• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Measurements on amplifiers with speakers and cables connected

cjf

Active Member
Joined
Apr 20, 2016
Messages
160
Likes
58
Location
CO
Hrmm...Here is an interesting read regarding the issue I described previously.

"Audio signal transmission system with noise suppression network" US Patent 4,954,787

The inventor and Patent itself is from Brisson; Bruce A. of MIT Cables Fame.

It looks like Mr Brisson has witnessed the same issue I have described and created a "fix". Not sure which of his cables incorporate this "Fix" but I'll be doing some more research to find out.

Here is the snipet of the Patent that piqued my interest. It pretty much describes my problem to a tee: :)

In transmitting music and other audio signals having a broad range of frequencies from an audio signal source to a speaker, it was discovered that under certain conditions audible noise is generated which can adversely affect the quality of the music heard from the speaker. For example, in one investigation of the phenomenon using a frequency sweep generator, it was discovered that a 200 Hz signal produced noise on a cable comprising a frequency of about 8.3 kHz. In another investigation, a 310 Hz signal was found to produce noise on a cable having a different length comprising a frequency of 6.6 kHz. These frequencies are clearly within the audio frequency range and, therefore, audible. Noise at such frequencies can emphasize or pre-emphasize certain sounds at unwanted times and is therefore undesirable.

From an investigation of the physical mechanisms which give rise to the noise, it appears that at certain audio frequencies and under certain conditions, the inductance and/or capacitance of the cable can cause the cable to momentarily appear as an open circuit to the signal source. Under these conditions it has been observed that the magnitude of the signal, e.g. the 200 Hz signal, applied to the cable increases to a maximum and then suddenly decays across the gap, i.e. the dielectric separating the positive and return lines of the cable, producing a first order distortion to the applied audio signal comprising the 8.3 KHz noise described above. The discharge or decay time of this noise is attributed to the RC time constant created by the cable's own dielectrics used to separate the positive and return conductors. These dielectrics control not only the capacitance of the cable via their materials and thickness, but also have an equivalent resistance. This capacitance and equivalent resistance constitute the RC time constant as outlined above.

After the initial discharge has taken place across the gap between the two conductors through the dielectrics, a second order distortion takes place. This second order distortion manifests itself as a sustained resonance. This resonance rings down with a period that is substantially longer in time than that of the initial gap discharge, adding yet a second distortion to the applied audio signal.

Following the above investigation it was determined that if one controls the amplitude and period/frequency of the first order distortion, the second order distortion, i.e. the sustained resonance, will automatically be controlled, i.e. reduced, if not eliminated. The noise thus produced, if not eliminated, is amplified and superimposed on the applied signal causing distortion.

In accordance with the present invention, an audible noise suppressing interface circuit is coupled across the output of the amplifier, i.e. across the input of the transmission cable. The interface circuit comprises a series coupled capacitor and resistor, i.e. an RC network. The value of the capacitor and the resistor in the network is chosen such that the RC time constant of the combination of the cable with the network is well above the audio frequency spectrum, but still within the bandwidth of the amplifier. For example, in one embodiment of the present invention wherein the amplifier is a pre-amplifier having a typical bandwidth of 20 Hz to 150 kHz and an output impedance of 1.5k ohms to 10k ohms and the transmission cable is approximately 1 meter long and has a distributed capacitance C1 of approximately 300 pF, a discrete capacitor C2 and resistor R1 are used. The value of the capacitor C2 and the resistor R1 in the interface circuit are 1000 pF and 10k ohms, respectively. These values produce a time constant with a period t of 13 microseconds and raise the frequency f of the decay time of the first order distortion to 77 KHz from the previous 6.6 KHz or 8.3 KHz as in the preceding examples in accordance with the following general equations for the combined cable and interface circuit:

where R=R1 and C=C1+C2

In another embodiment of the present invention, wherein the amplifier is a broad band power amplifier having a typical bandwidth of 20 Hz to 1 MHz and an output impedance of 4-16 ohms and the transmission cable is approximately 20 meters long and has a distributed capacitance C3 of 4.5nF/m, i.e. 90 nF, the interface network comprises a distributed capacitance C4 of 3000 pF which is provided by approximately 19.5 meters of unterminated coaxial cable and a discrete resistor R2 of 50 ohms. According to equations (1) and (2) above, the RC time constant (RE(C3+C4)) provides a frequency f approximately equal to 265 kHz for the combined cable and interface circuit, which is, once again, clearly above the audio frequency spectrum but well within the bandwidth of the amplifier.

