Hi Lars, We soon should have two ULD versions of the Elsinores up and running and reported on here. I can't wait. I hope it might generate more orders for you or through your dealers.
I reckon using eight Purifi drivers in the Elsinores, the question is which other DIY speaker out there is more ambitious? I do hope you get a chance to listen to them sometime.
the Elsinore Mk6 "ULD" have been working here for around three months and the tweaks to the crossover are working nice and settled upon. Those who have heard them have been full of praise. No doubt about the perception of very low distortion, effortlessness and remarks about stunning clarity, from bottom to top. If you ever do make a tweeter, consider it with an integrated waveguide, when implemented right can get amazing low distortion along with what you may apply.
I had Morris Swift here (a regular visitor) from manufacturer company Serhan Swift and he did an estimate and said that commercially done, we are talking $35.000 at least in the shops. He said it would be a sin not to manufacture them. Hmmm...
Cheers, Joe
I reckon using eight Purifi drivers in the Elsinores, the question is which other DIY speaker out there is more ambitious? I do hope you get a chance to listen to them sometime.
the Elsinore Mk6 "ULD" have been working here for around three months and the tweaks to the crossover are working nice and settled upon. Those who have heard them have been full of praise. No doubt about the perception of very low distortion, effortlessness and remarks about stunning clarity, from bottom to top. If you ever do make a tweeter, consider it with an integrated waveguide, when implemented right can get amazing low distortion along with what you may apply.
I had Morris Swift here (a regular visitor) from manufacturer company Serhan Swift and he did an estimate and said that commercially done, we are talking $35.000 at least in the shops. He said it would be a sin not to manufacture them. Hmmm...
Cheers, Joe
What's the voltage gain of the input stage? And is the second input of the LTP connected to anything?It's good for a cheap amp, the parts are cheap. But I am not sure if I would use it with the ULD version. But it is fun to build. The IC sounds better for it and I built it to prove that the Elsinores work with a current source amp with >200 Ohm o/put impedance.
I am fairly sure that the Pass F2J would surpass it though.
The Elsinores would work with an impedance up to infinity.
View attachment 1079635
Green is voltage source and Red is current source. With the latter, the enormous power handling at LF will only be slightly less. I don't you will even notice it.
But what I am listening to looks something like this, where the EL34 output tubes only have 1.5x gain, so almost running the tubes at unity gain:
View attachment 1079634
Bet you won't see a circuit like that every day. I would put this amp up against pretty much anything. The input converts voltage to current, which is the current that drives the Power Fet which is 'grounded' signal-wise (+35V). The non-linear Gate capacitance is suppressed because there is no signal on the gate, very wide bandwidth. I had somebody at Power Electronics in Segunda California (part of Tesla and SpaceX), computer model the circuit. It is phenomenally linear and low distortion and no feedback is required. I wish everybody here could hear this amp into a pair of Elsinores.
I am really forward to you finishing the ULD version.
Being a transconductance amplifier, it does not have voltage gain in the usual sense. The voltage gain has to be referenced to the load.
Take a look below and on the right it says 'STRUCTURE' which is a reference to gain structure:
The gain at 4 Ohm load is 20.3dB (10x) and at 8 Ohm it is 26.2dB (20x). This is with 100mV input.
Being virtually flat 7 Ohm, the ULD version will have a gain circa 18x or 25.1dB.
The MFC and NRX are flat 6 Ohm and hence 15.5x or 23.8dB.
(The ideal transconducatce amp will have a voltage gain of infinity when no load and that is why that 470R 1W resistor with Zobel is fitted.)
Let's have a bit more fun with numbers:
The voltage sensitivity of ULD and MFC versions are virtually the same.
Let us concervatively reference that to 90dB SPL with 2.83V @1KHz. The nominal 1 Watt (except it isn't, but that's another story).
MFC: It takes 472mA of current through the voice coil to produce 90dB SPL @ 1 Meter. Do the maths.
ULD: It takes 404mA of current through the voice coil to produce 90dB SPL @ 1 Meter. Again, do the maths.
