The "Elsinore Project" Thread

Yeah be careful with the wording not to lead imagination astray 🙂 it all happens practically simultaneously. If its a single circuit loop any current at any point in the loop is basically the same at the same time, voltages / currents / impedances sum and affect simultaneously. If there is voltage source and loudspeaker making one circuit the same current flows anywhere in the circuit at the same time. Claiming speaker backEMF affecting amplifier current or speaker current or circuit current or amplifier affecting loudspeaker current or what ever, they are all the same current and everything works together at instant. Things change if there is a shunt circuit(s) which make multiple current loops / paths possible. Still its simple maths, simple circuit analysis.
Back EMF also occurs with regular fixed inductors in response to normal signal related flux variations.
Yeah this is the impedance of an inductor called inductance 🙂 Input ac tone, or flick a switch to apply dc current, so that current flowing in the inductor changes. Change in current generates changing magnetic flux around the conductor which makes the backEMF voltage, which makes opposing current / reduces the current. This is the impedance of the inductor, the backEMF and happens at realtime. The impedance is as if there was a frequency dependent resistor which lowers voltage across the inductor and thus affects current but there is no resistor but the magnetic field, self inductance making opposing voltage, a voltage drop. Impedance is Z = V / I, voltage per current.

Can be simplified to imagination by thinking its separate voltage source with separate current generated over the impedance at some particular frequency, circuit analysis.
 
Last edited:
I would like to correct a few minor ouppsies in my text above...

are not effected by the speaker amplifier impedance (they win the tug) in the same way as e.g, a high output impedance amplifier (e.g. a typical tube amp or indeed a current drive amp) where the high output impedance amplifier will have a distorted FR which follows the drivers impedance response.

It was probably quite obvious but....

//
 
It is interesting that certain people criticise my language and then confirm what I have been saying all the time. What's going on?

How about this for language:

The amplifier has no control over the current. The load does, because it is the load that determines the current.

The fact is that the driver/speaker affects and corrupts the current in a "circulating" manner (this is the word that Esa Merilainen uses, not me). So clearly we are not talking about a one-way action, but something that can become continuous as long at the stimulus is there.

True or false?

Esa say true and that there is only one solution. Religiously you have to employ current-drive.

I also say true, but refuse to believe that there is only one solution.

I also say no to rejecting voltage sources.

So I am on the side of voltage sources, but we have a number of techniques we can employ that will reduce the corrupt "circulation" currents. It's the not so secret sauce in the Elsinores - they sound like current-drive when a good voltage source is used.

We must not take our eyes of the ball: How well the speaker performs is related to the current it processes, because the force-factor is related to the current of the amplifier and not the voltage of the amplifier. The voltage can only ever be a guide, but not the end result.

So what is wrong with that language?
 
Hi,

there doesn't even need to be an amplifier for current to circulate, only some conductor between driver terminals to make a complete circuit and you'd still measure the same phenomenon, the backEMF and the side-effects in the motor, current in the circuit and distortion on it if you move the cone mechanically. If you replace the short with a resistor, or an inductor, you'd measure less of the effect. You see, because its mostly due to the driver, not due to the amplifier, the amplifier could be abstracted away and you'd measure less distortion with increased impedance. This is the nuance that sticks out and triggers to write.

You keep preaching it like this:
Let's be clear, and I have said this many times, current-drive cannot lower driver distortion. The key to current-drive is that there is less amplifier distortion.

Once again I will be misinterpreted, but it goes with the job, so-to-speak.

I am not talking about voltage distortion of the amplifier, I am talking about the current distortion of the amplifier.
The text I've quoted completely downplays what happens in the driver motor, all the distortion mechanisms that actually get better with the increased circuit impedance. I think you are blaming wrong tree. Sorry being blunt but to me this reads lack of understanding whats going on in the circuit, what the distortion mechanisms actually are.

