If you took some time to step back and understand the material that is presented, you will find that, firstly you are confusing Ohm's Law with the transfer function, and secondly everything I have written pertains to driver measurements and their relation to "sound quality". Nothing barring one short paragraph that I deliberately enclosed in brackets is or was off-topic.
I have repeatedly asked you to supply an example of where I am incorrect, or an example of where my contributions are off-topic, and you have continually failed to do so. If driver measurements and their relevance to subjective assessments depend explicitly on the manner in which you conduct your measurements, then why is it off-topic?
I will for once allow myself an admittedly rather poor analogy and something of a thought experiment...
I am lucky enough to own a very fast sports car. If I want to measure how fast it is, I can measure the time it takes to go around a racing track. Is it then relevant whether it is me driving or Max Verstappen?
In a test environment or in the conveying of information contained in measurement data, context is everything.
Again you supply nothing more than conjecture - and conjecture that is errant too. So for the record, I absolutely welcome anyone's contribution that will help me develop my understanding of a subject.
As to my politeness, I respond with factual information in the best way that messaging here allows. Would smiley emojis change any of the information I am writing?
@soundbloke - it is clear to me that the impedance which a driver sees will have an effect on the distortion performance of the driver.
If a passive notch filter is used to reduce the amplitude of a cone breakup, the result will be less distortion than if DSP EQ is used to notch out the resonance. If the resulting frequency response is the same between the two cases, it means the voltage signal at the driver is the same. But the impedance that the driver sees is different, and the passive notch filter presents the driver with a high impedance load at the cone breakup frequency. As you pointed out, a high impedance suppresses motor-generated back current.
A couple of questions about current drive amps:
1) If the output impedance of an amp is significantly higher than the load, what is the power dissipation within the amp? It would seem to me that it would be high. If a 4 Ohm driver requires 2A , and the output impedance of the amplifier is 50 Ohm, it means the driver is dissipating 16W, and the amplifier is dissipating 200 W. Or am I missing something?
2) If I wanted to experiment with a current drive amplifier, what would be a good supplier? Who makes a good current driver amp?
Thanks!
J.
If a passive notch filter is used to reduce the amplitude of a cone breakup, the result will be less distortion than if DSP EQ is used to notch out the resonance. If the resulting frequency response is the same between the two cases, it means the voltage signal at the driver is the same. But the impedance that the driver sees is different, and the passive notch filter presents the driver with a high impedance load at the cone breakup frequency. As you pointed out, a high impedance suppresses motor-generated back current.
A couple of questions about current drive amps:
1) If the output impedance of an amp is significantly higher than the load, what is the power dissipation within the amp? It would seem to me that it would be high. If a 4 Ohm driver requires 2A , and the output impedance of the amplifier is 50 Ohm, it means the driver is dissipating 16W, and the amplifier is dissipating 200 W. Or am I missing something?
2) If I wanted to experiment with a current drive amplifier, what would be a good supplier? Who makes a good current driver amp?
Thanks!
J.
Yup. And I tried a few times discussing the measurement. Got no answers. So to me this all stays moot.
@hifijim This again????
The passive notch filter presents the driver with a high impedance source (not load) at the cone breakup frequency.
You seem to get confused on this topic easily. I suggest to go back to Risbo's paper and read it again.
Regardless, the source and its driving characteristics have been declared off-topic by the moderator anyways.
Dave.
The passive notch filter presents the driver with a high impedance source (not load) at the cone breakup frequency.
You seem to get confused on this topic easily. I suggest to go back to Risbo's paper and read it again.
Regardless, the source and its driving characteristics have been declared off-topic by the moderator anyways.
Dave.
Last edited:
again what... have we discussed this before?
I am not too concerned about what anyone thinks should be discussed in a thread.... all of my threads have gone significantly off topic over the years, and I was never bothered by that.
These are very good questions. As you can see, the manufacturers of "higher quality" (i.e., higher cost) drivers have a vested interest (at least...you might think) to show the measured differences between cheap and expensive. Do they do this? No, not really.
Why? I would guess that legal liabilities in naming specific competitors' drivers would be the first reason to "pull their punches". That could very easily be remedied by those manufacturers anonymizing the drivers that they use to show the difference.
