For you LTP is improved CFB, for me not. Tested both in many ways, also very famous schematics from both sides, gave the same neutral approach to each. Stayed on CFA with a reason.
Hi David; sorry i missed your post.
You need the three amplifiers to prove that connecting the compensation capacitor from the collector of the second stage to the point between the feedback resistors does not constitute Miller compensation or even "input inclusive" Miller compensation, and that it is merely phase lead compensation with the output stage excluded. Note that this technique invariably requires additional input lag compensation and cannot make a stable amplifier, in general, on its own.
The Tian probe is so-called in LTSpice, but really works in accordance with Middlebrook's technique.
see:
http://www.ele.tut.fi/teaching/ele-3100/lk0809/tehol1/MiddleBrook75.pdf
Sorry Lazy Cat, but your statement that properly designed CFB amp with CCS for input bias current improve greatly over one with no CCS does not stand.
I simulated your VSSA with and without CCS. Output 36 Vpp.
VSSA no CCS - THD1k 0.001077%, THD20k 0.019868%
VSSA with CCS – THD1k 0.001099%, THD20k 0.020324%
I don’t see any improvement here. Neither of the two has low THD, not even close to the well design VFB amp.
Damir
I don't think Douglas will be pleased with this specific off topic.
Your models are wrong mister.
OMG so you can hear the difference between THD 1k 0,001% in CFA and 0,0001% in VFA? What about time domain? 🙄
You are right , i was about to point it , i got the same results
using Gaborbela s old amp wich is the same design and i also got
the resistive loaded version having lower distorsion in the same
order as your sims.
It s not a matter of models , since the better results of VFB
sims would be equaly flawed , hence , the results although
wrong in absolute values would still be relevant for a relative
comparison.
When I had to write for a living, my bosses told me to write about the things that our readers felt were important. Not just about the things that i, alone, felt worthy of airtime.
"Pandering to the masses" I asked ? "Critically address, not simply agree....." came the answer.
That said, I think that you do this more for fun than anything else, so I guess ignoring questions that the masses (appear) to want answered is a completely reasonable option.
From my small perch it looks like there's a fair amount of love for CFB amplifiers - the F5 thread alone has had more views than anything else on diyaudio but perhaps it just best to ignore what the market wants/demands. What do they know anyway.
I can't see anyone else (Cordell aside)investing time to write books on our arcane hobby so I think you're pretty much entitled to do what you like Douglas.
"Pandering to the masses" I asked ? "Critically address, not simply agree....." came the answer.
That said, I think that you do this more for fun than anything else, so I guess ignoring questions that the masses (appear) to want answered is a completely reasonable option.
From my small perch it looks like there's a fair amount of love for CFB amplifiers - the F5 thread alone has had more views than anything else on diyaudio but perhaps it just best to ignore what the market wants/demands. What do they know anyway.
I can't see anyone else (Cordell aside)investing time to write books on our arcane hobby so I think you're pretty much entitled to do what you like Douglas.
Last edited:
Its easy sometimes....
One of our forum member "Sonny" designed this CFA amplifier and many members built it also.
http://www.diyaudio.com/forums/solid-state/139439-mirand-a1.html
Better have a look.
Last edited:
Hi Douglas,
I'll answer you later, as -at the moment- I have serious troubles with my PC (an insidious bug raised during the installation of new hard- and software).
Cheers,
E.
Never said I could, if you are talking about slew-rate etc.
I do, and it has no downside. But wide bandwidth can easily be an embarrassment, leading to destructive oscillation. Some of the "fast" Japanese amplifiers were notorious for this.
I do. Nobody wants an amplifier that hums like a beehive.
But why? The implication is that you want a very fast slew-rate, but as I said upthread, what is really needed is the optimal slew-rate- faster than required for audio by ten times, say, for a good safety margin, but no more, so that if HF oscillation occurs we are not left cringing under a shower of molten silicon.
It's good to talk...
Completely untrue.
The non-linearities of the two devices in a LTP cancel even-order distortion in a very elegant way; they do not reinforce it.
This is not exactly new knowledge.
