• 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

Don Hills

Addicted to Fun and Learning
Joined
Mar 1, 2016
Messages
708
Likes
464
Location
Wellington, New Zealand
As you said, for all practical cable lengths and audio frequencies, the cable can be largely ignored provided it's a sensible geometry and heavy enough gauge. The fun comes at frequencies in the RF range, if the amplifier is unstable at those frequencies. The gain of an amplifier falls off as the frequency increases. This is compensated for by (negative) feedback within the amplifier, where part of the output signal is inverted and mixed with the input signal. If the output differs from the input due to distortion or gain changing, this applies a correction. The problem is that a signal takes time to pass through the amplifier. If the signal happens to be delayed by half of a cycle, the output signal is now "inverted" compared to the input. Feeding back an inverted version of this signal results in "positive" feedback, where the output is now in phase with the input. The output increases, the feedback increases, the mix at the input increases... This situation is called an oscillator. Designers have to ensure that the gain of the amplifier falls off below 1 before this frequency is reached.
Also, connecting resistors, capacitors and inductors across the amplifier output can cause this instability to occur at lower frequencies. Other effects can occur too. For example, putting a high value of capacitance across the amplifier output will "short circuit" it at high frequencies and on transients containing high frequencies. As well as triggering the amplifier overload protection, this shorts out the feedback signal. The amplifier tries to run at full gain. The usual result is, again, high frequency oscillation. As well as possible IM effects down in the audible range, output transistors are not designed for RF use and can rapidly overheat.

(Yes, you EEs, the above is deliberately oversimplified.)
 

cjf

Active Member
Joined
Apr 20, 2016
Messages
160
Likes
58
Location
CO
I appreciate everyones continued feedback.

One thing (among many) that is curious to me is this. I see that the speaker appears to dip into the depths of hell @ 20kHZ (.5ohm) but the noise I spoke about initially occurs as soon as the amp is taken out of Bypass mode even before any signal is present. I can remove the XLR Input cable from the amp and just have a speaker cable connected to the speaker terminals while taking the Amp out of Bypass Mode and the results are the same noise.

With that said, wouldn't the only Loads at play upon initial boot up of the amp, before the signal is present, be the Nominal Impedance of the speaker and the Characteristic Impedance (CI) of the speaker cable?

Doing some quick maths based on the published specs of three different cables I have on hand I see the following:
  • 10ga Marshall/Mogami Soundrunner (ie..Zip Cord) has a Characteristic Impedance = 98ohms
  • Cardas Clear has a (CI) = 8ohms

  • Cardas Neutral Reference has a (CI) = 17ohms
The Magico S3 has a published Nominal Impedance = 4ohms
The Mola Mola Kaluga Mono Block Amps have an Output Impedance = <0.002 ohm (DF>4000), all frequencies
The Mola Mola Kaluga Mono Block Amps have a published Bandwidth = >50 kHz

If in fact the only Impedances at play (with no XLR Input) but speaker cable connected are the Nominal value for the speaker itself and the (CI) of the speaker cable then I cant help but wonder how oscillation and/or Amp instability come into play in terms of the noise described? What am I missing?

Based on previous tests it appears that all Zip style cables that I tried are all in the range of 98ohms (CI). As soon as a cable is attached that has a (CI) in the 8-17ohm range the noise doesn't go away at all, even using lengths up to 13ft.

Going back to my past readings that talk about "Ideal Transmission Lines" it is said that when the (CI) of the Cable and the Load at the far end of the cable (speaker in this case) are "matched" then it can be considered an "Ideal Transmission Line" which in turn results in what is considered a cable of infinite length where all signal and noise heading towards the Load are Used/Absorbed fully.

One thing I have not tried as of yet was to use a cable that meets the above "Ideal Transmission Line" description between the Amp and Speaker. Assuming the theory is true, cable length should not come into the picture at all meaning I could try one that is, for example, 5ft long verses the 10-12ft long cables I've been using. The challenge here is finding cables that have a (CI) as close to 4ohm as possible. I've only found one so far and it ain't cheap :eek:. The cable I am referring to is the Cardas Clear Beyond which looks to have a (CI) of 4.7ohms. But at $7500 for a 1.5m length I cant help but be "taken a back" a few steps o_O. If the opportunity presents itself I will try a test with them just to see what happens using a short 1.5m length. You know.....just in case it works and I'm forced to buy it :eek:

Goertz cables claim to have a 3-4ohm (CI) but I've seen tests of those over on Audioholics that shows them to be more like 8ohm cables instead of the claimed 3-4ohm.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
How have you calculated the CI of the cables? Why is that relevant at audio frequencies? As explained transmission line issues dont become an issue until you are at RF frequencies. This only happens in a normal amp if it is already unstable. Can you find any other examples on the net where these amps are unstable?

