2-pole is used commercially, in diy designs here, may still benefit from some VAS anti-saturation measures but seems practically usable
additional local loops in nested/multiloop arrangements also exist in Audio Power Amps, Halcro is an infamous example with some commercial success
from Cherry's nested feedback to Bob's Mosfet Powers amp, to multiloop LM3886 there are people playing with added loop gain, "conditionally stable" designs, working examples
the BJ Lurie Control site refs I gave a few pages back do talk about real limits - in a bit more detail: http://www.diyaudio.com/forums/loun...ch-preamplifier-part-ii-8469.html#post4782512
additional local loops in nested/multiloop arrangements also exist in Audio Power Amps, Halcro is an infamous example with some commercial success
from Cherry's nested feedback to Bob's Mosfet Powers amp, to multiloop LM3886 there are people playing with added loop gain, "conditionally stable" designs, working examples
the BJ Lurie Control site refs I gave a few pages back do talk about real limits - in a bit more detail: http://www.diyaudio.com/forums/loun...ch-preamplifier-part-ii-8469.html#post4782512
Last edited:
"When I measure speakers and headphones I never have seen a situation where the electronic's distortion swamped that of the speaker."
With the majority of speakers I would assume we would be talking about 2nd and 3rd harmonic distortion and non-linear distortion mechanisms in the suspension systems that would as you say swamp out the electronic distortion. The higher order distortion from the electronics should be at least a factor less in amplitude if not much more than that. So I can see how you can measure the electronic higher order distortion products by electronic means but to say you can easily detect these through the speakers is where I must question what that limit must be? How high must the level of that distortion reach to be audible over the typical speaker distortion from most commercial speaker systems and not just in a very few low distortion designs or say something like a high efficiency low distortion compression driver?
With the majority of speakers I would assume we would be talking about 2nd and 3rd harmonic distortion and non-linear distortion mechanisms in the suspension systems that would as you say swamp out the electronic distortion. The higher order distortion from the electronics should be at least a factor less in amplitude if not much more than that. So I can see how you can measure the electronic higher order distortion products by electronic means but to say you can easily detect these through the speakers is where I must question what that limit must be? How high must the level of that distortion reach to be audible over the typical speaker distortion from most commercial speaker systems and not just in a very few low distortion designs or say something like a high efficiency low distortion compression driver?
Time and time again, as happened today while listening to a MBL system: I listen first, notice a brittle top end. I look to see no TT or the TT is idle. Another demo de rigueur with music pumped from a music server or worse yet a laptop. I know where that high frequency distortion is coming from.
Not with current recording/playback chains. The resonances surround you and so would require an ambisonics type of arrangement to capture. One of the cases where proper surround sound would actually make a difference.
My position is that every attempt to setup up controlled tests for audibility have shown that Peter Walker was right all those years ago. Once you get below 0.1% THD there is no evidence you can tell them apart without peeking. The fact that eyes trump all other senses unless you are blind should be obvious without having to read too much literature. People believe they have magic hearing and that's fine as long as the snake oilers don't take advantage of them and they dont try and persuade the rest of us that we are deaf. It's a hobby after all!
Hi end stuff sells on the bling and the marketing story. Never on the specs any more. I have an amplifier on my restore list that will be 50 years old this year and still measures better than a lot of boutique products.
Yes. My standard (maximum acceptable) for speakers is -50dB (0.3% for any harmonics). Today it is so easy to get much lower than that for amplifiers (my design threshold is 0.02% THD). But I feel amps still sound different.
Is that speaker or amplifier? Amplifier of course. Why do you think there is no other variable other than THD that can affect sound? There is high order versus low order (or phase related variable). Try class-A versus class-B amplifiers. Or class-A versus class-D amplifiers.
I don't know whether people can tell them apart or not, but my intuition is that I CAN. Of course I'm very interested to know if I'm right or wrong.
Put aside any inferiority (or superiority) complex. Aren't you all interested to know the truth? Don't you think that some sound clips and Foobar ABX is sufficient for the test?
Four clips: A, B, C, D. Two class-D (or class-B) and two class-A. Arrange such amplitude is one (slightly) higher and one lower for each amp so that amplitude difference has no effect on the result. Listener should be able to tell which ones are the class-D...
From above test I expect to hear fatigue from amplifiers that generate high order distortion, even tho it is down 60dB and beyond 20kHz...
Another test is slew rate test. One with low SR and the other with high SR. The music must be complex (many instruments) and fast. Here I expect the low SR amp will fall apart...
Yes - I agree.
50 transistor amplifiers always make me chuckle - it does not show sophistication. Just the inability to make sensible engineering tradeoff's.
TPC or TMC will get you to c. 45-50 dB LG at 20 kHz comfortably. But you then still have to take other precautions. Its about as far as I will go for an audio amp. I just don't see any reason for higher order comp networks and the attendant grief it brings. Keep it simple - the ear does not mind a bit of 2nd and 3rd introduced by the amp that's not present on the source material. Tube amps attest to that.
Wow - that's old school. I remember reading a review (already a few years old at that time) in about 1970 while still at school . . .
Attachments
![01-Restored-STA25[1].jpg](/community/data/attachments/506/506711-a60f7dbea40161dd01b3199e4208bbda.jpg)
Last edited:
You cannot say you are interested in the truth if you then pepper your text with ' I feel, expect and inuition'. You have pre-supposed there must be a difference and you are the one to find it. But if you can get to the right frame of mind for the test and a good protocol you might be suprised by the results.
Yeah bought it as a student in the 80s with a blown transformer. Was my first upgrade from the ubiquitous NAD3020. 26dB GNFB, which is a lot for a tube amp of the era, unless I missed some other designs (entirely possible)
All the spec's look the same... so... duh.
Hmm do I want this 100w amplifier or this one? I see one has 0.001% distortion and the other one has 0.0013%. Real useful.
You hear something and you see something at the same time and draw the conclusion that they must be linked. Gestalts like that often occur and can be very wrong.
I would rather have (complete) measurements to work from, including such useful things as clippling and recovery from clipping behaviour that someone telling me how a myrtle wood block under a power cable transformed the sound. I am not the target market for boutique audio though. I just want an amplifier that, against the currently known limits of human hearing and for the load I present is at least 10dB better than the audibility thresholds. If on a 32 tone test (which I think is useful although really a fast production QA check) the grot is all >100dB down then I would be hard pressed to come up with any logical reason for that not being completely transparent unless abused.
Attachments
- Status
- Not open for further replies.