Last edited:
Great book, but since I am a novice at this stuff, I have only one request - more pictures if possible.
But this is not one of those times. The quoted spec for this rather complex design is:
Slewrate: +/-120V/usec.
Bandwidth w/o filter : 1MHz.
Bandwidth w filter : 350KHz.
Power supply : +/-30V <=> +/-50V.
Output into 8R@+/-42V: 80Watt.
Distortion at 56W/8R: 0.02%
0.02% THD (at an unknown frequency) is not Blameless performance.
The book will be something like 50% larger, with illustrations increased proportionally.
I'm glad you liked the earlier versions. Any more suggestions for improvement?
Last edited:
Douglas,
I am not sure how out of context it would be but how about a section on the interaction of the NFB section and a real loudspeaker load with both resistive and reactive components. A pole that is outside the feedback loop and how it affects the operation of an amplifiers feedback circuit. I am very interested in correcting the load impedance in a loudspeaker using a conjugate LCR circuit directly at the speaker terminals, but most of the time this is not done and the amplifier does see some horrendous impedance curves and back emf that the NFB circuit doesn't seem capable of dealing with. I am also a reader of your electronic crossover book and so I would only use a passive network to correct the impedance problems, so a combination of electronic crossover and passive impedance correction is my preferred method.
I am not sure how out of context it would be but how about a section on the interaction of the NFB section and a real loudspeaker load with both resistive and reactive components. A pole that is outside the feedback loop and how it affects the operation of an amplifiers feedback circuit. I am very interested in correcting the load impedance in a loudspeaker using a conjugate LCR circuit directly at the speaker terminals, but most of the time this is not done and the amplifier does see some horrendous impedance curves and back emf that the NFB circuit doesn't seem capable of dealing with. I am also a reader of your electronic crossover book and so I would only use a passive network to correct the impedance problems, so a combination of electronic crossover and passive impedance correction is my preferred method.
Doug,
I think that it is wrong to reject and completely ignore CFA in your book. This engineering field is supposed to serve human perception, not only lab instruments. CF amps simply sound much better than VF amps.
I am not sure how many readers of this thread can read German but I shall retell what one can read in Black Devil current feedback amplifier diy instructions (this is hardly great amp concerning lab specs, capacitor coupled output, single rail supply, etc). These amps were given together with competent VF amp designs to professional (classical) musicians for comparisons. The conclusion was that even such simple CF circuits sound MUCH better than VF amps.
I think that audio power amplifier design should not be the subject of techno dogmatism. It is wrong to reject this excellent topology for technical reason only.
I think that it is wrong to reject and completely ignore CFA in your book. This engineering field is supposed to serve human perception, not only lab instruments. CF amps simply sound much better than VF amps.
I am not sure how many readers of this thread can read German but I shall retell what one can read in Black Devil current feedback amplifier diy instructions (this is hardly great amp concerning lab specs, capacitor coupled output, single rail supply, etc). These amps were given together with competent VF amp designs to professional (classical) musicians for comparisons. The conclusion was that even such simple CF circuits sound MUCH better than VF amps.
I think that audio power amplifier design should not be the subject of techno dogmatism. It is wrong to reject this excellent topology for technical reason only.
There were already a lot of discussion about the so-called current feedback, eight years ago :
http://www.diyaudio.com/forums/soli...voltage-feedback-how-do-i-see-difference.html
http://www.diyaudio.com/forums/solid-state/172720-alexander-amp.html#post2290224
If I remember well, somewhere else, Baxandall is again quoted, saying it's a misnommed and unfortunate expression.
I read good ways to explain why on this forum :
- voltage and current can't be substracted.
- a single transistor is a differential device.
http://www.diyaudio.com/forums/soli...voltage-feedback-how-do-i-see-difference.html
http://www.diyaudio.com/forums/solid-state/172720-alexander-amp.html#post2290224
If I remember well, somewhere else, Baxandall is again quoted, saying it's a misnommed and unfortunate expression.
I read good ways to explain why on this forum :
- voltage and current can't be substracted.
- a single transistor is a differential device.