|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
209.137.226.179
In Reply to: RE: The road ahead posted by Tre' on June 24, 2012 at 11:48:28
Good advice for anyone. I wouldn't, however, make too many assumptions about bandwidth realities in my applications-- you haven't seen or tested these items.
Some of my application info. is proprietary, but I think you can find out from others that your figures are wildly-- not applicable to this particular item.
While the formula is correct, the assumptions being used to insert figures into that formula-- in this case-- are incorrect.
---Dennis---
Follow Ups:
"Some of my application info. is proprietary, but I think you can find out from others that your figures are wildly-- not applicable to this particular item."I very seriously doubt what you have said.
Having said that, I was speaking not about the frequency/phase response of the amplifier as a whole. Just the response of the filter that is very much reality and has to be dealt with.
How you may or may not go about correcting this problem after the fact is not my concern.
I feel that it is always best to avoid this types of problems to start with rather than compensate for them after the fact.
"While the formula is correct, the assumptions being used to insert figures into that formula-- in this case-- are incorrect."
The output impedance of your 7b4 driver stage is easy to measure or calculate.
The Miller can be easily calculated once the interelectrode capacitance's and the mu are measured, so your use of the term "wildly" is not called for.
If my numbers are off, please feel free to correct me and give us the correct numbers.
In any event, with a 7b4 driver tube with a plate resistor the output impedance will be too high for a DHT's Miller capacitance. The resultant low pass filter will have consequences in the audio band. There's just no way around that.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 06/24/12 06/24/12
You're not entirely wrong here. Unfortunately, it's your emphasis on only one part of an entire design that is way over the top.The Miller Capacitance of a DHT is a very real parameter, as you state, and it must be considered, as you correctly point out.
When I build an amplifier, that design is supposed to replicate a studio recording in the best way that I can build.
ALL amplifiers are sets of compromises, and in all cases, one item in a design comprising many parts and calculations may have to be changed in one or another direction-- but not too far!-- in order to allow for another important consideration-- to NOT go ignored!
In this case, I'm not willing to give up the transparency and speed of a High-mu, low plate current driver just to satisfy a desire for the ultimate in powering Miller Effect inside a DHT, although I do consider DHT bandwidth very important, as I do slew rate.
I fully understand the truly awesome performance that comes from a high-mu, low plate current driver-- as opposed to higher-current stages which are much harder to drive themselves AND THUS SEVERELY COMPROMISE THE SIGNALS COMING INTO AN AMPLIFIER FROM MUSICAL SOURCES-- that are used to drive amplifiers..
In the case of the High-Mu driver stage, it is very important to avoid overloading its output... and I recognize that you have noted this.
Accordingly, I allow wide voltage swings with high-impedance loading on the driver stage output. Certainly, the Miller Effect of the DHT being driven is also loading this stage, but I am able to get the bandwidth I need anyhow by carefully evaluating all of the factors that impinge on it.
So, the simple calculation of the filter network that is always part of a driver stage's loading, is not the only consideration there.
In your own case, I would be cautious of overestimating Miller Effect in DHTs-- different brands are very different, and plate currents and voltages that are used in a circuit operating the DHT are also contributors to the same tube exhibiting DIFFERENT Miller Effects under different operating conditions.
Miller Effect Capacitance is NOT a numerical figure that one can simply plug into a formula-- it changes with operating conditions, and with brands of tubes.
Overall Balance is everything.
In this case, Miller Effects have been recognized, but the much larger other factors present in the circuit design have also been considered-- and kept close to THEIR optimums-- all the while still keeping in mind the limitations imposed by Miller Effect, and not going overboard on that, either.
We want ideal TOTAL loading-- of the driver stage as much as is practical so that Miller Effect-- which really isn't that bad at all-- is really the only thing left that negatively affects driver stage performance.
By using such a comprehensive, overall approach, the Miller Effect that is present here simply isn't a problem.
The final result of an entire amplifier's performance is what I'm after, and in this case I have it right where I want it.
---Dennis---
Edits: 06/24/12 06/24/12
You state this above :
"The final result of an entire amplifier's performance is what I'm after, and in this case I have it right where I want it."
That says it ALL, and it sure SOUNDS that way to me and many others who listen.
I respectfully suggest you do not waste precious time on "non issues". Your circuit and your most thoughtful product implementation speaks so eloquently for itself, and it always has !!
