I am reading Esa Meriläinens book "current driving of loudspeakers" . He uses voltage controlled current sources (transconductance amps) to drive the loudspeakers. If he is right with the idea that this is the physically correct operation method, this could probably gain lower distortion in the range of 3 to 26dB (depending on the speakers used an the frequency).
But all amplifiers you can buy are voltage controlled voltage sources - one has to build the voltage controlled current source by doing it yourself. So my question here is: is this worth the effort? It is not only the amp that is different, the way to build the speaker has also to be adapted to the other way of driving the speaker....
for example: if you build your speaker with Qt=0,7 you know that the level is down 3dB at the spakers resonance - when driven with a voltage source.
If driven with a current source it is different as the current source draws its current without regard to the speakers impedance - and therefore the impedance curve shows up on the speakers level curve. The consequence is,that the speaker is louder at its resonance and gets louder as the frequency rises - following the impedance-curve.
If one will compare both ways to drive a speaker the speakers equalisation has to be done individually with respect to the driving method. Assuming the equalisation is done in a way that both types measure the same in sound pressure level.... : Is a current driven speaker delivering a better sound quality? Or is it just different / the same as driven in a conventional way?
All hints for building a high quality (state of the art) current amp are welcome, and so are measurements on real speakers that show the difference between the driving methods.
Bernd
But all amplifiers you can buy are voltage controlled voltage sources - one has to build the voltage controlled current source by doing it yourself. So my question here is: is this worth the effort? It is not only the amp that is different, the way to build the speaker has also to be adapted to the other way of driving the speaker....
for example: if you build your speaker with Qt=0,7 you know that the level is down 3dB at the spakers resonance - when driven with a voltage source.
If driven with a current source it is different as the current source draws its current without regard to the speakers impedance - and therefore the impedance curve shows up on the speakers level curve. The consequence is,that the speaker is louder at its resonance and gets louder as the frequency rises - following the impedance-curve.
If one will compare both ways to drive a speaker the speakers equalisation has to be done individually with respect to the driving method. Assuming the equalisation is done in a way that both types measure the same in sound pressure level.... : Is a current driven speaker delivering a better sound quality? Or is it just different / the same as driven in a conventional way?
All hints for building a high quality (state of the art) current amp are welcome, and so are measurements on real speakers that show the difference between the driving methods.
Bernd
Last edited:
"Is it worth the effort?"
Can only reply with anecdote. Personally, I like and prefer the sound when my full range and 2-way speakers are driven with a current source amp, and when the 2-way uses a series crossover, over when they were driven by amps with feedback. But it's a taste thing. this whole audio thing is basically about taste. When the technical details are out of the way, e.g. impedance and other source-load issues are correct, then mixing and matching gear is a bit like wine tasting. They're all different, usually in subtle ways, and what one person likes, another may not. The only answer for the ambitious is to jump in and try.
"The consequence is,that the speaker is louder at its resonance and gets louder as the frequency rises - following the impedance-curve."
This is the phenomenon tamped down to some extent by the shunt resistor in parallel with the driver, in a series crossover setup (maybe it also works in a parallel crossover? Idk). As the impedance of the driver goes up, a greater proportion of the current is shunted through the resistor. Thus the value of the resistor determines the level at which the shunting through it predominates over current through the driver. Also a matter of taste imo.
Can only reply with anecdote. Personally, I like and prefer the sound when my full range and 2-way speakers are driven with a current source amp, and when the 2-way uses a series crossover, over when they were driven by amps with feedback. But it's a taste thing. this whole audio thing is basically about taste. When the technical details are out of the way, e.g. impedance and other source-load issues are correct, then mixing and matching gear is a bit like wine tasting. They're all different, usually in subtle ways, and what one person likes, another may not. The only answer for the ambitious is to jump in and try.
"The consequence is,that the speaker is louder at its resonance and gets louder as the frequency rises - following the impedance-curve."