While the value chosen for the capacitor C and a resistor R in the interface circuit may vary widely within the constraints discussed above, it is preferable that the resistor R have a value which is equal to or greater than the output impedance of the amplifier so as not to unduly load the amplifier down at high frequencies and that the value of C be as low as possible so as to save space.
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,985
Location
Riverview FL
"The interface circuit comprises a series coupled capacitor and resistor, i.e. an RC network."

What's the price of those cables ?
 

cjf

Active Member
Joined
Apr 20, 2016
Messages
160
Likes
58
Location
CO
What's the price of those cables ?

Hello Ray,

At this point I'm still unsure which of their cables, if any, include this contraption. I've called them and left a message yesterday to find out but no response yet.

Going just on the Patent number alone none of the ones I can afford listed this Patent number in their literature.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
As described in the long thread on WBF my issue was in summary as follows:

ANY speaker cable I tried (many) connected between my Amp and Speaker that is shorter than 12ft (depending on gauge) results in a noise like this coming out of my speaker:
https://www.flickr.com/photos/137435336@N08/25030695085/

As the cable length is increased the noise goes away. Wire gauge does seem to play into this as well assuming due to the varying levels of LCRZ..etc that different wires have. I currently have a 4m length of Cardas Clear on hand which still results in noise presumably due to either A:\ its very low Resistance (6.5ga), B:\ its very high Capacitance, C:\ its very low Inductance or D:\ its Impedance

Based on my research the issue smells a little like what NAIM users deal with when using the wrong wire in there setup except mine doesn't cause the amp to shut down or anything as nasty as that. Just noise but I cant say if/how this noise could be detrimental to the hardware if it remained present after continued usage.

I do wonder if its more Inductance related than anything else because bog standard ZIP Cord (parallel wire run geometry) seems to offer the most success in avoiding the issue compared to the much more complex geometry used in the Cardas cables I have tried. A 10ga run of ZIP Cord works with no issues at a length of 12ft. A 14ga run of ZIP Cord works with no issue at a length of 10.5ft. A 12ga run of ZIP Cord works with no issue at a length of 11.5ft. :eek::eek::eek:o_O

Lastly, if the speaker is changed out for something else using the same Amps and short wire the issue goes away. If the Amps are changed out to a Class A/B model of a any kind using the same speaker and short wire the noise goes away. Funny thing is, according to Magico they have a Hypex NC1200 amp on hand that they use for testing and claim they don't have the issue I describe. BUT....when I queried them about trying a test with a short wire I received a rather dismissive response that "We don't have any wire shorter than 12ft on hand to test with". :mad::rolleyes:

When I queried Hypex about the problem they were clueless on what the issue could be but did state that the crossover network Magico uses could be a contributing factor.


First thing, the cardas clear you mentioned does indeed have high capacitance and low inductance. their website says 278 pf and 0.017 uh per ft. Compared to Belden 5000 series at 24 pf and 0.15 uh per ft. I appreciate you may well have spent a hideous amount of money on the cardas, but I would strongly recommend something with a more "normal" range of electrical parameters.

You mention naim. It's normal and common for amplifiers to have an inductor in their output. This stops them becoming unstable and oscillating into certain capacitive loads. Old naim amps did not have this inductor because apparantly it negatively affected the sound. Instead they relied on the inherent speaker cable inductance for this stability. Hence the risk of the amps going wonky with the wrong cable/speaker combo.