Why don't they do this? A very, very interesting question. The answer to that is key to the whole discussion , I think, and some of the the answers are probably not going to be liked by potential customers. Let's continue with your questions, however, to make sure we're addressing them all...
Note that I have no real interest in direct radiating midrange drivers since I use horn-loaded drivers exclusively nowadays. But the questions you ask are still relevant to the type of drivers that I buy/use.
What are the relevant distortions and limitations of drivers in this frequency range that distinguish them in subjective listening--and which measurements are applicable? What is the threshold of audibility of each distortion type? These are questions that few audio engineers or psychoacoustics researchers are really addressing in their technical studies (at least, those that are published and not perhaps held by these companies as "competitive sensitive").
Finally, you ask the best questions of all (IMHO). The answers are not good ones, I believe. Most driver manufacturers I think fear losing their base of buyers if they bring up specific distortion type subjects and their thresholds of audibility, and some of the cheaper drivers they might make are found to have problems. So they stay silent on these issues--even for their better performing/more expensive drivers. So it's basically a stalemate.
Other manufacturers that make only higher priced drivers that also have better performance should have an incentive to illuminate these issues. Do they? By and large, they do not. Why?
Well, I've worked in engineering organizations where the perceptions of the customers was of utmost importance, and these customers were, in some cases, as well trained/educated on the relevant technical issues as the technical staff of the companies engineering and producing the products, i.e., they were educated customers. Did that keep the organizations that I worked for completely honest and open? Well, no, not when two particular subjects came into focus:
1) in highly competitive engineering environments, there is a lot that is considered "proprietary" and attention to these subjects by a customer is kept at arm's length--so they don't "weedle out" any proprietary details from the engineers...and,
2) Among those in our organization that still considered the customer community to be ignorant in important ways (and these individuals tended to be managerial people), they believed firmly in the notion of "there's a sucker born every minute" (attr. P.T. Barnum). They didn't wish to make the customer representatives any smarter.
The bottom line is that both of these groups distrust their customers, who seem to have an insatiable appetite for "something for nothing". I've seen that many times. And most "proprietary" items are not usually protected by patents, rather they are held as "trade secrets". This is especially true of manufacturing methods and tooling, which often cannot be patented due to "prior art".
________________________________________________________________________________________
So what is our response, as customers? Are we setting up independent measurement means to catalogue the performance of drivers independent of the manufacturers?
In some limited cases on diyAudio--yes. Is this good? It can be if the people and equipment doing the testing know what they're doing. Having a Klippel NFS R&D system is a good place to start, and there are at least two people that have laid out $10K-->$100K USD to do just that.
Do they have the required technical backgrounds to do the job? My experience has been that these two particular individuals really do not have the background that you would find in some chief engineer working at a driver manufacturing firm, or at a loudspeaker design/manufacturing firm--and so they continue to get the wrong questions and the wrong answers.
In addition, even among these chief engineers working at driver manufacturers, there are issues and blind spots in their technical backgrounds that lead them to do things in "probably not the best technically-driven manner".
Chris
[I will answer even though the questions are off-topic, because I have also been accused of being impolite, and not answering would just add support to that unwarranted criticism too]
Subject to the cone break-up manifesting itself as a minimum phase entity, passive EQ can reduce the problem, yes. But also to remember is that cone break-up can be reflected in the coil impedance too, such that there will also be a further non-linearity that the linear passive filter will not circumvent. (I believe the motivation of the odd looking cone surrounds on the Purifi drivers are an attempt to ameliorate such non-linearities?).
A perfect current source has an infinite output resistance, that in practice is accompanied by a finite (but relatively large) resistance in parallel. This is exactly equivalent to a perfect voltage source having a finite (but relatively small) resistance in series. (I think this might be known as Thevenin's Law??). Nevertheless, all else being equal, the actual amplifier will exhibit much the same efficiency in either mode. Having said that, a current driving amplifier will likely require an additional current sensing resistance that will add to the dissipation. "All else being equal" is not, however, an easy result to achieve with OTS amplifiers. Maybe configuring op-amp voltage and current sources and scaling resistances will prove insightful instead?
