Depends on how you define NFB: it can be global, as in a normal amplifier where the overall gain is defined by the ratio of the feedback network, it can be local for linearizing or stabilization purposes on two stages f.e., and it can be inherent: a follower configuration of an active element is the typical example for that.
If you rule out inherent NFB, you are going to make your life very difficult.
If you just exclude local and global NFB, it is still possible to achieve vanishingly small levels of THD without excessive complexity: I have a working example of such an amp, it's just lazyness that prevented me from publishing and sharing it, but I will do it, some day....
In the mean time, you can have a look at the Tringlinator, it is an headphone amp, but it fits the bill.
Every amplifier schematic I've seen that said it used "zero-feedback" actually did have feedback, often several types. That includes design by so-called gurus of whom frequent this site. Feedback is not a problem, nor anything to fear so long that the collective design is intelligently executed. Even if someone does fear it, they will have to gather themself for the truth - all amplifiers use some type of feedback in order to operate. Audio gear designers have a common ego trait of modifying the facts to make themselves appear as though they know how to design in ways that anyone else can't.
Last edited:
Headphone amp or 1kW doesn't matter: power amplification is defined as a role, not a numeric level.
An hearing aid amp may give a few hundreds µW at most, but it is a power amp. The line out stage of a professional audio equipment will be able to put out much more than that, yet it is categorized as signal amp.
Without degeneration, you won't be able to define gain or impedances and you will end up with a very informal amplifier.
Additionally, if you really want to hunt down degenerations, you will have to tackle the inherent degeneration of active devices: with tubes, it will be a huge task, because the degeneration is ~1/gm, and tubes have notoriously low gm's.
With FETs, it will be somewhat easier, and BJT's will end up the big winners of that contest.
Unless of course you categorize inherent degeneration as OK, the evil being in external degeneration, in which case rational people can do little to help you...
An hearing aid amp may give a few hundreds µW at most, but it is a power amp. The line out stage of a professional audio equipment will be able to put out much more than that, yet it is categorized as signal amp.
Without degeneration, you won't be able to define gain or impedances and you will end up with a very informal amplifier.
Additionally, if you really want to hunt down degenerations, you will have to tackle the inherent degeneration of active devices: with tubes, it will be a huge task, because the degeneration is ~1/gm, and tubes have notoriously low gm's.
With FETs, it will be somewhat easier, and BJT's will end up the big winners of that contest.
Unless of course you categorize inherent degeneration as OK, the evil being in external degeneration, in which case rational people can do little to help you...
It is an interesting topic. 
Most folks look at a schematic and if they don't see what looks like a "global" loop they assume it's a no-feedback design.
Here's one from a "guru" to ponder:
FIRST WATT M2
Cheers,
Dave.
Most folks look at a schematic and if they don't see what looks like a "global" loop they assume it's a no-feedback design.Here's one from a "guru" to ponder:
FIRST WATT M2
Cheers,
Dave.
I've seem this method of amplification before, around 1998. An step-up autoformer provided the voltage gain while the active devices provide the current to sustain the pending voltage magnitudes. The design mitigates the devices upper poles, but has high distortion and bandwidth limiting, leading to slew induced distortion, ie: SID, that would otherwise not exist in the use of active voltage gain topologies with a sufficiently high corner frequency. First Watt's designer appears to among the group that still believes in TIM (appropriately known as TID, transitory intermodulation distortion; Roddam, 1953) to be the result of feedback, when it is actually the result of slew rate limiting by either miller compensation, or another low-pass limiter such as an inductive medium. This has been known for at least five decades. The input jfet and output mosfet arrangment in the M2 looks like a play on some of the earlier Yamaha pro designs that operated in the same fashion. Too bad there is no complete schematic in that link to fill in the omitted portions. That's where the details reside.
Last edited:
Then, to a purist this does not classify as "No FB" since you have to use a follower of some kind.
In Speedskater's optic, the only acceptable type of topology is common base, but to arrive at a working amplifier with common base only would require transformers everywhere, which might not be the surest path to high reproduction quality.
The second best would be a string of non-degenerated common emitters, but it is not "perfect", because of the internal degeneration, and transformers would also seriously help, even if they aren't absolutely mandatory
A common-base stage can be seen as a common-emitter stage with current follower feedback around it. That is why it has a much reduced current gain that is close to unity.
A common-emitter, common-source or common-cathode stage is the only stage with no feedback at all except for some parasitic internal effects such as collector-base (drain-gate, anode-grid) capacitance, emitter bulk resistance or the effect of anode voltage on the field in the grid-cathode region. Fortunately for Harold Black, no-one considered these effects to be feedback before he got his patent.
By the way, there are some common-source amplifiers on http://firstwatt.com/pdf/art_sony_vfet_pt1.pdf You would have to redesign the supplies, though, as they do use feedback.
A common-emitter, common-source or common-cathode stage is the only stage with no feedback at all except for some parasitic internal effects such as collector-base (drain-gate, anode-grid) capacitance, emitter bulk resistance or the effect of anode voltage on the field in the grid-cathode region. Fortunately for Harold Black, no-one considered these effects to be feedback before he got his patent.
By the way, there are some common-source amplifiers on http://firstwatt.com/pdf/art_sony_vfet_pt1.pdf You would have to redesign the supplies, though, as they do use feedback.
Last edited:
actually negative feedback was known, used in electronics - but only to "neutralize" positive feedback that might cause oscillation
Black's patent is about using large amounts, previously unexplored in amplifier theory to improve distortion performance and reduce sensitivity to device variations/aging
of course followers existed previously too
and other systems likes the centrifugal governor for steam engines employed negative feedback
that wasn't at all a bar to a legitimate patenting of negative feedback in "wave amplifiers" - the holdup at the patent office was their disbelief in the theory, that such performance improvement really was possible
Black's patent is about using large amounts, previously unexplored in amplifier theory to improve distortion performance and reduce sensitivity to device variations/aging
of course followers existed previously too
and other systems likes the centrifugal governor for steam engines employed negative feedback
that wasn't at all a bar to a legitimate patenting of negative feedback in "wave amplifiers" - the holdup at the patent office was their disbelief in the theory, that such performance improvement really was possible
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.