Jeff Medwin
The Miller capacitance of a tube is the (static) grid to plate capacitance multiplied by the mu of the tube plus the (static) grid to cathode capacitance.The output impedance of a resistor loaded driver stage is the plate resistance (at the operating point) in parallel with the plate load.
You used a lot of words to say very little.
"Certainly, the Miller Effect of the DHT being driven is also loading this stage, but I am able to get the bandwidth I need anyhow by carefully evaluating all of the factors that impinge on it. "
If you would just say what the output impedance of you driver stage is (and how you arrived at that number) and the Miller capacitance of your output tube we could settle this.
I don't see how the output impedance of a resistor loaded 7b4 driver stage can be low enough to prevent the Miller capacitance of a DHT from causing problems within the audio band.
You say you have done this, I say you can not have done this.
Unless you are willing to explain, in plain English, how you prevent this low pass filter from negatively effecting the audio band I will go on believing that it does.
I think that is fair and reasonable.
BTW You have said in the past that your amplifiers start to roll off at 15kHz. You also stated that with NOS tubes this roll off can start as low as 9kHz do to the increased Miller of some of the older tubes.
That sounds like an admission that the high output impedance of your driver tube is causing (in all cause, even with the new tubes that have lower Miller) problems in the audio band.
If, when you say that you get "the bandwidth I need", includes roll off and phase disturbance within the audio band, I would propose that the bandwidth you need may not be the same as the bandwidth myself and most everyone else needs.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 06/24/12
All tube amps have some rolloff.
Tubes themselves have a lot of it. Just broken-in tubes often greatly outperform tubes with more hours on them.
A good question would be "why do we bother with tubes at all?".
I think we like the transparency INTO the musical layers of a soundscape that tubes are better at, than are more mathematically flat-curved components.
No one here would argue that DHTs have perfect, flat powerbands. We all realize that pentode operation or BiPolar S/S devices, or even Mosfets can cream all DHTs for bandwidth.
The arguments will go on for millenia. Some even think that DHTs should be "corrected" by circuitry to deliver more "flat" response. NO! ALL devices sound BEST when NOT pushed, molested, or tinkered with in any way!
My own experience is that once a certain threshold of information that the human hearing system needs to form a certain sonic picture is reached-- then adding more bandwidth to that threshold produces nothing more of value.
All of us have experienced what happens when you make a tape recording of a tape recording. The amount of complete sonic information that the human hearing system needs is not met in the second copy, so the result is a drastic drop in fidelity in the second copy.
Then, also-- we know what happened with the older CD players. Although these had plenty of bandwidth and low distortion specs, they fell VERY short on usable sonic information that the Human Body can use--and simply sounded like trash when compared to vinyl systems.
At this time, we are also experiencing this same phenomenon in a sort of reversed way-- in Digital computer audio.
On LESSER systems (ones that, like the tape recorders of old, lose too much information THAT THE HUMAN BODY CAN USE)-- sound better when High-Resolution material is played, rather than the old Redbook CD standard-- 16 bits @ 44.1 KHZ.
YET-- when really excellent equipment is used to play a well-recorded CD that is only 16/44.1-- it sounds OUTSTANDING! AND, higher resolution material may very well NOT sound any better at all-- in fact some of it is MUCH WORSE!
Well, what is going on here? It's simpler than it appears: once the Human Body gets what it actually needs to form a sonic impression completely, then THINGS WORK, and NOTHING MORE IS NEEDED, and in fact, SHOULD NOT BE PROVIDED..
Take things BELOW that level, and things rapidly fall apart musically, regardless of specs or bandwidth.
SATISFY that level, and anything that you do BEYOND that level will produce very little-- if any-- improvement.
The moral of the story is just like life itself: DO WHAT ACTUALLY WORKS.
---Dennis---
Dennis
I agree on some things and disagree on others.
> > All tube amps have some rolloff.
Disagree
You CAN design for 10-200KHz bandwidth
and so rolloff would be negligible.
> > Just broken-in tubes often greatly outperform
> > tubes with more hours on them.
Agree
> > ALL devices sound BEST when NOT pushed,
> > molested, or tinkered with in any way!
Agree
> > All of us have experienced what happens when
> > you make a tape recording of a tape recording...
> > the result is a drastic drop in fidelity
> > in the second copy.
Disgree
I had excellant 3 head tape decks back in the 70's
that had adjustable bias so when I made copes they
were indistinguishable from the originals.
A well recorded tape will sound better than
well recorded CDs when each are played on
equally high quality equipment.