However, as I mentioned previously, remember these are switching amps which do output a high frequency. Its clear now that the speaker design and XOs are horribly reactive and are almost certainly the cause of the problem.

I appreciate you are trying to find reasons to explain and rectify all this but simply temporarily swap the speakers out and see if the problem goes away. If so, change them; yes I would change the speakers before the amp.

Hope this helps :)
 
Last edited:

cjf

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

(CI) or more correctly "Zo" was calculated as follows:

Zo=SQRT(L/C)

Honestly I believe I may be one of only a handful of people using such a combination on the planet.:eek:

I've only personally seen one other person who said they were running a Magico Q Series with a Hypex Amp (Mola Mola Kaluga also) reply on the WBF thread. They didn't report any issues but they also didn't say what cables they were using or how long they were. Magico themselves said they use an NC1200 Amp for testing with no issues....But...also blew me off when I asked them to try short cables with their combo. I suspect they did and found the same problem but didn't want to get involved. Hypex said they didn't anticipate ever having a Magico speaker/crossover on hand as a tester so that was the end of the conversation.

And before anyone asks...Yes I did actually take my amps with me to demo the speakers before I bought them..There was no noise of course but as it turns out I used the dealers speaker wire which was at least 25ft or more long which at the time I was oblivious to how it would be of any importance.

In terms of swapping speakers, I have done that with these flavor amps and the issues does indeed go away. Same story if I remove these amps with the same speaker. The issue goes away. Certainly changing either component for something else will end the misery but, as you can imagine, will also carry a significant penalty in terms of $$$ lost.

I'm sure changing the speaker would carry less of a penalty then changing the amps (cant seem to even give Class D Amps away these days due to the embedded Old School thoughts on them) but its still not something I am ready/willing to do at this stage. At the end of the day, using the Zip Cord still sounds better than any system combo I've owned to date so this journey for an answer is purely me wanting to know the Root cause with some level of certainty while leaving no stone unturned.

I have taken on the task of finding a "Fix" as more of a hobby/project during my free time. It keeps me occupied, intrigued and frustrated all at the same time ;) . Its also taken me down multiple rabbit holes and sent me on journeys reaching both ends of the Internet. All the while I learned new things about topics I would have otherwise never even thought about so all is not lost.

My journey will continue mostly likely due to being more stubborn than a Mule and a Llama combined. Because of this I'm confident I will find an answer :D
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Hello,

(CI) or more correctly "Zo" was calculated as follows:

Zo=SQRT(L/C)

Honestly I believe I may be one of only a handful of people using such a combination on the planet.:eek:

I've only personally seen one other person who said they were running a Magico Q Series with a Hypex Amp (Mola Mola Kaluga also) reply on the WBF thread. They didn't report any issues but they also didn't say what cables they were using or how long they were. Magico themselves said they use an NC1200 Amp for testing with no issues....But...also blew me off when I asked them to try short cables with their combo. I suspect they did and found the same problem but didn't want to get involved. Hypex said they didn't anticipate ever having a Magico speaker/crossover on hand as a tester so that was the end of the conversation.

And before anyone asks...Yes I did actually take my amps with me to demo the speakers before I bought them..There was no noise of course but as it turns out I used the dealers speaker wire which was at least 25ft or more long which at the time I was oblivious to how it would be of any importance.

In terms of swapping speakers, I have done that with these flavor amps and the issues does indeed go away. Same story if I remove these amps with the same speaker. The issue goes away. Certainly changing either component for something else will end the misery but, as you can imagine, will also carry a significant penalty in terms of $$$ lost.

I'm sure changing the speaker would carry less of a penalty then changing the amps (cant seem to even give Class D Amps away these days due to the embedded Old School thoughts on them) but its still not something I am ready/willing to do at this stage. At the end of the day, using the Zip Cord still sounds better than any system combo I've owned to date so this journey for an answer is purely me wanting to know the Root cause with some level of certainty while leaving no stone unturned.

