 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
108.213.92.252
In Reply to: RE: Does an individual power tranny for driver stage improve posted by drlowmu on July 03, 2012 at 09:15:53
used and listened to a quality (Pimm) CCS ? ? ? ......
Perhaps you should before making claims, otherwise you are poopooing
something based on your theories, with no experience, which is exactly what you complain about in other's critiques of your experiences without the same. Hello?
Follow Ups:
It should be pointed out that while a CCS does isolate the circuit from the power supply that's not it's main function.
The main reason for using a CCS as a plate load is to insure a horizontal load line for the triode.
With a triode, the more horizontal the load line the lower the harmonic distortion is and the higher the gain is.
I'm not telling you anything you didn't already know.
Thanks.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
No Grant, I have not listened to a Pimm CCS.I am reasonably smart, and have lots of experience. I HAVE the right approach to start with, developed over thirty years, and I do not need to add lesser approaches and complexities to my KISS designs.
A CCS can be very good for cheap amps, or, if you don't know how to build the supply well to start with.
"If the CCS improves your amp, the supply is not built as best as it can be built."
My Power Transformers are custom built to my spec, under 9 Ohms DCR, and my L1 and L2 chokes, when added in series, total under 7 Ohms (DCR - for L1 PLUS L2 )!!
Steve Brown, an experienced Forum member, posted a response to you, and it appears as though he has built both approaches. His UNSOLICITED final conclusion is :
" On a day to day basis, and for pure musical enjoyment, the 2A3 with low DCR supply is the one that wins out. By the way, I've also had two friends who are musicians and very familiar with my equipment listen and they agree as well. "
Cheers,
Jeff Medwin
Edits: 07/03/12
they are complimentary as each delivers its own set of benefits. It just so happens that one of the CCS benefits overlaps into the domain of the PSU.
Naz
So what you are claiming is that it is not possible that anyone can build a better sounding amp than your design.
How discouraging to the rest of us. We might as well quit.
I have. I have several of Gary's CCSs (various revs) and have used them to load drivers for a number of both DC and cap coupled SETs. Yes they sound great. Currently I'm using a direct coupled 6SL7 to 2A3 (MQ transformers) and a power supply modeled after Jeff's advice. I can currently compare it to two other SE amps (45 with driver loaded by a Pimm CCS, and a more conventional 300B design similar to JE Labs). I can also compare the 2A3 amp to several PP designs. On a day to day basis, and for pure musical enjoyment, the 2A3 with low DCR supply is the one that wins out. By the way, I've also had two friends who are musicians and very familiar with my equipment listen and they agree as well.
Using a CCS to load a driver and a low DCR PS can exist in the same amplifier BTW. The possible benefits of the supply can act on the 2A3,
while leaving the performance of the CCS loaded driver unchanged, and perhaps unmatched. There's little comparison between a 12AX7 loaded with 280K and a 5687 loaded with a CCS, in terms of distortion.
The same unanswered question remains:
Is it "true" or somehow "provable" that a low current hi-mu hi-Rp driver is more sensitive to low level information than a high current lo-mu lo-Rp version ? ? ? ?
We've heard the "opinions", but so far each opinion has included far to many "other" variables, attenuation, how and how much, being one of the larger ones IMHO. I wish I had the test equipment to set up a test for this, but I don't. I bet somebody does
"There's little comparison between a 12AX7 loaded with 280K and a 5687 loaded with a CCS, in terms of distortion."
I agree and you don't necessarily need test equipment other than your ears.
The math says so, the test equipment would say so (if you had it) and your ears say so.
When all three agree, it's a good thing.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
You have the test equipment, it is on either side of your head...called ears. Listen to Dennis amps and you will know what performs.
Jeff Medwin
I never seem to get get the question I pose to be specifically addressed.
IMHO, even though distortion differences can be measured or "proven" with math, as well as "sometimes" heard, (in the two scenarios stated above),
Dennis has claimed that a lo-mu high current driver ("unless driven hard")
will not reproduce low level sonic information as well as a 12AX7.
Now, keep this statement in mind, and NOTHING else for a moment please.
If a test was to be conducted, how would we do it?
Thoughts:
First pick an output level, or power dissipated, let's say 0.25W which
will drive an efficient speaker (95-100db) to plenty of volume.
Then, because we have different mu's we have to do some math and figure out how much drive signal must be applied to each driver scenario to get the 0.25W final power out. We must try to remove attenuation in this test
"the best we can". Then carefully pick various frequencies with a signal generator and progressively lower each's output proportionately in each case until one scenario "doesn't reproduce when the other one still does"
or something like that?
