Joe, Maybe it would help in thinking about some of this stuff to recall something very basic from physics: the extent to which the voltage and current in a speaker are in phase is an indication that power is being dissipated inside the speaker system. That could either by because of frictional or resistive losses, or because the sound waves it produces are launching or coupling energy outside and away from the speaker itself.
Likewise, to the extent voltage and current are out of phase, it means energy is being temporarily stored in the speaker system (and maybe its local environment) until it is released back out (perhaps only a fraction a cycle later). In addition, when a speaker has a resonance energy storage/release process can occur over a much longer time period.
Hopefully, that more or less makes sense. Of course, if I misspoke at all please do correct me. The idea is to get to some some shared understanding of the actual physics.
Once we can agree on some basic principles, we have a big picture and we know our more detailed, lower level understanding will have to be consistent with that.
Likewise, to the extent voltage and current are out of phase, it means energy is being temporarily stored in the speaker system (and maybe its local environment) until it is released back out (perhaps only a fraction a cycle later). In addition, when a speaker has a resonance energy storage/release process can occur over a much longer time period.
Hopefully, that more or less makes sense. Of course, if I misspoke at all please do correct me. The idea is to get to some some shared understanding of the actual physics.
Once we can agree on some basic principles, we have a big picture and we know our more detailed, lower level understanding will have to be consistent with that.
Last edited:
Hopefully not. I just don't want to make Joe feel unwelcome when he is starting to get more comfortable talking with us.
No, it's more fundamental. BTW if your still interested I thought of something with reference to the error correcting output stage, the difference in re-entrant distortion.
I hope you don't think I am anti-physics or whatever. But I can tell you that there is computer modeling and actual tests that shows a direct (and powerful) relationship with the non-volatile impedance that is the DC resistance and the voltage that appears across it and the dB-SPL variation. The method of extracting that using a current sense resistor not only works, but validates it. I have literally done this on my bench. The fact that SoundEasy, which is a pure maths based program and I would think above reproach, can actually model this as well - makes me very confident (and if you replicated it, so would you). It is a bit too complex to explain on a forum like this where (as JC pointed out) there is a tendency to shut down certain non- approved topics (I tried it elsewhere on diyaudio and it got 'the treatment' there too).
So look at an impedance plot, measure the DC resistance, which is constant with frequency. It is not frequency dependent, but it is thermally dependent, but that is a different topic. Yet the overall impedance is indeed frequency dependent, but clearly not all of it. So if Re is 6 Ohm, the RMS voltage is 6V, then under ideal conditions it should draw 1 Amp. And if current drive equivalence, by sending 1 Amp, the 6V should develop across the voice coil terminals. Yet the V/I conversion does not take place at the terminals if the impedance is above 6 Ohm. Please have an open mind and ask what is really then happening. The overall impedance determines the current drawn from the amplifier (I am surely on safe grounds there), so if the impedance is 12 Ohm, then we have 6 Ohm of stable resistive impedance and another 6 Ohm of volatile back-emf impedance. The current will be 500mA, half the current - the dB-SPL of the driver will change by 6dB and then the penny drops, the volt times current across the DC resistance accounts for that 6dB exactly. You model it and that is how it works, then you do it physically on the bench, and it does the same. The physics is clear - the DC non-volatile impedance times voltage = dB-SPL - and it is not just theory, but a verifiable fact that anyone with fairly basic equipment can do.
If you don't find that interesting, then don't blame me if I feel otherwise.
Last edited:
Yep 🙂
The point I was trying to make is that neither pure current drive nor standard voltage drive (let alone pure velocity drive) might be the best for a given driver, rather its all a tradeoff and we have an free variable in the game which not everyone is aware of, and that's the drive impedance profile vs. frequency.
By shaping this profile to get overall "best" distortion (HD and notably IMD) at any given range of usable frequency and operating point (excursion and power) we might improve the fidelity of the driver significantly. What the best distortion behavior is supposed to be is then another topic open to discuss.
With a given terminal voltage (or current, for that matter) the SPL output will be the same no matter what the drive impedance is, so for comparison we'd need to EQ this voltage/current curve to whatever fixed target. But "fine-print" in the sound will be different, notably when we're about to leave the linear range of any parameter involved.
Would you like his email? You could ask him yourself. 🙂
I can assure you, I didn't misquote him. He is Dutch, I seem to recall you not liking the British much (why not?), so that might help? 😀
What does that mean, exactly? The "frequency" of DC resistance is 0Hz, by definition.
Right, because the impedance of the driver is not resistive, it is a complex impedance that includes capactive and inductive components.
No, that is wrong, who told you that? That would only be true at DC (in which case you wouldn't be measuring voltage as RMS). I know you understand the difference between resistance and impedance, not sure why you are saying thiis.
Joe,
When you say things like, "...the non-volatile impedance that is the DC resistance," and, "The method of extracting that using a current sense resistor...," it makes me think we are not speaking the same language. Until we solve that problem we will always have trouble communicating. May I ask about your background in the areas of electronics and engineering? (No disrespect intended, I'm just trying to get a better picture of where you are coming from.)
When you say things like, "...the non-volatile impedance that is the DC resistance," and, "The method of extracting that using a current sense resistor...," it makes me think we are not speaking the same language. Until we solve that problem we will always have trouble communicating. May I ask about your background in the areas of electronics and engineering? (No disrespect intended, I'm just trying to get a better picture of where you are coming from.)
So IOW, Joe's discussion/question/experience/advice is how does driver accuracy (lack of distortions) and FR (acoustical output) change WRT change in source impedance vs frequency.
Dan.
Dan.
Last edited:
Ok. I was considering your question:
The speaker impedance does affect amplifier performance, pure resistance is typically more benign. You may have to explain what you meant by timing, although I see you may have withdrawn that now.
It would be a waste of time. Class A/B amplifiers (or any amplifier for that matter) supply current when the system/load demands it. The phase of the load moves load lines around, changes the instantaneous power, etc. all easily computed by those skilled in the art. The idea that class A/B amplifiers "switch" and somehow if the load has a non-zero phase they are "out of step" is a nonsense, I don't care who said it.
In fact and you can look for yourself I made the comment that some British audio engineers rarely reference any work outside of their "club". Nothing personal, there is ample evidence for this.
Joe, you make it very hard not to be rude to you when you seem to promote the idea that one needs no understanding of the underlying physics and math to do audio design. All the things you talk about are easily analysed analytically. If you don't understand complex algebra a lot of things will make no intuitive sense.
Yes, and trying to make a complex output impedance amplifier rather than something closer to an ideal voltage or current source is going to require extremely good characterization of the speaker system.
Edit in reply to this:
Probably better to look at dsp based error correction from look up if you want to move that way without getting into motional feedback.
Edit in reply to this:
Probably better to look at dsp based error correction from look up if you want to move that way without getting into motional feedback.
Last edited:
Well I tried a bunch of simulations with a pure cubic non-linearity inside the loop and even with mismatches and imperfect cancellation I can't get any 5th harmonic in any of the voltages inside or outside the loop. A cubic non-linearity inside a conventional feedback loop will always have monotonically decreasing odd order distortions due to re-entrantcy. (I'm sure that's not a word). 🙂
- Status
- Not open for further replies.