CFA Topology Audio Amplifiers

Is this an absolute in cfa design or can you in fact have more than one gain stage and still follow the feedback topology that has been the subject of this entire discussion?
Of course, you imagine what you want inside the open loop.
Each time you add a gain stage, you add a pole. So, you ruin slew-rate, bandwidth, stability (phase margin) and add blurr. For little benefit, but distortion numbers on paper. And, somewhere, aginst CFA"s nature.
But you can add some gain in the input stage with positive results.
That is my opinion after long years on it.
I would really like to see some complex CFAs designed in this thread build, and compared to simple ones in real world.

Do we really need this for audio
That's similar to my own amp, and i find some benefit in transparency, ease of reproduction, little dynamic details. While i'm not sure it is correlated by this specific point.
Anyway, it cannot hurt.
 
Last edited:
>FREQUENCY RESPONSE : DC - 1.0MHz (-3dB) dictated by input filter (without filter 10 Mega Hertz).
Do we really need this for audio 🙄 though nice for radio-amateurs as a 'linear'. 😉

Hello Edmond,

This Question is only valid until You "Actually Hear and better FEEL That Speed"
after that happen You'll never ever look back and ask that question again - Only Scintillas Can Do ... 🙂
But of course one can forget any others speakers ever build, They simply cannot do and cannot compete with Apogee Scintillas - Direct Biampl. driven (Four monos).
All the Rest speakers - Dynamic and any kind of electrostats are pure joke Scooters comparing with.
You have to click to Bugatti Veyron blue text (link) on my previous post and when the page opens click on the big right arrow to understand what I mean ..... 😉

Cheers
 
So, you ruin slew-rate, bandwidth, stability (phase margin) and add blurr.

Really? There is about a gazillion of 2 gain stages amplifiers, audio (including some audiophile famous names, the John Curl name comes in handy), instrumentation, op amps, etc... out there, and never heard anybody complaining about such problems.

Single high gain stage CFAs (able to compete with the VFAs in terms of distortions) are more difficult to compensate for stability, they usually need extra Miller loop compensation. To add insult to injury, some techniques largely used in VFAs (like adding a zero in the forward loop gain by using a lead compensation across the feedback resistor) are, for stability reasons, big nono's for CFAs.

It is though true that 2nd order compensated CFAs (not necessary for audio, the high closed loop gain may preclude this) are easy to be designed as unconditionally stable. Up from certain ULGF frequencies, VFAs are always only conditionally stable.
 
Sweeeeeet ........ :spin:


http://hometheaterreview.com/metaxas-opulence-preamp-and-soliloquy-power-amps-reviewed/?page=2


TECHNICAL DESIGN

Input Impendance

The 130kOhms/82pF input impendance using BIPOLAR inputs, ensures that there are no loading effects to the output stage of the preceeding Preamplifier and severe H.F. roll-off with long interconnect cables. We use direct wiring to electrically connect the input to the Main Circuit Board.

Power Supply

The SOLILOQUY uses one of the most sophisticated, over designed Power Supply systems in a commercial amplifier.

Transformers: Each chanel sports a 1,600 Watt continuous transformer
with 12 Amps continuous current capability to the Output Stages and 2 Amps to the Input Stages.

Primary
Filtering: This is then bridge rectified and fed to the +/-33,000uF Filtering Supplies for the OUTPUT STAGE and the +/- 10,000 Filtering supplies for the INPUT STAGE. The Voltage Gain Stage and Output Current Gain Stage operate independently.

Voltage
Regulation: The Regulation Circuits are formed by totally discrete, high speed bipolar devices (42 matched and selected devices), which differ only in their 'Series Pass' transistors based on the current requirements. The Output Stage Regulators use the same ultra-fast Bipolar Power Transistors as used in the Output Stage, with it's of 60 Megahertz and current capabilities of 12-15 Amps smaller Power Device with higher Master Circuit' features purely differential sensing elements with active current loading and a precision voltage reference. The negative circuit tracks the Positive 'Master' to ensure balanced operation under all condition. These circuits ensure ripple-free totally symmetrical and stable DC which is unaffected by current demands.