I would also suggest that Hypex are correct and that the magico xo is of a particularly reactive design that accentuates the issue. Remember the Hypex are self oscillating, you see the frequency, about 480kHz, on a scope if you connect it across the output terminals. I can well imagine an extreme capacitive load causing issues.

So, overall I would suggest you are correct that it is a capacitance and inductance issue and an unfortunate combo of speaker. It's not a resistance issue, try and keep that low.

EDIT:

After just reading this, I do strongly suspect Magico are doing something "unusual" with their crossovers.
http://magico.net/technology/crossovers.php
 
Last edited:

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
Resistance will dampen the Q of the circuit and make it less likely to misbehave.

Is there an impedance curve (magnitude and phase versus frequency) of the speakers available someplace?
 

cjf

Active Member
Joined
Apr 20, 2016
Messages
160
Likes
58
Location
CO

cjf

Active Member
Joined
Apr 20, 2016
Messages
160
Likes
58
Location
CO
First thing, the cardas clear you mentioned does indeed have high capacitance and low inductance. their website says 278 pf and 0.017 uh per ft. Compared to Belden 5000 series at 24 pf and 0.15 uh per ft. I appreciate you may well have spent a hideous amount of money on the cardas, but I would strongly recommend something with a more "normal" range of electrical parameters.

I would also suggest that Hypex are correct and that the magico xo is of a particularly reactive design that accentuates the issue. Remember the Hypex are self oscillating, you see the frequency, about 480kHz, on a scope if you connect it across the output terminals. I can well imagine an extreme capacitive load causing issues.

So, overall I would suggest you are correct that it is a capacitance and inductance issue and an unfortunate combo of speaker. It's not a resistance issue, try and keep that low.

EDIT:

After just reading this, I do strongly suspect Magico are doing something "unusual" with their crossovers.
http://magico.net/technology/crossovers.php

Thanks for your response.

I'm going to try and get my meat hooks on a few other brands of cable who's geometry more closely matches a standard Zip Cord "Parallel" run since that geometry has been shown to work for me. I will probably try something from Chord Co who's designs use a loosly wrapped twisted pair of separated wires. Kinda like the "Shotgun" approach. This should give a cable with higher than normal Inductance and low Cap and Resistance. If all else fails I will just use bog standard zip cord and call it a loss :eek:

In terms of the Amps I use now (Mola Mola Kaluga), they are from what I understand slightly different in their design from the standard Hypex flavors. I believe I've read that their switching Freq is rather low for a Class D amp in that they are in the 100kHZ range verses the 400+kHZ range for the other models. Not that I think this matters here because they all behave the same in this case with this speaker. Though I have found the Mola's to be even more sensitive to the wire used than the other Hypex amps I have on hand.

Yes I agree the Magico Crossover is indeed "unusual" and I guess I'm the poor slob who wins the set of steak knives while trying to figure out how to make this combo work in harmony. Despite the noise with certain cables this combo does work very well and if it were not for that fact I would have already thrown in the towel and moved on.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Thanks for your response.

I'm going to try and get my meat hooks on a few other brands of cable who's geometry more closely matches a standard Zip Cord "Parallel" run since that geometry has been shown to work for me. I will probably try something from Chord Co who's designs use a loosly wrapped twisted pair of separated wires. Kinda like the "Shotgun" approach. This should give a cable with higher than normal Inductance and low Cap and Resistance. If all else fails I will just use bog standard zip cord and call it a loss :eek:

In terms of the Amps I use now (Mola Mola Kaluga), they are from what I understand slightly different in their design from the standard Hypex flavors. I believe I've read that their switching Freq is rather low for a Class D amp in that they are in the 100kHZ range verses the 400+kHZ range for the other models. Not that I think this matters here because they all behave the same in this case with this speaker. Though I have found the Mola's to be even more sensitive to the wire used than the other Hypex amps I have on hand.

