diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Solid State (http://www.diyaudio.com/forums/solid-state/)
-   -   CFA Topology Audio Amplifiers (http://www.diyaudio.com/forums/solid-state/240712-cfa-topology-audio-amplifiers.html)

Bonsai 13th August 2013 02:46 AM

CFA Topology Audio Amplifiers
 
This thread is where we can discuss CFA and compare it to VFA in a friendly and constructive manner.

RNMarsh 13th August 2013 02:52 AM

Great.... There are some miss-conceptions and significant differences might start with.

-RNM

Bonsai 13th August 2013 02:58 AM

A quick summary

CFA: amplifier topology that inherently supports fast rise/fall times and wide bandwidths. Distortion performance with the basic topology incl. EF3 can meet sub 10 ppm levels at 20 kHz. PSRR not as good as VFA topology, requiring additional filtering and regulation especially of the front end buffer/level shifter stage. Considered easy to compensate. Of interest in audio are the wide loop bandwidths, in some cases up to 40 or 50 kHz. Typical upper loop gain in the 60 dB range.

VFA: amplifier topology that in MC guise supports moderate to fast slew rates. LF dominant pole required to ensure unconditional stability. Alternative comp schemes allow slew rates and rise times to match CFA - eg MIC. PSRR performance of the basic topology better than CFA; noise performance considered better in some cases. Distortion performance when using advanced comp techniques and EF3 can achieve low single digit ppm levels (20 kHz). Typical LF OLG in 100dB plus range, dropping at 20 dB/decade such that at 20 kHz, 50 dB of loop gain may be available using standard MC. This figure may be increased using alternative comp schemes like MIC or TMC.

michaelkiwanuka 13th August 2013 10:13 AM

Quote:

Originally Posted by Bonsai (Post 3593994)
Alternative comp schemes allow slew rates and rise times to match CFA - eg MIC


Typical LF OLG in 100dB plus range, dropping at 20 dB/decade such that at 20 kHz, 50 dB of loop gain may be available using standard MC. This figure may be increased using alternative comp schemes like MIC or TMC.

Actually, MIC, or input inclusive compensation does not allow more major loop gain than ordinary Miller compensation.

Secondly, I suspect, but haven't examined this yet, that the compensation network that is required to make the input inclusive compensation loop stable may worsen PSRR dramatically.

My suspicions are founded in the fact that the similar shunt compensation network required to make the minor loop stable with Jolly-McCharles input cascode inclusive compensation ruins the PSRR.

buzzforb 13th August 2013 11:44 AM

Esperado,
In your summation of how you want the thread to go, you seemed to close the argument on the possibility that the VFA could sound better than the CFA. While I am not saying whether this is true or not, an openminded conversation that begins with a close minded statement seems interesting. I have listened to both the TSSA1.6 from member Sonya and the VSSA that I proto'ed using laterals. I preferred the TSSA over the VSSA. I agree that the amp "sound" extremely fast, with good pace and lots of air. I did not think that either performed as well in terms of tone or overall body in comparison to some Class A VFB amps I have built. I will proto the SSA next to see if it comes closer in this regard. I thought the VSSA would step closer than the TSSA, but it seemed to only give up some of its better capabilities in comparison, without gaining anything in return. This is not to say it is not an excellent amp, as I am just talking about personal taste. Of course we all know that everyone likes different colors, although I believe blue is the best.;)

Bob Cordell 13th August 2013 11:53 AM

Quote:

Originally Posted by Bonsai (Post 3593986)
This thread is where we can discuss CFA and compare it to VFA in a friendly and constructive manner.

Thanks for opening up this thread, Bonsai. It is a good topic for discussion.

Cheers,
Bob

mcd99uk 13th August 2013 12:02 PM

All we need now are the posts moved from D Self's thread to here.

MagicBox 13th August 2013 12:04 PM

Somehow I suspect if I would put my MF80 VFA project (which had been designed for speed as well, with a fast output stage configuration and is capable of slewing in excess of 200V/us) inside a black box with a label saying "CFA MOSFET AMP", and invited all CFA proponents to listen to it, that the amp would receive universal acclaim.

Bob Cordell 13th August 2013 12:06 PM

Quote:

Originally Posted by michaelkiwanuka (Post 3594257)
Actually, MIC, or input inclusive compensation does not allow more major loop gain than ordinary Miller compensation.

Secondly, I suspect, but haven't examined this yet, that the compensation network that is required to make the input inclusive compensation loop stable may worsen PSRR dramatically.

My suspicions are founded in the fact that the similar shunt compensation network required to make the minor loop stable with Jolly-McCharles input cascode inclusive compensation ruins the PSRR.

I believe you are correct in that MIC does not allow greater global loop ULGF than ordinary MC.

MIC does not necessarily destroy PSRR, and in some cases may improve it. Take a look at the MIC implementation in my MOSFET power amplifier with error correction, http://www.cordellaudio.com/papers/MOSFET_Power_Amp.pdf. The MIC loop compensation is accomplished with a differential shunt series R-C across the differential outputs of the IPS. This thus does not compromise PSRR. Moreover, the MIC Cdom is referenced to signal ground on both sides, which eliminates one of the major causes of PSRR degradation in standard MC, namely that one side of Cdom is referenced to the power supply rail while the other side is referenced to signal ground (i.e., the VAS output node).

Cheers,
Bob

Mooly 13th August 2013 12:17 PM

1 Attachment(s)
Carried over :)

Quote:

Originally Posted by michaelkiwanuka (Post 3594241)
There is absolutely no way your amp. can have 80dB of loop gain at 20KHz and be stable. No way at all.:D

Mine has :D and more and is 100% stable.


All times are GMT. The time now is 06:46 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2