Voltage Amplifier


The Input Stage of the SOLILOQUY features a pure complementary cascoded differential BIPOLAR Stage with constant current regulating the biasing current. The devices utilise a considerable amount of 'emitter degeneration' (local feedback) which increases their input impendance, linearity and large signal slow rate. The second voltage gain stage utilises complementary cascoded high speed BIPOLAR devices which drive their quiescent biasing from the input stage current source. The voltage amplifier consumes 12 selected and matched devices.


Current Amplifier
(Output Stage)

The Output Stage of the SOLILOQUY features a complementary Darlington Stage Triple idling at 30w in the Class A mode and switching to Class AB up to 100w. This biasing maintains the optimum heating temperature for the Power Devices to yield their maximum bandwidth and HFE. It utilises ultra fast (60 MegaHertz) Bipolar Devices which have low input capacitance (100pF) and high current capabilities (12-15 amps each). These devices eliminate the need for output ZOBEL networks, INDUCTORS or lag compensation normally required to slow down the input stage for use with slower BIPOLAR (3-5MHz) or MOSFETS (10-13MHz).

Protection Circuits

The SOLILOQUY features two modes of protection - one which is inherent in the operation of the design which acts as 'current' protection and the other which uses associated sensing circuitry to monitor the outputs for DC and deactivate if DC is present.

If a high current condition arises, the positive supply fuse (5A fast blow) will preferentially blow instead of the negative fuse (4A fast blow). Because the negative supply 'tracks' the positive, if the positive fuse blows, the negative supply will always symmetrically discharge to avoid the possibility of DC voltage from reaching sensitive speakers.

If a DC voltage is sensed at the output (greater than 0.6V) a relay in series with the output will open .
 
Last edited:
Really? ...[snip]
Compilation of non sense, false informations (" adding a zero in the forward loop gain by using a lead compensation across the feedback resistor are, for stability reasons, big nono's for CFAs"), dropping names instead of demonstration, noise, as habit...

There is a lot of cfas using caps across feedback resistance, to minimize phase difference between signal and feedback, reducing distortion, smoothing square waves.

I love you, the objectivist, named John Curl, who promote the advantage of a 10cm 'burned' silver wire, 'oriented' in a preamp.

...Keep on barking to the moon.
 
Last edited:
Since the original definition of current feedback was that the current is the controlled variable, and you are running with the marketing version of the term, we'll never see eye to eye. On this.😉

Jan

😎🙂

I assumed we would be talking about the original definition in the beginning. You, know, the one the IC industry uses. And, if IT would have any identifiable benefit to audio. Silly me. My kak-kah. OK. So we are sliding over to the consumer definition. What ever. The consumer/marketing definition/design is a good topology also and in fact, is one of the more popular one in use today.
Clearly, not all consumer CFA's are current feedback/control.... I mean, for example -- If tubes are still voltage controlled devices... how can it ever be a CFA? Just because of where the feedback or control is placed? I don't think so. Will we ever see eye to eye is a good question. Might be time to throw in the towel and go with the flow? We don't actually really need to know how various topologies actually work, do we. Well? Do we? Why not just copy the other guy and mess with it in SIM and Bingo! we got a great working design to go build.


Thx-RNMarsh
 
Last edited:
Sweeeeeet ........ :spin:


Stereo Amplifier Reviews: Metaxas Opulence Preamp and Soliloquy Power Amps Reviewed


TECHNICAL DESIGN

Input Impendance

The 130kOhms/82pF input impendance using BIPOLAR inputs, ensures that there are no loading effects to the output stage of the preceeding Preamplifier and severe H.F. roll-off with long interconnect cables. We use direct wiring to electrically connect the input to the Main Circuit Board.

Power Supply

The SOLILOQUY uses one of the most sophisticated, over designed Power Supply systems in a commercial amplifier.