This is the phenomenon tamped down to some extent by the shunt resistor in parallel with the driver, in a series crossover setup (maybe it also works in a parallel crossover? Idk). As the impedance of the driver goes up, a greater proportion of the current is shunted through the resistor. Thus the value of the resistor determines the level at which the shunting through it predominates over current through the driver. Also a matter of taste imo.
Depends largely on the driver:
The big benefit isn't the response changes. If anything those are a problem but that can be fixed with some EQ.
The main benefit is less distortion caused by inductance in the motor:
Depends on each specific driver but for most normal cone driver the reduction starts at around 2 octaves above fs:
In this post I posted measurements of 3 different drivers: Monacor which is a cheap 6.5" midwoofer, AE TD6M which is an overkill midrange and 6ND430 which is a high efficiency pro midrange: https://www.diyaudio.com/community/threads/hypex-ncore.190434/post-4095122
If you have a driver without copper rings in the motor (which help reduce distortion, but decrease efficiency. Usually cheaper drivers have less of them) then the difference can be huge.
An extreme example of this (which is in the post I linked) is when I measured the 6ND430. Since it is a pro high efficiency midrange driver without copper rings. It has an fs of ~ 75 hz and above 400 hz the 3rd order harmonic distortion drops like a rock when driven with a current source amp vs a voltage source.
An extreme example in the other way is the AE TD6M which has a huge well designed and very expensive motor. It already has almost no inductance disortion so I couldn't measure any difference of voltage vs current drive.
A more reasonable example is the SB65WBAC25-4. It has copper in the motor to linearize inductance but a more reasonable amount. The benefit isn't as big as the 6ND430 but still a reduction on the top end.
The big benefit isn't the response changes. If anything those are a problem but that can be fixed with some EQ.
The main benefit is less distortion caused by inductance in the motor:
Depends on each specific driver but for most normal cone driver the reduction starts at around 2 octaves above fs:
In this post I posted measurements of 3 different drivers: Monacor which is a cheap 6.5" midwoofer, AE TD6M which is an overkill midrange and 6ND430 which is a high efficiency pro midrange: https://www.diyaudio.com/community/threads/hypex-ncore.190434/post-4095122
If you have a driver without copper rings in the motor (which help reduce distortion, but decrease efficiency. Usually cheaper drivers have less of them) then the difference can be huge.
An extreme example of this (which is in the post I linked) is when I measured the 6ND430. Since it is a pro high efficiency midrange driver without copper rings. It has an fs of ~ 75 hz and above 400 hz the 3rd order harmonic distortion drops like a rock when driven with a current source amp vs a voltage source.
An extreme example in the other way is the AE TD6M which has a huge well designed and very expensive motor. It already has almost no inductance disortion so I couldn't measure any difference of voltage vs current drive.
A more reasonable example is the SB65WBAC25-4. It has copper in the motor to linearize inductance but a more reasonable amount. The benefit isn't as big as the 6ND430 but still a reduction on the top end.
Attachments
It depends.
Using full range speakers, up to 6 to 10dB boost at resonance and continually rising boost from,say, 1kHz up can be seen as sounding better.
In the niche world of Guitar amps/speakers that certainly holds true.
Now in a 2 or 3 way system, impedance varies wildly across the audible range, unexpected peaks and dips within the passband can become ugly
You sure you want that huge peak around 1500 Hz?
Even worse in a 3 way.
Using full range speakers, up to 6 to 10dB boost at resonance and continually rising boost from,say, 1kHz up can be seen as sounding better.
In the niche world of Guitar amps/speakers that certainly holds true.
Now in a 2 or 3 way system, impedance varies wildly across the audible range, unexpected peaks and dips within the passband can become ugly
You sure you want that huge peak around 1500 Hz?
Even worse in a 3 way.