Nobody of whom I am aware supplies current drive amplifiers. And for good reason too because the amplifier requires to be engineered for or calibrated to a specific driver. Neither manufacturers of amplifiers or manufacturers of drivers would be interested as they can generate satisfactory results with voltage drive and maintaining the status quo. Manufacturers of active loudspeakers have the option available, but they are not going to split apart their products for any reason I can fathom. What it means is that current drive will likely remain the luxury of DIY constructors...
Last edited:
I answered your contribution in post #86. I thought your unanswered question in post #18 stood perfectly well as a rhetorical question. The best answer anyone could give to that particular question is to simply take care that problems are not created. What question has been missed?
Just a couple of weeks ago in a different thread. Memory faulty as well.
I don't disagree with your take on off-topic threads. But, in this particular thread the moderator has mandated the thread to stay on-topic!
My two cents on "current-drive" amplifiers. I agree with @soundbloke , there aren't any available that I'm aware of. There are numerous amplifiers marketed as "current drive".....but these are not actually that.
If interested in real-world transconductance amplifiers you might check out the Pass F1. It's not actually a current-drive amplifier, but it's on the way.
Dave.
Usual current drive amps are still voltage sources themselves, but they are current controlled.
It's just like a current limited lab supply: the voltage is adjusted to get the desired current. The "virtual" output impedance is created via feedback loop.
taking the same lab supply analogy: in ancient times current sources (high impedance output) were created using very high voltage plus high resistances. In this case the power dissipation was indeed high on the current source side.
@soundbloke - thanks for your response.
I was trying to understand your discussion with @anatech. It seems (to my limited understanding) that his position was that if the voltage applied to the driver is constant, then the amp source impedence was not a factor in how the driver responded. Your position (again to my limited understanding), was that source impedance was an important factor.
I was aware of the Purifi white paper that discussed passive notch filters, and I in fact used a passive notch filter in an active DSP design very recently for the same reason, to notch out a cone resonance in the stop band. So your position made sense to me, again based on my limited understanding.
I asked someone else in another thread about current drive amplifiers, and I got a similar answer... I was hoping you might have a different answer.
It is a shame. I was hoping this whole discussion would lead to some practical application for real-world speaker design, particularly active speaker design. But it seems that a current-drive amp is fairly hypothetical. It could exist, if someone would just make one.
j.
I was trying to understand your discussion with @anatech. It seems (to my limited understanding) that his position was that if the voltage applied to the driver is constant, then the amp source impedence was not a factor in how the driver responded. Your position (again to my limited understanding), was that source impedance was an important factor.
I was aware of the Purifi white paper that discussed passive notch filters, and I in fact used a passive notch filter in an active DSP design very recently for the same reason, to notch out a cone resonance in the stop band. So your position made sense to me, again based on my limited understanding.
I asked someone else in another thread about current drive amplifiers, and I got a similar answer... I was hoping you might have a different answer.
It is a shame. I was hoping this whole discussion would lead to some practical application for real-world speaker design, particularly active speaker design. But it seems that a current-drive amp is fairly hypothetical. It could exist, if someone would just make one.
j.
It's debatable whether it's a load or a source. After all the loudspeaker induces the current, if i'm not mistaken?
You and @hifijim just disagreed on the wording, right?
Last edited:
The impedance of a moving-coil driver is not linear. The output of a moving-coil driver has therefore an explicit dependence on the impedance of the source driving it. A measurement or subjective assessment of a moving-coil driver will therefore also have some relation to the said source impedance.
I mentioned Purefi surrounds because I believe the intention with these designs was to prevent non-linearities being reflected in the coil impedance (and so obviating too any need for EQ).
Hi Jim
The normal way to do such things is to use an ordinary amp and build another feedback loop around it that is using current feedback via a very small shunt resistor. Depending on the function of the feedback network a broad variety of output impedances is possible. In this case the amp itself does not actually have a high output impedance if it is set up as current source but it behaves as if it had one. Therfore there is no incrased dissipation.
Just two simple examples: If you use negative current feedback, your amp has a positive series output resistance. I.e. the higher the load the more it goes "onto its knees". If you have positive current feedback your series output resistance will get negative. I.e. the more you load it the higher the output voltage. These two cases are sometimes used to rise or lower Qes of a driver by virtually increasing or lowering Re.
