line. High Tension lines are transmission lines as well. It's all about getting a signal from point A to B through a solid medium and the complications in doing so. Speaker cables are indeed transmission lines. There are even balanced and unbalanced lines. I.e. Balanced microphones that use XLR connectors. In the RF world, ladder line is a balanced line technology and coax is unbalanced. All very neat stuff!

That's not what I was taught.  Transmission line theory is supposed to apply to signals with a wavelength shorter than the length of the cable in question.  Analogue audio signals max out well below 100kHz, which has a wavelength of 3 kilometers.  10kHz is 30 km.  So I don't see how transmission line effects are even remotely relevant for speaker cables.

The story for digital audio cables is different - I think the carrier wave for S/PDIF is in the MHz range (correct me if I'm wrong), so that's at least getting close.

[Niteshade]

Transmission Lines:

Look here: http://en.wikipedia.org/wiki/Transmission_line

[jneutron]

The capacitance of a cable is not always constant. The measurement depends on the frequency at which it is measured.

Agreed.  That is why I mentioned the fact that the article linked did indeed have the capacitance and inductance measured from 100 hz to 1 Mhz.

I see we now have 2 people who think they know T-line theory. But what the hell do I know? At my (old) day job, I was only the in-house "expert" on transmission lines.

Then it should not be difficult for you to understand what was said.  I will admit however, that I have to assume your T-line expertise is confined to that typical of RF/microwave, and not within the realm of what was mentioned.

Consider T-lines which are a very small fraction of a wavelength..

When the line impedance matches that of the load, the amplifier will see that load impedance.  And the line will store within it the exact same amount of stored energy that propogates along that line at it's prop velocity.  And the inductive and capacitive energy storage is equal.

When the load does not match the line impedance,  the energy stored within the cable will be higher due either to more inductive storage (load less than cable impedance), or more capacitive storage (load greater than cable impedance).    In the case of a mismatch, the cable length and reflection coefficient of course come into play (as you know) , but in the short-short cable case as we speak of, the slew rate is far below the transit speeds, so edge fidelity means nothing..

The fact that you are an "expert" (your quotes), does not necessarily mean that you normally worked with extremely small wavelength cables..  As such, I can understand why you would question the assertions.  I'd be unhappy if you didn't..

You know what.........believe what you guys like. Try designing a high-power amp that is unconditionally stable, into any load, no matter how bizarre it might be. Or unlikely to be encountered in the normal course of business. Then lecture us on how poor of a job guys like us do.

Perhaps you meant that for others, I've stated nothing that would demand that level of "sillyness".

I am through with this thread. It is pointless for me to devote any more time to it.

Pat

Quite honestly, you did come in awfully hot under the collar.  Is there a history here I am unaware of?

I'd certainly be happy to continue the T-line aspect, especially if you do indeed have the background.  Your attitude suprises me.

Cheers, John

[jneutron]

That's not what I was taught.  Transmission line theory is supposed to apply to signals with a wavelength shorter than the length of the cable in question.  Analogue audio signals max out well below 100kHz, which has a wavelength of 3 kilometers.  10kHz is 30 km.  So I don't see how transmission line effects are even remotely relevant for speaker cables.

T-line theory is generally taught under the assumption of signal propagation.  Signal propagation occurs at the prop velocity, and it makes the assumption that the energy stored within the magnetic field is exactly the same as that of the electric field.

As the wavelength to length ratio increases, the load to cable impedance ratio alters the dominating storage mechanism within the cable...this is not something that is taught in the RF domain.

Cheers, John

[Niteshade]

"When the load does not match the line impedance,  the energy stored within the cable will be higher due either to more inductive storage (load less than cable impedance), or more capacitive storage (load greater than cable impedance).    In the case of a mismatch, the cable length and reflection coefficient of course come into play (as you know) , but in the short-short cable case as we speak of, the slew rate is far below the transit speeds, so edge fidelity means nothing.."

Being a "Ham", I have dealt with SWR's allot. They can be a pain!

[jneutron]

aha..got it to work..

Here's the graph I mentioned which shows the effective dielectric constant of the cables in the AH article.  I haven't been here in a while, I forgot there was a gallery I could plant stuff in...

The DC on this graph is actually the effective dielectric constant.  It is by equation, equal to LC/1034.  L in nH per foot, C in pf per foot.

It also makes the assumption that the permeability of the construct is equal to 1.

If you do 1/sqr(EDC), you obtain the prop velocity of the cable (just a point of interest of course...)  This is completely analogous to the propagation of energy in free space, v = lightspeed/sqr(mu*epsilon)

Cheers, John

[jneutron]

Here's the energy storage vs line/load ratio.  Uses E = 1/2 LI^2 and E = 1/2 CV^2.



Cheers, John

[opaqueice]

Transmission Lines:

Look here: http://en.wikipedia.org/wiki/Transmission_line

That more or less coincides with what I recall about T-lines.  In particular, there's nothing on that page that applies to speaker cables carrying an analogue audio signal, for the reasons I gave above.

T-line theory is generally taught under the assumption of signal propagation.  Signal propagation occurs at the prop velocity, and it makes the assumption that the energy stored within the magnetic field is exactly the same as that of the electric field.

As the wavelength to length ratio increases, the load to cable impedance ratio alters the dominating storage mechanism within the cable...this is not something that is taught in the RF domain.

