In fact, it makes a lot of sense to have a far-end Zobel embedded in the cable (at least Goertz has got some cables of this kind). In case you disconnect a speaker from the cable while your amplifier is "on" with the cable attached to it, this kind of cable doesn't look like an "evil capacitor". Some amplifiers will be fine anyway, but some others may not like a non-HF-terminated cable, converting themselves into a "screaming oscillator" 
Last edited:
Thank you, Valery!
Our load is not clearly passive. Yes, it's impedance are complex and changes with frequency. But it's impedance is constant at one given frequency. The things are little more complicated. It have moving mass with some stiffness around. For sinusoidal moving above resonant frequency we need to actively pump energy inside this system and actively dissipate energy going from there. Actually like accelerating and braking the loaded/unloaded car. All kinetic energy must be dissipated by engine (for compressing air - to heat and compression), by brakes (mostly to heat) or by destroying any styffy elements.
Now we came to the question "what we could".
So you can easily try attached.
Clearly.
And if you can't - use any of the wire compensation methods.
Attachments
http://www.diyaudio.com/forums/atta...nce-harmonic-distortion-wire-compensation-gif
Thank you. We would appreciate a translation of the embedded text and a reference to the source for further discussion.
George
Actually, the impedance at any specific frequency is dependent on the velocity and acceleration of the coil in the gap. As such, a dynamic driver's impedance is dependent on signal amplitude as well as frequency content. When a coil is being driven hard at some frequency, the resultant spring forces on the coil impact what other drive frequencies see.
In the motion control world, gap position force nonlinearities I would address with adaptive PID algorithms. In a driver gap, I cannot see this happening in the short term.
Jn
It's just a three wire compensation setup. RTD's use the technique a lot. The gain of the circuit is adjusted to comp the drop of the two wires feeding the load.http://www.diyaudio.com/forums/atta...nce-harmonic-distortion-wire-compensation-gif
Thank you. We would appreciate a translation of the embedded text and a reference to the source for further discussion.
George
They drew the third wire close to the load wire, so I suspect they meant close proximity to reduce m field errors and loop pickup.
I prefer a 4 wire kelvin.
Jn
Hi, George, it’s obvious and doesn’t need much comments.
Text is easy and sets gain of 2x with corner freq above audioband.
Or something like Clarion FDS:
4-wire allows easy common mode filtering but also adds HF leaking.
While The different color vc's look really neat, from the text, they do not address the force nonlinearities.
Jn
It seems to me that evaluation of the issue should come first, before solutions.
There are qualitative descriptions, but no quantitative evaluation, so far. No evaluation either of the impact of this distortion measured at the end cable with respect to the distortion made by the speaker non linearities.
The source of the distortion is the speaker non linearity. The current is distorted even in the very best case of no distortion at the speaker terminals. Do cables significantly add to this unavoidable current distortion ?
Answers to this should come prior to solutions about an issue that is purely speculative.
There are qualitative descriptions, but no quantitative evaluation, so far. No evaluation either of the impact of this distortion measured at the end cable with respect to the distortion made by the speaker non linearities.
The source of the distortion is the speaker non linearity. The current is distorted even in the very best case of no distortion at the speaker terminals. Do cables significantly add to this unavoidable current distortion ?
Answers to this should come prior to solutions about an issue that is purely speculative.
This is actually an important thing. The point is that it looks constant if we apply, say, a sinusoidal signal of certain RMS voltage and measure an average impedance value.
As soon as we look at the momentary impedance values, as the signal travels throughout the sinusoidal curve - we will see the impedance is not constant anymore, exactly because the dynamic speaker's impedance is non-linear.
Particularly "the voice-coil inductance depends on both the displacement of the voice-coil and the current intensity flowing through it".
You explain it further in your post - I just felt the need to open up the "impedance is constant" statement a bit
This statement is misleading. We can minimise the visibility of the speaker's nonlinearity by using a lower resistance cable, but it doesn't change the effect.vzaichenko said:
Yes.
Yes.Tournesol said:
I know what you are saying, but those who do not understand the magnitude of the effects you describe will think you are supporting the 'true believers' when in fact you are gently smiling at them.Elvee said:
See comment above.mchambin said:
We should take measurements with quad wires, no wires with speaker terminal right to the speakers (as much as possible) and compare braided cables, single cable, and twisted 4pr or 8pr.
The test should have a speaker frequency response sweep, and IMD with two or more signals with different intensity signals.
Should measure with microphone, at the speaker terminals and into a resistor load too,
Comparing the resistor load to the speakers and eliminating the difference will reveal the cable contribution to sound or not.
The test should have a speaker frequency response sweep, and IMD with two or more signals with different intensity signals.
Should measure with microphone, at the speaker terminals and into a resistor load too,
Comparing the resistor load to the speakers and eliminating the difference will reveal the cable contribution to sound or not.
Probably "minimise" is not a good word in my statement - something like "reduce" would be better. However, my thinking is - a low-impedance cable does not only make the speaker's impedance non-linearity less visible.
Assuming the amplifier's Zout is low enough, it also improves the speaker's damping and reduces the voltage fluctuations, caused by the speaker's impedance non-linearities (the speaker is driven "harder" with a better cable).
As a result, I expect the total distortion, produced by the speaker with a lower-impedance cable, to be lower than the total distortion, produced by the speaker with a higher-impedance cable.
What do you think?
Attachments
I was the one who stumbled upon the higher distortion FFTs and alerted Valery to it. I think his data clearly shows that there are effects of the speaker cable on the FFT. It’s not the same as no cable for sure. Whether or not one wants to believe that the cable is the cause of the changes is something that is being argued ad infinitum here.
If you have a sound interface, connect it to your dummy load with and without cable. See what you get. Yes, take all proper precautions to ground properly. You will still see distortion. It’s not the cable that makes the distortion visible - imagine a very short 5cm cable. It won’t be the same as long cable - it will be close to same as no cable. The effect is caused by the cable not that the cable makes it easier to see.
I have seen distortion in an amp change via simple substitution of one brand of thin film resistor vs another (Vishay vs Panasonic) in the feedback or output device source resistor, and I get different harmonic profiles (relative H3 was higher va H2 on one and vice versa). That’s a single 2cm long part. It doesn’t surprise me that a 2m long multi stranded cable with insulation might have something non-linear going on.
If you have a sound interface, connect it to your dummy load with and without cable. See what you get. Yes, take all proper precautions to ground properly. You will still see distortion. It’s not the cable that makes the distortion visible - imagine a very short 5cm cable. It won’t be the same as long cable - it will be close to same as no cable. The effect is caused by the cable not that the cable makes it easier to see.
I have seen distortion in an amp change via simple substitution of one brand of thin film resistor vs another (Vishay vs Panasonic) in the feedback or output device source resistor, and I get different harmonic profiles (relative H3 was higher va H2 on one and vice versa). That’s a single 2cm long part. It doesn’t surprise me that a 2m long multi stranded cable with insulation might have something non-linear going on.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.