My DIY Project & a Forgotten LINK

[Eric]

<font color='#000000'><table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Quote </td></tr><tr><td id="QUOTE">
When anyone makes unsubstantiated and/or un-provable claims about cables we thrash them.  What do you guys think?
</td></tr></table>

I wouldn’t say you “trash them” but rather challenge their claims using engineering. Keep up the great work!  

</font>

 

[Guest]

<font color='#000000'>&quot;&quot;RE the Audio Review forum link, I think that site speaks for itself, it is a cable naysayer haven, and their reception to my DIY designs is not exactly going to be a warm welcome. &nbsp;Not proof of anything, including jneutron/John Escallier's nit-picking post that you link to.&quot;&quot;

Well, at least you spelled my name right

My post is not nitpicking...I have placed there a scientifically consistent, to the best of my knowledge (and my peers at work) taking apart of what I consider to be very unscientific information..

I am not a naysayer, not a yaysayer, any of that garbage..

What I have provided is science....and methodology..

I point out the items on you site that do not fall into that category...

I also invite technical debate on any of the points I've made, as that has always been my intent..



Cheers, John</font>

 

[Dan Banquer]

<font color='#000000'>Well Jon R, you have proposed a cable answer to improperly designed amps. Why not get to the root of the problem and fix the amp, and not add more zobel filters, along with more reactive cable? Why add more reactance in front of a loudspeaker crossover? This makes absolutely no sense.
As far as my use of RG 174 for line level cable internal to the equipment, Yeech is not exactly a  precise adjective. You have not proven that there is anything wrong with with RG174. You have not documented any possible audio problem, and neither has anyone else to my knowledge. So I will view your response of Yeech as simply your personal opinion, because you lack any credible evidence to do otherwise.</font>

 

[Jon Risch]

<table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>

Dan Banquer : <font color='#000000'>Well Jon R, you have proposed a cable answer to improperly designed amps. Why not get to the root of the problem and fix the amp, and not add more zobel filters, along with more reactive cable? Why add more reactance in front of a loudspeaker crossover? This makes absolutely no sense.</font>

<font color='#000000'>Most people have to treat the amp as a black box, they can not go in and fix it, or re-design the circuit or layout.

If they want to use low Z cables, they have to live with the cards they have been dealt, as most folks are not going to run out and buy a &quot;properly designed amp&quot;.

So using a Zobel at the speaker, in order to use the cables they want to, is not such a strange thing after all.

I note that Goertz is now offering a built-on Zobel for their ribbon sandwich cables at no charge upon customer request.

Jon Risch</font>

 

[gene]

<font color='#000000'>Jon;

Zobels are effective at making high inductive loads look resistive. &nbsp;Lets say for example you have a tweeter that impedance increases without bounds with frequency, this is where a zobel will help reduce this problem. &nbsp;As far as I know, the most effective way to prevent amplifier oscillations due to highly capacitive loads is to provide a series resistance or inductance, or both before the load. &nbsp;I suspect adding a zobel after the speaker cable does not provide effective isolation for the amplifier from the capacitive load. &nbsp;I am working on this and will hopefully have something on paper soon.

IF you have any measurments for these high capacitive cables, they would be helpful.</font>

 

[rode]

<font color='#000000'>Gene is quite right here. A &quot;far-end&quot; zobel has little or no effect on the amplifier with cables of typical lengths, since the reflections caused by the impedance mismatch are at frequencies well outside the amplifier's bandwidth. Should the amp's response still be above unity gain at around 20MHz or greater (unusual but possible), then a RC zobel at the speaker end is recommended. The resistive component should be equal to the cable's characteristic impedance - i.e. not the typical 100nF in series with 10 ohms (unless cable Zo is, in fact, 10 ohms).

A speaker end (far end) zobel will be useful to terminate the cable in areas with high RF interference, and is (IMO) a good idea - just don't expect it to have any effect on amplifier stability with most amps.

Amplifier-end zobels (series RL types) are essential to isolate the amplifier from very high capacitance cables to maintain a satisfactory phase margin. Where the amplifier is not fitted with an RL zobel (usually ~ 800nH || 10 ohms or so), then this should be supplied with the cable.

Naturally, the addition of an inductance to make the amp stable with the ultra low inductance cable is somewhat counter-productive, but strangely this doesn't seem to bother your run-of-the-mill snake oil vendor &nbsp;

I have done some initial simulation work on this topic, and will be working with Gene to produce the final article. There is a lot more to this than any snake-oil vendor tells us in the colour glossies - much more. &nbsp;(Oops - I meant to say &quot;cable&quot; - never mind, same difference.)

Naturally, this entire debate can be rendered obsolete by simply using powered loudspeakers, with a triamped system using electronic crossovers and separate amplifiers either built into a plinth or even in the speaker box itself.

Now that is the sensible way to do it, and eventually all loudspeakers will be made that way. It is already common with near-field studio monitors and small PA systems, and has also been done with hi-fi (with great success, I might add).