So 404mA versus 472mA and that mean the ULD are more efficient than the MFC, but they have the same voltage sensitity.
The above illustrates something to me, that there really is no such thing as voltage-drive or current-drive. Here is why:
With the ULD version, it will always be 404mA to produce 90dB SPL. That will never change.
With the MFC version, it will always be 472mA to produce 90dB SPL. That will never change.
It does not matter if it is a voltage source or current source - all amplfiers are current delivery systems.
The two 'drives' are just diffferent ways to deliver current. If the current reaches 404mA, you get 90dB SPL, every time.
I know, to others it might not be as obvious as it is to me.
Goodnight, Joe
PS: Here is another curious thought: We know that tube amps with high output impedances produce a change in frequency response with loudspeakers with non-flat impedances. With a current source, this effect will be maximum, the response will vary wildly. Many blame the the speaker for causing the non-flat response deviation. But a current source demonstrates something else. That the amplfiers gain is changed by the load.
So if using a tube amp with several Ohm output impedance, we say that the speaker is caused by what? Partly it is because the speaker as changed the gain of the amplifier.
Ergo...
When using a voltage source, it has no control over the current. The current can get bossed around by the speaker load.
When using a current source, it has no control over the voltage. Now the voltage can get bossed around by the speaker load.
That is how the speaker can affect the voltage gain when using a current source, but not when using a voltage source.
"Thinking current is hard" to paraphrase Menno Vanderveen. And he should know because he designs transformers.
I better get some slep, 'night!
OK, so the second input of the amp is connected to a current sense resistor tied between the speaker negative and ground. That makes sense, and also explains the overall gain of the amp/GNFB structure. I was under the impression there was no GNFB loop so the output voltage would be controlled by the load on the cascode FETs (x1.5 voltage gain of the output tubes). I am still curious what the open loop voltage gain is.
It is interesting the way sensitivity flips with a current source amp. The higher impedance speaker becomes more sensitive. You also capture the speaker in the global feedback loop...
One more thought: the "cheap current drive hack" of putting a resistor in series with the speaker is already implemented on a lot of tweeters in the crossover circuit to match sensitivity to woofers...
If Volts are English and Amps are Chinese then Watts are Greek to me. (joke)
I am trying to tighten up my thinking on all of this. Much of it seems to be fuzzy.
A transconductance amplifier is a Voltage Controlled Current Source.
If I can write it in MatLab code I have the definitions tight and the variable relationships correct.
Here is a good start to understanding Transconductance Amplifiers:
https://www.mathworks.com/help/physmod/sps/ref/operationaltransconductanceamplifier.html
I have been trying hard not to say "transconductance amplifier", I guess now that the cat is out of the bag I will.
Transconductance amplifiers output voltage and current with a variable output resistance. The variable output resistance holds the output current constant. This variable output resistance upends several things like Frequency Response, Quality Factor and damping factor.
No wonder Transconductance Amplifiers and Voltage Amplifiers sound different.
Looking foreword to seeing the measurements of the reference Vifa 6-1/2" driver. I took a look, I have a couple of them on the shelf.
Thanks DT
Last edited:
You have got the general idea. The open loop gain will still be what the LM3875T is rated at, that just becomes a maximum because there will be a voltage without any load and the voltage will go through the roof until it reaches the max rated open loop gain of the amp. At least that is how I see it. But it is rather academic. That is why that 470R 1W resistor has to be there, to put a ceiling on it.
Ahah! This is a pet subject of mine. Big time!
Current drive of tweeter when the amp is a voltage source?
YES!
In an earlier post yesterday, I mentioned I like the use of waveguides to lower distortion. This 'hack' works along with that series resistance 'hack' you mentioned.
The number one reason to use a waveguide? Distortion!
Less of it. Below, this is 15 degrees off-axis tweeter with waveguide, not the rise or 'waveguide gain' below 10KHz:
Calibrated 2.83V @ 1M.
