john curl said:
I did? Hmmm. Perhaps you can explain what these posts of yours have to do with answering the original poster's questions vis a vis single versus dual differential inputs?
http://www.diyaudio.com/forums/showthread.php?postid=262844#post262844
http://www.diyaudio.com/forums/showthread.php?postid=262918#post262918
http://www.diyaudio.com/forums/showthread.php?postid=262984#post262984
http://www.diyaudio.com/forums/showthread.php?postid=263560#post263560
http://www.diyaudio.com/forums/showthread.php?postid=265347#post265347
http://www.diyaudio.com/forums/showthread.php?postid=265441#post265441
http://www.diyaudio.com/forums/showthread.php?postid=265448#post265448
http://www.diyaudio.com/forums/showthread.php?postid=265586#post265586
http://www.diyaudio.com/forums/showthread.php?postid=265735#post265735
http://www.diyaudio.com/forums/showthread.php?postid=265835#post265835
http://www.diyaudio.com/forums/showthread.php?postid=266223#post266223
http://www.diyaudio.com/forums/showthread.php?postid=266384#post266384
se
Bricolo said:
Bricolo, can you see any global feedback in the circuit of post no. 159 and have you seen the specs (www button below). And not only specs, but listening tests with well-known amps? (Quite difficult to perform here, I know).
This is not only a question of feedback or non-feedback, but the circuit.
PMA said:
Hehehe. Sorry. I woke up at 4:00AM and couldn't get back to sleep so... 🙂)
Well, I guess it depends on how one defines "better." Better in pure technical terms doesn't necessarily mean better in subjective sound terms. If that were the case, we'd never have to bother to listen to anything. We'd all just pick whatever measures best.
se
Subjective terms, Steve. But I do not speak about the case when you need to mask imperfection of the audio chain by coloured amp's sound 😉
PMA said:
But subjective terms are ultimately individual terms. And no one individual's terms are inherently superior to another's.
But if one ultimately prefers that case, who is anyone else to judge?
se
PMA said:
No.
But in fact, it was more a "general" question, after having read your other posts here, I thought that you dislike amps without global feedback. And I wanted to know why
It would be a never ending story, Steve.
In case you need distortion to get more "pleasant" sound, then something wrong is in your chain, and unfortunatelly this happens in 98% of cases. This is not about preferences. Take a group of people and give them a choice (listening test). In 99% they will prefer better solution - that's my experience.
I give up further argueing😉
In case you need distortion to get more "pleasant" sound, then something wrong is in your chain, and unfortunatelly this happens in 98% of cases. This is not about preferences. Take a group of people and give them a choice (listening test). In 99% they will prefer better solution - that's my experience.
I give up further argueing😉
Steve Eddy said:
Because (we've done a mini full circle here)
In order to have a 'reasonable' input Z
(say 10K), and a 'reasonable' Av (say 10)
there are no xfmr/output device combos
that will work. (unless we add more devices
(than the simple 2 xsistor follower (in which
case the xfmr becomes unneeded)))
IMHO ......... mike
I had hoped that we would stick to input stages, primarily those that could be normally driven from external sources like preamps, cd's etc. I recently saw some interesting input, but it is now effectively lost in subsequent posts.
That said, I would like to clarify a few random assertions.
I have NOT worked with Mark Levinson since 1976. Therefore I did not design the 6,7,etc products produced later. However, most design concepts are based on 3 or 4 design topologies. The JC-2 used 2 of these topologies. Some of these later ML designs use either these 2 topologies or something similar in many products.
I am neither for or against negative feedback. I just don't like to use an output stage feedback pair inside of a global feedback loop. There is justification for this that was debated extensively on a previous thread, somewhere.
That said, I would like to clarify a few random assertions.
I have NOT worked with Mark Levinson since 1976. Therefore I did not design the 6,7,etc products produced later. However, most design concepts are based on 3 or 4 design topologies. The JC-2 used 2 of these topologies. Some of these later ML designs use either these 2 topologies or something similar in many products.