Yes I agree the Magico Crossover is indeed "unusual" and I guess I'm the poor slob who wins the set of steak knives while trying to figure out how to make this combo work in harmony. Despite the noise with certain cables this combo does work very well and if it were not for that fact I would have already thrown in the towel and moved on.


I always recommend the Belden 5T00UP 10 AWG from Blue Jeans cable with the ultrasonically welded terminations, spade preferable. Cheap as chips and fit for purpose.

https://www.bluejeanscable.com/store/speaker/
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Here is the only one I've ever been able to find. Not sure how reliable this rag is but they show measurements towards the end of the .pdf:

http://www.absolutesounds.com/pdf/main/press/S3_HFN_NOV2014.pdf

Yep, they are an absolute bitch to drive. So really on balance I would say its not the amps, its not the cables, its the speaker thats the real issue here.

upload_2016-7-8_12-32-58.png
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
Yah. A high phase angle and very low impedance at very high frequencies is going to cause issues with a lot of amps, and IME (limited) switch-mode amplifiers are more sensitive to HF impedance dips. The output impedance increases while feedback factor and intrinsic stability of any amp decrease with rising frequency. A lot of ribbon and ESL speakers also have very low HF impedance.

The rising HF response is interesting; would not be my choice in a speaker design but then again I am not a speaker designer. However, it seems destined to cause problems for a lot of amps, audible or not.

There is no easy fix for amp/speaker interaction like this. Using series resistors affects the output impedance and will change the overall frequency response. A larger series inductor will isolate the load better at HF but again can change the frequency response of the speaker well into the audible range (or maybe not at my age ;) ). I would see what your amp manufacturer has to say, what Magico has to say, and at worst consider a different amplifier.

IMO - Don
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Yah. A high phase angle and very low impedance at very high frequencies is going to cause issues with a lot of amps, and IME (limited) switch-mode amplifiers are more sensitive to HF impedance dips. The output impedance increases while feedback factor and intrinsic stability of any amp decrease with rising frequency. A lot of ribbon and ESL speakers also have very low HF impedance.

The rising HF response is interesting; would not be my choice in a speaker design but then again I am not a speaker designer. However, it seems destined to cause problems for a lot of amps, audible or not.

There is no easy fix for amp/speaker interaction like this. Using series resistors affects the output impedance and will change the overall frequency response. A larger series inductor will isolate the load better at HF but again can change the frequency response of the speaker well into the audible range (or maybe not at my age ;) ). I would see what your amp manufacturer has to say, what Magico has to say, and at worst consider a different amplifier.

IMO - Don

Yep, Im with you on this. It may not be what CJF wants to hear but IMO the speaker design is somewhat flawed to be presenting that sort of load to an amp. You are asking for trouble and unpredictable results. Less of an amp and more of an arc welder is required. There are plenty of great sounding speakers out there that dont do this, so I think I would personally move them on. :(
 

cjf

Active Member
Joined
Apr 20, 2016
Messages
160
Likes
58
Location
CO
Yah. A high phase angle and very low impedance at very high frequencies is going to cause issues with a lot of amps, and IME (limited) switch-mode amplifiers are more sensitive to HF impedance dips. The output impedance increases while feedback factor and intrinsic stability of any amp decrease with rising frequency. A lot of ribbon and ESL speakers also have very low HF impedance.

The rising HF response is interesting; would not be my choice in a speaker design but then again I am not a speaker designer. However, it seems destined to cause problems for a lot of amps, audible or not.

There is no easy fix for amp/speaker interaction like this. Using series resistors affects the output impedance and will change the overall frequency response. A larger series inductor will isolate the load better at HF but again can change the frequency response of the speaker well into the audible range (or maybe not at my age ;) ). I would see what your amp manufacturer has to say, what Magico has to say, and at worst consider a different amplifier.