Cheers, &nbsp; &nbsp;Rod</font>

 

[Jon Risch]

<table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>

rode : <font color='#000000'>Gene is quite right here. A &quot;far-end&quot; zobel has little or no effect on the amplifier with cables of typical lengths, since the reflections caused by the impedance mismatch are at frequencies well outside the amplifier's bandwidth. Should the amp's response still be above unity gain at around 20MHz or greater (unusual but possible), then a RC zobel at the speaker end is recommended. The resistive component should be equal to the cable's characteristic impedance - i.e. not the typical 100nF in series with 10 ohms (unless cable Zo is, in fact, 10 ohms).


Amplifier-end zobels (series RL types) are essential to isolate the amplifier from very high capacitance cables to maintain a satisfactory phase margin. Where the amplifier is not fitted with an RL zobel (usually ~ 800nH || 10 ohms or so), then this should be supplied with the cable.</font>

<font color='#000000'>Reply to Gene and Rod:

With regard to speaker cable far end ZObel networks, I am not talking about theory here, this is real world experience, and it works. &nbsp;

In a few desperate instances, some folks have placed a series resistance in line to help stabilize an amp, but this is obviously a last resort, and is most often used with electrostatic speakers, which are already a difficult load unto themselves.

As to why the far end Zobel works, this should be obvious to Rod. &nbsp;The typical speaker cable 'Zobel' values of 0.22 uF and 8-10 ohms tends to swamp out the cable C, which is usually on the order of from 0.01 uF to .003 uF to reach instability. &nbsp;The R provides a way to keep this amount of C from pushing the amp over the edge.
In technical terms, it keeps the amp phase margin from exceeding the oscillation threshold before unity gain is reached.

RE the so-called amp zobel being an RL, I have never heard this refered to in this manner, a Zobel network has always been an RC network, in so far as I am aware. &nbsp;I have heard the series RL &nbsp;network referred to as an isolation network, but never a zobel.

Is this just your own use of the term in this manner, or is there any sort of readily cited professional use of the term Zobel for a series RL? &nbsp;

Strictly speaking, I am not sure that the typical speaker cable RC network actually qualifies as a true Zobel either, but at least it has the same nominal location and circuit as a Zobel, and so can be excused as being referred to by many as a Zobel network.

Jon Risch</font>

 

[Bprest0n]

<font color='#000000'>Jon;

I think I would agree with some of what you said if the &quot;R&quot; was in series with the load, not in shunt.  Perhaps Audioholics could get some amplifier vendors to comment here.  BTW, the discussions here lately have given me little or no desire to participate in other forum venues.  Good stuff.

All of you guys have been debating with a high degree of professionalism that in my experience has been unmatched by other audio forum discussions.</font>

 

[rode]

<font color='#000000'>Jon,
Somewhere along the line, theory and practice should find common ground. If this does not happen, then one disipline or the other is obviously at fault. &nbsp;In the case of a &quot;far-end&quot; zobel, it can have a profound effect on the signal at the speaker end, but will (usually) have a minor effect as regards the phase margin of the amplifier.

Where an inductor is used at the amp output, the effects are non-existent to the amplifier, since the cable is decoupled from the amp by the inductance at those frequencies.

You are right about the inductor/resistor BTW - it is not (strictly speaking) a &quot;zobel network&quot;, so I suppose I had better stop referring to it as such to avoid any confusion

In the various simulations I have done so far, using a &quot;lossy&quot; transmission line (with R, L and C), the optimum value of the far end zobel resistance is equal to the line impedance. Since it is a rare speaker cable that has a characteristic impedance of 10 ohms, then that value is clearly not optimum as a &quot;generic&quot; fix. Although it reduces the amplitude of the primary reflection, it does not (and cannot) eliminate it. Instead, it just moves the peak somewhere else (typically to a much lower frequency)

Transmission line theory (and practice) indicates that this should be the case - a mismatch will always cause reflections - only a perfectly matched transmission line is free of all reflected energy (standing waves).

Having looked at this more closely than before, I feel that there is good reason to use a far-end zobel, matched to the line impedance for RF. Doing so will make the speaker lead less likely to act as an antenna for stray RF fields, and minimising standing waves will ensure that a very wide bandwidth amp (without a series inductor) will not be affected. How many such amps are available? I honestly don't know, but for a 3 metre (10') cable, (Zo of 43 ohms) the first major reflection appears at around 5MHz - this is generally well outside the danger zone for most amps.

Predictably, longer cables will lower the frequency, but again, the use of 100nF and 10 ohms is ill advised unless the Zo of the cable is 10 ohms. Matching networks should be matching networks, not &quot;different mismatching&quot; networks.

I am still gathering data on the characteristic impedance (and other parameters) for a selection of cables for my simulations, and hope to be able to publish some results shortly.</font>

 

[Guest]

<font color='#000000'><table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Quote </td></tr><tr><td id="QUOTE">While I am not normally that anal about mispellings or typos, it would be a cable recipe, not receipt.</td></tr></table>

While I am not normally that anal about those who erroneously correct mispellings or typos, receipt is a synonym of recipe, Jon.

And so you don't think I came over here just to pick on you, the real reason I came here is to express my concern over the curious phenomenon of unregistered users popping up like clockwork over on Audio Asylum within days or less of an Audioholics article being published to tell everyone about the article.

If there is any sockpuppeting going on on the part of anyone associated with Audioholics, knock it off. This whole thing's already beginning to get just as childish as the Stereophile/TAS rivalry.

se</font>

 

[Guest]

<font color='#000000'><table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Quote </td></tr><tr><td id="QUOTE">RE the so-called amp zobel being an RL, I have never heard this refered to in this manner, a Zobel network has always been an RC network, in so far as I am aware.  I have heard the series RL  network referred to as an isolation network, but never a zobel.

Is this just your own use of the term in this manner, or is there any sort of readily cited professional use of the term Zobel for a series RL?</td></tr></table>

Jon, a Zobel network is simply an impedance compensation network, intended to present a resistive load to the source by way of neutralizing the reactance causing the load's impedance to change with respect to frequency. Whether you use an RC network or an RL network depends on what you're compensating for. If you're compensating for inductance, you use an RC network. If you're compensating for capacitance, you use an LC network.

See Zobel, OJ, &quot;Distortion Correction in Electrical Circuits with Constant-Resistance Networks,&quot; Bell Systems Technical Journal, July 1982, p. 438.

That reference is also cited in the following article that's more readily accessible:

Zobel network tames reactive loads

You really make it hard not to seem like you're being singled out to be picked on, Jon.

se</font>

 

[gene]

<font color='#000000'>Steve;

Welcome to our forums.

<table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Quote </td></tr><tr><td id="QUOTE">If there is any sockpuppeting going on on the part of anyone associated with Audioholics, knock it off. This whole thing's already beginning to get just as childish as the Stereophile/TAS rivalry.</td></tr></table>

While I appreciate your participation on these forums, it will not be welcomed if you continue to make such accusations.
Your accusations are both incorrect and unwarranted.  While I can't speak for all of our forum members, I can assure you that our staff have better things to do then sit at AA promoting our articles.    What would be the gain?  (Especially since the majority of AA members don’t agree with our viewpoints)  

May I suggest that you ask the AA moderators to ban the IP address violators.  If certain Audioholics forum members are in fact doing as you say, then please report it to us and we will warn them. &nbsp;IF you send us the IP addresses, we can verify if any of the AH forum members have similar addresses.

Thank you for understanding and your very informative second post with accompanying link.</font>

 

[Guest]

<font color='#000000'><table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Quote </td></tr><tr><td id="QUOTE">While I appreciate your participation on these forums, it will not be welcomed if you continue to make such accusations.</td></tr></table>

Perhaps you missed the word &quot;if&quot; which began my statement that you quoted. It's a little word I know and sometimes easily overlooked but I made sure to put it there so as to make clear that I was making no accusation.

se</font>

 

[gene]

<font color='#000000'>Point well taken Steve. &nbsp;I apologize if my response came off rather strong. &nbsp;Unlike many other audio forums, we don't tolerate flame wars and/or personal attacks against each other or other websites,companies, etc. &nbsp;I was attempting to eliminate that from starting with my original post. &nbsp;However, I can see you are a reasonable person. &nbsp;Again thanks for your useful contributions. &nbsp;Rod and I are currently working on an article about Zobel terminations that should prove a most interesting read to those who are interested in this topic.</font>

 

[Guest]

<font color='#000000'>No apologies necessary. I can come off rather strong myself. Just ask Jon.

se</font>

 

[rode]

<font color='#000000'><table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>

Guest : Jon, a Zobel network is simply an impedance compensation network, intended to present a resistive load to the source by way of neutralizing the reactance causing the load's impedance to change with respect to frequency. Whether you use an RC network or an RL network depends on what you're compensating for. If you're compensating for inductance, you use an RC network. If you're compensating for capacitance, you use an LC network.

Hi Steve,
I saw the reference you indicated in your post, but try as I may, could not find any reference anywhere to a series RL network being referred to as a Zobel. Parallel, yes, but not series.

Technically (as an impedance compensation network), the term is probably appropriate, but common usage shall prevail, and I have determined to refrain from calling it a Zobel.

While I must admit that Jon and I rarely see eye-to-eye, to prevent any confusion it is probably best referred to as an &quot;isolation network&quot; or similar as he suggested.</font>

 

[Guest]

<font color='#000000'><table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Quote </td></tr><tr><td id="QUOTE">I saw the reference you indicated in your post, but try as I may, could not find any reference anywhere to a series RL network being referred to as a Zobel. Parallel, yes, but not series.</td></tr></table>

Well I guess the acid test will be Zobel's article. I'll see if I can round up a copy.

se</font>