Hello Joe Rasmussen,
I can speculate what the inductor function is.
In terms of reduced distortion, why is the inductor there?
Thanks DT
Dear Mr. Rasmussen,
thank you for providing such a DIY project. I have read all 4.800 posts (took me around 15 Hours)
I am preparing to order parts and looking if I can make some changes to encloser.
Can you please advise if some tilt on front panel could be possible (that will not affect sound greatly):
1. 5° tilt on uper driver and 5° tilt on 3 drivers down, like on picture (all other encloser spec will remain the same as in project)
or
2. 5° tilt on complete front panel (all other encloser spec will remain the same as in project)
thank you very much.
thank you for providing such a DIY project. I have read all 4.800 posts (took me around 15 Hours)
I am preparing to order parts and looking if I can make some changes to encloser.
Can you please advise if some tilt on front panel could be possible (that will not affect sound greatly):
1. 5° tilt on uper driver and 5° tilt on 3 drivers down, like on picture (all other encloser spec will remain the same as in project)
or
2. 5° tilt on complete front panel (all other encloser spec will remain the same as in project)
thank you very much.
How many words do I have to type? A few short explanations or dig deeper? This is quite a subject.
And I might get into trouble airing stuff?
For one, the inductor does not work on its own, but it is part of a carefully adjusted LCR tuned to near 70 Hertz. Together with other LCR/Zobel, it equalises (EQ) the current, so that the impedance (Blue) looks like this below, very flat. But note also the electrical phase angle (Red):
If up have a current source amplifier with high output impedance, the response deviation is close to nil, and with high impedance tube amps, even less so. So being impervious to amplifiers and incredibly amplifier friendly, that alone would make it work.
But there is also more thinking behind it. The Elsinores were designed very much with a current in mind. The current of what? The speaker? That would seem the obvious answer, but it is in fact the current of the amplifier.
This I know will sound silly to some. But... ideally, the distortion on the voltage amplifier should be the same distortion we see on the current side of the amplifier, except hardly anybody looks there. If the load of the loudspeaker (let's say a simple near-perfect driver) would be totally resistive, then that would be true.
In reality, it is not. Indeed far from it. Here is a crazy, and some may say radical thought, that all amplifiers are really just different types of current delivery systems.
Should we measure, not just voltage distortion of the amplifier, but also the current distortion of the amplifier.
Nobody is doing that.
OK, might get into trouble, but here goes.
There really is no such thing as voltage drive,
Maybe a case can be given for current drive, but I can even think of reasons why that is not so. Indeed the idea of 'drive' or 'driving' loudspeakers can be seriously questioned. That word implies some kind of control, but I don't see it. It is somewhat more complex than that, and yet it can also be simplified into more fundamental basically correct explanations. One of the reasons why we should look upon this as current delivery rather than any drive is because the force factor that produces sound, only current and not voltage, that any deviation between that force factor and the current (of the amplifier) is distortion.
Are not all amplifiers not really just different "current delivery systems?"
That is the view I have adopted. But this is of course just the tip of the iceberg and in no way can I cover it all here. I can only point to some things. A fuller discussion paper is being prepared, but it will not be rushed, but hopefully this year something will come out in 'print' so to speak. And it will produce distortion measurements to back things up, but electrical and acoustic distortion measurements. Presented logically and supported by solid arguments.
Back to that inductor.
That inductor and LCR stabilise the current in the LF area and keep the current phase angle as low as possible. The array of four drivers reduces excursion, improves both sensitivity and efficiency (they are not the same), and reduces distortion at frequencies above the fundamental notes that create excursion.
Nobody listens to sine waves, but instruments that are tuned at different frequencies, as well as harmonics. As Purifi themselves have pointed out, this leads to distortions related to non-linear excursions (and non-linear force factor), but these also modulate the current of the amplifier, and the reason is simple.
As the force factor does not stay stable, the impedance of the driver is not stable with excursion. Ouch!
This does not just affect the bass, but distortion is the critical midrange. This is where the ULD drivers are intended to shine. But they also have less effect on any possible errant behaviour of the amplifier:
Impedance modulations = current modulations = Distortion.
Yes, the ULD is an improvement above the SB drivers. The latter does have one advantage, lower inductance than ULD. But not as stable with excursion as ULD. So the ULD has an inherent advantage.
External 'added' series inductance can reduce modulated distortions.
Using a single series inductor that is larger than the driver's non-linear inductance, that suits the SB drivers because the ratio is higher (the ULD also benefits from this 'trick'). If the inductor value is higher than the internal inductance, then the total inductance flattens out with excursion. This is how the SB drivers work and it works really well. Yet the ULD is already ahead because it is already stable to start with. Whichever driver is used, they both benefits and is part of the explanation why the Elsinore still sounds so good even with the SB drivers. The ULD are of course better and use them if you can afford them. But if not, the SB drivers are already, in this design, giving you plenty of the same benefits as the ULD option. So don't dither, just build them.