This isn't exactly my area, but I'd say it differently.  I thought that treating something as a transmission line was necessary when you had to worry about finite wave propagation speed, and along with it reflections, impedance matching, etc.  But for speaker cables I don't see why we can't set the speed of light to infinity and just model them by their equivalent RLC circuit.  No need to think about waves at all.

[jneutron]

That more or less coincides with what I recall about T-lines.  In particular, there's nothing on that page that applies to speaker cables carrying an analogue audio signal, for the reasons I gave above.

Sigh.  Eventually I'll have to modify that wiki page to include the realm of line much shorter than wavelength scenario's.  I've bandied back and forth as to whether or not it dilutes the primary goal of the wiki, which is the standard RF view.  The statement that propagation occurs with inductive and capacitive energy storage equality is another point that I'm not sure I want to put there....  I've held off probably way too long..

This isn't exactly my area, but I'd say it differently.  I thought that treating something as a transmission line was necessary when you had to worry about finite wave propagation speed, and along with it reflections, impedance matching, etc.  But for speaker cables I don't see why we can't set the speed of light to infinity and just model them by their equivalent RLC circuit.  No need to think about waves at all.

It can be done either way, and if one is careful, it can be correct either way (course, it can also be done incorrecty either way).  Using t-line does have the concern that there will be people who think that higher prop speeds are a reason for a change, when it could be something as simple as energy storage that is the factor.

Cheers, John

[dyohn]

Here's the energy storage vs line/load ratio.  Uses E = 1/2 LI^2 and E = 1/2 CV^2.



Cheers, John

Wow, worst case condition is about 12-millionths of a watt-second.  How have I not been hearing that all these years??   

[jneutron]

Wow, worst case condition is about 12-millionths of a watt-second.  How have I not been hearing that all these years??   

Hmmm.

Perhaps you should ask the context of the levels first??   Is it normalized, is it temporal, is it capable of 2.to 5 uSec delay, or .5 dB or so level shift?

Context is important..

I presented the graph simply to demonstrate the minima in cable energy storage as a result of the deviation of the load from the cable's characteristic impedance...nothing more, nothing less.

Cheers, John

[Niteshade]

Getting back to the speaker cable problem- I have a simple idea that may help those with touchy amps: In my Motors and Controls class we studied Power Factor Correction. A big term for making reactive loads look more resistive. So- if someone has cable that looks capacitive, why not throw an inductor in series with it? If it looks inductive, want not add some capacitance in parallel with it?

Matching networks can add losses in the system, but I have a feeling that the matching necessary is so minute that the losses would be minimal and never noticed.

Somebody a while ago mentioned the amp designer who's amps don't like capacitive loads doesn't use  inductors on his outputs. The fix may be to add inductors and do whatever circuit changes are necessary to make it work properly.

[Daryl]

Perhaps you meant that for others, I've stated nothing that would demand that level of "sillyness".

Quite honestly, you did come in awfully hot under the collar.  Is there a history here I am unaware of?

Your attitude suprises me.

I know exactly what's going on here and I think you do also.

[Daryl]

Transmission Lines:

Look here: http://en.wikipedia.org/wiki/Transmission_line

That more or less coincides with what I recall about T-lines.  In particular, there's nothing on that page that applies to speaker cables carrying an analogue audio signal, for the reasons I gave above.

T-line theory is generally taught under the assumption of signal propagation.  Signal propagation occurs at the prop velocity, and it makes the assumption that the energy stored within the magnetic field is exactly the same as that of the electric field.

As the wavelength to length ratio increases, the load to cable impedance ratio alters the dominating storage mechanism within the cable...this is not something that is taught in the RF domain.

This isn't exactly my area, but I'd say it differently.  I thought that treating something as a transmission line was necessary when you had to worry about finite wave propagation speed, and along with it reflections, impedance matching, etc.  But for speaker cables I don't see why we can't set the speed of light to infinity and just model them by their equivalent RLC circuit.  No need to think about waves at all.

No need to speak of standing waves and such but characteristic impedance is key to the problem PLMONROE is describing and why it only occurs in certain situations.

[PLMONROE]

This enormous amount of information that this thread has produced is nothing short of fantastic!. It is evidence that the knowledge of those on this forum is truly mind boggling!    As a practical observation I have to assume that my problem is a somewhat rare occurrence. Otherwise it would seem that manufacturers of both amplifiers and speaker cables would take steps to insure such a situation would not occur. Also if it were more common it shouldn't have taken me a year to resolve the issues, although perhaps I was not looking in the right places. Thank you all.

Am I to assume that insertion of the -6db attenuator into the amps input, which allowed the amp to operate with the Acoustic Zen cables without shutting itself down, somehow kept the amplifier from becoming unstable by its reduction of the input level???

Paul

[avahifi]

Even if the amp is not shutting down now, the excessive capacitive load is still doing very bad things to the linearity of the amplifier.  The internally oscillations are slightly less now, but still there.  Making a change at the input will not affect the bad effects of the capacitive load on the output, unless there is a big interaction in the internal feedback loops affecting the dynamic input impedance of the amp.  Then it is possible that the different input load could be damping the internal oscillations.  It is interesting to put an amp misbehaving like yours on the test bench, and move the scope probe from the output when it is oscillating to the input, and find the same oscillations showing up there!

Please please go back to simple two conductor speaker cables, available at Home Depot for about 25 cents per foot.

Regards,

Frank Van Alstine