Transformers: Each chanel sports a 1,600 Watt continuous transformer
with 12 Amps continuous current capability to the Output Stages and 2 Amps to the Input Stages.

Primary
Filtering: This is then bridge rectified and fed to the +/-33,000uF Filtering Supplies for the OUTPUT STAGE and the +/- 10,000 Filtering supplies for the INPUT STAGE. The Voltage Gain Stage and Output Current Gain Stage operate independently.

Voltage
Regulation: The Regulation Circuits are formed by totally discrete, high speed bipolar devices (42 matched and selected devices), which differ only in their 'Series Pass' transistors based on the current requirements. The Output Stage Regulators use the same ultra-fast Bipolar Power Transistors as used in the Output Stage, with it's of 60 Megahertz and current capabilities of 12-15 Amps smaller Power Device with higher Master Circuit' features purely differential sensing elements with active current loading and a precision voltage reference. The negative circuit tracks the Positive 'Master' to ensure balanced operation under all condition. These circuits ensure ripple-free totally symmetrical and stable DC which is unaffected by current demands.

Voltage Amplifier


The Input Stage of the SOLILOQUY features a pure complementary cascoded differential BIPOLAR Stage with constant current regulating the biasing current. The devices utilise a considerable amount of 'emitter degeneration' (local feedback) which increases their input impendance, linearity and large signal slow rate. The second voltage gain stage utilises complementary cascoded high speed BIPOLAR devices which drive their quiescent biasing from the input stage current source. The voltage amplifier consumes 12 selected and matched devices.


Current Amplifier
(Output Stage)

The Output Stage of the SOLILOQUY features a complementary Darlington Stage Triple idling at 30w in the Class A mode and switching to Class AB up to 100w. This biasing maintains the optimum heating temperature for the Power Devices to yield their maximum bandwidth and HFE. It utilises ultra fast (60 MegaHertz) Bipolar Devices which have low input capacitance (100pF) and high current capabilities (12-15 amps each). These devices eliminate the need for output ZOBEL networks, INDUCTORS or lag compensation normally required to slow down the input stage for use with slower BIPOLAR (3-5MHz) or MOSFETS (10-13MHz).

Protection Circuits

The SOLILOQUY features two modes of protection - one which is inherent in the operation of the design which acts as 'current' protection and the other which uses associated sensing circuitry to monitor the outputs for DC and deactivate if DC is present.

If a high current condition arises, the positive supply fuse (5A fast blow) will preferentially blow instead of the negative fuse (4A fast blow). Because the negative supply 'tracks' the positive, if the positive fuse blows, the negative supply will always symmetrically discharge to avoid the possibility of DC voltage from reaching sensitive speakers.

If a DC voltage is sensed at the output (greater than 0.6V) a relay in series with the output will open .

For the moment I was thinking that you described my amp, but then, my was never built and it is VFA. http://www.diyaudio.com/forums/solid-state/216780-tt-amp-200w-8ohm-701w-2ohm-3.html#post3119028
To tell the truth it was inspired with Metaxas amps.
 
Compilation of non sense, false informations (" adding a zero in the forward loop gain by using a lead compensation across the feedback resistor are, for stability reasons, big nono's for CFAs"), dropping names instead of demonstration, noise, as habit...

There is a lot of cfas using caps across feedback resistance, to minimize phase difference between signal and feedback, reducing distortion.

I love you, the objectivist, named John Curl, who promote the advantage of a 10cm 'burned' silver wire, 'oriented' in a preamp.

...Keep on barking to the moon.