Vinyl beats them both though.
> > when really excellent equipment is used to play
> > a well-recorded CD that is only 16/44.1--
> > it sounds OUTSTANDING!
Agree
I use NOS DACs because (as stated above)
they don't "molest" the signal 8^D
> > > higher resolution material may very well NOT sound
> > any better at all-- in fact some of it is MUCH WORSE!
Agree
Because they tend to "molest" the signal.
> > > once the Human Body gets what it actually needs
> > > to form a sonic impression completely, then
> > > THINGS WORK, and NOTHING MORE IS NEEDED,
> > > and in fact, SHOULD NOT BE PROVIDED..
"There's the rub" (Shakespeare)
But that is different for everybody.
I have hearing out beyond 20KHz so
my level of HF energy needed is less
than those who only hear to 15KHz.
My room may be different ...
My speakers may be different ..
My musical taste may be different ...
My listening volume may be different ...
So to even suggest that one product,
no matter how well designed it is,
can satisfy everybody is foolish.
> > DO WHAT ACTUALLY WORKS.
I would add - "for you" to that.
DO WHAT ACTUALLY WORKS FOR YOU.
DanL
Some really astute observations here!
Ok-- you wonder what I'm thinking about this:
"All tube amps have some rolloff". I should have said all naturally operating tube amps have rolloff. I left out the application of negative feedback, the application of controlled positive feedback, and the reduction of frequency peaks which will result in frequency troughs appearing larger because the peaks have been reduced by comparison to the troughs. I wouldn't recommend any of these things to anyone who is actually trying to listen to music, however.
"I had excellent 3-head tape decks back in the 70's"...... Agreed, I did also-- The large Ampexes running one-inch tape at 32 IPS, the Revoxes, The commercial 10 3/4 inch Sonys and the large Tascams, etc. These were all superior to the best vinyl-- it was badly compromised by dynamic compression, which included the infamous RIAA curve. It still sounds OK at it's best, but it's no longer anywhere near the standard today. Can you make-- today-- a really great L.P.? You can, and some do! I have no argument there at all.
And, I agree with you some more. No one appliance is going to please all the women in the kitchen. They're going to get what they like regardless of you or I. We can, however, seek to allow maximum dynamics to get through, to allow max. transparency to get through, and to build amplifiers that can actually drive speakers properly. We can do a lot.
After we've done all that we can, the Owner is still the Boss.
---Dennis---
"Many typical audio tubes are good to 20 or 30 MHz with proper RF circuit design, assuming that you are actually making an RF amp. The old Metal 6L6 will work on the 20 meter ham band (14 MHz) and put out considerable power. Even the worst type of tube should cover up to a few MHz, so frequency response of the tube itself is not a factor in an audio amp."
From tubelab.com answering the question "Is there any available data sheet that shows output tubes' frequency response?"
You know Dennis, you really do live in your own little world.
Your post did not answer my question in any way.
It did, however, give us insight into your thinking.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
To me, talk is just that, 'talk'. I'd like to see (hear) a 'showdown of the ideas' and put one of the sides to rest, kinda. (Its not that either of you are totally wrong.) That's one of the shortcomings of this online forum. This 'talk' can go on forever. Break out the 'guns' (amps) and put this thing to rest.
When the word 'religion' is used, the first thought that comes to my mind is 'yuck', next is dogma, indoctrination, stiffnecked, hardhearted, the systematizing of error, unyielding, conscience seared with a hot iron (soldering iron?)......anyway, you get my point. God doesn't like religion and neither do I, concerning anything. IT AIN'T GOOD. When you become religious about anything, you have shut off the learning aspect of it. I believe neither of you are totally religious, you're both intelligent, but someone has to be wrong at some point. I vote for a showdown.
L.D.
Tre and I have locked horns plenty enough in the past, but mostly over semantics rather than anything else.
The math is the math though, and try as one might to ignore it, the fact of the math remains and does manifest in any design. Sounds to me like Dennis hand-picks his tubes to get bandwidth, if he does any measurements.
I asked a question of Dennis in a thread below which went unanswered. It had to do with heating in the VC of a loudspeaker. Maybe he could answer it here....
If not, I have to assume that the VC heating thing isn't real.
I found that there are different ways that an amplifier can drive a speaker. It depends on output impedance and the like, and has a lot to do with how the speaker is designed. Once I sorted out some of the differences in the design rules I wrote a short paper explaining what was going on. I have posted a link.
