What suspension of laws are we talking about?
What a strange thing to say!
Too many errors there for a start, to even correct.
A simplified way of speaking. I can speak that way to my physics professor buddies and they get the drift. Obviously, cause and effect don't work exactly that way in reality.
I'm sure we have people here capable of understanding any technical claims you make. So far nobody has stepped up to say Joe is exactly right, everybody should listen to him. Why not?
I'm sure we have people here capable of understanding any technical claims you make. So far nobody has stepped up to say Joe is exactly right, everybody should listen to him. Why not?
Last edited:
Clearly you didn't even read my posts. I didn't suggest current drive. Far from it. The truth is that we are using a source that controls voltage and not current. I am using amplifiers that are have no control over the current - most of us are. Unless you are listening to electrostatics.
So if the amplifier outputs voltage and the driver tracks the current (over which the amplifier has no control), then IF you are experts, or think you are, then tell me how this works?
What claim did I make? This sounds awfully like straw man stuff.
Joe, I know. Understood. Amplifiers use voltage feedback to make them function as dynamic voltage regulators. Also, understood you do not recommend going to current feedback to make them into dynamic current regulators. One could also describe such things as approximations of ideal controlled voltage or current sources, respectively.
Didn't you say something about wanting networks in parallel with speakers to make the effective speaker load look resistive to the amplifier?
Not much actually, mostly a bunch of arm waving with one equation. In other words, more like a salesman than an engineer. I understand some people wear more than one hat quite well, but we appreciate engineering talk here.
No fog here, when a terminal voltage is impressed the current is generated via the static voice coil impedance. The current produced at any given moment is the terminal voltage minus microphonic voltage ("back-emf") divided by this impedance. Since the current (and only the current) drives the motion which in turn causes the microphonic voltage, we have a feedback system, the lower the static impedance the larger the feedback, degenerative feedback that is. The static impedance is the system's V->I transfer impedance.
Alas, both the "static" impedance isn't that static and microphonic voltage isn't purely proportional to velocity, that's where all the problems come from and a certain impedance at any given frequency will give best results overall.
You did read my posts #2977 and #2984, didn't you?
Last edited:
Absolutely correct, the exact point I am making, they 'regulate' voltage.
Thank you, I want something that works in the real world and current driving is not the answer either - but whenever I bring this up, instantly I get put into the school and I mildly protest.
Yes, I said something really revolutionary there (I am being cynical of course). But think about it, if constant current drive is not the answer, and the reasons we all know are very numerous, then if the amplifier cannot control the current and indeed the amplifier just obliges with any current that the load demands, then why not force the amplifier into supplying constant current with frequency? Yep, that revolutionary idea (being cynical again) of making the effective speaker load resistive accomplishes exactly that goal. Are there benefits?
Why do you say stuff like that? Actually, I was told that a scientist trumps an engineer by 'Julf' a few years ago. Well, in the above suggestion I was actually channeling somebody who is both an engineer and a scientist. So I strongly suggest that not only was my 'talk' based sound on engineering, and ask why you think not?
The truth is that the use of EQ of current is becoming, very gradually, something that some speaker designers are already doing now, some in a very crude way and others in more sophisticated ways. Call it progress or whatever.
The question though is this: Does the amplifier sound better when voltage and current are in sync when leaving the amplifier - and I mean a voltage amplifier? The above scientist says "yes"! - and especially with Class A-B amplifiers where the output devices has to switch the current on and off. His argument is sound, because if the load is not resistive, then you have a timing problem - and hello, is that so hard to see?
But maybe less beneficial to Class A and Class D amps.
But is there also a benefit to the speaker?
That discussion is definitely happening - so both amplifier (A-B type) and speaker, which is still the most common combination most of us listen to, then we have a potential double-benefit. That would be nice.
Personally, I think it is fair to say that instinctively, if the current isn't right, and the drivers used are definitely current devices, then perhaps we could find something? Has it all been thought out?
