Thank you. But what merit? How would you describe it? Asking questions in good faith.
Explain how? Another reasonable question. Please?
The explanation is very well thought out in consultation with others.
Note that "impedance compensation" is another way of saying "EQ the current" is a powerful descriptor that also tells us that the current of the amplifier is the same at all frequencies. Think F=BLi and start analysing what is really happening. Now part of the current goes through the EQ network and part of the driver. What affect does this have on the amplifier? What does it now not do, that it did before?
I am now asking the questions (six just here) and I have every right to.
Au contraire, the opposite is happening.
The criticism I am getting here is by a small minority. Not elsewhere. Can you explain that? Many are receptive to have this discussion, there is an open mind.
But then, this isn't about me, even if there are some here who wants to damage my credibility (sorry, that is not going to happen). I note that everybody ignores F=BLi and there is a deathly silence. This is a topic where we have been lazy too long.
We all have!
And:
Isn't good that we are talking about speakers? We all need them.
The EE is correct. Any electrical damping (driver's Ze + amp's Zout introduces local feedback in a driver, the more of this feedback is applied the more the driver operates in a velocity controlled operation. F=BL*i is only one part of it.
The other, way more relevant, is (dare I repeat?) : e(t) = k*v(t) + (Ze+Zout)*i(t).
Any difference between the amp voltage e(t) and the velocity voltage v(t), which happens to be distorted (mostly from BL being nonlinear in various regards), re-enters the Force impressed via F=BL*i(t), again nonlinear, BL(x) mostly. Ze+Zout being the transfer resistance which converts the voltage difference to a current... alas, this resistance also is nonlinear.
I would think it can be seen quite easily that the amount of re-entrant error, which is controlled by total termination impedance (Ze+Zout), yields two *different* distortion profiles under normal voltage and current drive, when EQ'd to the same SPL output curve. "Distortion profile" is a complex set, covering any kind of small signal errors, expansion of those with growing levels and reaching large signal behavior (steady-state and transient) as well as overload recovery. From that it looks possible to find a optimum "drive impedance profile" vs. frequency that does the best job overall for a given driver one could commonly agree on.
The external Impedance Correction adds in here, too. With a high-Z amp it modulates the frequency response to flat (main goal) and changes the damping to some reasonable levels around impedance -- that's the discussed profiling here, a bit crude but easily established with passive components.
For the amp itself it helps present a predictable and mostly ohmic load which result in a more predictable behaviour, eg. the distortion (in the output voltage) is the same as that of a true resistor. But also introduces additional load.
Yet, it is debatable if the more constant but larger overall distortion is perceptually more benign than the uncompensated error, for a standard low to medium (few ohms) Zout amplifier. With highly distorting amps it might actually happen from time to time, though...
Last edited:
I don't know as I'm probably not part of that minority, but I can think logically so may be I know.
1. Logic: If someone doesn't use the standard terminologies, that is most probably because he's not trained well and/or has read insufficient technical books.
2. If someone doesn't know what other (more knowledgeable) people already know, there will be possibility that he thinks he has invented something new and big, while actually he's not. He might also complicate things that are actually simple.
Pay attention that most critics are from knowledgeable people. You cannot assume that they don't know what you know. Au contraire, you have to know that they know more than you know about the exact topic. What you think as a fancy mechanism can be just simple Ohm Law or Kirchoff for them. Note this: The only thing that they don't know is the audibility of things!!
So you're talking about impedance compensation. You think it's very important, right? It has been done by many people decades ago, yet as you know, it is not in everyone's design trick repertoire. Why is that?
So you have a speaker designed with such philosophy. Is it 'that' good? You use FOUR expensive woofers. With or without your compensation circuit, it ought to sound good!!
Many people 'protect' the feedback network from the effect of speaker coil by adding resistor after the FB take-off point. JLH did it. But not everyone.
Last edited:
We are talking 1992 Versus 2007 !!!
Apology please?
ARE YOU ACCUSING ME OF BEING A THIEF?
(Thank you, at least you gave me a name where others slander anonymously - as this is tantamount to slander.)
OK, I have counted to 10 and calmed down. I have designed a number of different microphone preamps, both solid stage and tube. Allen made a version later and that was used by Peter Gabriel. Billshurv also knows that I did a custom six channel front-end called the "Tube Station" and that too has been used by a number of famous artists. Now that another brand is using "Tube Station" - will I also be accused in relation to that too?
Please, anybody who knows me also knows how much I miss Allen. I was there when hard times really hit him and when he recovered, both a good weather and bad weather friend. Nobody has defended Allen's legacy stronger than I have. I was also the one who registered and owned the domain name Vacuum State - High End Hifi Equipment and signed it over to his widow Hanni for ten years after he died. The website will be there at least till 2021. I have ceased to be part of the business.
My advice: Always try to think of the best in people. The times that works out will be more than when not.
Are you sure of your facts here? I find it hard to believe that this was not done in the 50s and it's easy enough to check.
I was not aware of that. You must be confusing me with someone else.
Could you please make up your mind whether the speaker changes the amplifier current (as you say here) or the speaker determines the amplifier current (as you correctly said in an earlier paragraph).Joe Rasmussen said:
Yes. Evidence of FUD.
I think you meant to say (in normal EE language) 'the impedance of the load is the same at all frequencies'. If you want to communicate with EEs why not learn their language?Joe Rasmussen said:
Maybe the people on this forum are smarter and understand EE better than those elsewhere?
On the contrary, I and others have reminded you more than once that Force does not equal Sound. You just don't listen to what we say but instead repeat your confusion. I am not sure if you believe it will become more true with repetition or you are just hoping to wear us down into silence so then you can tell people on other forums that even the experts at DIYaudio could not find fault with your ramblings.
This question has come up before so let's get to the bottom of this. In your post #27778, you state that:
"I was there, the RTP preamp with M-S matrix internally was commissioned for the job. I was told that told that it had never been done that way before."
And two lines later:
"The irony here is that all of you here would have almost certainly heard recordings where one of my custom microphone preamps have been used."
The RTP preamp is Allen Wright's design and very well documented to being so.
To come back to your question in bold:
Based on your own statement, yes, but not a very clever one.
Please prove us wrong, by just showing one of your original designs.
Last edited:
That's interesting because the EE does not agree with you.
I was thinking about revealing who he is, but that would just lead to another character assassination.
Thanks for the effort in reply and while I don't have much to pick on, it is because you are conflating issues together - but I can also see why that is understandable.
In fact that he calls this a 'feedback' is not to be the best description, but I understand what he means by it. I know that it is not what you thought it meant.
You say that EQ is crude, but in fact, putting a parallel resistor is. I speak to designers and one very famous loudspeaker designer confided that he did, but just did not tell, hush-hush. Then visiting SB Acoustics I was told by their chief designer (again I shall not mention his name) that he knew for a fact that many loudspeaker brands did this, hush-hush.
BTW, the EE I mentioned also said that it reduces noise and hinted at jitter, but he favoured adding impedance.
What do you think? A noise floor?
There are many who have heard the improvement in sound by adding a crude resistor. Richard Marsh has mentioned it several times, I recall. What is the answer?
But EQ is frequency selective and not crude (passive components as you say) and you hear the improvement even more so. But most manufacturers won't bother - trust me, I have asked. Resistors also mean you get a lowe impedance where you often don't want it. EQ on the other hand, with intelligent crossovers, you end up with a flat 6 Ohm with 8 Ohm system and 3-4 Ohm with 4 Ohm system. I have done both and extremely amplifier friendly.
(But is it easier on the amplifier other than just being resistive?)
Other than saying the improvement is all in the head (as some will say here), how would you explain it?
If the audible improvement is real, then the mechanism is also real. What is that mechanism? I would be interested in your opinion.
A similar improvement is also achieved by adding series R, but we know that series will disturb FR and a voltage amp will not.
As a thought experiment, imagine having a perfect voltage source with an infinite current capability and connect it up to a driver. Now also connect a 1 Ohm resistor or even a 0.1 Ohm resistor. The current will be humongous, but what will the behaviour of the driver be?
Here the amp will indeed see a resistive load, that is for sure. Anything else. Not a trick question or trying to trip anyone up.
PS: I am not sure, did you see the 'Equivalence' test I posted a while back?
Guys - please refrain from using color, bold or large fonts in your posts. Stop shouting! This will be your only warning.
I understand his language and he understands mine, I'm quite sure.
He used a text description of what actually happens "inside" of the formula I posted, e(t) = k*v(t) + (Ze+Zout)*i(t):
I see no point of disagreement anywhere between him and especially his explicit notion of the feedback mechanism is worthwhile because it is a key point in the amp<-->loudspeaker interface.
If otherwise, he should join the discussion personally.
You used the work perfect so now it's a fair question, but you should answer it.
No way, it's too toxic here, if you know what I mean? But I have no beef with you and in time things will work itself out.
Have you seen my 'equivalence' test? It shows a direct proportionality between current changes and dBSPL of the driver. Also between dissipation and dBSPL. It can both be tested and it can be modeled in SoundEasy (are you aware of this software?). Nobody here even tried to understand it. The maths are also surprisingly straight forward. The 'equivalence' part of it is the reference point, and why it is so powerful in demonstrating that isolating current as a value, it can be proved precisely that it is the force. SoundEasy compatibility is also important because it means that we have something that can be integrated into existing knowledge. So existing equations have to be reconciled with it and I am totally confident that it will.
What was surprising to me was that heat dissipation, because the driver is 0.5% efficient (as opposed to sensitivity), it is also proportional to the heat in the voice coil to within 1%.