IMO - Don

Hello,

It does seem super strange that the Impedance of this speaker is shown to go waaay DOWN at that high of a Freq (20kHZ) which based on almost everything I've read to date shows that the Impedance usually goes UP at higher Freq. So basically this speaker is doing the complete opposite of what is normally seen. o_O

During my Internet travels I have come across quite a bit of material discussing a phenomena known as "Signal Reflections Found Within A Transmission Line" which occur when the source signal encounters an Impedance mismatch somewhere along its journey down the cable and into the Load/Speaker. Some of these discussions refer to the importance of the "Characteristic Impedance" of the cable in use being matched to the Load on the speaker end. Supposedly when the Characteristic Impedance of the cable is matched to the Load, the signal see's the cable as an ideal Transmission Line of infinite length which in turn results in the total consumption of the original signal by the Cable/Load combination (ie..Reflection Free delivery).

On the other hand, in a less than ideal Cable/Load matching scenario, each time an Impedance Mismatch is encountered a reflection of the original signal occurs which heads back towards the source(amp) and embeds itself with the original signal. It then bounces back and forth from source/load until it is absorbed (which could be never).

(Very interesting video that talks a bit about these reflections and how they are formed)
https://www.bethesignal.com/bogatin/epsi0310-think-about-reflections-p-853.html?content=curriculum

It seems like in some cases a "fix" for this issue can be found by incorporating a Zobel Network on the speaker end terminals. The Zobel supposedly helps to smooth out the signal where the Impedance mismatching occurs which in turn supposedly eliminates or reduces the "Reflections" previously mentioned. They say that if the Zobel is built correctly it should have no affect on the sound within the audible band. I guess if this is true or not is another matter.

Personally I would much prefer to find a cable that can provide a "fix" by way of how its geometry/materials/length..etc are incorporated instead of using a Zobel. The hard part of course is finding the magic cable that is significantly better sounding than the Zip Cord I use now.

Building the perfect Zobel Network could be a good topic in a new thread :) . I have at least 10 billion questions I can think of on how to go about doing this in a system using a very powerful amp driving a speaker such as mine into very low Impedance Loads. Almost all the stuff on the web only talks about doing this into systems using relatively low powered amps (100w or less) which is 12X less power than what I'm working with per channel. :eek:
 

cjf

Active Member
Joined
Apr 20, 2016
Messages
160
Likes
58
Location
CO
Yep, Im with you on this. It may not be what CJF wants to hear but IMO the speaker design is somewhat flawed to be presenting that sort of load to an amp.so I think I would personally move them on. :(

:eek::eek::eek::eek::eek::eek::)
 

Don Hills

Addicted to Fun and Learning
Joined
Mar 1, 2016
Messages
708
Likes
464
Location
Wellington, New Zealand
...
During my Internet travels I have come across quite a bit of material discussing a phenomena known as "Signal Reflections Found Within A Transmission Line" which occur when the source signal encounters an Impedance mismatch somewhere along its journey down the cable and into the Load/Speaker. Some of these discussions refer to the importance of the "Characteristic Impedance" of the cable in use being matched to the Load on the speaker end. Supposedly when the Characteristic Impedance of the cable is matched to the Load, the signal see's the cable as an ideal Transmission Line of infinite length which in turn results in the total consumption of the original signal by the Cable/Load combination (ie..Reflection Free delivery).

On the other hand, in a less than ideal Cable/Load matching scenario, each time an Impedance Mismatch is encountered a reflection of the original signal occurs which heads back towards the source(amp) and embeds itself with the original signal. It then bounces back and forth from source/load until it is absorbed (which could be never).
...

o_Oo_Oo_O:rolleyes::rolleyes::rolleyes: This place needs a facepalm smilie. The above isn't actually wrong, but it doesn't apply to speakers unless the cables are many kilometres long.
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
Hello,

It does seem super strange that the Impedance of this speaker is shown to go waaay DOWN at that high of a Freq (20kHZ) which based on almost everything I've read to date shows that the Impedance usually goes UP at higher Freq. So basically this speaker is doing the complete opposite of what is normally seen. o_O