We're NOT talking about distortion in this case, the difference there
has NO DISPUTE! (unless you're a fool)
But, "lowest possible" distortion is NOT always "better" especially when the worst of the two being compared is still "low enough"...right?
As a listener, I will have to hope or assume that other listeners/builders
would like to get to the bottom of this, as low level detail is a desirable attribute. Though other factors, not discussed (yet) are ALSO contributing to low level performance, let's stick to this particular ONE of them, for now please.
For the sake of trying to put some objectivity into this discussion, distortion aside, what you are describing is dynamic range which can be both measured and heard.
We all agree that we can achieve exactly the same BW through design and we can verify this. So let's start by designing the two different drivers with equal BW to take that out of the equation.
Next, ensure that the gain is equal ... also easy to do.
The major difference between the resulting drivers is that their OP impedances and hence drive capability is nowhere near identical but even this can be largely mitigated by ensuring an extremely high impedance load.
If we were simply to measure the two driver stages in isolation, ie not driving an OP tube or any load it would be easy to measure the dynamic range at all frequencies within the band on equal terms. Assuming that the max OP swing to drive an OP stage to clipping is the same for both drivers, you could simply reduce the input signal to a level where the S/N ratio is considered to be near the limit. This will vary from tube to tube but generally, high gm tubes will easily win this comparison, with better S/N and hence better low level resolution.
The problem in the real world is that everything interacts, which makes it extremely difficult to compare apples with apples. In this case, connecting the driver to the OP tube already skews the result because Miller and the necessary grid leak R of the OP tube provide a load which is more significant to the driver with high OP impedance and which is elliptical and potentially very significant at HF. Of course, you could buffer the drivers but that adds a further variable.
In reality, assuming good design (taking load lines and BW etc into account), there are far too many factors at play to say categorically that any one topology is better than another in anything like all circumstances and IME more often than not there is greater difference between a good and bad sounding tube of the same type.
Other factors? Well, for many reasons, some tonal differences are perceived as providing greater or lesser resolution and this is particularly true at higher frequencies. Synergy between components is another factor both at a micro level (Rs Cs, Ls and tubes etc) or at a macro level (speakers and the like). Some components are complimentary but too much of a good thing can be a problem.
Naz
Hi Naz,The design in question, (a Loftin-White with half a 12AX7 DCed to the grid of a 2A3), is designed opposite of what 99.99 percent of the people are taught to do.
You stated the following assumption:
"Assuming that the max OP swing to drive an OP stage to clipping is the same for both drivers, "
This common approach, (assumed by your wording above), is designing for "maximum power at clipping" of the finals stage.
When operating, the Loftin White circuits of Isamu Asano (.43 mA.), Nobu Shishido ( 1 mA.) and Dennis Fraker ( <0.8 mA.) will have the driver stage softly clip BEFORE the finals reach maximum output, and it optimizes the amp's performance while it is running under one Watt, exactly where the High Efficiency speaker operates for 99% of one's listening time.
For 100 years, engineers have designed amplifiers for "maximum power".
In the case of these Loftin-Whites, we are talking about designing for "maximum FIDELITY under the operating conditions the amp SEES 99 percent of the time". This is one of the reasons why the L-W approach, (especially when implemented well as Dennis does), sounds so darn good to me and thousands of others around the world. !!
It renders some of these Forum discussions to be a waste of my time, and any talk of Miller as being stupid, archaic, and useless.
The proof of the pudding is in the eating. In the actual power range of where the L-W amp operates, one easily hears the adequacy of the design, so one can't really argue any !!
Your 845 amps are no doubt brilliantly designed to do about 15 Watts cleanly, because your speaker load is not highly efficient. That is mighty fine engineering Naz.
The amplifiers we are discussing are designed to do zero to one Watt cleanly, because that is precisely how the high efficiency speaker loads the amps, and where it was designed to play.
Simple isn't it.
Jeff Medwin
Edits: 07/05/12
- "Assuming that the max OP swing to drive an OP stage to clipping is the same for both drivers," was made purely in the context of providing the poster a direct answer to his question, something some here seem to ignore. The discussion was about TRYING to compare two drivers on an apples with apples basis but you obviously didn't get that.
I made no assertion as to how I would design the driver, especially the headroom that I would allow. I have my own views on that which I freely share with anyone who is interested, but do not push my beliefs with religious fanaticism upon anyone.