Last edited by a moderator:
I've only heard one transconductance amp and I wasn't impressed. On a full-range driver I found that the sound stage compressed significantly when switching from a near ideal voltage controlled voltage source (so ultra-low THD amp with a very low output impedance) to the transconductance amp. That said, this was in a sighted trial, N = 1, and I designed the voltage controlled voltage source amp so I can easily be accused of bias, but I also know that we were several in the room who perceived this. Do keep in mind that this was just one implementation of a transconductance amp. Maybe there are better ones out there.
Tom
Tom
Other benefit of current drive is supposedly better transient response, as feedback corrects signal after the speaker, but the web page showing the data is down.
https://www.diyaudio.com/community/...uctance-current-amplifier.239321/post-6291777
https://www.diyaudio.com/community/...uctance-current-amplifier.239321/post-6291777
Last edited:
The idea is to equalize them out, so you get whatever small-signal response you want, but with less distortion and less voice-coil-heating-related compression than with voltage drive.It depends.
Using full range speakers, up to 6 to 10dB boost at resonance and continually rising boost from,say, 1kHz up can be seen as sounding better.
In the niche world of Guitar amps/speakers that certainly holds true.
Now in a 2 or 3 way system, impedance varies wildly across the audible range, unexpected peaks and dips within the passband can become ugly
You sure you want that huge peak around 1500 Hz?
Even worse in a 3 way.
Thank you all for your input.
I did some simulation with ideal current and voltage sources and a complex RLC combination that approximates
the impedance of a speaker. What i see with the current source is that the current is in phase with the source
and the resulting voltage across the RLC combination shows a phaseshift in the voltage. If this is also true for
a voltage controlled current source, and assuming it is the current that generates the audio, the timing of the spectrum
of an instrument would be more correct, am i right?
Bernd
I did some simulation with ideal current and voltage sources and a complex RLC combination that approximates
the impedance of a speaker. What i see with the current source is that the current is in phase with the source
and the resulting voltage across the RLC combination shows a phaseshift in the voltage. If this is also true for
a voltage controlled current source, and assuming it is the current that generates the audio, the timing of the spectrum
of an instrument would be more correct, am i right?
Bernd
Isn't some of the benefits of CFA that the gain increases with increasing load impedance, the output level follows the speaker impedance curve, you get a boost both at low-end resonance and rising output at high frequency. This gives a scooped response. If you EQ, all this 'benefit' is EQ'ed away...?
This makes CFA more problematic with everything but single driver full range speakers, and pretty hopeless with many-way speakers with difficult XOs since the impedance and phase is all over the place, as has been mentioned already.
This amp is a class-A current source, most current feedback amps are just regular class AB amps with tons of open loop gain and requiring negative feedback for proper response/gain. I don't think they are directly comparable. Not only being pure class-A, the Pass F1J wont have such a crazy respons into a difficult multi-way speaker with difficult impedance since the gain is not that dependent on the load impedance. The gain, or level out, will vary with impedance, but not so wildly, and the lack of feedback means it tries not so hard to follow the impedance/phase of the load. Without the NFB the amp doesn't try so hard, said in a non technical way, since I don't know how else to say it...
That used to be regarded as an advantage in the 1930's, with valve radios that had ventilation holes in the back and the associated bass loss. With modern loudspeakers, it is a disadvantage, but one that can easily be solved with equalization - much easier than correcting the distortion and compression with voltage drive, anyway.
With voltage drive, the voltage is in phase with the source and the resulting current across the RLC combination shows a phase shift in the current. It's all complementary.
Jan
No global feedback makes this amp sound musical with great soundstage. I use is only from 150Hz up, not for woofers, and rarely need more than 1 watt. Lack of global feedback is completely different topic, do not want to bring it here, but i was exploring classA buffers without feedback and achieved similar holographic soundstage.
Hi,
as its voltage the amplifier commands but current is what makes voice coil to move, does this phase shift make into acoustic domain?
I guess if it does it shows in frequency response, in other words any effects into acoustic domain are matter of EQ?
Hi,
have you seen this https://www.researchgate.net/public...provement_in_the_current_coil_of_loudspeakers ?