Never mind that rise below 10KHz. You want that waveguide 'gain' and use it. So the above is pretty much what you want to see. Yes, this example is from the Elsinore tweeter/waveguide.
This is with 2.83V, and you can see it peaks around 97dB and the waveguide gain is around +5dB.
Now you can use a much smaller series capacitor in series with your matching sensitivity resistor. The ideal crossover point is 3KHz, in other words, don't use the waveguide to lower the crossover frequency (that will have some benefits), but I would rather use the waveguide to reduce distortion.
A few years ago I visited SB Acoustics in Herning, Denmark. Their designer Ulrik Schmidt was kind enough to spend three hours with this Danish vagabond from Australia.
But during the discussion, he really did floor me when he said that in his overall experience, with most drivers available, the natural crossover frequency is 3KHz. I told him how much it pleased me to hear him say that. Yes, they make tweeters that can be crossed over much lower than 3KHz, but if there is a market for them, you have to make them.So coming back to the waveguide, the smaller value series capacitor now required, will equalise the response flat. But it will also do more, it will be down -6dB at 3KHz. Do the math, that is an 11dB reduction at the crossover frequency.
Oh yes, that is a major reduction mechanism in distortion, because only a fraction of the power is now required for a flat system response at the crossover. But there is something more:
Now think about it, the value of the cap required is now around 1/3rd of normal. I use 1.8uF in the Elsinores, not 5-6-7uF. The reactance of that cap is tripled, around 30 Ohm with an 8 Ohm tweeter. That is a lot more than any series resistor you might add to match sensitivity.
I also make sure the current phase angle (of the amplifier) is zero (current EQ). Now around the critical crossover and the immediate frequencies below it, we effectively have something that approaches or mimics a pseudo-current source.
Trust me on this, it sounds really good. Below is the final result, note the Blue trace, which is our waveguide tweeter:
The reverse-phase test shows the 3KHz crossover point.
The tweeter is electrical first order near the crossover, but not a Butterworth and sum at -6dB and not -3dB. It gradually becomes very high order due to a carefully tweaked Q of an LC notch filter L3/C3 and looks like this part of the Elsinore crossover below:
With the chosen tweeter that is made by Scan-Speak, it was first designed for the Australian market (with input from our Russell Storey down here). It was initially called the Peerless HDS tweeter, but it was always made by Scan-Speak. Hence now, due to politics, it disappeared as Peerless HDS and is now the Scan-Speak "Discovery" Tweeter D2608/913000.
The waveguide is a conical design that does not suffer high-order modes, which give it that horn sound. Not here. It also has other advantages in how it interacts with the room. Very few negatives.
I have a friend in New Zealand who rebuilds electrostatic drivers. He says the Elsinore tweeter matches the ultra-low upper mid and treble distortion he craves. Yes, he has got the Elsinore NRX version and is seriously looking at the ULD version. I believe he is sure to do it.
OK, my waveguide rant is over... take a
break.I believe this is the first time I have described the Elsinore high-pass design in this detail. High-pass filters are potentially more critical than low-pass filters when it comes to distortion. Take a good look at that LC notch filter. The series cap is detrimental to damping and that LC restores it.
Cheers, Joe
hi Joe, congratulations with the design! I would love to hear and see them - give a shout if you ever get near Roskilde (with or without your speakers). 8 woofers should be amazing. just taking the step to our new 8" is quite shocking. We have a tweeter in the works and we plan for a variety of integral waveguides plus the option designing your own.
Cheers,
Lars
Interesting on the speaker side, I appreciate the depth. As for the OLG, I was asking what the OLG of Beyond Triode is.
Hi Lars, probably November next year when ETF 2023 will be on in the Netherlands for its second year. Kurt Steffensen we regard as the Founder as he was the first host when it was called the Aarhus Triode Festival, and we were there and put our hands up to call it the European Triode Festival, the late Allen Wright, the Christian Rintelen, Joe Roberts, Guido Tent, Emile Sprenger and more. It makes Kurt the Founder and the rest of us co-founders. Since then in 2002 I have tried to be there when I can, but it is a long road to travel and it was 2018 the last time, then Covid struck. Now I am looking forward to next year. BTW, the motorway that you are near, if you head into KBH and turn right at Hvidovrevej and the first apartments on your right is where my brother Flemming lives. My sister lives in Hellebæk, not far from Helsingør (for others, that is Elsinore in English).