During my Internet travels I have come across quite a bit of material discussing a phenomena known as "Signal Reflections Found Within A Transmission Line" which occur when the source signal encounters an Impedance mismatch somewhere along its journey down the cable and into the Load/Speaker
<elided>

As the other Don said, transmission line theory does not really apply to this situation. I have an article or two over over on WBF showing the results of t-line analysis for speakers and the more relevant case of serial digital signal transmission. Your case is a matter of amplifier (in)stability or resonance when presented with a particular load impedance. A long or lossy wire helps isolate the amp from the load (speaker) to tame the instability. A little extra resistance and/or inductance helps prevent the amplifier from "seeing" such a nasty speaker dip.

IMO - Don
 

cjf

Active Member
Joined
Apr 20, 2016
Messages
160
Likes
58
Location
CO
o_Oo_Oo_O:rolleyes::rolleyes::rolleyes: This place needs a facepalm smilie. The above isn't actually wrong, but it doesn't apply to speakers unless the cables are many kilometres long.

Maybe this Google Cached article is worth your time in reading. I have the full .pdf with all the graphs still present that Google cache seems to remove for some reason. The original files, several in fact, are located on the klipsch forum which is now locked to members only (wasn't yesterday). I hesitate to post the .pdf here as I'm not sure if its allowed.

Anyway, the article talks and shows that the reflections I mention do exist in shorter audio length cables which are not miles long. I do wonder though, after watching the video I linked to previously, how reflections seem to exist in anything considered a Transmission Line (as mentioned in the video). The presenter dude is some .PhD from MIT supposedly. Surely all these folks cant be smoking crack?

This article was written by a dude named Cyril Bateman (https://linearaudio.nl/welcome) FWIW

https://webcache.googleusercontent....ach&attach_id=86522+&cd=1&hl=en&ct=clnk&gl=us
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,204
Likes
16,985
Location
Riverview FL

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Maybe this Google Cached article is worth your time in reading. I have the full .pdf with all the graphs still present that Google cache seems to remove for some reason. The original files, several in fact, are located on the klipsch forum which is now locked to members only (wasn't yesterday). I hesitate to post the .pdf here as I'm not sure if its allowed.

Anyway, the article talks and shows that the reflections I mention do exist in shorter audio length cables which are not miles long. I do wonder though, after watching the video I linked to previously, how reflections seem to exist in anything considered a Transmission Line (as mentioned in the video). The presenter dude is some .PhD from MIT supposedly. Surely all these folks cant be smoking crack?

This article was written by a dude named Cyril Bateman (https://linearaudio.nl/welcome) FWIW

https://webcache.googleusercontent.com/search?q=cache:tgue-G36_WgJ:https://community.klipsch.com/index.php?app=core&module=attach&section=attach&attach_id=86522+&cd=1&hl=en&ct=clnk&gl=us

Hi. Im with Don and Don on this one. Your speakers are a truly horrible load which many amps will struggle with. We know the ncore designs work very well in most situations, in fact yours is the first I have heard of that seems to be problematic. So I really wouldnt be searching for more "esoteric" reasons or causes.

Simple test, beg borrow or steal another set of speakers and see if the problem persists. :)
 

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
7,835
Likes
16,497
Location
Monument, CO
The flight time in a typical audio cable is so short that any transmission line effects will be long gone by the time the speaker begins to move. In addition the rise times of audio signals are slow enough that it just doesn't matter, reflections are long gone by the time the signal starts to rise... Finally the output impedance of an amplifier is normally low enough that any reflections will be absorbed almost completely and will not cause triple-transit and other issues RF designers deal with daily. The flight time (propagation time) is about 6.67 ns/m assuming 0.5c for velocity (half the speed of light in vacuum). Using 30 m of cable (about 100 feet) means it takes about 0.2 us to arrive, and another 0.2 us to get back to the amp, about 0.4 us. That's the period of about a 2.5 MHz sine wave, and a quarter wave of about 625 kHz. I can't hear that high but undoubtedly some audiophiles can.
 
Top Bottom