If only could be that simple. I saw guys struggling building yours latest creations, far from sim predictions, also no guarantee for actual audio performance. VFA - are you joking?
Lets not go down this road again.
Both topologies are capable of outstanding performance measured and subjective.
We have learnt a lot as a community about CFA on this thread and I want to keep it that way.
Subjective assessments are a personal thing.
Leave it at that and build what you feel comfortable with.
Both topologies are capable of outstanding performance measured and subjective.
We have learnt a lot as a community about CFA on this thread and I want to keep it that way.
Subjective assessments are a personal thing.
Leave it at that and build what you feel comfortable with.
"Andrew (Bonsai), I'm really just making sure we keep Simple CFA on the table and yanking your chain a little"
Reminds of the Anglia Water Authority add from the late '80s (IIRC).
There was a picture of one of those overhead cisterns with the pull chain.
The tag line was 'Just pull the chain. Well take care of it from there'
Reminds of the Anglia Water Authority add from the late '80s (IIRC).
There was a picture of one of those overhead cisterns with the pull chain.
The tag line was 'Just pull the chain. Well take care of it from there'
Can-we all go back to the subject of this thread: What, is a CFA, how to design them, exploring all the solutions ?
As far i'm in concern, i'm absolutist bored by the endless controversies, turning round and round in circles with aggressive messages.
There is NO point where a CFA present issues that cannot be solved in an elegant and simple way (including PSRR).
There is no point where a VFA is not a good solution to design an amplifier (most of them are).
Some (i'm one of them) have the feeling CFA are more musical. Do your own experiences and figure out by yourself. But, please, let-us continue to exchange nicely on our favorite amplifier topology.
As far i'm in concern, i'm absolutist bored by the endless controversies, turning round and round in circles with aggressive messages.
There is NO point where a CFA present issues that cannot be solved in an elegant and simple way (including PSRR).
There is no point where a VFA is not a good solution to design an amplifier (most of them are).
Some (i'm one of them) have the feeling CFA are more musical. Do your own experiences and figure out by yourself. But, please, let-us continue to exchange nicely on our favorite amplifier topology.
For an amp output, back-EMF of a driver just looks as an increase of the impedance of the load.
Leach VFA question to OS
Which model Leach? LowTIM (100w) or SUPER-AMP (300w)? Have you built one with real parts on a bench and run tests? What "tweaks" are you referring to: matched transistors not only at the input dffamp but thoughout??? Thanks, GRH
Here's my2 cents.
At least in simulation , a "tweaked out" VFA "leach" can slam any of the
CFA's.
And not have offset/PSRR issues.
The IC makers have absolute matching by process ... so CFA is easy for them.
With our crude though- hole DIY construction , this is more difficult.
I'm new to symmetrical topology's , but the VFA approach can exceed
in all 3 categories (psrr/slew/offset correction).
250V/us 100db+PSRR/1mv offset is easy. I struggle to get the last 2
attributes with a CFA ... I might have to use 14 devices for IPS ... but it pays
off.
OS
Which model Leach? LowTIM (100w) or SUPER-AMP (300w)? Have you built one with real parts on a bench and run tests? What "tweaks" are you referring to: matched transistors not only at the input dffamp but thoughout??? Thanks, GRH
I'm not sure of this. (Not at all)
A loudspeaker is too a generator (a microphone), continue to resonate and move after the original signal has ended, and send this signal to the speaker's terminals -> feedback.
And even if there is no signal at the source, noises in the room will produce a signal in this feedback path, héhé.
OTA CFA
As I mentioned in the other thread, I have built a CFA with an input buffer based on the Operational Transconductance Amplifier (OTA) topology, purely based on current mirrors (Here, see post #5 for the final schematic and live pictures). It required to apply NFB in a special way, but proved to be a fully working high quality solution. Fully live-prototyped, tested, auditioned, etc.
If you study the schematic, you will see that IPS degeneration resistor (R4) is "imaged" to the output of the buffer via current mirrors (the point where NFB signal is coming). So first impression that the buffer operates without the global NFB is actually wrong. The overal gain = R39 / R4 + 1.
bimo, isn't it #3 in your list?
As I mentioned in the other thread, I have built a CFA with an input buffer based on the Operational Transconductance Amplifier (OTA) topology, purely based on current mirrors (Here, see post #5 for the final schematic and live pictures). It required to apply NFB in a special way, but proved to be a fully working high quality solution. Fully live-prototyped, tested, auditioned, etc.
If you study the schematic, you will see that IPS degeneration resistor (R4) is "imaged" to the output of the buffer via current mirrors (the point where NFB signal is coming). So first impression that the buffer operates without the global NFB is actually wrong. The overal gain = R39 / R4 + 1.
bimo, isn't it #3 in your list?
As I mentioned in the other thread, I have built a CFA with an input buffer based on the Operational Transconductance Amplifier (OTA) topology, purely based on current mirrors (Here, see post #5 for the final schematic and live pictures). It required to apply NFB in a special way, but proved to be a fully working high quality solution. Fully live-prototyped, tested, auditioned, etc.
If you study the schematic, you will see that IPS degeneration resistor (R4) is "imaged" to the output of the buffer via current mirrors (the point where NFB signal is coming). So first impression that the buffer operates without the global NFB is actually wrong. The overal gain = R39 / R4 + 1.
bimo, isn't it #3 in your list?
Yes, I forgot it. I wrote the another thread when I am on vacation. Thank you for remind me.
