Interesting perspective -- 1997. His history of watching audio evolve is similar to my own.... covering a 50 year period... 50 yrs ago I was building my first Hi-End tube audio preamp and amp.
Most interesting is his conclusion as it pertains to us here....he too preferred the CFB amp topology for sound. 😎🙂
THx- RNMarsh
Hi Mr. Marsh !
Yes , long time ago by reading this DePalma audio article I get some answer and confirmation why some of my pretty old SS / AB class (CFB) amplifiers sounds very good to me , and why some newest SS/AB class (VFB) Amps sounds so unnatural regardless to excellent specified numbers , this was one paradox which I accepted .
BTW , I was read that yours intention is to build 2 x 250W AB class CFB power amps , that is in summary 0,5KW ! , I guess that you have some relative inneficient loudspeakers when you need so much power from Amps .
My personal experience and opinion is that much better direction is to invest in some relative very efficient loudspeakers driven with relative low power CFB Amp`s around of 20 - 50 W / 16 ohm / CH .
Best Regards !
Last edited:
Did you have a look at the rest of DePalmas website ? Lots of "free energy" nonsense, I wouldn't believe a word of what he writes.
Well, I hope you all are finished ego pruning, arguing and pontificating. Why don't you all get busy and build something. The proof is in the listening and before I will commit to building another amp, I'm gona do a lot of listening to CFA s in my system before I will plunk down the money. I still haven't heard a solid state amp that sounds as good as my Tim de Paravacini LUX MB3045 tube amp. My LEACH solid state amps are driving sub woofers. And I don't think I'm one who likes to "listen to distortion"! GRH
Wrong, they dont use Bonsai s NX, if anything Bonsai uses their M 6000, and then not even close, what about MCS. Neither do they use Bonsai s fet protection switches, its probably the other way round.
Also your CFA is nearly exactly like their E 210 and 4200, not the other way round.
Accuphase amps technology predate your or Bonsai ideas by more than 13 years.
I did not make claims of originality about the nx-Amp design or anything else that I do for that matter. Its just not my style. For the record, there are VERY few really original designs on anything in semiconductors or electronics. A few gifted people move things forward, and the rest of us build on that and refine it. And its the same with Accuphase, Marantz or whoever else you want to bring up.
What I did do is build two amplifiers to compare the circuit configuration with two VFA's I had previously built and then start a serious discussion about the CFA topology on this forum.
People here have taken to calling my design the 'nx' or 'NX' (or 'sx') front end etc in their posts and analysis because the circuit is published openly, is free to use and it has generated a lot of interaction, which for the most part I am pleased about. We also talk about the Badger, the Woverine etc for the same reason: these are designs published here and free to use.
You make a good point here.
I've been meaning to separate the two aspects when specing an amp into 'small signal rise time' (SSRT) and SR which is clearly a large signal phenomena.
The SSRT on my sx-amp is about 50ns with the front end filter disabled and the SR is about 250 V/us - large signal step response.
These high SR's were not design targets for me (1~2 V/us per Vo pk is my general guideline) but simply a feature of CFA. For VFA, you need to consider it in your design - if you use straight MC you will need to make sure the LTP and LTP current source can meet this design guideline - esecially in the case of a big amp. But, I doubt many mainstream VFA's will be using MC today - seems we have moved on from that.
I could give a +1 or a thumbs up.... but... this is a main point --- you can not build a slow CFA without trying hard. It falls out from the Current-Mode of operation and topology. So....
😎🙂
Last edited:
Less current at feedback / more degeneration.
Why would one do this ? ... easier offset null with less servo current and
cool running FB resistors.
What is the price ? (typical CFA)
1k/39R ...250v SR - 20ppm 20k
2.2K/82R ... 200V - 30ppm
3.9K/150R... 150V - 38ppm
The last 2 sets of values can be negated somewhat with
higher gain devices or a smaller Re at the IP pair.
I noticed the OEM's I mentioned do this for those above reasons.
On the higher priced CFA's- higher current for "bragging rights".
Lower priced are geared for avoiding warranty and sourcing issues (cheaper ,
cooler resistors/cheaper servo's).
PS - I like the "middle value" - best compromise. Still beats most VFA's !
OS
The test might look like CFA x VFA with analog low pass filter (30kHz, Bessel, Butterworth, 4th order?) in front of the amplifier input. The input LPF would remove possible higher input SR that might be generated e.g. by high frequency spikes coming from digital player output. We would then compare CFA x VFA on music programme generated dv/dt only 😛
Natural sound
Guys, just want to share a link to CFA I designed and prototyped recently (Here)
Auditioning it for couple of days now, all I can say, as I mentioned in the other thread earlier - remarkably natural sound.
This is the first CFA I tested in my setup and it definitely stands out for level of detail in mid-highs.
A side advantage of this particular design - absolutely equal balance inputs with no NFB connected to one of them. You can use the amp as inverting, non-inverting or both (balanced). FETs make sure the input current is virtually zero, so you can ground the unnecessary input or leave it open - does not matter for overall balance of the system.
I don't want to say we all have to forget about VFA and start building CFAs, but this is something worth to try and explore.
BR,
Valery
Guys, just want to share a link to CFA I designed and prototyped recently (Here)
Auditioning it for couple of days now, all I can say, as I mentioned in the other thread earlier - remarkably natural sound.
This is the first CFA I tested in my setup and it definitely stands out for level of detail in mid-highs.
A side advantage of this particular design - absolutely equal balance inputs with no NFB connected to one of them. You can use the amp as inverting, non-inverting or both (balanced). FETs make sure the input current is virtually zero, so you can ground the unnecessary input or leave it open - does not matter for overall balance of the system.
I don't want to say we all have to forget about VFA and start building CFAs, but this is something worth to try and explore.
BR,
Valery
Hi... Thank you for telling us your listening impressions on your wonderful CFA. It's very important to get some more listening feedback (sic) as well as CFA designed circuits. Especially useful input from a designer.
Which parameters do you think are responsible for the improved sound? Do you have any ideas or tests as to why it sounds this way?
THx-RNMarsh
Last edited:
A triple mirror CFA.... use your own OPS and devices and see how it does.
THx-RNMarsh
View attachment 3x mirror.pdf
THx-RNMarsh
View attachment 3x mirror.pdf
In http://www.diyaudio.com/forums/solid-state/240712-cfa-topology-audio-amplifiers-3.html#post3595004 post #54 you use the terms compressive and expansive.
What do you mean by this, please.
Dan.
Compressive: VFA peak current into the VAS is limited by the LTP current and is 2x the quiescent or standing current in each half of the LTP.
Expansive: In a CFA, the peak current is controlled by the feedback resistor network and the loop gain - so called 'current on demand' behavior. My simulations show it can be as much as 8x the standing current.