I have also purposely refrained from arguments over what is best because I believe there are thousands of ways to achieve the same sound, though I do think that synergy deserves more credit.
As for your rubbish about any one BEST topology or design over another, well ... no need to repeat myself.
Regarding your comment about my 845 design or the efficiency of speakers in general ... again you know my views. Your opinion of the attributes or otherwise of mine (or anyone else’s) equipment or design that you haven’t seen, heard or mostly even understand merely reflects your arrogance (or is it ignorance).
What I do find curious is that on the one hand you throw out engineering as being useless and a waste of your time, eg your comment - "and any talk of Miller as being stupid, archaic, and useless" but on the other hand, ever since you learned how to use PSUD, you have pushed its benefits. So you selectively believe in engineering only when you think you understand it, it suits you or you believe it benefits your cause?
Don’t get me wrong here my foot is firmly in both camps, sound engineering first and then using a combination of engineering AND listening to get the best result.
The sad thing is that occasionally, when you refrain from technical argument and limit your comments to your opinion you actually make some sense. But then you can't help yourself and go and blow it all.
Anyway, this will not deteriorate into a slanging match because I don't have the time or the patience for it. I’ve said my piece and will leave it at that.
Naz
Naz,
Your comments and even-handedness are appreciated by many of us up here. First class, really.
Cheers,
Jeff
"Currently I'm using a direct coupled 6SL7 to 2A3 ......"
Is the 6sl7 CCS loaded?
If so, are you using the "mu" output of the CCS?
If you are using a CCS on the 6sl7 can you listen using the plate of the 6sl7 for the output and then listen using the "mu" output of the CCS and tell us what you hear.
Thanks.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
No, it is not CCS loaded, it is resistor loaded, sorry I should have clarified that. I'm not sure the Pimm CCSs will work with such low current (ca. 1ma) but it might be interesting to try.
Thanks anyway.It would have been a good test.
The output impedance of that 6sl7 stage is not low enough to fully drive the Miller capacitance of the 2a3.
The result is a small loss of high frequencies and a phase shift that extends well into the audio band.
If you had a CCS loading the 6sl7 plate you could use the mu output and have a drive impedance that would be low enough to fully drive the Miller capacitance of the 2a3. (but you still wouldn't have the necessary drive current so maybe the whole point is moot)
Oh well, maybe next time.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 07/03/12
Tre': You sound like a broken record, saying the same thing over and over. While I don't disagree with what you are saying, I don't believe that such designs are sonic junk because they don't have zero phase shift within the audible range. Have you seen the phase response on loudspeakers? 99.9% have excess phase shift and group delay that are MUCH worse than suspect amplifiers. And yet, many of these speakers sound good. The human ear and hearing mechanism is not the same as an AP or o'scope looking at square waves.
Donald North
I just said it would have been a good test.
Comparing that same circuit with the only change being the Miller fully driven vs. not.
If one likes that circuit they might like it more with the Miller fully driven.
"While I don't disagree with what you are saying, I don't believe that such designs are sonic junk because they don't have zero phase shift within the audible range."
Neither do I but I'm not buying into the opposite claims made so many times we've lost count.
I'm looking at one now.
http://www.audioasylum.com/forums/set/messages/7/71278.html
"Just design your equipment so it does what it's supposed to do-- AND NOTHING ELSE."
Delusions of grandeur when the thing doesn't even overcome Miller?
That's my point, not that the thing is junk and nobody should like the sound of it. Everybody likes different things.
That OK, I just wish Dennis would stop saying "it's the best amp ever made" and "everyone who doesn't understand that from a technical (common sense) point of view has no common sense".
From a technical point of view, the thing is flawed.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I'm sorry, are Steve and Dennis the same person?
As for the Miller issues and testing, does running the mu-out of a CCS alter current loops versus not using mu-out. Does the load line change with CCS versus R? Personally, I am not sure as I have not investigated CCS in detail.
Point is, reductionist thinking and experiments are helpful only when we use them effectively and draw appropriate (non)conclusions. Even then, they get us only so far before we need to return to other methods of investigation...
Cheers.
“As long as we have any intention to be right… we should be wary. So long as words have the slightest ego attachment, they are dishonest.” Charlotte Joko Beck
"As for the Miller issues and testing, does running the mu-out of a CCS alter current loops versus not using mu-out."Yes.
"Does the load line change with CCS versus R? "
Yes.
Even a poor CCS will represent a very high impedance. Gary Pimms Self Biased CCS represents more than 2 billion ohms mid band.