I have taken on the task of finding a "Fix" as more of a hobby/project during my free time. It keeps me occupied, intrigued and frustrated all at the same time ;) . Its also taken me down multiple rabbit holes and sent me on journeys reaching both ends of the Internet. All the while I learned new things about topics I would have otherwise never even thought about so all is not lost.

My journey will continue mostly likely due to being more stubborn than a Mule and a Llama combined. Because of this I'm confident I will find an answer :D


Fair enough, in that case I think your best bet is to add a decent length of "ordinary" speaker cable which hopefully gets the impedance the amp is seeing above the critical level. :)
 

cjf

Active Member
Joined
Apr 20, 2016
Messages
160
Likes
58
Location
CO
Returning from the dead to provide an update to my previously discussed "issue".

A real cure has finally been found! After installing a set of MIT Magnum series speaker cables into my system all noise issues ceased immediately. Not a fuckin peep ever since, under any condition. The cables in question are also only 8ft long which was previously not an option.

If my story hasn't provided a better case for the purpose/existance of a networked speaker cable then I don't know what else would. In addition, if this doesn't show hard proof to the tin ear 'd number only types that a cable can make a difference to the sound of a system then I don't what else to say other than ignorance is bliss.

I'll be the first to admit that I hate everything about the idea of putting a mystery box in the signal path but I'll be damned if these things dont sound absolutely stunning. No doubt, mostly due to my noise issue being truly fixed by way of the CVT Coupler verses just masked/hidden by using a band aide longer speaker cable.

Interestingly enough, it turns out that the CVT Coupler used in the MIT Magnum and UP cables is the actual device referred to in the MIT patent I posted previously.

Needless to say, I'm a happy man :D
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Returning from the dead to provide an update to my previously discussed "issue".

A real cure has finally been found! After installing a set of MIT Magnum series speaker cables into my system all noise issues ceased immediately. Not a fuckin peep ever since, under any condition. The cables in question are also only 8ft long which was previously not an option.

If my story hasn't provided a better case for the purpose/existance of a networked speaker cable then I don't know what else would. In addition, if this doesn't show hard proof to the tin ear 'd number only types that a cable can make a difference to the sound of a system then I don't what else to say other than ignorance is bliss.

I'll be the first to admit that I hate everything about the idea of putting a mystery box in the signal path but I'll be damned if these things dont sound absolutely stunning. No doubt, mostly due to my noise issue being truly fixed by way of the CVT Coupler verses just masked/hidden by using a band aide longer speaker cable.

Interestingly enough, it turns out that the CVT Coupler used in the MIT Magnum and UP cables is the actual device referred to in the MIT patent I posted previously.

Needless to say, I'm a happy man :D


No, with respect it has shown not to buy speakers with such wayward design parameters that they are a problem. Putting cables with impedance changing networks in them is not a solution IMO.

Are those cables really $8k?

http://www.mitcables.com/legacy-products/speaker-interfaces/magnum-ma-speaker-cable.html
 
Last edited:

Speedskater

Major Contributor
Joined
Mar 5, 2016
Messages
1,610
Likes
1,333
Location
Cleveland, Ohio USA
I had lost track of this thread.
a] The cable impedances listed in post #82 are all Radio Frequency Characteristic Impedances. The cables don't approach those impedances until about 100 kHz. At audio frequencies the Characteristic Impedance varies with frequency and is a very high number at low frequencies.
b] A well behaved transmission line requires that the output stage impedance, the cable impedance and the load impedance all be equal.
c] A interconnect system does not function as a transmission line until it's length approaches a goodly fraction of one wavelength. At 5 MHz that would be almost 100 feet.
d] A Cyril Bateman paper showed that a reasonable amplifier, reasonable cable & reasonable loudspeaker can oscillate at 5 MHz.
 

cjf

Active Member
Joined
Apr 20, 2016
Messages
160
Likes
58
Location
CO
No, with respect it has shown not to buy speakers with such wayward design parameters that they are a problem. Putting cables with impedance changing networks in them is not a solution IMO.