Ultimately this is also about benefitting the amplifiers as well. Will it produce lessen the distortion of the amplifier? Logic tells us it will. Distortion measurements will cement it. Already basic measurements have shown that a current source can reduce distortion. The general conclusion that current drive of a driver makes the driver more linear is not correct. The conclusion is understandable, but not right. Measure the current distortion (this requires a jig/setup) and you will see it there too. So current drive does not reduce the distortion of the drivers. Yes, distortion is definitely reduced, no question about that. But it is in fact the amplifier that is triggered into producing distortion and that shows up in acoustic distortion measurements. The driver distortion cannot really be reduced, but the total distortion can. Once again we are reminded that the driver quality is the key and that current drive cannot make a cheap driver or tweeter become better. Indeed the opposite happens. The bad drivers become even worse!
This may also help us to understand finally why amplifiers sound different. They absolutely do! But the driver that is triggering the amplifier to distort on the current side (and hence pollute the force factor), different amplifiers may produce different, and some worse, distortion profiles. Yes, bad amplifiers produce bad current distortion that directly messes up the current in the voice coil. We may be able to have objective facts to show the deniers.
Here is a thought for those with analysing capability:
Make an amplifier produce the same current at all frequencies.
Including LF.
Then the output impedance of the amplifier will have no effect on the frequency response and have no effect on the 'damping' in the bass. Indeed the so-called 'damping factor' of an amplifier becomes completely unnecessary. No, this is not an April 1st joke. It is a demonstrable fact. The Elsinores actually do this. I have an amplifier here with an output impedance of >200 Ohm and I can play that 40W amplifier through the Elsinores and there is no bass issues at all, the alignment stays put. In 1975, none other than Richard H. Small taught me a lesson: The bass alignment is entirely defined by the alignment and if it is a Butterworth alignment, then it will have the damping of a Butter worth alignment.
Some have said that so-called current drive cannot be done in the bass. Yes, it can. Make a sealed Q=0.707 Butterworth alignment, using an LCR to tune flat the Fc on the electrical side.
But nobody has done this commercially yet. But it can be done. I know, because I did it and it works great. Just supply the correct current versus frequency and it works. Note supply the correct current.
I will end here.
Cheers, Joe
PS: I recall Menno Vanderveen had a Hypex Class D amplifier (NCore?) that could be switched from near Zero, 3 Ohm and 18 Ohm output impedances. He did some presentations at ETF17 held in Denmark into a pair of speakers I believe belonged to Jan Didden. Pretty much all who heard it preferred 18 Ohm. I believe what they heard as an improvement was less current distortion from the amplifier because 18 Ohm series impedance would give a significant reduction in impedance/current modulations = less distortion heard. That's the key, not so-called current drive, more like better controlled current behaviour!
I hope nobody thinks any of the above is somehow heretical.
Last edited:
If you did this with the current Elsinore design and drivers, then the focus of the top three driver array will fall to a much shorter distance.
About amplifiers:
Amplifiers have power gain, could be voltage gain could be current gain or could be a combination of both. It is always an Ohm’s law thing, zero volts gets you zero amps and zero Watts.
Typically, normally TS/P’s are measured with a voltage source amplifier.
Start adding series resistance (the equivalent of adding amplifier output impedance) the apparent shape of the impedance curve will change. The shape of the frequency response curve will change. The shape of electrical phase curve will also change.
If you start adding inductance similar but different things will happen to the impedance curve, frequency response and the electrical phase curves. As you say this is textbook RCL kind of stuff.
In terms of reduced distortion because of an added inductor that will require some acoustic measurements to demonstrate.
Perhaps multi-tone and or IM tests.
Thanks DT
This is playing in the deep end of the pool.
Not a heretic.
Tilting by 5° starting at the edge of the waveguide changes the distance by approximately 10mm in my rough drawing (based on the MFC woofer).
At the crossover a quarter of a wavelength would be approximately 30mm (@2900Hz)...
Oh yes, Ohm's Law is very much at the heart of this topic and indeed an unstable impedance can modulate the current that flows through the voice coil when the amplifier is a voltage source. If the impedance changes (not stable), the current changes and you will have distortion. And yes, it is very much measurable.
But the concept of 'current distortion' is yet to catch on, but it will happen. My hope is that it will.
The voltage only creates a potential for current to flow.
But what exactly does the amplifier see? Here is a very basic concept where the two ends are the terminals of the speaker, what the amplifier sees as a load:
Part A is the DCR (Re) of the driver. It is largely constant with frequency. But Part B of the impedance, the "V' is not constant and varies with frequency.
The voltage only creates a potential for current to flow. BUT, how stable is that current?
Take a look at the impedance graph shown below, note that any impedance above Re is some form of back-EMF and that only at 0Hz, or DC, there will not be some kind of back-EMF:
Note that the impedance cannot drop below Re at any frequency and that there are three different types of back-EMF involved.
Why are we calling this "V" which sounds like a voltage?
Because any form of back-EMF is a voltage source in its own right.
The "F" in EMF, electromotive-Force, is a voltage.
This force acts against the current of the amplifier and becomes an impedance. One knowledgeable friend of mine, when I pointed this out, called in an anti-voltage that acts as an impedance. And yes, it acts as an impedance that has a value in Ohm.
This simple fact has gotten me in trouble in the past: A voltage can be an impedance and a value in Ohm.
Our impedance is not stable and hence the current of the amplifier is not stable.
So many things can cause this problem. The impedance can and does change with excursions. This is where Purifi has done their homework. But any irregularity will show up, and any aberrant behaviour will end up modulating that impedance plot we see above. Imagine that impedance plot line as being very fuzzy and you will visually have some idea. Motor deficiency, cone resonances, poor suspension performance, surrounds too (Purifi again), all these can bounce back into the system, they get reflected right back into the amplifier on the current side. How can we ignore this? Signs are that some at least are not ignoring it.