So, over to you guys, if the amp is a voltage device, is it beneficial to the speaker and not just the amplifier, to get the current that the amplifier supplies, with both current and voltage in sync, and that means EQ because the current must be the same over a broad span of frequencies?
Joe, It was my impression. However, I like what you just said much more.
Regarding people thinking you want current regulation, I believe the way you a approach the subject when talking about it tends to leave people with that impression. You talk about use of voltage amplifiers when force is due to current. You seem to focus on that early and persistently, and leave people to form their own conclusions about where you are going with it. It looks a lot like it probably would if you were going towards favoring current drive.
The other thing about making the load appear resistive is something you were much less clear about and also less persistent. At first there was a guessing game about square waves which turned off some people from interest in further listening. The parallel network was only mentioned very briefly in one post, IIRC.
So, while it is probably quite clear to you what you think and where you are going with your posts I don't think there was one post where you just laid it out all at once as well as you just did now.
Indeed I did and I think you took my 'fog' quote out of context.
Has it occurred to you that any impedance above the DC resistance is caused by back-emf, so take any frequency and measure the impedance, and if you also know the DC resistance, you have now made back-emf quantifiable. Effectively, the current (assuming a single driver) is determined by the load of the driver, but the dB-SPL that is produced by the driver at that frequency is relative to the DC resistance and the voltage that formed across it. This is in fact easy to prove in computer modeling (like SoundEasy) and hence it is perfectly in line with current that produces dB-SPL and not the voltage as such, the voltage becomes incidental to the current - and that voltage is actually not that of the voltage that appears across the driver's terminals. Use a 0.1R current sense resistor, and if the VC is 6 Ohm, multiply the voltage across formed across the 0.1R by sixty, and you have a reconstructed version of what is really happening and what affects the final dB-SPL. I have physically done this and very repeatable. Now we should be able to compare the voltage of the amplifier as the reference (in terms of volts, the amp is the reference) to that voltage times sixty. We can see the current phase shift there etc. As for the many driver imperfections, they modulate the back-emf impedance and not-so-nice things happen to the current and not hard to guess how the final dB-SPL is affected.
I am now aware that others are also looking at the above. Let's not be hasty if it leads to a better understanding of some kind. At least it can't hurt.
The real answer is to produce a better transducer - some sort of near perfect piston, perhaps with a new type of drive mechanism (digital in nature, perhaps?) with accurate motion feedback. I reckon we have the materials science and technology to do it, but the number of people who would care is so small that it would not be worth the money...
It's relatively easy to put DIY effort into vanishingly small levels of distortion in amps and DACS, electricity-to-sound transducers are hard to DIY so don't get the same attention...
It's relatively easy to put DIY effort into vanishingly small levels of distortion in amps and DACS, electricity-to-sound transducers are hard to DIY so don't get the same attention...
Yes, for instance an electrostatic speaker is capacitive and the voltage and current are 90 degrees out of phase. The fact that driving a complex impedance will always have an angle between voltage and current is not a "timing problem", it's physics. Back-emf does not "cause" impedance, it is better to consider separately the reciprocity/damping of the motor/generator system and work that into your model, because for sure short of some sort of motional feedback you are only going to be able to use passive R/L/C's in front of the speaker. No matter what you do the voltage and current in the voice coil will do what they want.
Some of these statements are kind of "wouldn't be nice if this was the way things worked". I'm reminded of an argument from long ago that a more efficient speaker "has" to sound better. There is no technical basis for that statement.
In the end I don't see what the problem is, make speakers that way and see what everyone thinks.
Last edited:
I think he may be getting confused about voltage feedback to a transconductance amplifier and sinusoidal phase shift at the output.
When the load is a capacitor rather than a resistor, the amp has to output a higher current to cause the same rate of voltage change, so the feedback has to cause more output current and hence the input/output voltage error has to be larger. Of course, this is not the same as a time delay.
Last edited:
OK, the physicist who said that led me up the wrong way. I see.
No.
- Status
- Not open for further replies.