There are a few articles mentioning the implications of current source amplifiers, as opposed to the prevalent voltage source amplifiers, but I have not seen any schematics. Funny thing is that I know how to rewire my speaker crossovers for a current source amplifier thanks to these fine articles, but I don't know how to build a current source amplifier. I'm not interested in the high power resistor method of converting a voltage source amplifier to behave like a current source amplifier even though that is interesting, but would prefer to learn about how to make an amplifier that is a current source.
I did try searching for this topic before posting, but mostly I find the well-known constant current source building block which improves the Class A voltage source amplifier by replacing the inefficient bias resistor. Actually, there are quite a few constant current source circuits around these parts, and they have more than one useful application in amplifiers. But that's not what I'm looking for.
A couple of articles looked promising. One asked whether the F4 is a current source amplifier, but the answer was no. Another asked about ZV8 and ZV9, but I don't see that they're current source amplifier circuits. Am I missing something?
I did try searching for this topic before posting, but mostly I find the well-known constant current source building block which improves the Class A voltage source amplifier by replacing the inefficient bias resistor. Actually, there are quite a few constant current source circuits around these parts, and they have more than one useful application in amplifiers. But that's not what I'm looking for.
A couple of articles looked promising. One asked whether the F4 is a current source amplifier, but the answer was no. Another asked about ZV8 and ZV9, but I don't see that they're current source amplifier circuits. Am I missing something?
There are plenty of the published circuits that can be modified for Current
Drive. They already have high open loop output impedance, and only the overall FB need to be rearranged from Voltage fb to Current fb.
These candidates will have:
Load connected to (complementary) drains/collectors - like in the F5. The cfb will be taken not from the top end of the load, but from the bottom end, connected to the small (a few tenths of Ohm), but powerful resistor (current sensor) to the Gnd.
Or,
Output is taken from the source/emitter follower arrangement with added capacitor bootstrap from the output. This mod will transform your Follower (Common Drain) into Common Source configuration with naturally high output impedance. Very good example is here:
http://www.diyaudio.com/forums/solid-state/3075-jlh-10-watt-class-amplifier-37.html#post2000098 Link, p.2, mosfet version
Again, you will have to remove Voltage fb to preserve high Zout.
Now, you may want to consider the need for the Current fb at all- after all, the Zout already will be sufficiently high, and the cfb will be reducing distortion. Maybe, just servo DC fb.
Drive. They already have high open loop output impedance, and only the overall FB need to be rearranged from Voltage fb to Current fb.
These candidates will have:
Load connected to (complementary) drains/collectors - like in the F5. The cfb will be taken not from the top end of the load, but from the bottom end, connected to the small (a few tenths of Ohm), but powerful resistor (current sensor) to the Gnd.
Or,
Output is taken from the source/emitter follower arrangement with added capacitor bootstrap from the output. This mod will transform your Follower (Common Drain) into Common Source configuration with naturally high output impedance. Very good example is here:
http://www.diyaudio.com/forums/solid-state/3075-jlh-10-watt-class-amplifier-37.html#post2000098 Link, p.2, mosfet version
Again, you will have to remove Voltage fb to preserve high Zout.
Now, you may want to consider the need for the Current fb at all- after all, the Zout already will be sufficiently high, and the cfb will be reducing distortion. Maybe, just servo DC fb.
Actually I believe almost any FET common source amp without NFB will be essentially a "current amp" or Transconductance amp. Such as the initial versions of ZV8 and ZV9. Trouble is, we ussually need at least 6-12db FB.
If I follow what StevenOH says, I believe he is describing the typical circuit for a "current forcing" "VI" card of a test system, as many of us may know it. Although a DAC is outputing a level to the amplifier section, rather than a music source, just useing a small sense resistor in the output leg for the FB and it will do the job. It is important to think about Over Voltage Protection when working with such circuits as when conditions are not readily met with the FB or the load, it will swing to the rail trying.
If I follow what StevenOH says, I believe he is describing the typical circuit for a "current forcing" "VI" card of a test system, as many of us may know it. Although a DAC is outputing a level to the amplifier section, rather than a music source, just useing a small sense resistor in the output leg for the FB and it will do the job. It is important to think about Over Voltage Protection when working with such circuits as when conditions are not readily met with the FB or the load, it will swing to the rail trying.
As pointed out by StevenOH, you can take about any amplifier with feedback, rearrange it, and you have the amplifier you need, see this thread http://www.diyaudio.com/forums/vendors-bazaar/157787-secret-tube-amplifiers-revealed-much-more.html the figure in post#1 has a sense resistor (0.4 - 0.5 ohms). For non-feedback amplifier, a traditional transconductance amplifier would be a nice choice, among these are the Pass' examples...
It is my experience that while you can get benefits from
a higher output impedance than 0 from an amplifier,
almost nobody ends up wanting a very high value. Most
of the time somewhere less than 10 ohms is judged optimal,
and if that's the case, you would expect that a series resistor
of the appropriate value would do the job. A little power
wasted, but a cheap experiment nevertheless.
It's interesting that no one seems to do this. I've heard
comments like "It sucks the life out of the music".
Why this would be is a mystery.
a higher output impedance than 0 from an amplifier,
almost nobody ends up wanting a very high value. Most
of the time somewhere less than 10 ohms is judged optimal,
and if that's the case, you would expect that a series resistor
of the appropriate value would do the job. A little power
wasted, but a cheap experiment nevertheless.
It's interesting that no one seems to do this. I've heard
comments like "It sucks the life out of the music".
Why this would be is a mystery.
Thanks for the replies so far. I'm going to need to spend some time digesting the comments.
It makes sense that most transistor configurations are initially current output. The typical single stage inverting circuit just uses a resistor from the positive rail to convert a current being pulled to ground via the transistor into a voltage that is inversely proportional to the input. Remove the resistor and you're left with the current, but only in one direction with a single transistor.
Can you simply place the speaker between the positive rail and the transistor? That would seem bad, since there'd only be unipolar current, which is like a DC offset.
I suppose the other option is to run common drain and place the speaker between the source and ground, with a small resistor between the speaker and ground for current sensing feedback. Again, that's still unipolar current, but at least the speaker terminal voltages are near ground rather than near the positive rail.
It seems that the only way to have bipolar current is to run a push-pull transistor pair in current mode, or to somehow have a balanced single-ended output that has transistors on both speaker terminals - but I don't quite see how the latter could work. Also, running push-pull seems to be getting away from Zen ... does current output require a more complex circuit with more transistors, and thus is less Zen?
Phew.
It makes sense that most transistor configurations are initially current output. The typical single stage inverting circuit just uses a resistor from the positive rail to convert a current being pulled to ground via the transistor into a voltage that is inversely proportional to the input. Remove the resistor and you're left with the current, but only in one direction with a single transistor.
Can you simply place the speaker between the positive rail and the transistor? That would seem bad, since there'd only be unipolar current, which is like a DC offset.
I suppose the other option is to run common drain and place the speaker between the source and ground, with a small resistor between the speaker and ground for current sensing feedback. Again, that's still unipolar current, but at least the speaker terminal voltages are near ground rather than near the positive rail.
It seems that the only way to have bipolar current is to run a push-pull transistor pair in current mode, or to somehow have a balanced single-ended output that has transistors on both speaker terminals - but I don't quite see how the latter could work. Also, running push-pull seems to be getting away from Zen ... does current output require a more complex circuit with more transistors, and thus is less Zen?
Phew.
Ok, I think one aspect of current source amplification seems clearer:
You can create a 'negative' voltage across speaker terminals via a pair of single-ended voltage output stages by having the (-) terminal at a higher potential than the (+) terminal. Thus, you do not strictly need a negative supply rail, nor do you need the typical push-pull transistor pair (although you still have a pair of transistors, at a minimum, because you need two output stages to operate in this balanced output mode).
It seems that you cannot create a 'negative' output current with single-ended transistor stages. So, if you want to design a current output amplifier, I have the impression that it must be based on a push-pull output circuit. If I'm missing something, please let me know.
You can create a 'negative' voltage across speaker terminals via a pair of single-ended voltage output stages by having the (-) terminal at a higher potential than the (+) terminal. Thus, you do not strictly need a negative supply rail, nor do you need the typical push-pull transistor pair (although you still have a pair of transistors, at a minimum, because you need two output stages to operate in this balanced output mode).
It seems that you cannot create a 'negative' output current with single-ended transistor stages. So, if you want to design a current output amplifier, I have the impression that it must be based on a push-pull output circuit. If I'm missing something, please let me know.
Thank you!
I read all of the articles and manuals in a whirlwind of an effort, and then cogitated on everything for a while. The result was that I was very interested in current sources, so I went back through all of the articles and found nothing (except the speaker and network articles). For some reason, I forgot the First Watt manuals. They do indeed answer my questions (at least as well as I understand my questions at the moment
....... Most
of the time somewhere less than 10 ohms is judged optimal,
and if that's the case, you would expect that a series resistor
of the appropriate value would do the job. A little power
wasted, but a cheap experiment nevertheless.
It's interesting that no one seems to do this. I've heard
comments like "It sucks the life out of the music".
Why this would be is a mystery.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.