The above is clearly simplified, but realistic view.
If one digs even deeper, then take a look at this:
Note the grey area are equivalent, the graph on the left correlates to the one on the right.
What we have here is the same driver as the impedance graph earlier, but now we are looking at the frequency response. Actually two responses, one is with a voltage source and the other is a current source.
The above can be analysed deeply, but will try to keep it simple(?) here. The SPL output is directly related to the current and only current explains the difference in both responses.
I have three frequencies shown above, but let us just look at 3KHz below:
Using the variable Vre we can see that the difference SPL corresponds to both current 353mA as well as dissipation in the voice coil, since efficiency is well under 1%. The reference voltage is 6 Volts and the reference current is 1 Amp respectively. The 6 Ohm Re is the determining factor for choosing these two.
I know the above takes a bit of an effort to digest and get a feel for it. Remember when the current source is used, we would have 1 Amp current at all frequencies. The voltage "V" is still present, but it will not impede the current. When a voltage source is used, then "V" will act upon the current and at 3KHz it will reduce to 353mA which is -9dB relative to 6V/1A/6W.
It's also important to understand that 6V and 1 Amp were chosen as useful equivalences because the Re, or DC resistance of the voice coil is 6 Ohm. Hence without the presence of "V" the voice coil would dissipate 6 Watt under both voltage source and current source conditions.
I think I will leave it there. I hope you found it interesting. There is of course far more to it and will put in down in written form soon, but I won't be rushed to print, so to speak.
The above is clearly simplified and this is just an open door for more work.
If one digs even deeper, then take a look at this Vre can be used by an audio analyser to measure current distortion. We can now simultaneously measure both voltage distortion of the amplifier and current distortion of the amplifier.
What if it turns out that amplifiers turn out to have significantly different current distortion profiles, that would help explain why amplifiers DO sound different. Would it also be crazy if it also showed that transformer-coupled amplifiers have an advantage in this respect?
Maybe?
Cheers, Joe
Yep! And that would be problematic when it comes to the Elsinore design, which is very close to that crossover frequency.
Blue is the four driver combo, Red is the tweeter and Green is the sum of Blue and Red.
Here is another view:
Note that the drivers are not fighting each other in time. This means that the crossover circa 3KHz will perform well, some cleanly at -6dB (optimal) and near perfect driver integration.
One either get it right or wrong.
ULD:
The null is when the tweeter is connected in the wrong phase, the crossover frequency is revealed at 3KHz.
Last edited:
I fully agree that amplifiers can and do sound different.
Over the years we have migrated to solid state amplifiers with near zero output impedance. These amplifiers are near perfect voltage sources. They have near zero impact on the Quality Factor (total Q) of the connected speakers. I lean in preference to the newfangled solid-state amplifiers because they have less expensive watts and most often they are much quieter. Little or no hum, buzz or hiss.
If you connect a speaker to an amplifier you can measure SPL, FR, distortion, current, electrical phase, Total Q and other variables as you choose. Things can get complicated pretty quickly.
You can change enclosure volume and ports to adjust the output to your preference. You can add and or remove resistance, capacitance and inductance to equalize.
For this concept of Back EMF I am going to require a bit more convincing.
I think of Back EMF a voltage generated by a motor as it rolls to a stop (it is a stored energy type of thing) there is also an associated phase change to go along with it.
If you look at the phase plot of a woofer at resonance there is stored energy. There is also a large amount of phase change stuff going on as well.Here comes ELI the ICE man. For Inductance, voltage leads current. For Capacitance, current leads voltage.
Look at the typical woofer impedance curve; you have the classic stored energy impedance resonance peak and Back EMF going on. As the frequency increases you see increased impedance due to inductance. Also as you would expect you see a gradual phase shift due to the inductance of the voice coil. There is no back EMF due to resonance.
I fully agree that amplifiers can and do sound different.
Over the years we have migrated to solid state amplifiers with near zero output impedance. These amplifiers are near perfect voltage sources. They have near zero impact on the Quality Factor (total Q) of the connected speakers. I lean in preference to the newfangled solid-state amplifiers because they have less expensive watts and most often they are much quieter. Little or no hum, buzz or hiss.
If you connect a speaker to an amplifier you can measure SPL, FR, distortion, current, electrical phase, Total Q and other variables as you choose. Things can get complicated pretty quickly.
You can change enclosure volume and ports to adjust the output to your preference. You can add and or remove resistance, capacitance and inductance to equalize.
For this concept of Back EMF I am going to require a bit more convincing.
I think of Back EMF a voltage generated by a motor as it rolls to a stop (it is a stored energy type of thing) there is also an associated phase change to go along with it.
If you look at the phase plot of a woofer at resonance there is stored energy. There is also a large amount of phase change stuff going on as well.
Here comes ELI the ICE man. For Inductance, voltage leads current. For Capacitance, current leads voltage.
Look at the typical woofer impedance curve; you have the classic stored energy impedance resonance peak and Back EMF going on. As the frequency increases you see increased impedance due to inductance. Also as you would expect you see a gradual phase shift due to the inductance of the voice coil. There is no back EMF due to resonance.
Thanks DT
I am not too sure about voltage or current distortion either. In a few days I will be back at my bench to test a few things.
The ClioFW01 sometimes does that, it is just a measurement artifact. It can safely be ignored. Notice that the tweeter is reversed electrically, but the output under 10KHz is in phase and is timed to sum upwards. That way the crossover becomes very clean with near 100% vector summing in and around the crossover. This is how you get the stitch right between the midrange and tweeter drivers.
Keep in touch with me as we go on in the near future. Perhaps the whole writeup and actually distortion measurements, that is what is required.
I know that the idea of current distortion seems foreign.
But once you figure out what it is, that becomes a game changer in one's thinking and cannot be undone.
Right now, with tweeters and midrange, in active speaker systems, there are models on the market that use current sources (they call it current-drive, but I don't think that is a correct description) to supply current and not voltage to those drivers. The reduction is distortion at extremely critical frequencies, which can be 20dB easily. It is quite an audile improvement and it is absolutely measurable. So current distortion is very real, even if it is not yet widely understood. I am just one who is being very direct and pointing it out. Others already know because their designs, again active speakers, are doing exactly that. So if I speak to them, I am actually reinforcing something they have already experienced and they will say "that's a good point."
Take a good look at these three pages/links:
http://pmacura.cz/speaker_dist1.htm
http://pmacura.cz/speaker_dist2.htm
http://pmacura.cz/speaker_dist3.htm
Note that he (Pavel) refers to "speaker current" and that he has measured it.
It is actually "amplifier current" and not speaker current.
I gently tried to point out to him (Pavel) that this is incorrect and he didn't take it well.
Then in one of the comments, he says:
"This is a direct influence of speaker non-linear impedance resulting in an increase of current distortion."
BINGO!
We agree! So I am not the only one saying current distortion."
And what is the cause of this current distortion?
I think it is plain: This is current distortion is produced by the amplifier but triggered by the unstable impedance I referred to earlier.
An impedance graph is not a straight line it looks, rather it is fuzzy in nature, and hence the current, based on good ol' Ohm's Law, that current is also fuzzy and... that is current distortion.
An unstable impedance causes the amplifier to produce distortion. Ohm's Law is not only satisfied, it is the explanation!
Study the above R1 resistor which must be a small fraction of SP1, this way you are measuring the current of the amplifier. The voltage formed across R1 will be proportional to the current that the amplifier is producing. The key word here is proportional. The voltage of R1 can now be fed through an audio analyser and we can measure that voltage as proportional to the current of the amplifier, and we get current distortion measurement.
My method is a little more elaborate and produces greater accuracy:
A low distortion signal amplifier, 60x above, will mean we can match the voltage that appears across Re of the driver.
We need to measure both sides of the amplifier, both voltage and current.
Since the amplifier is producing both voltage and current, it must also be producing both voltage distortion and current distortion. Do you see the point?
As Pavel has shown, the use of a current source limits the ability of the load SP1 from modulating the current of the amplifier, and that is what we have achieved when we measure lower distortion around 15-120dB of distortion.
How could Pavel get a reduction of 20dB in 3rd order harmonic distortions when the ratio was too poor for that? If I may postulate for a moment and say, maybe if we connected a different and maybe even better amplifier, that the difference would be much less than 20dB? In fact, that different amplifiers would react differently even if the load (trigger) is not changed.
If that postulate proves true, then we may have a measurable way of separating the boys from the adults, when it comes to amplifier designs.
This would require a driver in a box that is not too perfect and will trigger events in the amplifiers, that are different from each other.
Also, so-called current-drive does not reduce distortion produced by the driver. That distortion will always be there. But a good driver that behaves properly will have two sets of benefits, 1) lower driver distortion because it is a better design, and 2) will trigger less current distortion in the amplifier, making the overall distortion we hear worse.
Clear as mud?
The current source amplifier refuses to produce current distortion because it locks in the current and goes a long way to stabilise the current in the voice coil.
The Purifi drivers do something else again to gain that reduction of distortion.
The Purifi drivers will have a much less fuzzy impedance!
Think about that for a moment. That means we get the same or at least similar improvement with their ULD drivers. We don't need a current source amplifier to get that reduction in current distortion because the ULD drivers trigger level is much lower.
And here is the key:
The above flat inductances are in series with the DC resistance (Re) of the driver. The impedance is much more stable than the average driver which has significant non-flat inductance. The flat inductance also indicates more linear force-factor linearity.
So with this driver, the impedance seen by the amplifier will be much less fuzzy and hence we get lower distortion from the amplifier.
This driver will get you so-called current-drive advantage with a more conventional voltage source, so-called voltage-drive.
I know that the above requires a bit to get one's head around, but there is a logic in there.
To me the key is that all amplifiers are "current delivery systems" and we need to understand what that current does and keep it clean as possible.
Some amplifiers will join in with worse distortion than other amplifiers, explaining why amplifiers DO sound different.
It all harkens back to the voice coil, the force factor is entirely related to the current. If that is the case, it is largely ignoring the voltage that we measure across the speaker terminals. And keep in mind, in modern overhung voice coils, only about 30% or less of the voice coil is inside that gap where the force is generated, but 100% of the current is always present in that 30% gap. So if that current gets corrupted by the amplifier, then we will hear.
But some amplifiers will be better than others.
Hope it is clearer.
Cheers, Joe
Hello Joe,
A first responder will speak to mechanism of injury, as in what kind of injury will be caused by running your car into a tree. So if you drive your car into a tree they will take you to the ER and X-Ray looking for broken bones.
If there is distortion coming from your speakers, first what kind of distortion is it (a topic for later), and what caused it?
Is the distortion caused by the amplifier? Is the distortion caused by the speaker? Or is the distortion caused by an interaction of both? Can a good amplifier help a bad speaker or even the other way around?
I will start with one most can agree with. You can install a woofer in a sealed enclosure and use the air inside the enclosure as a part of the suspension of the bass speaker. Yeah sure it is done frequently. The problem is that air does not make a perfect spring. Air only sort of follows the ideal gas laws. If the sealed enclosure is small enough and the Woofer has a long enough excursion the nonlinear air compression may well be the primary cause of distortion for the air suspension woofer. Air compression is the mechanism of injury in this example.
Joe, you posted a figure that shows a “sensing resistor”. If we measure the voltage across the sensing resistor using Ohms Law we can calculate the current. This same configuration with software is how we measure/calculate speaker impedance curves, phase curves and TS/P’s. What is it about measuring delta Volts across the sensing resistor and doing an FFT that tells us that we are measuring current distortion? It could be voltage distortion, it could be impedance distorition, it could be phase distortion, it could be delta “Q” distortion, it could be inductive distortion or construct another name for distortion that we like better.
We do not know what causes the distortion, calling it current distortion may be too quick and easy.
What we need to do is to isolate some the potential causes of distortion (variables) and make some measurements.
For grins here are some plots of a HDS 830875 6 ½ inch woofer made with a ApX 1701 test amplifier with sensing resistors and AP500 software. The distortion measurements are acoustic, made with an AP 426M16 microphone.
Thanks DT
Maybe, I think I have gone as far as I can on social media. If we could discuss it in person, that would be different, but here, this is just too difficult. A whiteboard, or lots of paper, then it becomes possible.
Maybe all I can suggest is to read my posts on the subject again because your last post brings up things that are not directly related to the topic.
You have something right there. Yes, it tells us we are measuring current distortion. But this is not an LF issue. We can tolerate some distortion at LF that we cannot in the mids. Think of mixed signals, LF and midrange. One to, excite the position of the driver/voice coil.
May I suggest you read some of the Blogs on Purifi's website. That would be good.
BTW, the point you make about a sealed box (LF) and air/spring, are all very valid, if that was the topic.
But this is about distortion that shows up in the midrange, not really LF. It is about the voice coil itself and the force factor being related to the current. If voltage, force factor, and current, if they all lined up, then we would all be happy. Alas, any corruption of the force factor and you have distortion.
You cannot separate the link between the force factor and current. Any deviation is bad. Current distortion is bad. It causes a fuzzy load.
It is the current that creates the force factor and that is what we end up listening to. We must analyse current and that means measurement. We must measure it.
So I am hoping one day that will become a common thing to do, and DIY guys could lead the way. The discussion has started...
Convert the current of the amplifier into an accurate proportional voltage. Measure the distortion of that voltage and you have measured the distortion of the current. The keyword here is proportional.
That is what that 0R1 current sense resistor does.
For example, Purifi uses a combination of 40Hz (excursion) and 1KHz (no excursion as such). You can certainly measure this on the voltage side and they do.
It is not easy to identify a problem and get it down to its simplest components so that it can be analysed. This is a difficult mental task. Clarity is everything and that means weeding out the things that seem to be related, but they get in the way to get to the heart of the problem.
You also have not grasped the idea that the impedance (graph) is fuzzy and why it occurs. This is crucial.
Nice clean impedance lines? As the driver processes music and multiple signals, some creating LF excursions, others higher in frequency, then our impedance is not that clean, it is fuzzy and the inverse of a fuzzy impedance is fuzzy current that corrupts the force factor. Fuzzy equals distortion.
Think of being locked up in a bank vault and only hours to breathe oxygen. Something similar is happening here. You corrupt the air you are breathing.
Of to bed.... I am feeling fuzzy.
Maybe all I can suggest is to read my posts on the subject again because your last post brings up things that are not directly related to the topic.
You have something right there. Yes, it tells us we are measuring current distortion. But this is not an LF issue. We can tolerate some distortion at LF that we cannot in the mids. Think of mixed signals, LF and midrange. One to, excite the position of the driver/voice coil.
May I suggest you read some of the Blogs on Purifi's website. That would be good.
BTW, the point you make about a sealed box (LF) and air/spring, are all very valid, if that was the topic.
But this is about distortion that shows up in the midrange, not really LF. It is about the voice coil itself and the force factor being related to the current. If voltage, force factor, and current, if they all lined up, then we would all be happy. Alas, any corruption of the force factor and you have distortion.
You cannot separate the link between the force factor and current. Any deviation is bad. Current distortion is bad. It causes a fuzzy load.
It is the current that creates the force factor and that is what we end up listening to. We must analyse current and that means measurement. We must measure it.
So I am hoping one day that will become a common thing to do, and DIY guys could lead the way. The discussion has started...
Convert the current of the amplifier into an accurate proportional voltage. Measure the distortion of that voltage and you have measured the distortion of the current. The keyword here is proportional.
That is what that 0R1 current sense resistor does.
For example, Purifi uses a combination of 40Hz (excursion) and 1KHz (no excursion as such). You can certainly measure this on the voltage side and they do.
It is not easy to identify a problem and get it down to its simplest components so that it can be analysed. This is a difficult mental task. Clarity is everything and that means weeding out the things that seem to be related, but they get in the way to get to the heart of the problem.
You also have not grasped the idea that the impedance (graph) is fuzzy and why it occurs. This is crucial.
Nice clean impedance lines? As the driver processes music and multiple signals, some creating LF excursions, others higher in frequency, then our impedance is not that clean, it is fuzzy and the inverse of a fuzzy impedance is fuzzy current that corrupts the force factor. Fuzzy equals distortion.
Think of being locked up in a bank vault and only hours to breathe oxygen. Something similar is happening here. You corrupt the air you are breathing.
Of to bed.... I am feeling fuzzy.
Maybe, I think I have gone as far as I can on social media. If we could discuss it in person, that would be different, but here, this is just too difficult. A whiteboard, or lots of paper, then it becomes possible.
Maybe all I can suggest is to read my posts on the subject again because your last post brings up things that are not directly related to the topic.
You have something right there. Yes, it tells us we are measuring current distortion. But this is not an LF issue. We can tolerate some distortion at LF that we cannot in the mids. Think of mixed signals, LF and midrange. One to, excite the position of the driver/voice coil.
May I suggest you read some of the Blogs on Purifi's website. That would be good.
BTW, the point you make about a sealed box (LF) and air/spring, are all very valid, if that was the topic.
But this is about distortion that shows up in the midrange, not really LF. It is about the voice coil itself and the force factor being related to the current. If voltage, force factor, and current, if they all lined up, then we would all be happy. Alas, any corruption of the force factor and you have distortion.
You cannot separate the link between the force factor and current. Any deviation is bad. Current distortion is bad. It causes a fuzzy load.
It is the current that creates the force factor and that is what we end up listening to. We must analyse current and that means measurement. We must measure it.
So I am hoping one day that will become a common thing to do, and DIY guys could lead the way. The discussion has started...
Convert the current of the amplifier into an accurate proportional voltage. Measure the distortion of that voltage and you have measured the distortion of the current. The keyword here is proportional.
That is what that 0R1 current sense resistor does.
For example, Purifi uses a combination of 40Hz (excursion) and 1KHz (no excursion as such). You can certainly measure this on the voltage side and they do.
It is not easy to identify a problem and get it down to its simplest components so that it can be analysed. This is a difficult mental task. Clarity is everything and that means weeding out the things that seem to be related, but they get in the way to get to the heart of the problem.
You also have not grasped the idea that the impedance (graph) is fuzzy and why it occurs. This is crucial.
Nice clean impedance lines? As the driver processes music and multiple signals, some creating LF excursions, others higher in frequency, then our impedance is not that clean, it is fuzzy and the inverse of a fuzzy impedance is fuzzy current that corrupts the force factor. Fuzzy equals distortion.
Think of being locked up in a bank vault and only hours to breathe oxygen. Something similar is happening here. You corrupt the air you are breathing.
Of to bed.... I am feeling fuzzy.
Hello Joe,
Before I get too deep into this stuff I will do some measurements. I need to locate a 6 1/2 inch driver to test that Purifi has not fixed yet. I will be back later.
The discussion about air pressure caused distortion was only an example of a known distortion mechanism not intended to be a new topic.
Seems that so far multiple distortion mechanisms are all being lumped in under one heading "current"
Yes I know of fuzzy impedance. I have measured many driver curves, Low frequency and mid frequency cone speakers tend to have rather un-fuzzy impedance curves. The drivers that have the most fuzzy impedance curves are compression drivers with horns or waveguides. See the attached plots.
Even in the bi-amped M2 speaker JBL uses capacitors with parallel and series resistors to flatten and smooth the impedance curve.
We are introducing too many other factors and not focus on the real issue: Distortion!
THIS IS ABOUT 20dB OF DISTORTION!
Yes, fuzzy impedance is known, but it is understanding how it comes about and that a fuzzy impedance means that the current of the amplifier is also fuzzy.
One thing we need to get away from:
The pervasive view is that the voltage of the amplifier is all that really matters and current is secondary. Wrong! It should be the other way around.
This applies when the amp is a voltage source. Not when it is a current source. Will explain further...
Nobody says that current is not necessary!
Yes, we have to put power into the speaker's voice coil. So the view of current as being secondary and that all that it needs is headroom and that the amplifier can supply enough current.
What comes first? Voltage or current?
Current is thought of as secondary and about whether you have headroom or not.
As long as we have enough current up our leave, everything will look after itself. As long as the distortion does not rise on the voltage side, everything is OK. That view is not helpful.
You see this all the time. We measure distortion on the voltage side and then we are only concerned about difficult loads and asking the question, can the amplifier produce enough current?
I think that view is not correct and even more, it is highly misleading. And if we don't get things right, then:
THIS IS ABOUT 20dB OF DISTORTION!
Yes, this amount of distortion is running under the radar.
