andy_c said:The moon is made of green cheese. This is a fact.
If anyone thinks I'm wrong, the burden of proof is on you to demonstrate that my statement is false. To do this, you must construct a rocket ship, fly to the moon with it and sample the soil.
Sound reasonable?
Wrong albedo.
And you're risking SinBin by tempting me to break out into the Monty Python cheese shop routine.
O.K. I'll take the risk of being flamed and agree with what somone else already mentioned:
The darlington is consisting of two stages each with 100% feedback while the CFP has 100% feedback as well but this is forming a feedback loop around TWO stages. This leads to the disadvantage of decreased stability. The advantage however is that voltage gain can be made greater than 1, if feedback is reduced (and your circuit is made to take advantage of that).
I think none of them can be called better than the other. It is all depending on HOW they are used.
Regards
Charles
SE,
what is by your definition/subjective opinion an amplifier without any kind of global nor local feedback?
(for clarity I mean an amplifier which don't have any other kind of feedback more then around every each single stage which sets the gain for every each amplifier stage, just like for instance a VAS alone.)
Or don't they exsist?
what is by your definition/subjective opinion an amplifier without any kind of global nor local feedback?
(for clarity I mean an amplifier which don't have any other kind of feedback more then around every each single stage which sets the gain for every each amplifier stage, just like for instance a VAS alone.)
Or don't they exsist?
I would like to commend Jon Marsh for his comments of how Charles and I think about things. First, we have lots of experience, and we have tried lots of things at one time or another. For example, I first read about the 'feedback pair' being used as an output stage in the late 1960's, and in 1969 I built my first power amp with a complementary feedback pair output stage. What I found was that the feedback pair had problems with stability, and I actually had to compensate the pair (slow it down) in order to make it not oscillate. This gave me pause in using this output stage configuration in later designs, and I reverted to the more stable, complementary Darlington configuration in later designs.
This is why Charles and I often are surprised when people with a lot less experience demand proof about some opinion that we have. We are offering the results of our experience, as well as some objective info and perhaps some technical references, if we have have them. We can't 'prove' everything to everyone's satisfaction, it is not reasonable that we are required to do so, either.
This is why Charles and I often are surprised when people with a lot less experience demand proof about some opinion that we have. We are offering the results of our experience, as well as some objective info and perhaps some technical references, if we have have them. We can't 'prove' everything to everyone's satisfaction, it is not reasonable that we are required to do so, either.
That's right, John Curl! What are you trying to do, anyway? Confuse us with incorrect information? Well, I'm not going to stand for this, do you hear! You can just take your bogus information and keep it out of my forum! So there, you nasty old man!
I hate you,
Charles Hansen
PS -- just kidding, for those who can't tell...
I hate you,
Charles Hansen
PS -- just kidding, for those who can't tell...
john curl said:
I think people understand that no one can prove everything. That's pretty much a given.
A lot of times, statements here are made like facts. Take the LM317 thread where a poster stated that 1) it couldn't have been a soft start circuit; 2) the National guys screwed them up; and 3) a cap to the adj. pin cannot possibly improve ripple rejection.
apparently, the poor soul didn't take his own advice reading the datasheet, nor did he investigated it further, either in simulation nor in reality.
Sounds like he spoke from his vast experience and "expert" knowledge on the subject. But he was dead wrong on all those accounts.
I have no problem with people stating their experience ("I did it 10 years ago and this is what I got"), or their knowledge ("In my view, if A happens, B has to be true"). After all, how can anyone argue against experience or facts?
The debates, John, usually happen when people sell their "experience" or "expert" knowledge as universal facts, and ultimate truth. That, in my view, is a great disservice to the community at large.
The burden, in my view, is on the one making claims. they ought to be careful in their representation such that a one-time experience is presented as nothing but such.
Eva brought out a very good point. In some industries, two week old "experience" is useless (mine is such an example). so rather than saying "A is bad because I tried it", give your readers more context: when you tried it, how's the set up, etc.
hope it helps.
millwood said:
Millwood (stil no ^%$#**^% first name!),
As much as I hate to admit it, this time (this time only, you hear) I am in full agreement of your post above. Not I don't goof once in awhile, but I try to put my keyboard where my money is (huh? did I say that? Well, you get the point).
Jan Didden
Jan, "Mill" is his first name; he explained that in an earlier post. Though I don't know if he capitalizes it- like your alter ego "janneman", his shift key is informed by e.e. cummings.
Anyway, someone has already pointed out the Zeno Paradox aspect of feedback; I'll chip in with the observation that most of us can steer a car pretty effectively as long as it doesn't go too fast.
There, that tied cars and feedback together quite nicely, don't you think?