Now it's true I'm barking to the moon (sometimes wondering myself why bothering, as long as the moon is either unable or not willing to understand what I'm talking about), but perhaps others would find interesting that:

1. Regarding the feedback capacitance in true CFAs, perhaps reading this (in particular chapter 9, "Stability and the feedback capacitance") would help. I'm assuming the moon won't bother to read, so here's a quick quote from the conclusions following the analysis: "Adding capacitance to the inverting input node or across the feedback resistor usually results in instability"

2. Yes, I've seen feedback capacitances in audio CFA's. There's a single little problem, already mentioned before: because of the high closed loop gain, the "CFA" properties are long gone. They are VFAs by any EE metric, not even the marketing crowd that invented the CFA concept would dare call them "CFAs".

3. "caps across feedback resistance, to minimize phase difference between signal and feedback, reducing distortion", now that's what I call nonsense by any engineering metric. This is usually called "technobabble".

4. I am surprised you don't appreciate the oriented and burned silver wires in a preamp, you have all the traits and deep knowledge to. Anyways, JC work was just one example of commercial audio amplifiers using 2 gain stages. In fact, probably 90% do.
 
Last edited:
Transformers: Each chanel sports a 1,600 Watt continuous transformer
with 12 Amps continuous current capability to the Output Stages and 2 Amps to the Input Stages.

Primary
Filtering: This is then bridge rectified and fed to the +/-33,000uF Filtering Supplies for the OUTPUT STAGE and the +/- 10,000 Filtering supplies for the INPUT STAGE. The Voltage Gain Stage and Output Current Gain Stage operate independently.

Voltage
Regulation: The Regulation Circuits are formed by totally discrete, high speed bipolar devices (42 matched and selected devices), which differ only in their 'Series Pass' transistors based on the current requirements. The Output Stage Regulators use the same ultra-fast Bipolar Power Transistors as used in the Output Stage, with it's of 60 Megahertz and current capabilities of 12-15 Amps smaller Power Device with higher Master Circuit' features purely differential sensing elements with active current loading and a precision voltage reference. The negative circuit tracks the Positive 'Master' to ensure balanced operation under all condition. These circuits ensure ripple-free totally symmetrical and stable DC which is unaffected by current demands.

Voltage Amplifier


The Input Stage of the SOLILOQUY features a pure complementary cascoded differential BIPOLAR Stage with constant current regulating the biasing current. The devices utilise a considerable amount of 'emitter degeneration' (local feedback) which increases their input impendance, linearity and large signal slow rate. The second voltage gain stage utilises complementary cascoded high speed BIPOLAR devices which drive their quiescent biasing from the input stage current source. The voltage amplifier consumes 12 selected and matched devices.


Current Amplifier
(Output Stage)

The Output Stage of the SOLILOQUY features a complementary Darlington Stage Triple idling at 30w in the Class A mode and switching to Class AB up to 100w. This biasing maintains the optimum heating temperature for the Power Devices to yield their maximum bandwidth and HFE. It utilises ultra fast (60 MegaHertz) Bipolar Devices which have low input capacitance (100pF) and high current capabilities (12-15 amps each). These devices eliminate the need for output ZOBEL networks, INDUCTORS or lag compensation normally required to slow down the input stage for use with slower BIPOLAR (3-5MHz) or MOSFETS (10-13MHz).
Looks like not CFA
The output is similar to soulution710. Is it plagiat or clone?
what is the price (you get)?

Power supply and input impedance is just marketing.
 
Last edited:
For the moment I was thinking that you described my amp, but then, my was never built and it is VFA. http://www.diyaudio.com/forums/solid-state/216780-tt-amp-200w-8ohm-701w-2ohm-3.html#post3119028
To tell the truth it was inspired with Metaxas amps.


Only 4 pr output ? anyway build it and they will come ..... 🙂

Looks like not CFA
The output is similar to soulution710. Is it plagiat or clone?
what is the price (you get)?

Power supply and input impedance is just marketing.

look at the link posted, prices should be quoted in that article ....
 
Remember patents at best are only on a set of claims narrowed as much as necessary to be granted. CFA's as a genre date from the 40's with valves.