The grid resistor (if there is one) of the following stage will determine the load line.
In a direct coupled setup, a driver tube loaded with a CCS will operate into a absolute horizontal load line except where the reactance of the Miller capacitance dictates otherwise.
Because of the high impedance of the CCS, the power supply is not part of the (AC) current loop.
If the mu output is used, 1/2 of any current needed to drive the Miller of the following stage will have to come from the power supply in addition to and above the set current that is already being delivered to the tube through the CCS.
This places the power supply back in the (AC) current loop.
This is why Gary uses a 'CCS feed VR tube shunt regulated power supply' in these instances. (as do I)
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 07/05/12 07/05/12 07/05/12
Horses for courses....Miller means squat.
Be careful, don't go out at night, the BOOGIE MAN may get you.
Cheers.
Jeff
Edits: 07/05/12
Jeff, You quoted your own post, above, wherein you first wrote "Miller means squat". That hardly suffices for citing an independent source to corroborate your blanket statement. When you wrote it the first time, I assumed you meant with reference to a Loftin-White topology. Now, when you repeat it here, it seems more another "rule" that cannot be questioned. Can you in any way amplify on your statement? Why does "Miller mean squat", in your opinion?
Lew,
Well, by all means, Miller must be considered in a design. I just like to be dramatic at times, picks up the Forum from the summer time blahs.
But in terms of Loftin White, a half of a 12AX7 will beautifully drive a 2A3 if you design the amp the way Dennis does, as a total design.
I have had several others on this Forum build L-W 2A3 amps, and when they, or their audiophile buddies hear the amps, there is NEVER ever a discussion of Miller. The amps kick butt and they are very pleased.
Dennis is optimizing the first zero to one Watt...where the high efficiency speaker is operating 99.9 percent of the time, and he is not worried at all about how well it drives the finals at three point five watts.
One other thing, I was privy to what the EEs measured as far as frequency response of Dennis' amps at the University, and Lew, it was spectacularly wide band and linear. One can hear that at the RMAF show if one listens carefully, or, in private installations.
IF the amp is fed really good source materials, it responds and is exceedingly honest and neutral. In the last six years, I think Dennis has only had two instances when his digital (barf) source was up to snuff, 2005 when S. Harrell flipped-out in his Six Moons review, and in 2011, third day, Sunday, when the Berkeley and new computer server settled down.
This year, Dennis will bring a good digital source again, Dennis redid the computer's power supplies, and he'll tote his Berkeley to the show. With that, his amps seem to have no limits on good material. By Sunday, its FUN Lew, really FUN. Ask Dave Davenport privately, he took in 2011.
Jeff Medwin
(1) I had a feeling you were playing the gadfly (and it worked).
(2) My query was not implicitly meant to cast a slur on Dennis' amplifier. I have never heard it.
I try to learn from this forum. To do so, I have also to figure out what is hyperbole (e.g., your statement about Miller capacitance), what is just crap, and what it is that I should retain. It's not always easy. Thx for your more tempered explanation.
My own lifelong approach to audio has always been to start with the speakers I like best and work backwards from there. Choosing the speaker first places some constraints on the choice of amplifier. A one-Watt amplifier just will not cut it with my speakers of choice, even though I/we recently made a huge advancement in improving their efficiency via a mod to the crossover which in fact allows one to do away with the crossover entirely. High efficiency is a worthy goal in speaker design, for sure.
The mu out of the CCS cannot provide any more current even though it might be lower impedance so what makes you think it will drive the 2A3 any better?
cheers,
Stephen
It won't, that why I added "(but you still wouldn't have the necessary drive current so maybe the whole point is moot)" in my post to Steve Brown.http://www.audioasylum.com/forums/tubediy/messages/20/208642.html
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 07/04/12
Sorry, should have read better.
And yet Steve prefers such an amp over several that do not 'suffer' such issues. Perhaps the advantages outweigh the disadvantages, even if they are not all quantified (yet).
Interesting.
“As long as we have any intention to be right… we should be wary. So long as words have the slightest ego attachment, they are dishonest.” Charlotte Joko Beck
Then why is a resistor loaded 12AX7 the perfect tube to drive a 2A3?
sorry...
Chris
Because it isn't.
I much prefer the 6GK5 CCS loaded.
A 200H plate choke is better yet.
CB and others prefer the 6GK5.
DanL
Post a Followup:
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2013, Audio Asylum, all rights reserved ]
[ Take5 ]