I haven't read Esa's book but EDN articles linked from his home page give good summary I assume:
https://www.edn.com/loudspeaker-operation-the-superiority-of-current-drive-over-voltage-drive/
https://www.edn.com/loudspeaker-ope...y-of-current-drive-over-voltage-drive-part-2/
Also read everything on Purifi audio site https://purifi-audio.com/
From these materials it looks to me that the problem or solution is not type of the amplifier but impedance in the circuit where driver and amplifier both are part of. Oh well there is difference between the types of amplifiers but we are not limited to those two and can adapt impedance manipulation with voltage control amplifier to reduce driver distortion as well. There is many ways to reduce system distortion.
From Esa article linked above there is explanation where the driver impedance consists of and how back-EMF works out and why driver impedance shows up in frequency response. If you analyze amplifier / driver circuit backwards, think the driver (back-EMF) as voltage source in the circuit and replace the actual power amplifier with its output impedance its easy to see how much of a current back-EMF makes in the circuit.
Its current that turns into mechanical movement in the motor, and acoustic output. This means also any distortion source needs to make current in the circuit to get emitted acoustically. Imagine very simple circuit, woofer hooked directly to a typical voltage control amplifier which has very low output impedance. Analyzing the circuit backwards, load for driver back-EMF voltage, quickly shows that the load is almost a short!
Voltage control amplifier will provide maximal electronic damping on driver resonance as it provides very low impedance between driver terminals and keep frequency response nice on low frequency, low system Q. Higher up in frequency things get worse, all of the back-EMF voltage with non-linearities included turn into current maximally as well, all of the driver motor non-linearities turn into acoustic domain.
What if there was current amplifier with very high output impedance instead? Back-EMF voltage would still be there with all its non-linearities as its property of the driver, but now it doesn't turn into much of a current anymore because impedance in the circuit is high due to high output impedance of the power amplifier, not much of the motor non-linearities show up in acoustic domain.
What about voltage control amplifier with series inductor added in the circuit? Low driver load impedance on low frequencies and high load impedance for high frequencies: electrical damping on driver resonance and reduction of motor distortion higher up.
Looking from this perspective its quite simple to stem the knowledge to all kinds of systems like reason that current amplifier would work nice for fullrange driver system, or active multi-way systems, as long as damping is handled mechanically or acoustically, or EQ:d. Purifi tech paper "Low Distortion Filter for PTT6.5X04-NAA" demonstrates how to effectively reduce distortion in multiway speaker with voltage control. Basically, keep series impedance low for the driver at its resonance and make the impedance rise with frequency, or at least peak with driver cone breakups which would amplify any distortion current. Active system with voltage control amplifier can use inductor to low pass distortion, tweeter highpass with parallel inductor instead of series capacitor to provide electrical damping for low frequencies (assuming there is also pad resistor and high pass before amplifier) and what not.
I'm not sure if this line of thought captures all the benefits of current drive but considering distortion numbers of amplifiers are at least magnitude lower than what complete loudspeaker systems measure acoustically it makes sense to utilize what ever means to prevent driver distortion turn show up in acoustic domain. Current drive amplifier can work fine on some systems at least when the speaker is designed for current amplifier. Perhaps increasing series impedance with voltage control amplifiers is more practical way to get similar benefits on other situations. Don't forget multiplying cone area would reduce excursion and all driver non-linearities including mechanical ones, waveguides, better drivers, dividing the bandwidth into multiple ways... Many ways to make better sounding systems. Current drive seems very logical thing but don't forget all the other options that lend to low distortion systems
Last edited:
@tmuikku
You might find this an interesting read;
https://journals.pan.pl/Content/101305/PDF/aoa-2017-0008.pdf
Quite a nice read and paper about the whole thing.
You were kinda on the right path.
Current drive works well when there are only electrical changes, like the Le(x) for example.
As well as other issues like eddy currents and what not.
How well it works depends a little on what non-linearities are at play as well.
