Optimum coupling amplifiers to speakers.

[Goodman]

Anyone compared the distortion affect on speakers with voltage drive amplifiers vs current drive amplifiers?

 

[sergeauckland]

Why would anyone ever want a current-drive amplifier? Almost all loudspeakers have an impedance characteristic that varies greatly with frequency, and are expected to be driven from a voltage source of near-zero output impedance. Having a current source of high (theoretically infinite) output impedance would totally screw up the 'speaker's frequency response.

S.

 

[KSTR]

Anyone compared the distortion affect on speakers with voltage drive amplifiers vs current drive amplifiers?

Current drive, or more general, driven with a source impedance profile other that zero has proven beneficial for almost all loudspeaker drivers. Complete speaker with passive crossovers or other filter circuits less so.

Most voice-coil driven drivers like to see low to moderate impedance around system resonance, to provide some reasonable amount of damping. Elsewhere, as high impedance as possible impedance is better and the benefit is largest with less than stellar drivers (a Purifi driver gains very little from being current-driven, for example).
- Power compression is bascially absent
- Effect of modulation of voice-coil inductance vs excursion is bascially absent
That catches two major sources of distortion and/or non-stable behavior.

Other driver types like air motion transformers do respond well to pure current drive.

To sum up, current drive is no one-size-fits-all approach. When driver behavior is closely examined it turns out that for every driver in a given situation there is a optimum source impdance profile vs. frequency that achieves the best results overall (with regard top distortion but as well other aspects, like recovery from excursion overdrive etc). It almost never is zero. Sometimes it might even be negative (at and below resonance)!

Zero source impdance for amps was introduced not because it is technically better than everhing else, rather it was a way to standardize things so different speakers can be paired with different amps and not worry about frequency response changes.

@sergeauckland: EQ is the least of your problems here, even with pure current drive. Measure the impulse response of the current under voltage drive and voila, you directly obtain the required EQ to make it behave and sound the same wrt frequency response.

 

[DVDdoug]

Speakers are designed, tested, and specified with a constant* voltage source. A true constant current source would give you terrible frequency response because it will track the impedance curve. I assume you've seen some speaker impedance curves...

Some amplifiers are advertised as "high current" which is meant to suggest that they are very-good voltage sources, where the voltage can remain constant with low impedance ("high current") loads.

Some headphone outputs have "some" output resistance/impedance so the voltage depends on load. Since some headphones have a mid-bass bump in impedance, this makes a mid-bass bump in frequency response and some listeners like it. But, it's not how the headphones are designed to be used, it's unpredictable (different with different amplifiers and different headphones), and the headphones will have other impedance variations at other frequencies. And it's not a true constant-current output. It's closer to a constant-voltage source with the current influenced by the source impedance.

Speakers are usually "worse" than headphones because of the crossover & multiple drivers.


* Of course music & other program material is not really constant voltage but the voltage is independent of load impedance. And, frequency response measure measurements are done with a voltage that's the same at all frequencies.

 

[Tangband]

Current drive, or more general, driven with a source impedance profile other that zero has proven beneficial for almost all loudspeaker drivers. Complete speaker with passive crossovers or other filter circuits less so.

Most voice-coil driven drivers like to see low to moderate impedance around system resonance, to provide some reasonable amount of damping. Elsewhere, as high impedance as possible impedance is better and the benefit is largest with less than stellar drivers (a Purifi driver gains very little from being current-driven, for example).
- Power compression is bascially absent
- Effect of modulation of voice-coil inductance vs excursion is bascially absent
That catches two major sources of distortion and/or non-stable behavior.

Other driver types like air motion transformers do respond well to pure current drive.

To sum up, current drive is no one-size-fits-all approach. When driver behavior is closely examined it turns out that for every driver in a given situation there is a optimum source impdance profile vs. frequency that achieves the best results overall (with regard top distortion but as well other aspects, like recovery from excursion overdrive etc). It almost never is zero. Sometimes it might even be negative (at and below resonance)!

Zero source impdance for amps was introduced not because it is technically better than everhing else, rather it was a way to standardize things so different speakers can be paired with different amps and not worry about frequency response changes.

@sergeauckland: EQ is the least of your problems here, even with pure current drive. Measure the impulse response of the current under voltage drive and voila, you directly obtain the required EQ to make it behave and sound the same wrt frequency response.

Thanks for this teaching.

Ultra low impedance coupling is most beneficial at the EMF area, - the bass frequencys. This is the way active subwoofer works.

In some cases, a low driving impedance can sound better also in mid or treble area, sometimes not.
One has to test this with each driveunit.

With that said, most active highend manufacturers of studiomonitors with very low distortion and very good sound quality such as Genelec or Neumann or Linn products with their exakt technology - they all use direct coupling between voltage driven amplifiers and drivers, even for tweeters.

My own investigations shows that I prefer voltage driven amplifier with low output impedance for tweeter horns, bassunits and midranges. And I have a mixed experience with dome tweeters.

 

[KSTR]

Speakers are designed, tested, and specified with a constant* voltage source. A true constant current source would give you terrible frequency response because it will track the impedance curve.

While this is true, nobody would use current drive like that. So your bold-faced objection does not hold in practice.

Correctly used, current drive will most often reduce distortion in the frequency range where current drive is applicable, which is, as explained, most of the usable frequency range of a driver.
See my post #3 here.

 

[KSTR]

In some cases, a low driving impedance can sound better also in mid or treble area. One has to test this with each driveunit.