I am neither for or against negative feedback. I just don't like to use an output stage feedback pair inside of a global feedback loop. There is justification for this that was debated extensively on a previous thread, somewhere.
hitsware said:
Well, I don't see an input impedance of 1.5k unreasonable, except perhaps for some sources with tube output stages but they're the exception, not the rule. You offered up an amplifier that'd do about 6 watts into 8 ohms, and with a nominal 2 volt source such as a CD player, you wouldn't need a gain of 10 to achieve that.
se
john curl said:
Well, the circuit I put up can be normally driven from external sources like preamps, CDs, etc. And since this thread veered away from single versus dual differential long ago (with your complicity I might add), I fail to see how anything I posted was any more threadjacking than what you have been posting.
se
>Well, I don't see an input impedance of 1.5k unreasonable,
You may be right but most circuits (I've seen) have some sort of resistance ~600 or 1k ohm in series with the output.
My tuner has the outs off of a collector which I'm sure would go bunk with a 1.5K load. (though I've not tried it)
>You offered up an amplifier that'd do about 6 watts into 8 ohms, and with a nominal 2 volt source such as a CD player, you wouldn't need a gain of 10 to achieve that.
Any lower gain and (if'n I was to try to sell some) people would say "this ain't got no power". Though a misinformed stance, this is prevalent. People equate (to a great extent) power and fidelity with not turning the nob much for alot of gain. The x22 usually used is maybe alittle hotter than needed but not much......mike
You may be right but most circuits (I've seen) have some sort of resistance ~600 or 1k ohm in series with the output.
My tuner has the outs off of a collector which I'm sure would go bunk with a 1.5K load. (though I've not tried it)
>You offered up an amplifier that'd do about 6 watts into 8 ohms, and with a nominal 2 volt source such as a CD player, you wouldn't need a gain of 10 to achieve that.
Any lower gain and (if'n I was to try to sell some) people would say "this ain't got no power". Though a misinformed stance, this is prevalent. People equate (to a great extent) power and fidelity with not turning the nob much for alot of gain. The x22 usually used is maybe alittle hotter than needed but not much......mike
Before we were so rudely interupted.
Back to the original topic before the usual troublemaker derailed the entire thread again just to get attention.
"Fred, originally proposed a sophisicated and 'improved' biasing stage. I felt that it was not a 'universal' schematic, because it lacked placement of a source degeneration resistor in the N channel rail. However, Fred is correct in this one specific case with using J109's and K389's. Thanks Fred, I didn't notice the difference between Gm's before, in these two devices. However, if you were using K170's and J75's, the difference would be less, and almost unmeasurable. Personally, in a simple circuit like this, I prefer to use high Idss types and run them as hard as possible. This includes even slightly overbiasing the input devices in the forward direction, under transient peaks. Please think it through, before attacking my comment.'"
I would not dream of attacking your comment as I think it is good design advice and you have about a 20 year head start over me on amplifier design. The idea for splitting the bias resistor between the jfet pairs was for a pretty specific application. Many mosfets have differing transconductances between the P and N channel devices with the N channel having significantly higher Gm the P channel. I was looking for a simple way to compensate by reducing the gain of the P jfet pair to equalize the gain of each jfet/mosfet half of the circuit. I know that decreasing the bias for the jfets knocks down the transconductance of the jfet and requires a larger load resistor for a given voltage gain and the effects on bandwidth and source impedance must be considered. Thanks again for your excellent advice on running jfets near to there Idss values. I always appreciate your feedback positive or negative. One actually learns much more for being told what one is doing wrong and I enjoy the learning experience. Now that I think about it something similar to this was done for the front end of the Hafler XL-280, a Borbely design I believe.
http://www.hafler.com/techsupport/pdf/XL-280_amp_man.pdf
Back to the original topic before the usual troublemaker derailed the entire thread again just to get attention.
"Fred, originally proposed a sophisicated and 'improved' biasing stage. I felt that it was not a 'universal' schematic, because it lacked placement of a source degeneration resistor in the N channel rail. However, Fred is correct in this one specific case with using J109's and K389's. Thanks Fred, I didn't notice the difference between Gm's before, in these two devices. However, if you were using K170's and J75's, the difference would be less, and almost unmeasurable. Personally, in a simple circuit like this, I prefer to use high Idss types and run them as hard as possible. This includes even slightly overbiasing the input devices in the forward direction, under transient peaks. Please think it through, before attacking my comment.'"
I would not dream of attacking your comment as I think it is good design advice and you have about a 20 year head start over me on amplifier design. The idea for splitting the bias resistor between the jfet pairs was for a pretty specific application. Many mosfets have differing transconductances between the P and N channel devices with the N channel having significantly higher Gm the P channel. I was looking for a simple way to compensate by reducing the gain of the P jfet pair to equalize the gain of each jfet/mosfet half of the circuit. I know that decreasing the bias for the jfets knocks down the transconductance of the jfet and requires a larger load resistor for a given voltage gain and the effects on bandwidth and source impedance must be considered. Thanks again for your excellent advice on running jfets near to there Idss values. I always appreciate your feedback positive or negative. One actually learns much more for being told what one is doing wrong and I enjoy the learning experience. Now that I think about it something similar to this was done for the front end of the Hafler XL-280, a Borbely design I believe.
http://www.hafler.com/techsupport/pdf/XL-280_amp_man.pdf
Attachments
hitsware said:
Yes, though most tend to be around 150 ohms or so.
Not necessarily. The Apex power opamps I use have collector outputs but still have a low output impedance.
>You offered up an amplifier that'd do about 6 watts into 8 ohms, and with a nominal 2 volt source such as a CD player, you wouldn't need a gain of 10 to achieve that.
Any lower gain and (if'n I was to try to sell some) people would say "this ain't got no power". Though a misinformed stance, this is prevalent.[/B][/QUOTE]
So why perpetuate it?
Certainly so few watts won't statisfy most out there, but so what? There are still a lot of folks out there with high efficiency speakers (and that number seems to be growing with more recent renewed interest in horn loaded loudspeakers--c'mon, Mike, you hang out over at Melhuish's). Since when has high-end been about appeasing the lowest common demoninator? 🙂
Then why offer up circuits that'll only do 6 watts? 😀
se
Re: Before we were so rudely interupted.
If I'm just trying to get attention, then you and Curl with your remarks such as these are just playing into my hands and drawing even more attention to me.
Who's looking for attention here? These comments, instead of simply going on about what it is you wish to discuss, seem to be saying "Hey, don't look at Steve, look at me!"
se
Fred Dieckmann said:
If I'm just trying to get attention, then you and Curl with your remarks such as these are just playing into my hands and drawing even more attention to me.
Who's looking for attention here? These comments, instead of simply going on about what it is you wish to discuss, seem to be saying "Hey, don't look at Steve, look at me!"
se
>Then why offer up circuits that'll only do 6 watts?
My point is that whether you have 1Watt or 1kWatt (well maybe alittle exageration 🙂 ) you need the same gain for a 'comfortable' response from the nob. The other 999 Watts is overhead. As enlightened as I consider myself, I don't want to turn up something 3/4 of the way for normal listening. Just has a bad "feel" ........ But that's just me. (and customers in the past)
..........mike
My point is that whether you have 1Watt or 1kWatt (well maybe alittle exageration 🙂 ) you need the same gain for a 'comfortable' response from the nob. The other 999 Watts is overhead. As enlightened as I consider myself, I don't want to turn up something 3/4 of the way for normal listening. Just has a bad "feel" ........ But that's just me. (and customers in the past)
..........mike
I would unreservedly support the concept of global feedback. I believe that done right it cures far more problems than it makes. I also believe that all this wooden cross BS about global feedback is tantamount to ignorant sloganeering; the product of a fevered marketing department..... I see a future where recent work on distortion spectrum and harmonic distribution will yet be optimized, and the solution will probably be a GNFB amplifier.