Esperado, Forr, this with the speakers being their own signal generator is really one of my major concerns, and why I keep getting back to the suggestion of separating (feed back separation) so the backside of the input stage (or the CFA node) no longer has to deal with that as well as the errors generated in the amplifier itself. To drive the speakers I believe in a solid unit around that. Could be a gain of 1 OPS with low or variable output impedance, a good ground and plenty of capacitance with a star return right at the negative speaker wire.
why do you apparently disregard feedback theory well established > 1/2 a century ago, or the ~1980's Cherry, Cordell's JAES papers on "interface intermodulation distortion" as one of the guru's back then was trying to brand this nonissue
Blackman's Theorem - ring any bells? How about basic electromechanical transducer modeling, Reciprocity Theorem - watts the numbers when you square typical dynamic driver's few % electro-acoustic efficiency?
competently designed, adequately powered, biased output global negative feedback amps do not in fact have any issues with dynamic driver "back EMF, XO reactance causing amps of quadrature current
any "microphone mode" "disturbance" of feedback amplifier driving dynamic drivers is a laughable concept when you run the numbers
Blackman's Theorem - ring any bells? How about basic electromechanical transducer modeling, Reciprocity Theorem - watts the numbers when you square typical dynamic driver's few % electro-acoustic efficiency?
competently designed, adequately powered, biased output global negative feedback amps do not in fact have any issues with dynamic driver "back EMF, XO reactance causing amps of quadrature current
any "microphone mode" "disturbance" of feedback amplifier driving dynamic drivers is a laughable concept when you run the numbers
Last edited:
Then why does all amplifiers sound different..?? fact is that it's all tiny nuances the differences are buried.
Who said anything about disregarding feedback..?? I would consider feedback as an ideal solution, if it wasn't from the fact that devices are not very linear and there is delay through a circuit.
Who said anything about disregarding feedback..?? I would consider feedback as an ideal solution, if it wasn't from the fact that devices are not very linear and there is delay through a circuit.
output has to be sized for real world max current demand, can't be allowed to current starve, or have deadband, power supply has to sized accordingly - all basics for applying negative feedback
once those are in place your talk of separate unity stage, outside the loop is disregarding feedback theory
use global feedback up to reasonable stability margins and you get the lowest output Z for the available gain stages - separating feedback into series loops leads to higher output Z, more error
least interaction with speaker load variation, phase shifted and nonlinear load currents comes from global feedback
nested feedback like Cordell's MOSFET amp error correction lets you put more feedback around the output - inside a global feedback loop
you can claim you like "the sound" of less feedback from separated loops - but it can't be because of reduced load interaction - because the facts of feedback loop output Z reduction are exactly opposite your claim
once those are in place your talk of separate unity stage, outside the loop is disregarding feedback theory
use global feedback up to reasonable stability margins and you get the lowest output Z for the available gain stages - separating feedback into series loops leads to higher output Z, more error
least interaction with speaker load variation, phase shifted and nonlinear load currents comes from global feedback
nested feedback like Cordell's MOSFET amp error correction lets you put more feedback around the output - inside a global feedback loop
you can claim you like "the sound" of less feedback from separated loops - but it can't be because of reduced load interaction - because the facts of feedback loop output Z reduction are exactly opposite your claim
Last edited:
MiiB, you are on the right track. IIM is real. Feedback is what you say it is. MY solution to IIM is to have the LOWEST open loop impedance possible, so that the output stage itself tends to absorb the speaker current directly and almost instantly, using feedback just to lower output Z further. This way, the input stage is less stressed trying to keep the amp output at the proper voltage.
Since you are on the feedback concept right now what effect would there be by adding a third wire to the loudspeaker at the speaker terminal, after any cabling and used this as the NFB source point? This would seem to include everything possible into the feedback loop and would be your comparator point, complete global feedback.
I tried this approach in 80's (2-nd half of last sentury)
3-rd wire, so that speaker wires were within the loop. It worked, but did not bring any significant improvements to the sound quality (subjective). So I dropped that idea...
Just seen this.
I realise it's an OpAmp, but i didn't know it was a CFA ! Good specs.
**********
@ Kindhornman
Yeah the 3rd wire approach has been done before, but Good thinking
In fact some Pro Amps use this idea to monitor driver Voltage/Temp etc. This way you can set the limits etc more accurately
I realise it's an OpAmp, but i didn't know it was a CFA ! Good specs.
**********
@ Kindhornman
Yeah the 3rd wire approach has been done before, but Good thinking
In fact some Pro Amps use this idea to monitor driver Voltage/Temp etc. This way you can set the limits etc more accurately
vzaichenko,
Thank you for the response. So it does work but brought nothing to the party, I just wondered if it would be a better way to include all parasitic properties and counter the back EMF better than ignoring it.
Thanks Zero D. I haven't kept up with PA amps, been away from that for a bunch of years now.
Thank you for the response. So it does work but brought nothing to the party, I just wondered if it would be a better way to include all parasitic properties and counter the back EMF better than ignoring it.
Thanks Zero D. I haven't kept up with PA amps, been away from that for a bunch of years now.
Last edited:
JCX- how do you look at error correction..?? like local feedback or..??
what if there was a way to make that so efficient that you have PPM distortion in you OPS and output impedance can quite easily be trimmed, even to negative values.
I know that in feedback theorie you don't consider loop time as an issue, but it is in my book and a reason why I prefer or try to design with shorter localised loops.
what if there was a way to make that so efficient that you have PPM distortion in you OPS and output impedance can quite easily be trimmed, even to negative values.
I know that in feedback theorie you don't consider loop time as an issue, but it is in my book and a reason why I prefer or try to design with shorter localised loops.