But you can read all about that in the paper (I am not gonna summarize that here).
In some cases the measured distortion absolutely plummets!
It does not work very well when the loudspeaker operates around the control system region (up to roughly 2-3x fs).
So I think it will be really a good implementation in 2-way system from roughly 100-150Hz minimum.
The rest is being done by subwoofers anyway and leave those at a CV amplifier.
edit; attached paper before some links get broken again.
You might find this an interesting read;
https://journals.pan.pl/Content/101305/PDF/aoa-2017-0008.pdf
Quite a nice read and paper about the whole thing.
You were kinda on the right path.
Current drive works well when there are only electrical changes, like the Le(x) for example.
As well as other issues like eddy currents and what not.
How well it works depends a little on what non-linearities are at play as well.
But you can read all about that in the paper (I am not gonna summarize that here).
In some cases the measured distortion absolutely plummets!
It does not work very well when the loudspeaker operates around the control system region (up to roughly 2-3x fs).
So I think it will be really a good implementation in 2-way system from roughly 100-150Hz minimum.
The rest is being done by subwoofers anyway and leave those at a CV amplifier.
edit; attached paper before some links get broken again.
Yep, tried to keep the post short Purifi page has quite much content on driver distortions, some of which can be reduced by increasing series impedance in the circuit. Then there is stuff that can be good for one system and bad for another, current drive would prevent thermal compression for example. Perhaps nice for home system but would make thermal runaway and burn drivers in PA application.
I think inspecting circuit backwards is quite simple way to kind of approximate what the distortion current is and overall thought experiments on the issue gives insight what kind of systems would minimize distortion current, excursion, unnecessary movement.
My message is that thinking about which amplifier is better didn't provide much insight to me but expanding the idea of current control to fully include the driver and think both as a system gave tremendous amount of insight and potential to get better results in the end, than just wondering about should one build a current drive amp. Perhaps voltage amp for the low bass woofer and current for the rest as you say.
After all modern amplifiers seem to have very low distortion numbers. While current control amplifier makes the whole system distort less in acoustic domain than voltage control amp, the system distortion is still showing numbers magnitude or more than the amplifier alone. For this reason it looks to me we are not having amplifier problem but driver problem, or more like system problem, so it does not make sense to me to look the issue from amplifier perspective. I think we could assume any perfectly fine amplifier and concentrate on other aspects of a system to make distortions low, or what ever it is that makes better sounding system. Well, just use best amp for the application in the end.
Anyway, hope everyone got their imagination expanded a little After all all that matters is to get better sound and enjoying the hobby, having fun
Purifi page has quite much content on driver distortions, some of which can be reduced by increasing series impedance in the circuit. Then there is stuff that can be good for one system and bad for another, current drive would prevent thermal compression for example. Perhaps nice for home system but would make thermal runaway and burn drivers in PA application.I think inspecting circuit backwards is quite simple way to kind of approximate what the distortion current is and overall thought experiments on the issue gives insight what kind of systems would minimize distortion current, excursion, unnecessary movement.
My message is that thinking about which amplifier is better didn't provide much insight to me but expanding the idea of current control to fully include the driver and think both as a system gave tremendous amount of insight and potential to get better results in the end, than just wondering about should one build a current drive amp. Perhaps voltage amp for the low bass woofer and current for the rest as you say.
After all modern amplifiers seem to have very low distortion numbers. While current control amplifier makes the whole system distort less in acoustic domain than voltage control amp, the system distortion is still showing numbers magnitude or more than the amplifier alone. For this reason it looks to me we are not having amplifier problem but driver problem, or more like system problem, so it does not make sense to me to look the issue from amplifier perspective. I think we could assume any perfectly fine amplifier and concentrate on other aspects of a system to make distortions low, or what ever it is that makes better sounding system. Well, just use best amp for the application in the end.
Anyway, hope everyone got their imagination expanded a little
After all all that matters is to get better sound and enjoying the hobby, having fun
Last edited: