@ nigel & dvv,
of course the whole picture is a different (and a more complicated) thing, but it is quite interesting what different researchers did at that time and especially what Olson did to investigate the issue.
He talked about it in this interview 30 years later (Headline of the related paragraph is "Psychology of Sound"):
Oral-History:Harry F. Olson - GHN: IEEE Global History Network
and the detailed description of his experiments appears at page 30 in the pdf:
http://www.americanradiohistory.com/Archive-RCA-Broadcast-News/RCA-46.pdf
of course the whole picture is a different (and a more complicated) thing, but it is quite interesting what different researchers did at that time and especially what Olson did to investigate the issue.
He talked about it in this interview 30 years later (Headline of the related paragraph is "Psychology of Sound"):
Oral-History:Harry F. Olson - GHN: IEEE Global History Network
and the detailed description of his experiments appears at page 30 in the pdf:
http://www.americanradiohistory.com/Archive-RCA-Broadcast-News/RCA-46.pdf
"housekeeping" is a good way to put it. And it's true, I do pay much attention to that, and very probably overdo it sometimes. I tend to err on too much control, never too little. Probably my 25% of Austrian blood, and everybody knows Austrians are simply southern Germans. 😀
In effect, it's really Otala & Lohstroh underneath it all, just done my way. Why amplify something with a factor of 99 dB and then slug it with 74 dB of global NFB, when you can have it amplify say 40 dB, make much less errors along the way, and use just 20 dB of global NFB, which in that case really serves only to make a good thing better, definition of keeping house, equivalent to mopping up the floor? Solve problems as much as you can when and where they appear, don't put them off for global NFB.
As for your comments, as ever, I welcome comments. I learn from them infinitely more than from back slapping.
For 60 dB of NFB to work, you will need a lot more in open loop. No matter how hard you try, distorions of all kinds will always be greater than if you half the open loop gain.
I am usually able to obtain THD and IM figures of less than 0.4 %, 20-20,000 Hz into 4 Ohms, under open loop conditions, with an OL full power bandwidth of at least 75 kHz. In other words, realistically, I don't even need global NFB, but it always helps in ridding me of output stage residuals, which are inescapable when working into real world loudspeakers. Again usually, 20-26 dB of NFB does it just fine. I end up with THD of less than 0,2% even at 100 kHz (ref. nominal output at 1 kHz).
How's your hearing at 100 kHz, Wayne? 😛
There has been much criticism of dominant pole compensation arranged around the VAS ( TIS ) . It is logical as it does it in one neat chop . It also sorts out the linearity a tad in that area . One reason to dislike it is it has to be driven by the LTP which is like a dog chasing it's tail in terms of current required and capacitor to make it stable . This also causes slewing asymmetry ( like a one legged bicycle rider ) . Double LTP including double VAS cures that nicely and doubles the slew rate for the same input LTP current . It needs a circuit to mimic the bias side to get it to balance at > 20 kHz . Doubtless it needs tweaking for each amp made . In the book of Self that is 100% why you wouldn't do it . Fair enough if contract built and no quality control . Many op amps are double LTP .
Double VAS can work with one current mirror only and all other parts resistors . Tail rejection is 92dB with a resistor compared with 127dB with an optimum current source ( seldom is ) . The LED biased current source is simple and rather good . If the resistor is split and a capacitor added it is close to perfect . The voltage swing is almost zero so it can work . It isn't a VAS . We use a current source because it is cheap and does no great harm . I doubt it does much good . I have done a few listening test and the resistor won . The books say no way .
The double VAS has to be a Dvv type as it needs to sit 1 V above the rail . Thus it must have a shared Re if it is to be simple and elegant . Re = 100R is about right if 2 x 7 mA . That sits the 2 bases up at about 2 V which is nice . Balance without any special measures is about 2% . Critics often say very imperfect balance . Nonsense .
I must read it all . Love the lady from Atlantic City and her AM transmitter .
I moved the Magneplanars back into my room . They need a super tweeter as they don't play 78's any better than headphones . My OB speakers are rather good . With the super tweeter there is air and space . in theory only 4kHz if a 1926 78 . How very weird to need zero distortion and 50 kHz to play 1926 78's ! Those dam Williams sisters . I have fallen for them .
Phase shift displaces notes' leading edge? = loss of crispness? I think it's Griesinger has something to say about this.
I've developed a rule of thumb, fairly recently, that the amplifier should function correctly, in all areas, up to 200kHz - I've noted that amplifier designs of noted audio designers fail this test, quite badly, in simulation - so what chance has a real world example have of being of excellent quality?
Guess what? It's not, the "slightly bland" is the giveaway - distortion can make something sound sort of clean, but it is still most certainly distortion. Trouble is, standard measuring procedures don't pick it up ...
Would prefer not to - contributions by respected members of the forum who are very giving of their knowledge, and it wouldn't be fair "to make a thing of it", it's only something that's my POV ...
Of interest, much discussion of voltage feedback vs. current feedback amps in threads here - and the CF ones shine much brighter in this type of test, handle it with ease; the VF designs start to collapse.
Others would say that the only interesting things that happen at 200kHz are interference and oscillation.
I'm referring to the waveforms that occur in the leads of the components that make up the circuit: when the amplifier is not 'stressed' these show reasonable, expected characteristics; at the limit all sorts of nasty rubbish starts to take over, the amplifier is beginning to 'collapse' ...
One thing I know is rules for bipolar output stages and MOS FET's are different . MOSFET's don't measure too well until looking carefully . Everything nasty is stepped up a bit higher than a sub-optimally biased bipolar . None of the expected nastiest seem to result in harshness . One doesn't have to think much further and say Class D . My big shock listening to the the Hypex class D is mostly what you hear is PSU related . The amplifier is almost as neutral as a very good class AB . The spectrum analyzer fights hard to measure it . Unaided it can't . What I did notice is the sub 1 watt is very good and this goes down to the mW level . I didn't expect that . It rolls off slightly below 50 kHz . The sound doesn't suggest that . I suspect the slight residuals give a false brightness . One thing I like about MOSFET's is you can set them by ear and finger . It drives you mad because it is a compromise . Still nice that you can . Bias is a variable resistor ( <470R if 7 mA ) . That amp has just one current mirror . Mathematically it doesn't fall down . All the dog chasing it's tail stuff is the winning side of the argument . Slew rate is 35 V/ uS which for real music is fine ( 150 W 4R 100 W 8 R ) . The input is restricted at 3 MHz and that is it . VAS comp is 27pF . MOSFET's seem to like negative feedback and will respond to the more the better . Nothing you hear would suggest a high feedback amp . They sound like somewhere between tubes and bipolar . This as far as I can tell has nothing to do with the slight similarity between them and tubes .
Current feedback never seemed to get me anywhere and just made life harder . I dare say no one showed me the right thing to try . It worked on a regenerator . That goes for inverting amps also .
Current feedback never seemed to get me anywhere and just made life harder . I dare say no one showed me the right thing to try . It worked on a regenerator . That goes for inverting amps also .
Now answer me this . A Ceda processed 1926 78 sounds dull and boring on speakers that start to roll off at 15 kHz . These speakers have distortion like that of a 1950 hi fi amp ( - 60dB ) . My home made OB has two advantages . They throw an image many times wider and the will go to 50 kHz - 6dB ( room willing ) . The response is a bit bumpy , not bad for something done by maths alone with how they look placed slightly above how they work . That is the drive units central looks nicer although a bit wrong .
Why is it with speaker No1 playing 78's I hear a radio sound and the new speakers I am in the room where the music happened ? Having played with many speakers I can not explain this . I suspect it is the unusually large image that is doing it . The very big deal is you would swear you have 15 kHz . As the disc is Ceda processed there is no obvious sound of the abrasive powder that 78's had ( to burnish the needle I think ) . There is very little HF distortion which is even more unusual . Both speakers are open baffle so that is ruled out . As a problem it is one I love to own .
Why is it with speaker No1 playing 78's I hear a radio sound and the new speakers I am in the room where the music happened ? Having played with many speakers I can not explain this . I suspect it is the unusually large image that is doing it . The very big deal is you would swear you have 15 kHz . As the disc is Ceda processed there is no obvious sound of the abrasive powder that 78's had ( to burnish the needle I think ) . There is very little HF distortion which is even more unusual . Both speakers are open baffle so that is ruled out . As a problem it is one I love to own .
- Status
- Not open for further replies.