I vetted it with a certain Dr of Physics who is amazingly well credentialled. I will only mention names in a PM, definitely not here. You can understand that, right? But he is way more qualified than anyone here. So I am empowered by that. Now what is needed is a measurement and that is being worked out. That is what some here have been asking for and I believe they are right to. Just be kind is what I say, and patient.
Rgds, JR
I've often wondered what hell must be like. Flames, smoke, sulphur, people screaming in agony shaking their fists at a the giant goat-headed Beelzebub who sits on a huge hold throne, roaring with laughter at their predicament. For ever and ever my severely Calvinistic parents told me whenever I strayed off the thorny path.
Or, maybe its this thread . . .
Or, maybe its this thread . . .
Yes my imaginary friend is also well qualified.
Name dropping really doesn't work if you can't use the name!
So you've done no checking of your own? Just presented a story as fact?
Since my post was ignored, I will repeat it and elaborate.
Over the last 15 years give or take, I have had to deal with EE's who also make these simple mistakes in understanding. And as I stated earlier, it is in general one of the first three possibilities I mentioned; the fourth, that of misleading, is specific to the audio world it seems.
The EE's currently in the mix out there are not receiving an actual analog engineering education anymore but rather digital. As such, any attempt to simply replace an analog understanding with the comparable digital one will lead to erroneous conclusions such as Joe's "EE" reference and KSTR's agreement.
One need only reference "Control System Design, an introduction to state space methods", Bernard Friedland, Dover, 1986, page 3, eq (1-6) where a delay element is imposed on the transfer function. This is standard control system PID loop theory, where there is multiple processor cycle time delay as the loop math is calculated. As such, the system behaves as an IIR system (for you digital guys). Indeed, the entire chapter 1.1 "Mechanism of feedback" very quickly details how insertion of a delay element means that the system stability is dependent on the system gain, as the system will break into instability once the amplifier gain is increased sufficiently. He explains it in digital form, whereas the analog method required understanding Bode plots and phase margin. Unfortunately, Bode is not even taught to most EE's or physicists less than 45 years of age..
To arbitrarily use an IIR based thinking on an analog system is to go outside ones' area of expertise (or learnings).
I am currently involved in interviews to hire several EE's and ME's, including my replacement. It is very clear where the teaching over the last 15 years or so have fallen flat with respect to analog. (and don't even get me started on high power electronics, the best candidate so far believes 200 amp 480 volt systems are high power..
) It is no surprise that such erroneous thinking would show up in the audio world..even without trying to make a buck..
jn
I think Yamaha definitely and possibly Behringer would be in front of Sony.
Yamaha have made an amazingly diverse range of music / audio products and most have been pretty good.
Behringer... um.... let me think of something good to say about them.... they were probably solely responsible for keeping the hot melt glue people in business.
"Unfortunately, Bode is not even taught to most EE's or physicists less than 45 years of age"
"(and don't even get me started on high power electronics, the best candidate so far believes 200 amp 480 volt systems are high power.. ) "
--
I read what you write quite avidly because almost all of it provides some very good in-depth information on your area of expertise - EM theory and magnetism.
But the above two statements are off the mark IMV.
The education system in your neck of the woods must be pretty lacklustre if people are not touching on feedback theory and especially fundamentals like Bode plots. Are you pitching your recruitment ads at the right level? Seriously. You can pull down lecture notes and papers by Kenneth Kundert (MIT faculty) discussing bode plots and feedback theory in depth.
You operate in a highly, highly specialized field. If you call an electronic engineer in he will see 480V at 200 A as high power whereas a heavy current engineer won't and will see 50 or 100 kV at 50 kA as high power. Hell, I know EE's that thing 100mA at 5 V is 'high power'. (Remember the Philips ads where they ran a programmable gate array off two probes stuck into a Lemon?)
"(and don't even get me started on high power electronics, the best candidate so far believes 200 amp 480 volt systems are high power.. ) "
--
I read what you write quite avidly because almost all of it provides some very good in-depth information on your area of expertise - EM theory and magnetism.
But the above two statements are off the mark IMV.
The education system in your neck of the woods must be pretty lacklustre if people are not touching on feedback theory and especially fundamentals like Bode plots. Are you pitching your recruitment ads at the right level? Seriously. You can pull down lecture notes and papers by Kenneth Kundert (MIT faculty) discussing bode plots and feedback theory in depth.
You operate in a highly, highly specialized field. If you call an electronic engineer in he will see 480V at 200 A as high power whereas a heavy current engineer won't and will see 50 or 100 kV at 50 kA as high power. Hell, I know EE's that thing 100mA at 5 V is 'high power'. (Remember the Philips ads where they ran a programmable gate array off two probes stuck into a Lemon?)
- Status
- Not open for further replies.