Pavel, have you tried the prosaic double emitter follower output stage, particularly Self's Type II, in side by side listening comparisons, against your CFP? I have done tests with others some years back and concluded strongly in favor of the double emitter follower.
You make some challenging comments about RF injection into the amp too, which I'm looking into. Great work, Pavel!
Cheers,
Hugh
Pavel, have you tried the prosaic double emitter follower output stage, particularly Self's Type II, in side by side listening comparisons, against your CFP? I have done tests with others some years back and concluded strongly in favor of the double emitter follower.
You make some challenging comments about RF injection into the amp too, which I'm looking into. Great work, Pavel!
Cheers,
Hugh
Hi,
Ah...another 11 o'clock listener like myself....
I tend to agree with you in that it's good practice to have some gain/power left for better dynamic range.
Certainly. I reckon the fever originated from the obsession with figures some 30 years back which resulted in amps sounding as flat as a pancake amongst other oddities.
As with anything else, the truth sits somewhere in the middle; GNFB has been abused by incompetent designers and the ommission of it has been advocated by even less competent designers it seems.
No offense to any camp in particular but I feel that it's wise to design with reasonably linear blocks right from the start, GNFB would then only be the icing on the cake when implemented correctly.
Just my 2 cents, 😉
Ah...another 11 o'clock listener like myself....
I tend to agree with you in that it's good practice to have some gain/power left for better dynamic range.
Certainly. I reckon the fever originated from the obsession with figures some 30 years back which resulted in amps sounding as flat as a pancake amongst other oddities.
As with anything else, the truth sits somewhere in the middle; GNFB has been abused by incompetent designers and the ommission of it has been advocated by even less competent designers it seems.
No offense to any camp in particular but I feel that it's wise to design with reasonably linear blocks right from the start, GNFB would then only be the icing on the cake when implemented correctly.
Just my 2 cents, 😉
>No offense to any camp in particular but I feel that it's wise to design with reasonably linear blocks right from the start, GNFB would then only be the icing on the cake when implemented correctly.
Yes ! 🙂
Until recently I have always tinkered with amps with the feedback loop closed. On a trip of trying to emulate 'tube' characteristics (with SS) I remembered that most (I think) tubes amps are stable and operable without the feedback. (in fact at a place I worked years ago we did the entire QC (tube amp) procedure open loop and as a final step tacked in the feedback resistor). I then decided I didn't need no stinkin' feedback afterall, but your point is right on. Icing on the cake..................mike
Yes ! 🙂
Until recently I have always tinkered with amps with the feedback loop closed. On a trip of trying to emulate 'tube' characteristics (with SS) I remembered that most (I think) tubes amps are stable and operable without the feedback. (in fact at a place I worked years ago we did the entire QC (tube amp) procedure open loop and as a final step tacked in the feedback resistor). I then decided I didn't need no stinkin' feedback afterall, but your point is right on. Icing on the cake..................mike
Hitsware,
Tube amps are highly unstable with feedback, depending primarily on the transformer! IIRC the Williamson had around 16dB of global NFB and was acknowledged by the designer to be on the verge of instability; this at a time when global negative feedback was regarded as the cortisone cure-all of audio, just as NONE is now!
The problem with the transformer is phase shift, which turns negative feedback positive, seriously affecting the amp's ability to handle reactive loads and making a mighty fine oscillator. So, since almost all tube amps use transformers, that one is out. Believe me, if tube designers could use 60dB of NFB like many SS amps, they would!
Cheers,
Hugh
Tube amps are highly unstable with feedback, depending primarily on the transformer! IIRC the Williamson had around 16dB of global NFB and was acknowledged by the designer to be on the verge of instability; this at a time when global negative feedback was regarded as the cortisone cure-all of audio, just as NONE is now!
The problem with the transformer is phase shift, which turns negative feedback positive, seriously affecting the amp's ability to handle reactive loads and making a mighty fine oscillator. So, since almost all tube amps use transformers, that one is out. Believe me, if tube designers could use 60dB of NFB like many SS amps, they would!
Cheers,
Hugh
- Status
- Not open for further replies.