Are those cables really $8k?

http://www.mitcables.com/legacy-products/speaker-interfaces/magnum-ma-speaker-cable.html

Your right, I should have brought my mobile measurement lab equipment with me to the dealer and did a full host of tests on the speaker in their listening room prior to making a purchase. Things are always so clear after the fact and of course once you have been made aware that you are about to embark on a journey/purchase that has a 1 million chance of being problematic. :rolleyes:

Anyone familiar with the process of purchasing a big/heavy/expensive/built to order speaker would no doubt tell you that the logistics involved in having an in home demo prior to purchase is not very common or likely to happen or least not likely to happen as a kind gesture by the dealer/manufacturer on their dime just so you can see if every possible issue that could occur does not occur once that speaker is sitting in your room. In many cases the most one can do is bring some of their equipment with them to the dealer during the demo and hope to flush out any Voodo prior to making a purchase. I did just that in this case, but if anything the lesson here is to not take ANYTHING for granted in terms of the equipment involved during the demo including speaker wire, speaker wire length, speaker wire dielectric, AC outlets, wiring to the breaker box, power conditioning, power cables....and on ...and on...and on...etc. Hopefully by now you begin to see the bigger picture and that things are not alway so simple.

In terms of the MIT cables that fixed my problem its my understanding that they are not doing any kind of Impedance matching. At least they are not advertising that in any of their marketing material. The main component at play in the cable in terms of how it is fixing my issue is the "CVT Coupler" which, as you can see in the Patent I linked to previously, is a noise filter. The scenario shown in that Patent essentially describes my issue to a tee and then talks about how it addresses that problem, which as it turns out, is not related to Impedance dips or mismatches. I haven't a clue how or why that issue is occurring with this combo of equipment but clearly it is occurring for reasons unknown.

Yes, the cable is expensive but I'm not going to get into that here as it has no bearing on the conversation or the issue described.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Your right, I should have brought my mobile measurement lab equipment with me to the dealer and did a full host of tests on the speaker in their listening room prior to making a purchase. Things are always so clear after the fact and of course once you have been made aware that you are about to embark on a journey/purchase that has a 1 million chance of being problematic. :rolleyes:

Anyone familiar with the process of purchasing a big/heavy/expensive/built to order speaker would no doubt tell you that the logistics involved in having an in home demo prior to purchase is not very common or likely to happen or least not likely to happen as a kind gesture by the dealer/manufacturer on their dime just so you can see if every possible issue that could occur does not occur once that speaker is sitting in your room. In many cases the most one can do is bring some of their equipment with them to the dealer during the demo and hope to flush out any Voodo prior to making a purchase. I did just that in this case, but if anything the lesson here is to not take ANYTHING for granted in terms of the equipment involved during the demo including speaker wire, speaker wire length, speaker wire dielectric, AC outlets, wiring to the breaker box, power conditioning, power cables....and on ...and on...and on...etc. Hopefully by now you begin to see the bigger picture and that things are not alway so simple.

In terms of the MIT cables that fixed my problem its my understanding that they are not doing any kind of Impedance matching. At least they are not advertising that in any of their marketing material. The main component at play in the cable in terms of how it is fixing my issue is the "CVT Coupler" which, as you can see in the Patent I linked to previously, is a noise filter. The scenario shown in that Patent essentially describes my issue to a tee and then talks about how it addresses that problem, which as it turns out, is not related to Impedance dips or mismatches. I haven't a clue how or why that issue is occurring with this combo of equipment but clearly it is occurring for reasons unknown.

Yes, the cable is expensive but I'm not going to get into that here as it has no bearing on the conversation or the issue described.

Yep a filter will be changing the impedance at those frequencies, I didnt actually say impedance matching. Additional filtering is probably not a good thing and it doesnt solve the fundamental problem. The fundamental problem being that your amp is made unstable by what is, by anyones standards, a highly reactive and difficult speaker load. It is a wayward design that is going to cause problems for many amps.

Of course I appreciate that any of this would not have been apparent when auditioning the speakers at a dealer, and that it is a technical problem that few would understand or account for. However personally I do check out technical reviews of products before I purchase and the one you linked to highlighted the issue quite clearly.

BTW they are charging $8K for a box with a few caps, resistors and maybe inductors in?

images


attachment.php

attachment.php

mitoraclev2.1__.jpg

images
 
Last edited:
Top Bottom