Jorge,
As You may have noticed, I agree with You in what You say. I just wanted to point out, that this is a tradeoff. For a LOW BASS region resonance damping we are going through all this troubles generated by a low output impedance. We are not able to damp any cone resonances; box resonances; cone breakup resonances. But all this generates it's own EMF response, and the amp will try to react to all of them, too!! Not only to the fundamental resonance, in case of which it is effective. [In alterating it..]
And the crossover resonances, which are either damped or not by the actual speaker impedances. In any case, these crossover resonances ARE better damped in case of an existent, non zero output impedance of the driving source.
Then, that Booom is not a distortion, it's a tone control...
Ciao, George
As You may have noticed, I agree with You in what You say. I just wanted to point out, that this is a tradeoff. For a LOW BASS region resonance damping we are going through all this troubles generated by a low output impedance. We are not able to damp any cone resonances; box resonances; cone breakup resonances. But all this generates it's own EMF response, and the amp will try to react to all of them, too!! Not only to the fundamental resonance, in case of which it is effective. [In alterating it..]
And the crossover resonances, which are either damped or not by the actual speaker impedances. In any case, these crossover resonances ARE better damped in case of an existent, non zero output impedance of the driving source.
Then, that Booom is not a distortion, it's a tone control...
Ciao, George
Mauro
It took me some time to post a better looking plot (image was almost invisible...).
My conclusion is that your amp, with a more classical topology (not trying to increase output impedance, or almost like your first version), is a very nice implementation of a dual amp feedback loop, with lots of NFB.
There is a paper on the WEB about it somewhere.
It just may be that the output amp is happier with the added series resistor (more stable, whatever) and sounds better with it.
But not due to the increased output impedance. From the top of my head, the open loop output impedance has changed significantly between your first and second circuits with open NFB loop.
It took me some time to post a better looking plot (image was almost invisible...).
My conclusion is that your amp, with a more classical topology (not trying to increase output impedance, or almost like your first version), is a very nice implementation of a dual amp feedback loop, with lots of NFB.
There is a paper on the WEB about it somewhere.
It just may be that the output amp is happier with the added series resistor (more stable, whatever) and sounds better with it.
But not due to the increased output impedance. From the top of my head, the open loop output impedance has changed significantly between your first and second circuits with open NFB loop.
George
The way I see it:
Both high output impedance and low output impedance topologies (high and low under open loop conditions) can give the same damping value.
The high output impedance will require more feedback to achieve the same value of damping.
And the back EMF (whatever the reason may be) in the low output impedance case will generate less internal error signals in the low impedance case (for a given back EMF generator), so it shall be preffered.
The way I see it:
Both high output impedance and low output impedance topologies (high and low under open loop conditions) can give the same damping value.
The high output impedance will require more feedback to achieve the same value of damping.
And the back EMF (whatever the reason may be) in the low output impedance case will generate less internal error signals in the low impedance case (for a given back EMF generator), so it shall be preffered.
Joseph K said:
... but the path IS closed! Look in post 269, the nice model developed by Rodolfo. Somewhere in this model, it is not important where, there is a voltage generator that generates the EMF. I think so far we all agree. Where does the current from this generator go? Well, several paths. Surely through the components making up the speaker and xover. And, if possible, through whatever external impedance we have connected, like the power amp output Z. Now, if that output Z is infinite, there will be no current through that output stage as a result of the EMF. If that external Z is zero, like in a voltage amp, ALL current (1st approx) goes through it. We call that damping
.Can we agree that the resulting current through the speaker/xover components consists of current as a result of the current source amp, (which as you stated doesn't care whatelse we do, it just opumps out the required current), and the EMF?
Jan Didden
Mauro
There are so many factors envolved that it becomes difficult to give a straight answer.
My personal viewpoint, as posted above in a reply to George, is that a 'short circuit' is better.
But then the output stage has to be able to 'short circuit' without any significant distortion or instability under real loads.
If a given circuit is not so efficient as a 'short circuit' but generates less overal distortion and is more stable than a better 'short circuit' one, it will sound better.
There are so many factors envolved that it becomes difficult to give a straight answer.
My personal viewpoint, as posted above in a reply to George, is that a 'short circuit' is better.
But then the output stage has to be able to 'short circuit' without any significant distortion or instability under real loads.
If a given circuit is not so efficient as a 'short circuit' but generates less overal distortion and is more stable than a better 'short circuit' one, it will sound better.
Jan,
in my reply I was concentrating on that [clear] situation when only a speaker is connected to the amp. It is true that crossover parts are closing the circuit, and modifying the picture. [like in impedance correction network across speakers]. And also such elements are "stealing" current from the driving high Z out amplifier, so they are having an influence also in this meaning. But it's true also for the voltage out case.
I think it's important where the generated tension is presented in the model of Ingrast. I think, for example, that in case of the woofer it would be across R2; L2. In this way it becomes true that a high impedance drive would "switch off" the effect of L1, C1,R1,L3. [without crossover]
Ciao, George
in my reply I was concentrating on that [clear] situation when only a speaker is connected to the amp. It is true that crossover parts are closing the circuit, and modifying the picture. [like in impedance correction network across speakers]. And also such elements are "stealing" current from the driving high Z out amplifier, so they are having an influence also in this meaning. But it's true also for the voltage out case.
I think it's important where the generated tension is presented in the model of Ingrast. I think, for example, that in case of the woofer it would be across R2; L2. In this way it becomes true that a high impedance drive would "switch off" the effect of L1, C1,R1,L3. [without crossover]
Ciao, George
Joseph K said:
George,
Same difference. The speaker itself also closes the path, even when naked without any xover or networks across it. It is true that extra components modify the picture, but the basic tenet is that with a current source amp, the EMF may not disturb the amp, but it definitely DOES influence the speaker current and thus the sound. It is the same as saying that there is no damping, and the same mechanisms are involved.
Jan Didden
Jan,
With the second half of your statement I can agree. The coil movement generated tension rappresents a signal proportional to the REAL movement of the coil [not the cone..] The drive current in itself does not DEFINE the voice coil speed - it's only the excitation signal, and the mechanical response is uncontrolled. With a feedback through EMF we can try to actually control this, this will be called damping. Now the old question comes back: is it better open loop or feedback?
Ciao, George
With the second half of your statement I can agree. The coil movement generated tension rappresents a signal proportional to the REAL movement of the coil [not the cone..] The drive current in itself does not DEFINE the voice coil speed - it's only the excitation signal, and the mechanical response is uncontrolled. With a feedback through EMF we can try to actually control this, this will be called damping. Now the old question comes back: is it better open loop or feedback?
Ciao, George
Joseph K said:
Exactly. As to that 64k$ question, I would prefer voltage drive, because I think that the distortion resulting from the influence of the EMF on the nfb loop is much less than the influence of an uncontrolled flapping cone on the sound.
That said, I realise that the two influences are not the same. The low damping will tend to give a non-linear freq characteristic, while the influence on nfb could possibly generate amp internal effects more in the harmonic components domain, so you really can't compare the two.
But I think that in a competently designed amp, there is no audible effect from nfb influence, after all that EMF is limited, by physical reasons, to relatively low freq (certainly within the audio band) components, and the amp-nfb should have no problem with that. It's not that the speaker throws kilo-amps at the amp!
There may be specific combinations of current drive with specific types of speakers & enclosures that minimise EMF effects, but why look for trouble?
Jan Didden
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.