Hopefully, sometime before then, somebody in the larger KBH area or reasonably near will build Elsinore ULD? We can only wait and see.
Re waveguide and tweeters, I note manufacturers of speaker systems are using them increasingly, but availability for DIYers is still limited. The Morel CAT378 is one that I have used, it is none-too-shabby, it's OK and Devore uses them in their not-so-cheap speakers. Scan-speak makes one but does not have a lot of that 'waveguide gain' I look for. But at least they are both 6-8 Ohm and above 90dB sensitivity. The SB Sartori TW29BNWG-8 Beryllium looks interesting, even if I tend to be a soft/silk sort of guy. I am hoping you are working on will be 6-8 Ohm and well above 90dB. It can be 4 Ohm, but then the voltage sensitivity needs to be incredibly high. That means you can add series resistance and make it work as 6-8 Ohm that way. So I will be definitely be looking very at what you are coming up with. I know that it will be low distortion, that is your track record.
The kind of waveguide I prefer is the straight conical kind. The sides are 45° so that the other opposite side of the waveguides there are no reflections coming back because they are 90° perpendicular. So no HOM and no horn coluration.
D2608/913000:
Response Red 0° on-axis, Green 15° off-axis, and Blue 30° off axis.
The best response should just be a bit off-axis:
The light blue is the EQ line, adjust series cap and also we can see the 92dB sensitivity with 8 Ohm nominal impedance. Anything above that is the 'waveguide gain.'
So now you have a good idea of what I am looking for.
As regards an 8-incher, I shall certainly be looking out for that one.
Cheers, Joe
Joe,
Thanks for posting the comments and dimensions of your Tweeter with Waveguide. The Frequency Response looks like the assembly has a mass-break point starting at about 6kHz and gently rolls off from there.
Seems like fabric dome tweeters with waveguides are merging in the continuum with the current crop of Compression Drivers and attached shallow gentle sloping Waveguides. (They are no longer horns.)
Thanks DT
Thanks for posting the comments and dimensions of your Tweeter with Waveguide. The Frequency Response looks like the assembly has a mass-break point starting at about 6kHz and gently rolls off from there.
Seems like fabric dome tweeters with waveguides are merging in the continuum with the current crop of Compression Drivers and attached shallow gentle sloping Waveguides. (They are no longer horns.)
Thanks DT
Folks:
This may not be entirely appropriate, but it seemed worth mentioning here that I've placed an advertisement for a pair of professionally-manufactured Elsinore Mk 6 cabinets in the Swap Meet forum.
For anyone who is seriously considering building Elsinores but doesn't have the time or equipment needed to build the cabinets, this is a rare opportunity. And I'll sweeten the offer a little: if you're interested in the cabinets but are hesitant to commit because you haven't heard these remarkable speakers, you're welcome to audition my finished pair of Elsinores. The pair for sale is located in Northeast Philadelphia, PA (USA). No shipping -- local sales only. I am located about 40 minutes away. Northeast Philadelphia is about 2 hours from both NYC and Washington, D.C.
Regards.
This may not be entirely appropriate, but it seemed worth mentioning here that I've placed an advertisement for a pair of professionally-manufactured Elsinore Mk 6 cabinets in the Swap Meet forum.
For anyone who is seriously considering building Elsinores but doesn't have the time or equipment needed to build the cabinets, this is a rare opportunity. And I'll sweeten the offer a little: if you're interested in the cabinets but are hesitant to commit because you haven't heard these remarkable speakers, you're welcome to audition my finished pair of Elsinores. The pair for sale is located in Northeast Philadelphia, PA (USA). No shipping -- local sales only. I am located about 40 minutes away. Northeast Philadelphia is about 2 hours from both NYC and Washington, D.C.
Regards.
I did not expect this. Four pairs of crates for gluing landed in my tiny garage. I'm trying to control this chaos
Good to see more work happening in the workshop.
Th chaos looks well contained.
Good luck with build!Current amplification, but the delivery is unbelievable. The first, the best and so on
Yes, rather a lot of faffle about it.
Maybe this link will better work...otherwise https://avantgarde-acoustic.de/en/itron/
A little teaser of mine construction, front is 20 mm Panzerholz, body 25 mm beech multiplex. The socket is 15 cm high, that sounds the best for me. Everything ready, just the cabling has to be done. I intend to use 1 mm² LIFY cable, nothing very fancy or exotic. next weekend will see the drivers mounted and first musik.
Joe I have read this post more than a few times letting you meaning soak in soak in.
By the looks of it you have selected a reference driver to test and illustrate driver distortion.
It took a couple of times through identify sort out a bit the "distortion" that you were talking about.
The standard Log sweep (chirp) samples one frequency at a time and puts out distortion measurements in terms of Harmonic Distortions 2nd, 3rd harmonics and on up.
After reading your comments about the new QA403 analyzer and the lack of IM tests in the software it became more clear that you were also addressing IM distortions.
Reading the Purifi stuff I see IM test data included on the ULD driver data sheets. In the blogs I see mention of Amplitude Distortion and others
Joe, tell us more about the Distortion(s) that you expect to find in your driver distortion tests. Especially tell us where you expect to see "in the order of 20dB's" Distortion. I am guessing that it is not Harmonic Distortion you are talking about.
Thanks DT
Last edited:
Hi DT
I appreciate the effort. Thanks for re-reading.
I want to make a correction going back to #4782 where I made a mistake, I referred to FM distortion (Dopler) when it should have been AM (amplitude) distortion. I am scratching my head why I got that back-to-front. We are definitely talking about AM, amplitude modulated distortion. And this relates to physical excursions, which are of course amplitude.
Purifi drivers are stable in relationship to the load that they put on a voltage source amplifier, particularly when we have mixed LF with 1KHz for example.
My take: The Purifi ULP drivers produce less "current distortion" when the amplifier is a voltage source. So more explaining, so here goes...
Purifi has called referred to as "impedance modulation."
I then pointed out that impedance is related to current. So impedance modulations have to be current modulations. But what is being modulated?
The amplifier!
The current of the amplifier is being modulated. And that is not good.
That means the load that the speaker puts on the amplifier can induce current distortion and as a madman, I say YES!
This is what happens when the amplifier is a voltage source. But it does not when it is a current source. Then let us use a current source and all will be right? That is what some want you to believe, but the world is not getting converted to current-drive. Too impractical.
And maybe you don't have to anyway.
Just come up with different/better solutions that work with voltage sources.
I noted in a post here I note that Lars now called it "impedance/current modulation" and not just impedance modulations. Change impedance, and you change the current... of the amplifier.
The stock Elsinores already tackle FFM induced distortion by massively reducing LF excursions for a target SPL, sound pressure level. Then using the Purifi drivers, we can further improve and take the idea to the max. The Purifi drivers are even flatter FFM in the Elsinores for a target SPL, and they now have massive FFM headroom.
And that's the idea behind why I wanted to go for the Purifi rather than the suggested SB Sartori drivers. Were the guys willing to bear that cost and half a dozen put their hands up immediately and more to come.
Yes, you are right, this is IMD related, but have you read the Purifi blogs where LF (excursion, 40Hz from memory) and 1KHz mixed stimuli are used?
Please read the link below, it is about IMD/AM/FFM relationships, I think it will help put things in place.
Link: Low frequency harmonic distortion is almost inaudible. So what’s the point of low distortion drivers?
Yes, that is the crux of the matter, I will try to explain it as clearly as possible.
1. A voltage source can only control (regulate) the voltage, but in order to do that, it has to relinquish control over the current.
