Lets take the Purifi case over. When I write the speaker sees a large Z that is the Z of the series RLC. If a Parallell RLC is used the speaker sees a small Z.
It is a jargon used for example when discussing ladder filters
Maybe a better explanation is the Z you measure at the "speaker terminals" if the speaker is removed. That is the Z that the speaker sees
Last edited:
Its not clear to me that the Grimm Audio measurements relate to my own measurements. As I mentioned it is crucial when comparing High Z and Low Z drive that the end frequency responses are exactly the same. When doing so the distortion levels from most non-linearities are identical, the steel related current distortion being the one exception. I can't see how those results would be any different between any drive units available and from any era. It would be nice if someone wanted to duplicate the experiment.
Yeah, on/around that particular frequency you could say its roughly current drive. If impedance of the driver was 8ohms at that frequency, and you now had additional 32 ohms in series for easy math, you would have reduced current from back emf at that frequency five fold, you'd measure fove fold distortion reduction. Frequency response would drop five fold as well because also the input current dropped with constant voltage amp, unless you use five times voltage to maintain current and excursion and acoustic output. But in this particular case the breakup boosts it so its desirable to reduce current and acoustic output to EQ the breakup.
Is it high enough ratio to say it's current drive? depends who defines what it needs to be. But does it matter, you've just utilized same distortion reduction mechanism with voltage amplifier. Even if you had just doubled the driver impedance, (related) distortion would have slashed in half.
Is it high enough ratio to say it's current drive? depends who defines what it needs to be. But does it matter, you've just utilized same distortion reduction mechanism with voltage amplifier. Even if you had just doubled the driver impedance, (related) distortion would have slashed in half.
I find that a very strange and confusing way of looking at it.
It's all about either shunting/terminating the back EMF that's being reflected in sense of a constant current source, or absorbing it with a passive network.
The properties of a current driver system are and act fundamentally different.
Just the effect in this very specific case are kinda similar.
Emphasis on the word kinda.
A contant voltage source with a big internal resistance doesn't magically become a constant current source.
Only under certain very specific conditions we can get a decent approximation.
you are right, adding resistor doesn't make voltage amp to current amplifier, but the distortion reduction mechanism is the same for both, high impedance in series with the driver! That is why I've been writing forget about power amplifier, abstract it away with equivalent resistor to understand hownit plays out, how Le(x) shows in acoustic domain.
high circuit impedance prevents driver emitting (some of) it's own motor distortion into acoustic domain.
Think about it. If you hook up 0.01% distortion amplifier to a driver and measure 1% distortion woth mic. Add impedance with resistor and measure 0.1% distortion with mic. Do you think amplifier is responsible of the distortion reduction, or the added impedance? and why is that?
high circuit impedance prevents driver emitting (some of) it's own motor distortion into acoustic domain.
Think about it. If you hook up 0.01% distortion amplifier to a driver and measure 1% distortion woth mic. Add impedance with resistor and measure 0.1% distortion with mic. Do you think amplifier is responsible of the distortion reduction, or the added impedance? and why is that?
Last edited:
I literally just answered that in the previous post 😉
I would say both are just two different solutions to solve a problem.
But that doesn't make them the same.
Otherwise a towel, a t-shirt and some old pants would be also the same.
You can wipe grease with all of them.
Well, I'm not saying voltage amplifier is current amplifier. I'm using common denominator, generalize whats important so I can use the generalization to my advantage and not get confused in details.
I'd just say use fabric to swipe oil. Its irrelevat if it's t-shirt or towel as long as the grease got swept. Similarly, distortion drops with series impedance with driver.
You could use which ever, there is likely some other reason, or compromise involved why you'd use one or the other. If there is no towel at hand and must get rid of grease then t-shirt it is. If the grease is not that bad then perhaps leave it, clean it up later if it turns to be an issue.
I'd just say use fabric to swipe oil. Its irrelevat if it's t-shirt or towel as long as the grease got swept. Similarly, distortion drops with series impedance with driver.
You could use which ever, there is likely some other reason, or compromise involved why you'd use one or the other. If there is no towel at hand and must get rid of grease then t-shirt it is. If the grease is not that bad then perhaps leave it, clean it up later if it turns to be an issue.
For sure, thats obvious and has nothing to do wiping grease 🙂
I'm participating this dicussion with hope to help give insight to this issue, which seems to be complicated but really isn't. It gets into this silly post fight that doesn't reflect the intelligent people behind and doesn't bring value to anyone. I'm guilty of it, I haven't been able to summarize it well.
I'll stap aside for some time to leave room for others who might want to discuss about. There seems to be interest and knowledge on the subject like torgeirs and speaker dave have shown. Also you b_force, and allenB have show interest.
I'm participating this dicussion with hope to help give insight to this issue, which seems to be complicated but really isn't. It gets into this silly post fight that doesn't reflect the intelligent people behind and doesn't bring value to anyone. I'm guilty of it, I haven't been able to summarize it well.
I'll stap aside for some time to leave room for others who might want to discuss about. There seems to be interest and knowledge on the subject like torgeirs and speaker dave have shown. Also you b_force, and allenB have show interest.
Voltage drive is useful at low frequencies because it compensates for variations in net driving force due to cone resonance(s) and variations in BL (with excursion, i.e. BL(x)). This is because of the back-emf. So, where there's a reduction in BL as the voice-coil moves away from its rest position, back-emf will be reduced and drive current increases - and vice-versa. And when resonance reduces the mechanical load on the voice-coil, the tendency for voice-coil velocity to increase is reduced by the increase in back-emf, and consequent reduction in drive current.
At higher frequencies, variations in impedance due to drive current and voice-coil position (Z(i) and Z(x)) come more into play, so swapping voltage drive for current drive would work in maintaining constant drive force.
You can get the best of both worlds in the case of (most) OB designs, as a series inductor will not only provide the requisite EQ, but allow connection to a standard voltage-output amplifier to achieve voltage drive at low frequencies, moving to an approximation to current drive at higher frequencies.
Which might explain why this forum's Manzanita OB is said to sound better with a traditional inductor than with active EQ.
At higher frequencies, variations in impedance due to drive current and voice-coil position (Z(i) and Z(x)) come more into play, so swapping voltage drive for current drive would work in maintaining constant drive force.
You can get the best of both worlds in the case of (most) OB designs, as a series inductor will not only provide the requisite EQ, but allow connection to a standard voltage-output amplifier to achieve voltage drive at low frequencies, moving to an approximation to current drive at higher frequencies.
Which might explain why this forum's Manzanita OB is said to sound better with a traditional inductor than with active EQ.
Here is some of the core of the discussion and the case that I think is dificult to embrace. From a voltage perspective and transfer function that is the result I expect:
If voltage at a frequency f is reduced fivefold, distortion at f is reduced fivefold. But then you could instead use a active filter with the same fivefold notch and get the same effect.
The "magic" is that the distortion is reduces for f/3 with series filter and not with equivalent parallell filter. I am sure a active notch filter would give the same result as parallell filter. (No in band reduction with out of band filtering)
I now understand that it is reduced EMF and steel related current distortion (hysteresis) that explaines the whole thing.
And shunting the EMF and drive voltage to ground with parallell filter helps at f but not at f/3
Isn't that a bit obvious, since there is nothing to shunt because there IS no back EMF at f3?
So a usecase that valid in my opinion is a low price active system where the designer has some trouble with audible noise from mid driver at 3 to 6 kHz where the ear is very sensitive. It may be amp or peak in driver responce.
If the designer inserts a serial filter with -3dB at 1.5k (RLC or L) the noise will be decreases at 1.5 to 6 kHz, but the 3. harmonic distortion from 0.5k to 2 kHz stimuli will also be decreased the same amount as the noise at 1.5 to 6 kHz. Since it is an active filter the falling responce from 1.5 kHz can be corrected with a compensating active filter. Designer must then take a choise at what x over frequency to use based on noise + the other standard considerations.
Hopefully he will have inexpensiv silent mid with very low midband distortion as an added bonus.
If the designer inserts a serial filter with -3dB at 1.5k (RLC or L) the noise will be decreases at 1.5 to 6 kHz, but the 3. harmonic distortion from 0.5k to 2 kHz stimuli will also be decreased the same amount as the noise at 1.5 to 6 kHz. Since it is an active filter the falling responce from 1.5 kHz can be corrected with a compensating active filter. Designer must then take a choise at what x over frequency to use based on noise + the other standard considerations.
Hopefully he will have inexpensiv silent mid with very low midband distortion as an added bonus.
I mean stimuli at f/3. Where f is filter midfrequency. Thougth EMF followed movement of coil. That means fundamental f/3 + f*2/3, f, f*2 and other harmonics?
It's all about getting rid of the fundamental resonances of the cone itself, which is than being reflected back into the motor.
So if there is a substantial resonance at 3x fundamental cone resonance, it does something. Otherwise it won't.
The distortion harmonics are a result of, not the cause.
Agree that harmonics are a result and not the cause. But shunting harmonics to ground is done in DC decoupling so not so obvious for some.
I don't agree that this is intuitive and obvious even if it is a scientific fact.
I don't agree that this is intuitive and obvious even if it is a scientific fact.
So you think if there is no big resonance an equivalent aktive filter, parallell filter and serial filter will give the same 3.rd harmonic distortion result. And there are no conection to the experiment of speaker dave? (Dont think he mentioned there had to be a substantial cone resonance)