Anyway, someone has already pointed out the Zeno Paradox aspect of feedback; I'll chip in with the observation that most of us can steer a car pretty effectively as long as it doesn't go too fast.
There, that tied cars and feedback together quite nicely, don't you think?
I'd just like to point out that it was I and Jan who claimed load regulation would suffer with a cap between GND and ADJ, but we both admitted to being wrong. Fred adamatly stuck to it being helpful, as indeed it is.
Things would be much more productive around here if we could all agree to admit mistakes and involontary thread jacking, and just move on with the thread topics.
As for proof, I understand the difficulties with proving experience. However, for us less gifted/experienced, every bit of theoretical digression helps. If I'm told CFP-s - or what not - are bad, but given no shread of reasoning, I'm really none the wiser, as I'll have to go through pretty much the same process as the person who told me it was bad before I really know anything substantial. Not a bad thing in itself, but if that was how progress in general worked, we'd hardly be having amps to discuss in the first place...
I'm greatful for the breadcrumbs that come my way, and I do realize I can't expect to become a proficient audio designer just by hanging on DIYAudio. I'll probably have to read a few books and build a couple of hundred amps before I'm there, but the books lack the wonderful ability to talk back that this forum has.
Rune
Things would be much more productive around here if we could all agree to admit mistakes and involontary thread jacking, and just move on with the thread topics.
As for proof, I understand the difficulties with proving experience. However, for us less gifted/experienced, every bit of theoretical digression helps. If I'm told CFP-s - or what not - are bad, but given no shread of reasoning, I'm really none the wiser, as I'll have to go through pretty much the same process as the person who told me it was bad before I really know anything substantial. Not a bad thing in itself, but if that was how progress in general worked, we'd hardly be having amps to discuss in the first place...
I'm greatful for the breadcrumbs that come my way, and I do realize I can't expect to become a proficient audio designer just by hanging on DIYAudio. I'll probably have to read a few books and build a couple of hundred amps before I'm there, but the books lack the wonderful ability to talk back that this forum has.
Rune
I largely agree with Millwood too. I think in this latest post he was mostly right, although I still feel that more onus should be on the reader to sift and sort the information for his own special case. This means experimentation, building, and listening; but it does no harm and enrichs the experience of DIY anyway. The emphasis, after all, should be on the word 'DO'.
Let me give my experience on CFPs from some years back. This story is recounted as close as I remember it. It reflects my concern for how it sounds, rather than how it measures. It's my opinion that measurement can only go so far; this stuff is too complicated to be described solely by measuring things we know about.
I built two conventional, PP Class AB SS bipolar device amplifiers of classic design. Nothing fancy, a couple of kitsets from a local magazine.
One had the conventional double emitter follower, with Vbe multiplier for bias generation.
The other, identical in every respect except for the output stage, used a CFP, with 100R from emitter to base of the output device, and the same Vbe multiplier. I used MJE340/350 drivers, and 2SC3281/2SA1302 outputs on both amplifiers. Biasing was the same.
I set up a listening test with identical speakers and the same CD, switched. The test was mono. I invited the neighbours in.
Perhaps the best comment came from an 82 year old neighbour with marvellous ears for his age who remarked:
"The DEF sounds OK. The CFP sounds mechanical, like a machine, there's something strange I don't like." The others agreed with him that this described adequately what they were hearing.
So I pulled down the CFP amp, placed it on the workbench, and on the CRO with 1KHz at 20Vpp noted short term oscillation, of the order of half a volt, around crossover. It was pretty fast, I'd guess around 5MHz. I placed 470pF across the base/collector of the PNP driver (negative rail), and it was gone.
I resumed the test. The 'strangeness' was gone from the CFP, and it sounded OK. But there was no life in it, and someone remarked (we were using a vocal track at the time) that it was '...as though the singer was suddenly very tired'.
Thus emboldened, I concluded that by the time you fixed the short term transient instabilities of the CFP in Class AB, you'd killed the music stone dead. I now exclusively use the DEF, although on my Class A SE amp I use a CFP with bipolar/mosfet. Go figure.......
I have all sorts of theories about why CFPs sound dead in Class AB, and most are subject to varied interpretation, but I'm really only interested here in the outcome. Explaining this has cost time and effort, and I like to think I have no obligation to do it, but people are interested, for what it's worth, it's my experience, and interestingly it closely aligns with JC's experience. When people want chapter and verse I'm a little reluctant to roll up my sleeves and make the effort because time is always short.
BTW, I almost fell of my perch reading Charles' hilarious comment!
Cheers,
Hugh
Let me give my experience on CFPs from some years back. This story is recounted as close as I remember it. It reflects my concern for how it sounds, rather than how it measures. It's my opinion that measurement can only go so far; this stuff is too complicated to be described solely by measuring things we know about.