Within a recent thread where such things were also discussed I posted an example of a frequency dependant currwent loop that allows the emulatuion of a series inductor.
An extreme example for a frequency dependant current fedback loop is the ACE bass circuit that even allows to "change" Vas, fs and Qts of a woofer.
Regards
Charles
Last edited:
By the way, it's clear to me (at least) that the audio engineering industry is driven by "lower cost" solutions, and not nearly so much by higher performance products.
So all acoustic driver performance, as viewed by the manufacturers, is seemingly always accompanied by the notion of "at this price point" discussions. Many industries are in this sort of paradigm of thinking about their largely undifferentiated product lines. They don't want you to differentiate their products based on performance, but rather price. So they don't answer your questions about what to measure and just audible distortion levels.
So to talk about measured differences between drivers, and what those differences may be, you are really going against the grain of the manufacturers' intentions for their products. I.e.,
Only "boutique" acoustic driver companies try to vie for the very small market share that relates to diyAudio loudspeaker builders. I've found most of these type of enterprises operate on a shoestring budget in terms of real engineering capabilities; e.g., basically none of them own something as capable as a Klippel R&D system to educate themselves on "just audible distortion levels" and what each measurement type subjectively correlates to. They will typically not discuss these subjects with you.
YMMV.
Chris
So all acoustic driver performance, as viewed by the manufacturers, is seemingly always accompanied by the notion of "at this price point" discussions. Many industries are in this sort of paradigm of thinking about their largely undifferentiated product lines. They don't want you to differentiate their products based on performance, but rather price. So they don't answer your questions about what to measure and just audible distortion levels.
So to talk about measured differences between drivers, and what those differences may be, you are really going against the grain of the manufacturers' intentions for their products. I.e.,
Only "boutique" acoustic driver companies try to vie for the very small market share that relates to diyAudio loudspeaker builders. I've found most of these type of enterprises operate on a shoestring budget in terms of real engineering capabilities; e.g., basically none of them own something as capable as a Klippel R&D system to educate themselves on "just audible distortion levels" and what each measurement type subjectively correlates to. They will typically not discuss these subjects with you.
YMMV.
Chris
I fully agree with you.
I just wanted to point out that the definition of source and load (sink?) is not so clear here.
Hi Cask05,
The problem with smaller industrial manufacturers making a name is that they don't have access to as much in engineering resources as companies like JBL or Klipsch. So often they do their best and go out on a limb, but typically showcase a "star" designer and fancy story. Typically coupled with very overpriced products. There is your "high end" market, and it applies to electronics as well. The established companies also have part procurement and manufacturing down so they don't waste money. The small guys simply can't do that.
It is the case that some mass market manufacturers do make a superior product at much lower prices, coupled with better warranty support and occasionally (these days) after warranty support.
The truth of the matter is simple, every product has to have good performance "for the price". NAD isn't very good, but if you read reviews, it was good "for the price". But even a military system (attack or otherwise) has to be competitive for a price point. Same for industrial goods. A passenger jet has even more price conscious price targets to meet.
The real problem throughout all industry iss that when a company begins satisfying shareholders, their main business ceases to be the prime consideration. Just look around you these days.
The problem with smaller industrial manufacturers making a name is that they don't have access to as much in engineering resources as companies like JBL or Klipsch. So often they do their best and go out on a limb, but typically showcase a "star" designer and fancy story. Typically coupled with very overpriced products. There is your "high end" market, and it applies to electronics as well. The established companies also have part procurement and manufacturing down so they don't waste money. The small guys simply can't do that.
It is the case that some mass market manufacturers do make a superior product at much lower prices, coupled with better warranty support and occasionally (these days) after warranty support.
The truth of the matter is simple, every product has to have good performance "for the price". NAD isn't very good, but if you read reviews, it was good "for the price". But even a military system (attack or otherwise) has to be competitive for a price point. Same for industrial goods. A passenger jet has even more price conscious price targets to meet.
The real problem throughout all industry iss that when a company begins satisfying shareholders, their main business ceases to be the prime consideration. Just look around you these days.