A TRICK SOME USE:
One way to smooth out and suppress fuzzy impedance is to put parallel resistors across the driver terminals. Now you have reduced fuzzy distortion. I tell you a secret, Wilson loudspeakers do this. Linn loudspeakers have done this in the past and maybe still do. A number of loudspeaker manufacturers worldwide do this, but will not say anything publically. By causing the amplifier to produce less fuzzy distortion due to fuzzy distortion caused by the amplifier reacting to fuzzy impedance.
So there are funny things going on out there. Wilson uses some weird language to obscure what they are doing, they say that their crossover has less jitter and uses funny word games. This is not helpful. Incredibly obscure, but good advertising. But it is a sledgehammer approach and leads to difficult loads, ouch!
CURRENT IS NOT ABOUT HAVING ENOUGH HEADROOM.
The Elsinores reduce fuzzy current-related distortion by equalising the current of the amplifier.
No crude parallel resistors are used, but carefully tailored LCR (2) and Zobel (1) smooth out the impedance.
Yep, and it is so easily audible. Make the amplifier produce the same current at all frequencies, the current phase angle goes flat and there is a massive reduction in audible distortion.
THERE! YOU NOW KNOW A MAJOR 'SECRET' WHY SO MANY PEOPLE FEEL THAT THE ELSIONOES SOUND SO DIFFERENT.
The key is less distortion! It is not magic, it is targeted at lower distortion and then the music starts to sound right.
But less us get back to the topic of how current is looked at in the wrong way:
Only current can produce the sound we hear. So we must examine what the current does in order to explain what we hear.
NOTE THIS: The voltage only plays a role when the amplifier is a voltage source. This will always make up the majority of the amplifiers we use. The so-called current-drive is not going to take over the world. It will find a place in active loudspeakers and that is already happening. The rest of us will be listening to voltage sources. We need the reduction we would get in a current source but don't get in a voltage source. Most of us use voltage sources.
REPEAT: THIS IS ABOUT 20dB OF DISTORTION!
Voltage drive: 20dB of current distortion.
Current drive: That 20dB of distortion is suppressed.
Conclusion: We must reduce that 20dB of distortion when using a voltage source.
I am trying to make this as clear as possible. I want less distortion!
MY TRICK:
I don't use parallel resistors. I use current EQ, I equalise the current. But this is not the only way to do it. Use a waveguide the right way and you can reduce induced fuzzy current distortion further. But that is another topic for later. Waveguides can be used to reduce distortion, absolutely, if you know how.
AGAIN: THIS IS ABOUT 20dB OF DISTORTION!
Do you want the benefit of a 20dB reduction of distortion you hear?
I can't be clearer than that. We need to focus on this, we need to focus on how to do that with a voltage source. The problem does not exist if the amp was a current source. In theory, there cannot be fuzzy current distortion with a pure current source.
We want to get low distortion with voltage sources that otherwise would only be available with a current source!
Now before I go on, is the above understood?
Let us park here for a moment, establish that and we can move forward step by step. This is like a ladder, one step at a time.
Did I mention that this is about a potential 20dB reduction in distortion? Getting a bit sick of me saying it? Yet...
FOCUS ON THAT!
THIS IS ABOUT 20dB OF DISTORTION!
Yes, fuzzy impedance is known, but it is understanding how it comes about and that a fuzzy impedance means that the current of the amplifier is also fuzzy.
One thing we need to get away from:
The pervasive view is that the voltage of the amplifier is all that really matters and current is secondary. Wrong! It should be the other way around.
This applies when the amp is a voltage source. Not when it is a current source. Will explain further...
Nobody says that current is not necessary!
Yes, we have to put power into the speaker's voice coil. So the view of current as being secondary and that all that it needs is headroom and that the amplifier can supply enough current.
What comes first? Voltage or current?
Current is thought of as secondary and about whether you have headroom or not.
As long as we have enough current up our leave, everything will look after itself. As long as the distortion does not rise on the voltage side, everything is OK. That view is not helpful.
You see this all the time. We measure distortion on the voltage side and then we are only concerned about difficult loads and asking the question, can the amplifier produce enough current?
I think that view is not correct and even more, it is highly misleading. And if we don't get things right, then:
THIS IS ABOUT 20dB OF DISTORTION!
Yes, this amount of distortion is running under the radar.
A TRICK SOME USE:
One way to smooth out and suppress fuzzy impedance is to put parallel resistors across the driver terminals. Now you have reduced fuzzy distortion. I tell you a secret, Wilson loudspeakers do this. Linn loudspeakers have done this in the past and maybe still do. A number of loudspeaker manufacturers worldwide do this, but will not say anything publically. By causing the amplifier to produce less fuzzy distortion due to fuzzy distortion caused by the amplifier reacting to fuzzy impedance.
So there are funny things going on out there. Wilson uses some weird language to obscure what they are doing, they say that their crossover has less jitter and uses funny word games. This is not helpful. Incredibly obscure, but good advertising. But it is a sledgehammer approach and leads to difficult loads, ouch!
CURRENT IS NOT ABOUT HAVING ENOUGH HEADROOM.
The Elsinores reduce fuzzy current-related distortion by equalising the current of the amplifier.
No crude parallel resistors are used, but carefully tailored LCR (2) and Zobel (1) smooth out the impedance.
Yep, and it is so easily audible. Make the amplifier produce the same current at all frequencies, the current phase angle goes flat and there is a massive reduction in audible distortion.