This is why the amplifier has to have a very low impedance and for the ideal voltage source that would be zero Ohm.
2. A current source can only control (regulate) the current, but in order to do that, it has to relinquish control over the voltage.
If point 1) is hard to understand, then 2) is perhaps even harder?
The ideal current source needs to have a very high impedance, and for the ideal current that would be infinity Ohm.
Do you see the whole juxtaposition? Things are totally opposites in a fascinating way.
Indeed all amplifiers can be modeled this way:
Yes, all amplifiers can be modeled this way, as a voltage source followed by a series impedance.
The voltage source, it will be zero.
The current source, it will be infinity.
With tube amplifiers, it will have a value typically up to 5 Ohm, sometimes more.
Knowing the above, can we see where there is a mechanism by which a current source will have a 20dB (or significant) distortion advantage?
Yes we can. With a current source, the speaker load is not able to boss around the current like it can when a voltage source is connected to or speaker load.
Current sources produce less "current distortion."
The driver cannot boss around the current of the amplifier and hence lower current distortion.
Just as Purifi wants FFM to the forefront in spec, I want current distortion to become mainstream too.
When using current, the distortion shows up on the voltage side and not the current side, and causes virtually zero issues.
I know that this is not easy to grapple with and get one's head around.
The dynamic driver only reproduces the resultant current that comes from the amplifier. If that current gets corrupted, then you will hear distortion. This is what current-drive guys are saying they are hearing.
We need to come up with better solutions when using a voltage source. They are here to stay.
Let me know if you have followed so far?
I want to find ways of reducing distortion when connecting a speaker (current) to a voltage source.
We want the same potential low distortion with voltage sources. Make the loudspeaker behave better!
Cheers, Joe
PS: In his book on current-drive, Esa mentions current distortion in an indirect but very interesting way:
"The loudspeaker circuit operating on voltage drive exhibits a feedback effect where EMF deriving from voice coil motion summates directly with the voltage supplied to the driver. so that the resulting current is a mixture of the original signal and a spurious signal corrupted by the speaker's own mechanical, electrical, and acoustic properties and circulated in the feedback process."
(Yes, I was among the earliest to get his book and have a three-way discussion with him and John Curl.)
Feedback here is feeding back because the amplifier is distorting and corrupting the current. Not like feedback in an amplifier. The messy aspect is that it circulates and smears out in time. He uses complicated language, but once the mechanism he describes is understood, it changes everything and cannot be unlearned. It also highlights another aspect of current distortion, it is time smeared.
The effect he is discussing, it can be measured and what else can you call it but "current distortion."
Note the EMF? That it is a voltage? Remember this:
All the bad stuff mentioned by Esa above, ends up in Part B above, causing it to moderate the entire impedance because it is in series with Re and causes corrupt current. So Part A + Part B is not a stable impedance and hence unstable current that reverberates around like 'feeding back and forth.'
I appreciate the effort. Thanks for re-reading.
I want to make a correction going back to #4782 where I made a mistake, I referred to FM distortion (Dopler) when it should have been AM (amplitude) distortion. I am scratching my head why I got that back-to-front. We are definitely talking about AM, amplitude modulated distortion. And this relates to physical excursions, which are of course amplitude.
Purifi drivers are stable in relationship to the load that they put on a voltage source amplifier, particularly when we have mixed LF with 1KHz for example.
My take: The Purifi ULP drivers produce less "current distortion" when the amplifier is a voltage source. So more explaining, so here goes...
Purifi has called referred to as "impedance modulation."
I then pointed out that impedance is related to current. So impedance modulations have to be current modulations. But what is being modulated?
The amplifier!
The current of the amplifier is being modulated. And that is not good.
That means the load that the speaker puts on the amplifier can induce current distortion and as a madman, I say YES!
This is what happens when the amplifier is a voltage source. But it does not when it is a current source. Then let us use a current source and all will be right? That is what some want you to believe, but the world is not getting converted to current-drive. Too impractical.
