Please master, we are but ignorant children, have we done wrong that you will not lead us into the light? Have mercy oh wise one, reveal unto us the great truth that is hidden
What a weird thing to say.
Some very strange reactions here. One, two, three.. any more? What next? Who will bring the rope?
There is a wonderful phrase: "Intellectual Phase-Locking, a condition that results when dogmatic assumptions inhibit inquiry."
Sure. Not sure if novelty comes into it. Is that a new 'new' that is frowned upon here? Seriously, I am familiar with the term as it is used in some sciences.
But has anybody suggested that the previous work over the last 40 years is going to become invalid overnight? I certainly haven't and know of nobody who has. Rather, it should illuminate, not eliminate, right?
There is nothing threatening here. It shouldn't be viewed that way.
At least I have the safety of knowing I am not sticking my neck out entirely on my own, but there is traction on this, and by others. Not anywhere online, social media has its limitations, ahem. I can even guarantee it will not come up on Twitter.
See? That was a joke. Smile, OK?
However, F is not the only contributor to affect Acoustic output and its associated distortion at frequency range of interest. Moreover, why insist on voltage source if it does not provide enough current control for you? Nobody prohibits you to use capable current source.
Indeed, the back-EMF impedance has a large effect, any resonance or flaw will modulate the back-EMF, alter the current that is ultimately seen by the stable Re impedance of the voice coil. So you are both right and wrong, just a matter of how you look at it. But in the end, the altered current is the F=Bli that we listen to. There is no voltage in that equation. Have you an alternative equation?
Quite right, but what if we could make a speaker that manages the current in such a way that you can use both current and voltage drives, the same speaker can accommodate both. Such is possible and has been done DIY, but not commercially.
Certainly, as those links furnished by PMA that shows a reduction in distortion by using current drive. This is beyond dispute, I have confirmed this myself, so the links were informative, but what they conveyed is not new to me. In fact I predicted the result even before the test. Current drive means that any current flowing in the voice coil will become non-reactive, not totally because current drive that deals with LF resonances/alignment issues means compromises to making current drive possible. And if you want to do that, then indeed nothing prohibits you doing that. But it will be compromised current drive, not pure current drive. I suspect you are familiar with Esa Merelainen's work and got no issues with that.
What is suggested is an alternative to current drive, and that is current EQ is another way of lowering the distortion that some have indeed measured (and my congrats for doing that), but it is painful to put into practice and if non-commercial solutions are used, then you will have speakers, but what retailer will sell to current/transconductance amplifiers? It won't happen.
EQ the current of the amplifier, stop it producing reactive current and you have a solution that is compatible with both current and voltage drive. Suppress the source of back-EMF using crossover techniques that does just that (they are similar but not the same as Esa's suggestions in his book).
I attended ETF17 in Denmark last year and just come back from attending ETF in Normandy, France, earlier this month. There is a discussion going on and there is every reason to see further traction, new ideas are being expressed (that is a dangerous word here), that illuminate what we know so far and does not eliminate what has gone previously.
But more than just a single person needs to work on this. Social media is poison, so it is done quietly and discretely. When a consensus/agreement is made, then it will be presented in full form. I was told "not to rush this, let us get it right" was the paraphrased thinking. I hope it will result in an AES paper soon, but not up to me exactly what year, but 2019 would be great.
I am basically just reporting here. OK?
Last edited:
Yup- bi- or tri- amping is the only way to go, with decent dsp filtering earlier in the chain. Allow the opportunity to play with pre-distortion too.
That raises an important point, one that has been raised by Hans van Maanen of Temporal Coherence who has a 3-Way tri-amped design. He designs his own electronics including the six Class AB amplifiers that drives the six drivers. But he adds current EQ, passive inductors, capacitors, resistors, and EQs the current, so that when the transition between Class A and B of the amplifier, both current and voltage lines up 'temporally' in time. This prevents the feedback loop from trying to compensate that causes burst of high order distortions in the amplifiers.
