The "Elsinore Project" Thread

Indeed they are! But using any series impedance like that intelligently and by design, can be used to reduce distortion on the current side.

But there is a trade-off that must always be kept in mind.

It will improve 'control' of the current at the expense of having less control over the voltage.
hello All,

Interesting thoughts.

Using a voltage amplifier to supply a 4R resistor in series the 4R XT25 tweeter loses nothing in voltage control.

The voltage amplifier still outputs a controlled fixed gain voltage. The voltage drop is divided between the added resistor and tweeter. The current is reduced because of the series resistance.

A transconductance amplifier is a different story. Output impedance would decrease. Output voltage would increase. Current would remain unchanged.

Thanks DT
 
Using a voltage amplifier to supply a 4R resistor in series the 4R XT25 tweeter loses nothing in voltage control.

The voltage amplifier still outputs a controlled fixed gain voltage. The voltage drop is divided between the added resistor and tweeter. The current is reduced because of the series resistance.

A transconductance amplifier is a different story. Output impedance would decrease. Output voltage would increase. Current would remain unchanged.

Ah, but it does.

The voltage amplifier still outputs a controlled fixed gain voltage. The voltage drop is divided between the added resistor and tweeter. The current is reduced because of the series resistance.
Correct!

A transconductance amplifier is a different story. Output impedance would decrease. Output voltage would increase. Current would remain unchanged.

Incorrect on the first part, but the rest is correct. Output voltage will increase because the reactive (usually inductive) impedance, and yes, the current is unchanged, which really means the current is fixed and the voltage is not.

The output impedance of a current source amplifier will be many times that of the load impedance and thus increase, not decrease as you said.

I know this is not easy. Think about it, we can use a resistor to model what an output impedance is, we can even use an actual resistor to make a voltage source behave like a current source. But then we need a voltage source that can swing huge voltages because that resistor will also react as a voltage divider.

But an output impedance is not always the same as resistance.

We can use feedback and make it have zero actual series output resistance and yet have high output impedance (as my transconductance design does, by placing the loudspeaker inside the feedback loop with a current sense resistor). In that case, the full voltage swing will be available to the load (40 Watt with my design). Yet the speaker looking back will see hundreds of Ohm output impedance.

The ratio between the source impedance and the load impedance must be high. You lose control over the voltage, but you gain control over the current.

Welcome to the strange world of "thinking current.

But keep thinking about it, over and over again, and you will eventually figure it out. I know because I went through the same mental process.

Cheers, Joe
 
  • Like
Reactions: 1 user
Ah, but it does.


Correct!



Incorrect on the first part, but the rest is correct. Output voltage will increase because the reactive (usually inductive) impedance, and yes, the current is unchanged, which really means the current is fixed and the voltage is not.

The output impedance of a current source amplifier will be many times that of the load impedance and thus increase, not decrease as you said.

I know this is not easy. Think about it, we can use a resistor to model what an output impedance is, we can even use an actual resistor to make a voltage source behave like a current source. But then we need a voltage source that can swing huge voltages because that resistor will also react as a voltage divider.

But an output impedance is not always the same as resistance.

We can use feedback and make it have zero actual series output resistance and yet have high output impedance (as my transconductance design does, by placing the loudspeaker inside the feedback loop with a current sense resistor). In that case, the full voltage swing will be available to the load (40 Watt with my design). Yet the speaker looking back will see hundreds of Ohm output impedance.

The ratio between the source impedance and the load impedance must be high. You lose control over the voltage, but you gain control over the current.

Welcome to the strange world of "thinking current.

But keep thinking about it, over and over again, and you will eventually figure it out. I know because I went through the same mental process.

Cheers, Joe

Transconductance amplifiers use current-voltage feedback. Current through a sense resistor makes a voltage drop across the resistor. That delta V is feedback at the +input of the current output amplifier. This feedback loop is shown in your design.

With increased load impedance the current output amplifier compensates with decreased output impedance. The output voltage gets a tap on the shoulder and increases to maintain the output current.

Thanks DT
 
Transconductance amplifiers use current-voltage feedback. Current through a sense resistor makes a voltage drop across the resistor. That delta V is feedback at the +input of the current output amplifier. This feedback loop is shown in your design.

With increased load impedance the current output amplifier compensates with decreased output impedance. The output voltage gets a tap on the shoulder and increases to maintain the output current.

Thanks DT

As you know, I built one. Not to listen to, but as a tool. Don't get me wrong, it does not sound as bad as many other amps (I have been spoilt listening to a really good amplifier).

The output impedance could be described as a feedback trick. The irony is that to get extremely low output impedances, you need a lot of feedback and if you want high output impedances, then you need the same - high open loop gain.

But that is not my preferred way.

If you EQ the current of the amplifier, then it does not matter what the output impedance is. From zero to infinity.

So I only use amplifiers that have no global or interstage feedback at all. I don't get very low output impedance and don't need it. It is not high enough to be a current source. We are generally talking about 1.5 Ohm to 5 Ohm. The bass is just great. I mean, really great!
 
The output impedance could be described as a feedback trick. The irony is that to get extremely low output impedances, you need a lot of feedback and if you want high output impedances, then you need the same - high open loop gain.

Hello Joe,

The transconductance amplifier is intended to have high output impedance.

I see the current-voltage feedback as being there to control the output current, reducing the output impedance only enough to maintain the output current.

Thanks DT
 
You raise a number of things. Yes there are designs about there where the driver has no low-pass filter, sometimes the driver is designed exactly for that purpose. I can't really comment too much, and while adding an inductor and lower distortion, it might also upset the crossover point. Typically you could end up with a suckout in the upper mid or lower treble area. If you have a way to measure the response and add the 0.33mH inductance and see what happens, then it might be worth playing around with it. But beyond that, I can't say too much.
Fair enough - unfortunately I do not have the equipment to do that.. thanks a lot for taking the time to respond.
 
Addition to my yesterday post: Problem is fixed, it was the easiest thing ever, I confused the polarity of the tweeter.
Now everthing is like it should be and sounds as I hopped it would do. Or even a lot better.

It's like some guys allways write , go and build something, preferably a set of Elsinores ULD. Guess this is the end of my long journey for the best speakers.

Good to know that you were able to identify the problem quickly and fix it.

After you got some more listening done please share your impressions. Enjoy!
 
Hello Joe,

The transconductance amplifier is intended to have high output impedance.

I see the current-voltage feedback as being there to control the output current, reducing the output impedance only enough to maintain the output current.

Thanks DT

That's right, but in order to have control over the current, the current source amp must relinquish control over the voltage. The higher the output impedance, the more control over the current.

OTOH, a voltage source amplifier controls the voltage at the expense of having no control over the current. The max control over the voltage is achieved then the output impedance is zero Ohm.

So adding 4 Ohm in series with a tweeter, you have really modified the output impedance to just over 4 Ohm.

As you vary the frequency and measure across the voice coil terminals, the voltage will vary. Adding that resistor and you have started going towards the amplifier being a current source but value needs to be higher.

Let us now add 40 Ohm and that is pretty much a current source, 10 x 4 Ohm = 40 Ohm. Now you have even less, or no, control over the voltage across the voice coil.

But if we look at the current source, now the current through the voice coils changes little with varying frequency. And the higher that impedance the less change there will be to the current through the voice coil.

Voltage is a value that appears across the voice coil.

Current is a value that appears going through the voice coil.

From this we can also deduce:

A voltage source controls the voltage across the voice coil, but to do that it has to relinquish control over the current (through the voice coil).

A current source controls the current through the voice coil, but to do that it has to relinquish control over the voltage (across the voice coil).

Maybe the above will help.
 
That's right, but in order to have control over the current, the current source amp must relinquish control over the voltage. The higher the output impedance, the more control over the current.
This is incorrect. The output voltage of a transconductance amplifier is precisely controlled by current-voltage feedback to maintain the transconductance amplifier output current into a variable impedance driver load.
A current source controls the current through the voice coil, but to do that it has to relinquish control over the voltage (across the voice coil).
This is incorrect. The same comment as above.

The output voltage of a transconductance amplifier is precisely controlled by current-voltage feedback to maintain the output current into a variable impedance load.

We will save vectors, phase angles, inductive reactance and resistance both measured in Ohms for another day.

Thanks DT
 
This is incorrect. The output voltage of a transconductance amplifier is precisely controlled by current-voltage feedback to maintain the transconductance amplifier output current into a variable impedance driver load.

Hang on, that's weird. How can I be incorrect about a statement that I did not make?

Of course, it "maintains the transconductance amplifier output current into a variable impedance driver load."

I never said that it did not! That's the definition of a current source.


Clearly, you have entirely misunderstood what I have said and need to go back and read it carefully again.

BTW, please don't lecture me on how transconductance amplifiers work, I have designed a number of them over many years. Some use feedback, some have eschewed feedback. There are gain block circuits that I have used that create a very high impedance effect that does not use negative feedback.

I am not discussing how such an amplifier works, but how it interacts with the load. Let us keep on topic.

Ask yourself, looking back what does the driver see?

What does the voice coil see?

When the amplifier is the ideal voltage source, it sees an impedance that ideally should be zero.

When the amplifier is the ideal current source, it sees an impedance that ideally should be infinitely high.

Hence:

The voltage of the amplifier will not sag when the amplifier is a voltage source. Ideally, when the output impedance is zero, the voltage cannot drop. In fact, it could be said that in any moment's time, the voltage is regulated.

Think about it, if that is true, then the converse must be true with respect to the amplifier being a current source!

Now the current will only be maintained if the output impedance is infinite! Now the current is fixed (controlled) and the voltage is not.

THEREFORE THE FOLLOWING MUST BE TRUE (I REPEAT AGAIN):

A voltage source controls the voltage across the voice coil, but to do that it has to relinquish control over the current (through the voice coil).

A current source controls the current through the voice coil, but to do that it has to relinquish control over the voltage (across the voice coil).


That's it and don't say that it is incorrect. If you say that, you have not understood it.
 
Joe,

This is what you said:

"That's right, but in order to have control over the current, the current source amp must relinquish control over the voltage. The higher the output impedance, the more control over the current."

This is still not correct:

A transconductance amplifier maintains precise control of the output voltage. The the output voltage is what regulates the output current into a variable impedance load.
A transconductance amplifier is a voltage controlled current source. The voltage across the amplifier differential inputs is what regulates and maintains the current output.

The amplifier does not relinquish control over the voltage, quite the opposite in fact.

Thanks DT
 
@DualTriode doesn't this line Joe added to the reply help:
A current source controls the current through the voice coil, but to do that it has to relinquish control over the voltage (across the voice coil).
He didn't mean the voltage control at the amplifier, but the voltage at the voice coil. At least, that's how I read it.
 
When an amplifier controls the current it cannot control the voltage. It is a current source.

When an amplifier controls the voltage it cannot control the current. It is a voltage source.

Something has to give. You can only control one of them, choose which one. It is really that simple.

But then again I am not surprised thinking current has to start somewhere. But once you have that basic rule in mind, many other things start to fall into place.
 
  • Like
Reactions: 1 user
He didn't mean the voltage control at the amplifier, but the voltage at the voice coil. At least, that's how I read it.

Yep. Maybe I will give it a go one more time...

Let us say that the amplifier is producing 10V, then that 10V is fixed by the amp, it is not going to budge, it is not going to sag below the 10V. We measure the current and if it is 1A, then we know the impedance is 10 Ohm.

But what if we made the load 5 Ohm? We measure the current and it is now 2A. Yes, the amplifier determines the voltage because it has a zero output impedance. But change the load and the current changes. It is a voltage source, but it has no control over the current. The load influences it all over the place, the current cannot be fixed by the amplifier. It is a voltage source.

But now, let us say that the amplifier is producing 1A of current, then that 1A is fixed by the amp, it is not going to budge the current. It is a current source. We now measure the voltage and if it is 10V, then we know the impedance is 10 Ohm, just like before.

But what if we made the load 5 Ohm? We measure the current and it is now 1A. The voltage across the voice coil will read 5V. Now the load can no longer influence the current, and now the voltage cannot be controlled, it can be changed by the load.

Ergo:

A voltage source has to relinquish control over the current. The load influences the current.

A current source has to relinquish control over the voltage. The load influences the voltage.

There is no other way.
____________________

BTW, I haven't forgotten, I will get back to you and point re the current EQ question. I will make the effort when I get a bit of time.
 
  • Like
Reactions: 1 user
What about your hybrid way, series inductor which increases impedance for high frequencies and makes current in control on high frequencies? And leaves voltage in control for the low frequencies?

The key here is to understand what the current does and then formulate techniques that stabilises the impedance modulations caused by the driver, which are in fact current modulations. This is not about using it as a hybrid solution. Since we will not be using current sources (they will never take off despite claims that it is the only 'correct' method), we must make what people use better.

Since the inductor makes improvements from a few hundred Hertz upwards, even before it becomes a filter slope in frequency terms, it is not a hybrid drive at all, but an elegant way of improving the situation.

Please understand that we are discussing techniques that have been applied in the Elsinores or otherwise we are off-topic. As is always the case, the proof is in the pudding, the Elsinores stand as a testament that these techniques work. And they are not the only designs that I have done.
 
What about your hybrid way, series inductor which increases impedance for high frequencies and makes current in control on high frequencies? And leaves voltage in control for the low frequencies?
Yes that is what people do. Just i like a steeper slope in transition from low to high frequencies. Voltage drive or something with a few ohms output impedance works best at resonance and after 2*Fs pure current drive seems to be prefered.
 
  • Like
Reactions: 1 user