I built two conventional, PP Class AB SS bipolar device amplifiers of classic design. Nothing fancy, a couple of kitsets from a local magazine.
One had the conventional double emitter follower, with Vbe multiplier for bias generation.
The other, identical in every respect except for the output stage, used a CFP, with 100R from emitter to base of the output device, and the same Vbe multiplier. I used MJE340/350 drivers, and 2SC3281/2SA1302 outputs on both amplifiers. Biasing was the same.
I set up a listening test with identical speakers and the same CD, switched. The test was mono. I invited the neighbours in.
Perhaps the best comment came from an 82 year old neighbour with marvellous ears for his age who remarked:
"The DEF sounds OK. The CFP sounds mechanical, like a machine, there's something strange I don't like." The others agreed with him that this described adequately what they were hearing.
So I pulled down the CFP amp, placed it on the workbench, and on the CRO with 1KHz at 20Vpp noted short term oscillation, of the order of half a volt, around crossover. It was pretty fast, I'd guess around 5MHz. I placed 470pF across the base/collector of the PNP driver (negative rail), and it was gone.
I resumed the test. The 'strangeness' was gone from the CFP, and it sounded OK. But there was no life in it, and someone remarked (we were using a vocal track at the time) that it was '...as though the singer was suddenly very tired'.
Thus emboldened, I concluded that by the time you fixed the short term transient instabilities of the CFP in Class AB, you'd killed the music stone dead. I now exclusively use the DEF, although on my Class A SE amp I use a CFP with bipolar/mosfet. Go figure.......
I have all sorts of theories about why CFPs sound dead in Class AB, and most are subject to varied interpretation, but I'm really only interested here in the outcome. Explaining this has cost time and effort, and I like to think I have no obligation to do it, but people are interested, for what it's worth, it's my experience, and interestingly it closely aligns with JC's experience. When people want chapter and verse I'm a little reluctant to roll up my sleeves and make the effort because time is always short.
BTW, I almost fell of my perch reading Charles' hilarious comment!
Cheers,
Hugh
This is almost funny to me. Yes, 1969 is a long time ago. Yes, some power devices have improved, yet others have stayed about the same. What about the 2N3055? Some folks use them, even today.
Today's devices can be faster, but the feedback pair creates the same potential problem, whatever the devices. It can be a problem, BECAUSE it has a feedback loop, AND it can be unstable without a dominant pole being created. The Darlington does not have this problem.
Today's devices can be faster, but the feedback pair creates the same potential problem, whatever the devices. It can be a problem, BECAUSE it has a feedback loop, AND it can be unstable without a dominant pole being created. The Darlington does not have this problem.
Every CFP output stage that I have built has a tendency to oscillate in the 2-5 MHz range. And I have built lots of them. I use different tricks other than slowing them down.
Anyway.......I became tired of messing with them, and changed to a diamond buffer output stage.......which means open-loop all the way.
My customers.......(the key phrase here).......all felt that the diamond buffer sounded better than the CFP version, using the same front end.
As long as they like the results, and their check clears, I really don't care what the resident non-believer thinks.
BTW......Charlie........your e-mail bounces with the account that I am using in the interim. Wonder if it will work when I return home and use the business account again.
How long did it take to fix that snafu?????
Jocko
Anyway.......I became tired of messing with them, and changed to a diamond buffer output stage.......which means open-loop all the way.
My customers.......(the key phrase here).......all felt that the diamond buffer sounded better than the CFP version, using the same front end.
As long as they like the results, and their check clears, I really don't care what the resident non-believer thinks.
BTW......Charlie........your e-mail bounces with the account that I am using in the interim. Wonder if it will work when I return home and use the business account again.
How long did it take to fix that snafu?????
Jocko
Old but good...
To me the more important thing that i have read from you...is the use of a mosfet as the driving element for a bipolar output stage...because the mosfet provide isolation from the variation of impedance across the frequency range...and help to isolate the VAS stage from that variation..
This can be old...but good and true!
Cheers
john curl said:
To me the more important thing that i have read from you...is the use of a mosfet as the driving element for a bipolar output stage...because the mosfet provide isolation from the variation of impedance across the frequency range...and help to isolate the VAS stage from that variation..
This can be old...but good and true!
Cheers
Re: Old but good...
I said that too, in the JLH thread. But they shot me down,
Tube_Dude said:
To me the more important thing that i have read from you...is the use of a mosfet as the driving element for a bipolar output stage...because the mosfet provide isolation from the variation of impedance across the frequency range...and help to isolate the VAS stage from that variation..
This can be old...but good and true!
Cheers
I said that too, in the JLH thread. But they shot me down,
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.