So by all means make bi- or tri-amped speakers, but EQ the current is a good idea. It will also mean that the reactive part of the impedance of the drivers causing the amplifier from supply reactive current (because the amp will be supply zero degree current phase angle) and the increasing understanding that zero control of current in a voltage source will cause current chasing its own tail. Esa Merelainen calls this a "feedback" problem, because it feeds back on itself. I give him credit for illuminating that in his current-driving book.
So there are alternative techniques being used. It's early days, but it is happening.
What DSP Hans is using, if using it, I cannot comment on.
I can understand that viewpoint, but a better understanding what the 'back-EMF impedance' of the individual driver points to other solutions as well.
Last edited:
I have given a lot of thought to that as well. Like about six years ago. It seems to explain a lot of things and I went along with it for a while. Then in discussing it with a knowledgeable friend of mine and getting better understanding of how the back-EMF impedance of the driver affects the current and its relation to the Re of the driver, and a much better current angle/view emerged. So no, Thevenin is a basically voltage model, we need a current model explanation and we have one. The change in dB-SPL caused by the back-EMF impedance and the relation to current through the VC, the maths works perfectly.
If an official paper comes out (I believe it will), then I think the current model will have the much better explanation. The maths has already been done and is far more convincing. Current is the explanation, not the voltage.
After 12, so time to turn in. I am in a time zone ahead of most of you guys and in the city where Thiele-Small Parameters originated from.
Why do you insist on repeating history. Sigh.
What Joe (and others) is saying is that by impedance equalising the loudspeaker box as a system, there are no reactive circulating currents between amplifier and loudspeaker.
This has beneficial implications for noise reduction on more than one front.
This investigation is not necessarily revelatory (RF and AC Power systems strive to eliminate reflected power) but importantly bringing to light the benefits of pure resistive loading of CV or CC amplifiers.
The downside is the costs of incorporating significant extra RLC components inside the loudspeaker....for high-end this is justifiable if not required.
Dan.
This has beneficial implications for noise reduction on more than one front.
This investigation is not necessarily revelatory (RF and AC Power systems strive to eliminate reflected power) but importantly bringing to light the benefits of pure resistive loading of CV or CC amplifiers.
The downside is the costs of incorporating significant extra RLC components inside the loudspeaker....for high-end this is justifiable if not required.
Dan.
No, F is not what we listen to. I mean that we listen to the mechanical air wave in which the F applied to a driver is one of contributing factor in electro-mechanical conversion. Mechanical, couping, resonance, frequency dependent conversion efficacy and perhaps a host of other effects are not being considered. On some drivers, Earl Geddes observed worse and uneven response to a constant current drive frequency sweep, more consistent with Acoustic output = F + multiplicity of effect from other lores.
What is the benefit of resistive loading vs reactive loading on a CC amplifier?
Last edited:
There are always circulating currents or there is no sound. The idea that a pure voltage amplifier loses "control" for any current not at zero phase with the voltage is nonsense.
But appealing!
There are no limits for new discoveries in this thread, and you should have an open mind!
That sentence makes no sense. The only way current and voltage line up temporally is for a pure resistive load.
edit: and it must be pointed out that it is the driver relation that is important, not the amp output. So making the load resistive via elements, while keeping the amp out of quadrant 2 and 4, does nothing to remove the energy stored within the driver/cabinet combination.
The balance put here just for continuity, so that my reply makes sense.
He is thrashing about wildly. I see the same for so many other designers, a very simple misconception or lack of understanding. This is EXACTLY the same problem I am addressing in my motion control work.
When the response plot of a system is dependent on either the position, velocity, or acceleration of the system, the bode plot will be changing. As such, the phase margin will constantly change depending on the stimulus.
You can drop the feedback to accommodate the "loosest" state of the system to stop oscillation bursts, or you can try to adaptively change the feedback parameters by riding the system state.
I've never seen this work prior, so I may be on the leading edge. (bleeding edge perhaps)
.If such is the case, please try and keep up. (yah, I got my do anything badge..)
"Back EMF impedance" is a non-sequitor.
jn
Wow. I go to use quotes from several of my favorite movies as well as from NOMAD.
Last edited:
- Status
- Not open for further replies.