Exactly. You have to find the best source impedance (in a frequency range) by measurement and critical listening.

I mentioned negative drive impedance... an example are compression drivers with rather stiff and low-excursion surrounds and a underhung VC. At or below resonance the increased internal motional feedback helps linearize the driver. Same goes for some subwoofer designs.

In the end, the drive impedance only controls the amount of internal motional feedback in a driver, using the voice coil's signal as the velocity sensor. Depending on construction there is a sweet spot for that intrinsic feedback and the effect it has on the driver, the damping.

With that said, most high end manufacturers of studiomonitor with very low distortion and very good sound such as Genelec or Neumann or Linn products with their exakt technology - they all use direct coupling between amplifiers and drivers, even for tweeters.

There are examples of current drive in commercial speakers. Kii Three's midrange driver is current-driven. Tweeters in ADAM F-series were current drive (and woofers were mixed-impedance drive) -- I designed those, incidentially ;-) Some German servo-controlled speakers use current drive on the tweeter section as well.

 

[pma]

Anyone compared the distortion affect on speakers with voltage drive amplifiers vs current drive amplifiers?

Did some research, nothing unexpected

Current drive of speakers and speaker distortion

 

[KSTR]

A paramount, even though actually trivial thing to consider is the follwing:
When the voltage on the driver terminals is the same, the frequency response and overall behavior is the same, too, regardless of what the drive impdance is. It cannot be any other way.

That's the way you actually must compare different drive impedance, dialing in always the same voltage at the driver for same SPL response (as otherwise the frequency response difference will dominate).
The true difference is in the fine-print: Distortion and large-signal behavior (error recovery).

 

[sergeauckland]

A paramount, even though actually trivial thing to consider is the follwing:
When the voltage on the driver terminals is the same, the frequency response and overall behavior is the same, too, regardless of what the drive impdance is. It cannot be any other way.

That's the way you actually must compare different drive impedance, dialing in always the same voltage at the driver for same SPL response (as otherwise the frequency response difference will dominate).
The true difference is in the fine-print: Distortion and large-signal behavior (error recovery).

In which case, what's the difference between current drive and voltage drive?

We must have very different definitions of voltage drive (i.e. zero source impedance) and current drive (infinite source impedance)

S.

 

[KSTR]

In which case, what's the difference between current drive and voltage drive?

We must have very different definitions of voltage drive (i.e. zero source impedance) and current drive (infinite source impedance).

I think I have just explained it here (and other places): The difference is in the "fine-print" once the frequency response is dialed in to be the same with whatever source impedance one is using.

Speaker drivers are current transducers. With voltage drive the static voice-coil impedance is used as the transfer mechanism from terminal voltage to exiting current. When that transfer impedance is nonlinear (Lex(x)) and unstable (thermals) the internal feedback may not improve behavior, actually degrade it (reentrant distortion).

Simplified, with current drive the static transfer impedance is made irrelevant, the driver is operating in pure force-steered mode (F= BL * i). With a negative drive impedance of -(Re+Le) we have maximum velocity feedback, the voltage (across the actuator part of the voice coil as the static impedance has been cancelled) now is proportional to cone velocity (as measured by the VC istself) alone and any deviation from it will cause a huge correcting current to establish the velocity. This is a true velocity control feedback loop.
It relies on absolute stability of Re and Le which we don't have...
Standard voltage drive is in between these extremes, applying partial velocity feedback. Highest feedback is around resonance where VC current is lowest and develops only little voltage across the static VC impedance.

As mentioned, when you just swap a voltage drive amp with flat frequency response of voltage for a current drive amp with flat frequency response of current that's not going to work as the SPL frequency response will be different for obvious reasons.
Therefore, the terms voltage drive and current drive shall not automatically imply flat frequency response of the amp.

 

[sergeauckland]

Thanks for the explanation, but what you're suggesting isn't a way of driving passive loudspeakers, only something that can be engineered as an active system, including DSP control. I may be mistaken, but I didn't think that was what the OP asked, and was how I answered the question.

S.

 

[DrCWO]

You can find much more on it here.

For fulrange speakers in an active system it seems to be a good idea as it reduces distortions, especially 3. and 5. degree. I plan to do some measurements on my own with my fulrange horns this year. But as I don’t have a current drive amp I will introduce a resistor before the driver.

 

[Tangband]

You can find much more on it here.

For fulrange speakers in an active system it seems to be a good idea as it reduces distortions, especially 3. and 5. degree. I plan to do some measurements on my own with my fulrange horns this year. But as I don’t have a current drive amp I will introduce a resistor before the driver.

I have tried feeding a jbl compression driver with a high impedance compensating the frequency response with a dsp and the sound didnt get any better. So best try and listen/measure in specific cases If its worth it . I tested this also in the HYBRID project, with a dometweeter in a waveguide and didnt get any soundquality gain, and the sound was by some listeners perceived as worse.

My conclusion so far is that the sound gets better in most cases with directly coupled low impedance power amps in an active system, even at higher frequencies. But ….I have seen measurements thats beneficial with high impedance driving in certain combinations ( slightly less distortion ) so it might be better in some cases.

 

[DrCWO]

My conclusion so far is that the sound gets better in most cases with directly coupled low impedance power amps in an active system, even at higher frequencies.

I run these drivers in horns and all I read regarding those high efficiency full range drivers they like current drive. I will report back if I found time to test.