This is the reasoning me and others have posted numerous times in the thread. Its also in the Esa's work talk about this as well as far as I understand. Esa promotes current-drive amplifier because the driver is the culprit for distortion, it fixes distortion in the driver and the mechanism is increased impedance in the circuit and backEMF voltage does not generate as much current into the circuit. Any current in the circuit shows up in acoustic domain. Its not the amplifier that is great in Esa's work, its the circuit impedance, reducing current generated by backEMF, reducing acoustic distortion from backEMF.

Its the loudspeaker driver that makes most of the distortion not the amplifier. Its the loudspeaker motor distortion that reduces with the increased circuit impedance, not the amplifier distortion. Although, the amplifier distortion probably reduces as well but the bulk is from the driver you ignore the driver distortion completely, attributing it all to the amplifier. An amplifier works as designed as long as there is loudspeaker as its load, not a short, its designed to work with typical speaker loads and should be pretty linear in its operation, distortion numbers 0,001% or alike with the rated load of a loudspeaker, getting even better when the load impedance gets easier. A poor driver can add several magnitudes of distortion over the amplifier distortion and its mostly that what gets fixed with the circuit impedance.

When you alter the circuit by increasing impedance by adding an inductor in between an amplifier and a driver the amplifier works as good as before, with the rated distortion, but the driver generated distortion is reduced because the backEMF voltage makes less effect on the current in the circuit when there is higher impedance in the circuit. Exactly same thing is at play when there is current amplifier, which raises the circuit impedance with its output impedance. Exactly same thing happens if you use series resistor with voltage amplifier. Its the circuit impedance, it doesn't matter what the amplifier is. The amplifiers contribution to this driver distortion is only in its output impedance.

The same current flows through amplifier and the driver so you could say its "amplifier current" or what ever, but you seem to underline on every post that its the amplifier and ignoring the driver backEMF altogether and how it works.

As I've said before as long as you get the results it doesn't matter, you get to same end results, reduced system distortion measured acoustically. Also your customers get the real deal, real performance, so there is no harm in the discussion for outsiders. But as long as you preach ignoring the backEMF there is going to be someone asking you why do you think so and the discussion goes on.
 
Last edited:
  • Like
Reactions: AllenB
I've got to be thankful of your posts Joe, because they got me thinking about the subject in the first place.

After reading your posts last year the question that got to my head was something along this:
Why would system THD measured with mic drop from 1% to 0.5% when I add an inductor in series? I've got great amplifier with 0.001% distortion so how come amplifier distortion could have anything to do with it?

Reading Esa's article, purifi articles, it is logical to come to a conclusion that the distortion source is something in the driver motor and how the circuit impedance affects it. It was tough as I couldn't find much on the subject on the forums other than noise, you were the most active member with the subject but the logic didn't match whats in the papers I found. Papers can be tough to understand so it took some time to build confidence with it. Now I'm not afraid to post about the subject because I'm confident on it.

And I'm sorry to write about all this but I feel there needs to be some hints for anyone reading it that there is more out there, go on and find out themselves. I'm fine with you and your view, I'm concerned about other learners and tinkerers like me finding out the stuff that they don't believe what they read and get confused like I was. Our brains are different and process stuff in various ways, thought processes differ, providing various ways to look at things hopefully helps anyone to navigate their own path. In the end we don't know much about how stuff actually works, the bottom of the science, so all that matters is anyone to get to the results, better sound.
 
Last edited:
  • Like
Reactions: JarBar
Lars, if you are reading this, arguably the behaviour above 1KHz to near 5KHz made my job easier - the consistency between 0°, 15° and 30° is the best I have seen and I was thus able to used an increased series inductor value there. There is something you did to that cone, on top of the magnetics and the 'funky' surround, the way that I believe you distributed the mass has a lot to do with it. Just my guess.
Thank you. its the result of optimization of the cone geometry. A combination of Matlab code and comsol FEA is trying thousands of variation of the cone profile. The object function favours both flat on axis response as well as smooth off axis response. The funky surround helps control the usual cone edge breakup (in addition to eliminating modulation of the radiating area)
 
feedback: for voltage drive we have motional feedback around the driver resonance. current controls force and force is integrated to velocity. The velocity causes back EMF (velocity times Bl) and we get a negative feedback loop where increased velocity causes a decrease in current. For a low Qts driver this effectively works as motional feedback control of the velocity where the amp voltage controls the velocity and hence a 6dB/octave response around fs. This feedback can actually lower distortion caused by the suspension and we can measure current distortion that exceeds the acoustical distortion (because nonlinear current is needed to give linear motion). This obviously only works to the extend that Bl(x) is constant
 
Thank you. its the result of optimization of the cone geometry.

Well, it works. Brilliantly! That inductor I use is key to the effect that I am trying to achieve. If the off-axis falls off sharper, then I would have had to reduce the inductor value, something I am happy that I didn't have to.

BTW, I hope to have one of Alan March's amps here soon.
 
  • Like
Reactions: lrisbo
The text I've quoted completely downplays what happens in the driver motor, all the distortion mechanisms that actually get better with the increased circuit impedance. I think you are blaming wrong tree. Sorry being blunt but to me this reads lack of understanding whats going on in the circuit, what the distortion mechanisms actually are.

WHAT?

Why would you think that I am downplaying what happens in the driver motor? Nothing that I have said would justify such a comment... sigh.

I suggest that you have not thought through what I have said. Then to accuse me of lack of understanding? I wonder how much I have to endure?

Over and over I have said that the driver quality is paramount - that there is no substitute and that current-drive cannot overcome a bad driver. Are you not reading my mail???

So please, you need to see the whole picture that I am presenting, and when you do I am optimistic that you will end up agreeing with me. And maybe even apologise.... sheesh!

And the evidence I hope to present should be able to do that, but I feel no undue reason to rush until get it right.
 
Here is the quote again:
Let's be clear, and I have said this many times, current-drive cannot lower driver distortion. The key to current-drive is that there is less amplifier distortion.
...

Let me rephrase, perhaps there is misunderstanding in the words:

Current-drive lowers driver motor distortion reaching acoustic domain. High output impedance of curren-drive amp lifts impedance in the circuit up, which reduces distortion current from backEMF voltage, also kills electronic damping current, which is also from the backEMF. Key thing, as I understand you posts you clearly deny this in the quote.

The quote continues attributing all distortion reduction to amplifier, which is not true, its drivers contribution in acoustic distortion that reduces when current drive, or mixed drive.

You simply cannot say driver distortion is not affected, because that is not true. Driver distortion reduces in acoustic domain.

Perhaps here is the misunderstanding?
The distortion mechanism in the motor is still there as it is physical property of the driver and no amplifier or impedance circuit can change it. But all we are interested is distortion in acoustic domain , which we listen to, and that is reduced with current drive.

This driver distortion is eliminated/reduced reaching acoustic domain, with current-drive, or mixed drive, the circuit impedance outside the driver. The driver still makes same backEMF voltage, variation in impedance, appear but that does not turn as much into current because of high circuit impedance.

See, when the driver distortion mechanism makes less current into the circuit, there is less distortion in acoustic domain. Any current through voice coil turns into force in the motor and acoustic domain.
 
Last edited:
Reading Esa's article, purifi articles, it is logical to come to a conclusion that the distortion source is something in the driver motor and how the circuit impedance affects it. It was tough as I couldn't find much on the subject on the forums other than noise, you were the most active member with the subject but the logic didn't match whats in the papers I found. Papers can be tough to understand so it took some time to build confidence with it. Now I'm not afraid to post about the subject because I'm confident on it.

Yes. it is very challenging. Esa's is describing a mechanism that we have discussed and we agree on it. Asking others, most do not really understand what it is all about. You have not been privy to that interchange with Esa. It happened on the Vendor's Forum when the book came out and I remember also John Curl joining the conversation. I was among the first to hastily secure his book and I have said to many that I consider it an important book. Esa's language has also changed a bit here and there, and his English has improved too.

But kudos for trying - I have been there too.

So the reason why you find it challenging may also be the reason that you have so misunderstood my side of it. This is a tough subject. And you are right, there is a lot of uncoordinated noise on the net. Ain't that the truth.

The difference between Esa and I is that he believes that current-drive is the only solution. But think about it, if we had perfect drivers, would current-drive achieve anything? No, not really. Voltage-drive would work perfectly well and we would not need current-drive. So driver quality must come first.

My solution is to stick with current-drive or amplifiers with relative low output impedances, definitely below 10 Ohm.

I believe I can do just as well as current-drive and not need it.

So please, on this point at least, see that I am coming from a different angle than Esa is.

Now to pick on a single sentence saying that the amplifier is distorting, you could get a very wrong idea where I am coming from.

Still friends?
 
Yes, when one understands that its the circuit impedance affecting how driver motor works and how electronic signals turn into acoustic sound its easy to see through his message. It is very clear to me how to utilize the information.

I think too that Esa's message is colored and provoking and the tone on the article (haven't red the book) is daunting, but the science is clear and easy to understand, just look through the unnecessarily colorful tone he uses. His article shows how impedance in the circuit can be utilized to reduce distortion and utilize electronic damping if needed. What he is talking about is really the circuit impedance, and if he would have actually wrote the book (article) from that perspective it would be more easily readable I think, at least the article would. His words are concentrating too much on cursing industry to abyss and repelling most of the readers I think.

Type of amplifier is irrelevant in this sense, impedance in the circuit is only relevant thing to understand, to be able to utilize a driver with best suitable way with any amplifier you happen to have. You are right, voltage amplifier with low output impedance and a series coil is the ticket: it makes low impedance circuit for low frequencies, for electronic damping to happen at drivers main resonance if its needed, and increases circuit impedance higher up reducing distortion from driver motor non-linearities turning into acoustic sound. But it all changes if you introduce shunt components, but that is fine because it can be adapted again for best compromize. If driver motor is very good then there is not much distortion to reduce so all this is most usable for bad drivers, or applications where excursion is high. Besides distortion there is the damping at resonance.

The message and science is there, just ignore the current-drive-amplifier-war-against-the-world point of view and think the circuit impedance.

I haven't read the book, but all the important core information is here in this single article.
 
Last edited:
The message and science

Please don't think that I disagree with the science. If you have understood it so far, my hat's off to you.

But you have to understand that you have gotten where you are because of looking at it through Esa's eyes, and not mine. I am not sure if you understand where I am coming from. His view is coloured by his religiously advocating current-drive. I am not! When you look at it from a different angle, then chances are you will see something else. It is that kind of a subject.

So please keep you powder dry. There is an experiment that I am seeking to do and with which I have been consulting with others, that when you see the results, that it may revise your thinking about what I said (which you misunderstood anyway). It might even revise Esa's view? I think he is too wedded to the idea. But if this experiment works out, then a hypothesis will survive.

I need to ask for your patience. That is also a good thing to have in science.

BTW, the use of current-drive makes a lot more sense in active speakers. And that is already in some commercially available loudspeakers. But nobody has done current-drive on the bass (except using mechanical only damping) by maintaining the electrical Q. But there is a way... EQ the current like I do in the Elsinores - you can use current-drive with the Elsinore loudspeakers, bass and all.

BTW, been watching Finnish news lately - I hail from CPH Metro.
 
If you read my post again, read Esa's article again just by thinking impedance and ignoring the whole current-drive mentality. Just forget amplifiers altogether because they are almost irrelevant with the subject, and you'll see it I'm not for current-drive or Esa's view on things. I'm not on any amplifier view, amplifiers are not important thing here, at all.

I'm promoting understanding of the circuit impedance. I'm not promoting current drive nor voltage drive, they are both just part of the circuit impedance and affect acoustic outcome with their impedance together what ever else is there in the circuit. By circuit I mean circuit that consist of amplifier, a loudspeaker(driver) and possible passive network components in between. I'm promoting basic circuit analysis to analyze how backEMF voltage turns into current, which turns into acoustic domain. Hopefully impedance was mentioned enough times 😀
 
Last edited:
  • Like
Reactions: TNT
Your words saying current-drive (in other words elevated circuit impedance) does not affect driver distortion, when it does. It is exactly what you also do by introducing series inductor in the circuit with a voltage amp, you are elevating impedance in the circuit for backEMF. You are doing exactly this, and then on another sentence kind of deny it doesn't happen, twist and turn amplifiers into the mix, which makes very confusing message from you. BackEMF is distortion voltage source within the driver and it makes lots of current into low impedance circuit. Elevate the impedance, current is reduced, acoustic distortion is reduced.

To limit possibility misunderstanding I'll repeat something from in earlier post: the driver distortion mechnism, the backEMF, stays intact no matter what the circuit impedance is because it is physical property of the driver, and for this your words are true. I understand "reduced distortion" or "distortion reduction" as reduced distortion in acoustic domain, acoustic distortion is all that matters after all. Increasing circuit impedance prevents these backEMF driver distortion mechanism to turn into acoustic distortion. Distortion of the driver is voltage, the backEMF, and it needs to turn into current to show up in acoustic domain. Increasing circuit impedance reduces this current so less of the motor distortion ends up into the acoustic domain, to ear, to mic.

There are multiple distortion mechanisms with loudspeakers, its just few with the motor that are affected which appear in the backEMF, hysteresis and some other perhaps. I'm not too into it. But all we have to know its current that turns into acoustic sound. To reduce current we need to increase impedance, simple as that.

If you had a driver hooked to an amplifier and manipulated impedance in the circuit with a resistor or a coil and measure changes in acoustic distortion it would be mostly the driver distortion reduction showing up. Amplifier originating distortion would be minor, modern amplifiers have 0.001% or something.
 
Last edited:
Simplest example:
1kHz tone to driver. Nonlinearity within the driver motor, for example Le(x) or something, makes third order distortion component appear as backEMF voltage. It is distortion voltage varying at 3kHz, a distortion tone voltage amplifier inside the driver if you will.

But this is only in electronic domain and not yet in acoustic domain. The driver does not emit this 3rd order distortion voltage as audible sound unless the voltage turns into current, which would move voice coil. The distortion voltage turns into current in the circuit over the load impedance it has, impedance in series with the voice coil, including impedance of the voice coil itself of course. The current makes force moving voice coil and the distortion mechanism appears now as acoustic sound. And all of this happens simultaneously in real time.

The driver creates and emits its own distortion if circuit impedance is low. This has nothing to do with amplifier other than the amplifier completes the circuit with its output impedance. Current flows simultaneously on every spot of a circuit, if you think about it backEMF voltage makes current that immediately emits acoustically. It doesn't have to go through the amplifier first or anything like that, its all happening there within the driver in real time.

No matter what the circuit impedance is, the backEMF voltage distortion voltage would measure the same but we can adjust how much of a current it turns into, which is directly proportional to amount of the acoustic distortion you would measure with mic. The current and thus the acoustic distortion is reduced if you increase impedance in the circuit that is in series with voice coil at the 3kHz.

How to manipulate impedance in series with the voice coil? You could have a parallel notch filter in series to elevate circuit impedance at 3kHz. Or if you have current-drive amplifier whose high output impedance makes very high circuit impedance. If you had a voltage amplifier in there instead, whose output impedance is very very low, and no additional passive components to increase circuit impedance at 3kHz, you would have maximum amount of this distortion turn to acoustic domain. Or, if you have current amplifier to make high circuit impedance but then introduce a low impedance shunt circuit to EQ the system, the distortion current can flow again through the shunt. The amplifier is just an impedance for this distortion mechanism.

Homework: how long and thin cables, or short and thick cables would affect what? How about even or odd order low pass filter, how would it affect woofers acoustic distortion, or damping?
 
Last edited: