CFA I like it.
Daughter and I are making a very similar all bjt version (but more conventional output followers) for ipod speakers (I cant stand seeing earbuds or even headphones on for hours at a time).
Thanks
-Antonio
Daughter and I are making a very similar all bjt version (but more conventional output followers) for ipod speakers (I cant stand seeing earbuds or even headphones on for hours at a time).
Thanks
-Antonio
This is very promising CF circuit, distortion is most certainly very low. Somewhat overdone circuit, a bit more complicated than necessary, and finding J and K output mosfets specified is now difficult. But I like it, too.
BTW, I think that your proposal from other threads to use inverting input in VF amps has some merit and should be examined in more detail. At least, with power amps, it is easy to solve the problem of absolute phase by inverting connections at one side of loudspeaker cable. A number of people reported that VF amps sound better when inverting mode is used. Few years ago there were many amps with power ICs that used that kind of circuit. And John Linsley Hood promoted idea of using shunt feedback in MM phono RIAA stages as better sounding in spite of having 10dB more noise.
BTW, I think that your proposal from other threads to use inverting input in VF amps has some merit and should be examined in more detail. At least, with power amps, it is easy to solve the problem of absolute phase by inverting connections at one side of loudspeaker cable. A number of people reported that VF amps sound better when inverting mode is used. Few years ago there were many amps with power ICs that used that kind of circuit. And John Linsley Hood promoted idea of using shunt feedback in MM phono RIAA stages as better sounding in spite of having 10dB more noise.
Damir,
Is this some kind of trick question? I am not EE, and circuit is drawn in such way that it is somewhat puzzling to me.
Is this some kind of trick question? I am not EE, and circuit is drawn in such way that it is somewhat puzzling to me.
Clearly CFA. Q11, Q12 are simply just bias providers for Q3, Q4 tracking cascodes. Nothing special about the circuit, only to complex for resulting performance. A big question remains how it would sounds, so build and listening test is of utmost necessity. 😉
If I may add, circuits with this level of complexity are notoriously difficult to stabilize. This circuit asks for SMD pcb implementation which is not diy friendly. I wish that you could have a look inside Cyrus Straight Line amp (complicated CF topology). PCB is more like some PC card, you need magnifying glass to see all tiny SMD parts. That is why we, audiophiles with some diy inclinations, prefer simple CF circuits.
This is very promising CF circuit, distortion is most certainly very low. Somewhat overdone circuit, a bit more complicated than necessary, and finding J and K output mosfets specified is now difficult. But I like it, too.
BTW, I think that your proposal from other threads to use inverting input in VF amps has some merit and should be examined in more detail. At least, with power amps, it is easy to solve the problem of absolute phase by inverting connections at one side of loudspeaker cable. A number of people reported that VF amps sound better when inverting mode is used. Few years ago there were many amps with power ICs that used that kind of circuit. And John Linsley Hood promoted idea of using shunt feedback in MM phono RIAA stages as better sounding in spite of having 10dB more noise.
I built inverted gain clone long time ago(non inverting too) and used input buffer to overcome low input impedance and a source impedance gain influence(JLH simple buffer).
I can't say which one sounded better, never did listening comparison, both sound quite good.
Damir
Damir,
Is this some kind of trick question? I am not EE, and circuit is drawn in such way that it is somewhat puzzling to me.
It is not a trick question, I am not sure were to put it as the gain part is a current conveyor. A current conveyor has very low intrinsic distortion and to find equaly good output stage it's not easy.
This is just simulated project not ready to be built, main problem is to get close to zero volt at the output, probably neads servo to solve that.
Clearly CFA. Q11, Q12 are simply just bias providers for Q3, Q4 tracking cascodes. Nothing special about the circuit, only to complex for resulting performance. A big question remains how it would sounds, so build and listening test is of utmost necessity. 😉
You are correct nothing special, just my try to use current conveyor as a gain part.
Here is modified VSSA with JFETs, also nothig special. I was in preprations to convert my JLH MOSFET amp(I have two) to CFA to compare it with JLH VFA type, but not desided yet on the circuit.
Attachments
Dadod, what I would do on last sch is to leave BJT pair at the input, cascoded with j-fets, BJT VAS/TIS with 470 ohm base and 10 ohm emitter resistors and no gate stoppers at output mosfets. Compensation with smallest possible Miller caps and remaining necessary cap compensation to GND or to OUT, just enough to stabilize it. 😉
Dadod, what I would do on last sch is to leave BJT pair at the input, cascoded with j-fets, BJT VAS/TIS with 470 ohm base and 10 ohm emitter resistors and no gate stoppers at output mosfets. Compensation with smallest possible Miller caps and remaining necessary cap compensation to GND or to OUT, just enough to stabilize it. 😉
I don't get it, what advantage to cascode BJT with JFET?
I have to point out that I intended to use existing JLH power supplay with +-55V for input, TIS/VAS and +-50V for output and two output pairs(not very well mached, 2SK135/2SJ50), so probably need source resistors(I hope you meant source resistors not gate stoppers).
dadod
In latest schema I see you use TMC error correction - symmetric.
I have tried this in simulation for a current fb amp similar to yours.
It works well and lowered distortion with 12-13 dB.
In latest schema I see you use TMC error correction - symmetric.
I have tried this in simulation for a current fb amp similar to yours.
It works well and lowered distortion with 12-13 dB.
dadod
In latest schema I see you use TMC error correction - symmetric.
I have tried this in simulation for a current fb amp similar to yours.
It works well and lowered distortion with 12-13 dB.
I've got only 2dB at 20kHz lower distortion. I think this VAS/TIS has to low gain to get more. Could you show your circuit?
Thanks for asking dadod 🙂I've got only 2dB at 20kHz lower distortion. I think this VAS/TIS has to low gain to get more. Could you show your circuit?
I took VSSA and squeezed it. Then I squeezed it a bit more ...
Yes, I think my circuit has more OL gain.
Because I have to use like 220p for compensation.
Without TMC it has THD 0.00175%
With TMC it has THD 0.00037%. Both at 1 Watt 1kHz 8 Ohm.
At 25 Watt it has THD 0.00031%
As you can see TMC is best at 470p, 470p, 1.2k in this circuit.
Usuallý C3 is 3-5 times bigger, but not in this case.
My amplifier is targetted to give at least 25 Watt RMS at 2x25VDC.
--
I have another circuit which has even more gain and less distortion with TMC.
It uses darlington VAS = two transistors at each half.
Otherwise it is the same as this one.
Attachments
Good job dadod, now you're almost at the same conditions as in VSSA I have here. Four compensation capacitors are correct for the compensation job done properly. Although I manage to pass with a little lower values 2 x 5,6 pF and 2 x 100 pF, maybe a little more on the edge. Nevertheless this amp will bring music in your living room, enjoy. 

Great result!With LazyCat VSSA and TMC implemented I could get 14dB distortion decrease at 20kHz.
So I quite NFET input and probably tray this one.
This confirms the magnitude improvement is like 13-14dB, with TMC.
Great result!
This confirms the magnitude improvement is like 13-14dB, with TMC.
Indeed it is.

I did this compensation calibration by measuring 3 MHz signal and rotating all four trimmer caps, so I manage to find the sweet spot that amp was stable and without distorting 3 MHz sine signal. It was really easy. 😎
Indeed it is.
I did this compensation calibration by measuring 3 MHz signal and rotating all four trimmer caps, so I manage to find the sweet spot that amp was stable and without distorting 3 MHz sine signal. It was really easy. 😎
In your VSSA shematic you show only two compesating caps. Could you show were did you connect other two.
Thanks for asking dadod 🙂
I took VSSA and squeezed it. Then I squeezed it a bit more ...
Yes, I think my circuit has more OL gain.
Because I have to use like 220p for compensation.
Without TMC it has THD 0.00175%
With TMC it has THD 0.00037%. Both at 1 Watt 1kHz 8 Ohm.
At 25 Watt it has THD 0.00031%
As you can see TMC is best at 470p, 470p, 1.2k in this circuit.
Usuallý C3 is 3-5 times bigger, but not in this case.
My amplifier is targetted to give at least 25 Watt RMS at 2x25VDC.
--
I have another circuit which has even more gain and less distortion with TMC.
It uses darlington VAS = two transistors at each half.
Otherwise it is the same as this one.
Those are very good results lineup. So you've got the same distortion at 1W and 25W?
Could you show THD20k at 25W also?
- Status
- Not open for further replies.
- Home
- Amplifiers
- Solid State
- What about this amp-CFA or VFA?