THERE! YOU NOW KNOW A MAJOR 'SECRET' WHY SO MANY PEOPLE FEEL THAT THE ELSIONOES SOUND SO DIFFERENT.
The key is less distortion! It is not magic, it is targeted at lower distortion and then the music starts to sound right.
But less us get back to the topic of how current is looked at in the wrong way:
Only current can produce the sound we hear. So we must examine what the current does in order to explain what we hear.
NOTE THIS: The voltage only plays a role when the amplifier is a voltage source. This will always make up the majority of the amplifiers we use. The so-called current-drive is not going to take over the world. It will find a place in active loudspeakers and that is already happening. The rest of us will be listening to voltage sources. We need the reduction we would get in a current source but don't get in a voltage source. Most of us use voltage sources.
REPEAT: THIS IS ABOUT 20dB OF DISTORTION!
Voltage drive: 20dB of current distortion.
Current drive: That 20dB of distortion is suppressed.
Conclusion: We must reduce that 20dB of distortion when using a voltage source.
I am trying to make this as clear as possible. I want less distortion!
MY TRICK:
I don't use parallel resistors. I use current EQ, I equalise the current. But this is not the only way to do it. Use a waveguide the right way and you can reduce induced fuzzy current distortion further. But that is another topic for later. Waveguides can be used to reduce distortion, absolutely, if you know how.
AGAIN: THIS IS ABOUT 20dB OF DISTORTION!
Do you want the benefit of a 20dB reduction of distortion you hear?
I can't be clearer than that. We need to focus on this, we need to focus on how to do that with a voltage source. The problem does not exist if the amp was a current source. In theory, there cannot be fuzzy current distortion with a pure current source.
We want to get low distortion with voltage sources that otherwise would only be available with a current source!
Now before I go on, is the above understood?
Let us park here for a moment, establish that and we can move forward step by step. This is like a ladder, one step at a time.
Did I mention that this is about a potential 20dB reduction in distortion? Getting a bit sick of me saying it? Yet...
FOCUS ON THAT!
"THERE! YOU NOW KNOW A MAJOR 'SECRET' WHY SO MANY PEOPLE FEEL THAT THE ELSIONOES SOUND SO DIFFERENT.
The key is less distortion! It is not magic, it is targeted at lower distortion and then the music starts to sound right."'
Can you please present 2 measurements, with the mic untouched between the sweeps, where
Meas 1: speaker as is.
Meas 2: remove 2 LRC and 1 zobel (as per above)
Present normalised distorsion graphs - 1m mic distance, 90 dB.
(there will be FR differences but it can be ignored as the claim is that it is distorsion that makes the difference - there are settings in REW that normalise FR for distorsion graph.)
It would be very interesting to see - thank you!
//
The key is less distortion! It is not magic, it is targeted at lower distortion and then the music starts to sound right."'
Can you please present 2 measurements, with the mic untouched between the sweeps, where
Meas 1: speaker as is.
Meas 2: remove 2 LRC and 1 zobel (as per above)
Present normalised distorsion graphs - 1m mic distance, 90 dB.
(there will be FR differences but it can be ignored as the claim is that it is distorsion that makes the difference - there are settings in REW that normalise FR for distorsion graph.)
It would be very interesting to see - thank you!
//
The Elsinores are about low distortion, they sound that way. Every detail about them is about targeting as to what kind of distortion(s) that matters. On this there may be differences of opinion in some minds, then so be it. I am confident in what I target.
The type of distortion I want to reduce is similar to those heard when current source is used, such as in fullrange drivers and in active loudspeakers on the midrange and treble. Once you have cottoned on to that sound, you recognise and like it.
Except I want it when the amp is a voltage source and not a current source. A fuzzy impedance creating fuzzy current is suppressed in a current source. We need to develop ways to suppress it and get similar benefits with voltage sources.
In time, in the context of explaining the above, I will produce more than amble distortion measurements. These will not be your regular THD that does not show the more important things up, but mixed signals combining LF and mid frequencies. Both acoustic and electrical (electrical here means measuring bot voltage AND current). The correlation will be obvious and easily compared. Excursions excite current related distortions.
I don't have a REW setup. I am sure that it will be useful, but the usual measurements I am sure would be interesting, but not the type that will show force factor relative to current distortions.
Again, as we are dealing with Purifi drivers, may I recommend to you read Purifi's blogs.
The type of distortion I want to reduce is similar to those heard when current source is used, such as in fullrange drivers and in active loudspeakers on the midrange and treble. Once you have cottoned on to that sound, you recognise and like it.
Except I want it when the amp is a voltage source and not a current source. A fuzzy impedance creating fuzzy current is suppressed in a current source. We need to develop ways to suppress it and get similar benefits with voltage sources.
In time, in the context of explaining the above, I will produce more than amble distortion measurements. These will not be your regular THD that does not show the more important things up, but mixed signals combining LF and mid frequencies. Both acoustic and electrical (electrical here means measuring bot voltage AND current). The correlation will be obvious and easily compared. Excursions excite current related distortions.
I don't have a REW setup. I am sure that it will be useful, but the usual measurements I am sure would be interesting, but not the type that will show force factor relative to current distortions.
Again, as we are dealing with Purifi drivers, may I recommend to you read Purifi's blogs.