And maybe you don't have to anyway.
Just come up with different/better solutions that work with voltage sources.
I noted in a post here I note that Lars now called it "impedance/current modulation" and not just impedance modulations. Change impedance, and you change the current... of the amplifier.
The stock Elsinores already tackle FFM induced distortion by massively reducing LF excursions for a target SPL, sound pressure level. Then using the Purifi drivers, we can further improve and take the idea to the max. The Purifi drivers are even flatter FFM in the Elsinores for a target SPL, and they now have massive FFM headroom.
And that's the idea behind why I wanted to go for the Purifi rather than the suggested SB Sartori drivers. Were the guys willing to bear that cost and half a dozen put their hands up immediately and more to come.
Yes, you are right, this is IMD related, but have you read the Purifi blogs where LF (excursion, 40Hz from memory) and 1KHz mixed stimuli are used?
Please read the link below, it is about IMD/AM/FFM relationships, I think it will help put things in place.
Link: Low frequency harmonic distortion is almost inaudible. So what’s the point of low distortion drivers?
Yes, that is the crux of the matter, I will try to explain it as clearly as possible.
1. A voltage source can only control (regulate) the voltage, but in order to do that, it has to relinquish control over the current.
This is why the amplifier has to have a very low impedance and for the ideal voltage source that would be zero Ohm.
2. A current source can only control (regulate) the current, but in order to do that, it has to relinquish control over the voltage.
If point 1) is hard to understand, then 2) is perhaps even harder?
The ideal current source needs to have a very high impedance, and for the ideal current that would be infinity Ohm.
Do you see the whole juxtaposition? Things are totally opposites in a fascinating way.
Indeed all amplifiers can be modeled this way:
Yes, all amplifiers can be modeled this way, as a voltage source followed by a series impedance.
The voltage source, it will be zero.
The current source, it will be infinity.
With tube amplifiers, it will have a value typically up to 5 Ohm, sometimes more.
Knowing the above, can we see where there is a mechanism by which a current source will have a 20dB (or significant) distortion advantage?
Yes we can. With a current source, the speaker load is not able to boss around the current like it can when a voltage source is connected to or speaker load.
Current sources produce less "current distortion."
The driver cannot boss around the current of the amplifier and hence lower current distortion.
Just as Purifi wants FFM to the forefront in spec, I want current distortion to become mainstream too.
When using current, the distortion shows up on the voltage side and not the current side, and causes virtually zero issues.
I know that this is not easy to grapple with and get one's head around.
The dynamic driver only reproduces the resultant current that comes from the amplifier. If that current gets corrupted, then you will hear distortion. This is what current-drive guys are saying they are hearing.
We need to come up with better solutions when using a voltage source. They are here to stay.
Let me know if you have followed so far?
I want to find ways of reducing distortion when connecting a speaker (current) to a voltage source.
We want the same potential low distortion with voltage sources. Make the loudspeaker behave better!
Cheers, Joe
PS: In his book on current-drive, Esa mentions current distortion in an indirect but very interesting way:
"The loudspeaker circuit operating on voltage drive exhibits a feedback effect where EMF deriving from voice coil motion summates directly with the voltage supplied to the driver. so that the resulting current is a mixture of the original signal and a spurious signal corrupted by the speaker's own mechanical, electrical, and acoustic properties and circulated in the feedback process."
Feedback here is feeding back because the amplifier is distorting and corrupting the current. Not like feedback in an amplifier. The messy aspect is that it circulates and smears out in time. He uses complicated language, but once the mechanism he describes is understood, it changes everything and cannot be unlearned. It also highlights another aspect of current distortion, it is time smeared.
The effect he is discussing, it can be measured and what else can you call it but "current distortion."
Note the EMF? That it is a voltage? Remember this:
All the bad stuff mentioned by Esa above, ends up in Part B above, causing it to moderate the entire impedance because it is in series with Re and causes corrupt current. So Part A + Part B is not a stable impedance and hence unstable current that reverberates around like 'feeding back and forth.'