While this is true, it in fact was pioneer that pioneered the diamond input stage for CFAs. This was long before Comlinear, it took 3 years for pioneer to have their patents granted in the USA, one has to wonder why, when Comlinear patents were granted within months. Pioneer patents were rejected due to technicalities and they had to file revised version before they were granted. I have copies of both Pioneer and Comlinear patents and there is nothing wrong with the first Pioneer patent that was requested. Comlinear patents should never have been granted in the first place as they clearly infringe on Pioneers patents.
 
I just looked at schematic of Panasonic MZ1 (1980) and some areas of the schematic is just shown as a block diagram (Japanese writing inside). Might be a current-mirror? Cant tell from their service manual schematic.
But circuit still used electrolytic cap in gain setting feedback (220mfd bypassed with 2.2mfd). Triple OPS used a separate follower driver per each final output transistor.

Thx-RNMarsh

See my previous post why Pioneer kept it secret, inside those block diagrams is diamond buffers and current mirrors. See The Pioneer patents.

Harmon Kardon copied the concept from pioneer with their version around 1979 with the Citation XX. The Citation XX is regarded as the best sounding Harmon ever by their fans. This amp is also typical diamond buffer input CFA and uses so called hawksford cascode vas (also long before Hawksford showed the concept). This is another reason Comlinear patents should never have been granted. When you see the schematics of this amp youll also see how Harmon were decieving the industry as to its true circuitry.
 
Last edited:
I'm familiar with NAD and Accuphase, Sony and Pioneer i have never heard one worthy to take home. Nad's are OK, there really big amplifier is not bad, Accuphase is in a different league and IMO , excellent stuff.
What amplifier is that ...?
Interesting ....:bulb:

Do find examples, you are going to be pleasantly surprised at how good audio was already back then. Try any of Pioneers M series amps, they give any high end amps of today and at any price, very stiff competion. As late as 2008 Pioneer won awards for a cheap near entry level amp featuring Jfet CFA design.
I get a big grin when I see someone like Pass post designs using Jfet CFAs, everyone is in admiration, these deigns were in production by Pioneer in the late 1970s and their good sound quality is already well known.

Accuphase is regarded as special High End amps but when you see the schematics its basically just a plain jane CFA design, very simple, much like Bonsai s design. They just use a darlington vas, note not a EF vas.
 
It is though true that 2nd order compensated CFAs (not necessary for audio, the high closed loop gain may preclude this) are easy to be designed as unconditionally stable. Up from certain ULGF frequencies, VFAs are always only conditionally stable.
AT LAST! A post from Waly that might actually help us design better CFAs .. instead of the MikeK 'yus are all idiots & deaf' stuff .. 🙂

Oh Guru Waly, could you enlighten us further 😱 And please for the sake of da unwashed & indigent beach bums among us, don't just refer to $250 textbooks.


Waly, please feel free to substitute 'non-LTP input amps which might loosely be called CFAs' for any appearance of 'CFA' in this thread 😀
 
Last edited:
Do find examples, you are going to be pleasantly surprised at how good audio was already back then. Try any of Pioneers M series amps, they give any high end amps of today and at any price, very stiff competion. As late as 2008 Pioneer won awards for a cheap near entry level amp featuring Jfet CFA design.

I have heard the pioneer M amplifiers, sorry they are not at the level you perceive and definitely not better than accuphase i have heard ...

I get a big grin when I see someone like Pass post designs using Jfet CFAs, everyone is in admiration, these deigns were in production by Pioneer in the late 1970s and their good sound quality is already well known.

You don't know Nelson very well, He's always given credit where it's due .....

Accuphase is regarded as special High End amps but when you see the schematics its basically just a plain jane CFA design, very simple, much like Bonsai s design. They just use a darlington vas, note not a EF vas.

Yes like Bonsai's designs ...... 🙄

Unfortunately ... special stuff is sometimes about smoke and mirrors - and this gear got quite a reputation down under with very easily going into the 'smoke' department ...

Just Google 'metaxas' and 'reliability' ... 😉

Easy now, the man said sound, nothing about smoke .... 🙂