The paper ignores the current-drive method as that never really developed into what I would call a paradigm. But there is plenty of evidence for the Power Paradigm and the Voltage Paradigm is well-known. Most any SET, BTW, is a Power Paradigm device. I wonder if this is what Dennis was talking about with his VC heating comments?
The showdown would not be between two amplifiers.Shortcomings at one stage of an amplifier can be compensated for in the next stage.
What I'm talking about is the low pass filter created between the output impedance of the driver stage and the Miller capacitance of the output tube.
This low pass filter is real. There is no need for a showdown.
Dennis has already admitted that his amplifier suffers the effects of this filter in the audio band.
I would say the showdown already happened.
Dennis thinks his amplifiers sound good as they are.
"My own experience is that once a certain threshold of information that the human hearing system needs to form a certain sonic picture is reached-- then adding more bandwidth to that threshold produces nothing more of value."
I believe he should take some steps to mitigate this problem.
If he did, I believe he would find that his amps would sound even better to him, assuming his hearing will allow for that.
But this back and forth has never been, for me, about how his amps sound.
That is too subjective. For instance a lot of people think that 1970's Japanese receivers and Bose speakers are great.This back and forth is about Dennis' amps, that he claims are the best ever built, yet they have at least two real world flaws in their design.
Oh, and his use of junk science to explain it away.
Do you remember the question, "What is feeding the circuit when the diodes are not conducting because the input choke does not meet critical inductance that would keep the diodes conducting in a constant manner"?
The answers were always full of junk science and the question was never really answered. BTW The answer is....the last cap in the filter.
That is the opposite of the stated goal of his power supply design.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 06/24/12
I hear you. It boggles the mind how these threads just keep coming up. No one ever mentions how the DC coupled 2A3 amps in question sound, the arguments are always about the unproven claims of "performance", ridiculus invented "scientific" terms, unverifyable testing results and procedures, and junk science. I don't understand how Dennis and Jeff can't see that no one ever questioned how their amps sound. I'm really starting to believe this is some type of marketing scheme to gain product visiblity.
That said, have you ever thought about building one? If you look at SET amps in general, their distortion numbers are horrible. Yet people seem to love their sound. I know I do. Maybe this is another case where poor distortion sounds good to some people. At times I've thought about building one of them. Even thought the facts are that the HF will be rolled off from Miller effects, and that peuney sub-critical choke loaded power supply won't regulate worth a crap. Maybe all that signal filtering and distortion sounds good. High gm drivers allow for more current, but tend to pass more RF. Maybe that 12ax7 just doesn't allow higher freqencies to get through. I don't know, but sometimes I think about trying it out just to see.
I just long for the day when people stop all the nonsense claims about these products. I've given up on arguing with Jeff about it. In the end he'll just resort to namecalling.
Dan
A friend of mine built this with the help of Dennis, etc. The component values were all given to him.
I ask Dennis over the phone what the differences were between this circuit and his and he wouldn't really say. He did say that he uses multiple small by pass caps on the cathodes. My friend used only one but he used the right brand and the right size main cathode bypass caps.
Dennis complained that the current was a tiny bit to high in both the driver tube and the output tube and that the voltages were a tiny bit (and I mean a tiny bit) too.....I can't remember, I did take notes, I will have to see if I still have them.....well, high or low.
What he did say lead me to believe that it was very (my word) close.
When push came to shove Dennis said that 1" of the wrong wire will ruin the whole amplifier???!!!
I had it here for a few days. I circuit traced it and drew these schematics myself.
It's sounded awful.
On the test bench the PS was unstable even when not playing and the HF rolled off starting at 11kHz. Dennis has said (here on the AA) that his amp starts to roll off at 15kHz. He said the tubes he uses have less Miller capacitance than NOS. Dennis said (here on the AA) that his amps will start to roll off at 11kHz with Sylvania NOS 2a3's and 9kHz with RCA.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Got quite a few mistakes in there, Guy.No wonder you don't think it works-- I don't think THIS schematic does, either!
I hope nobody built this thing-- what you have here is a clock-radio amp.
There are wrong parts values and entire sections are left out... or the proper connections into circuits are in the wrong places.
Voltages are also off.
Hey! I don't mean to be mean here-- part of my technology is proprietary-- it isn't a freebie, so if you don't have it, it's not your fault.
The operating principles that I have shared and the concepts are what's really valuable anyway.
Good builders will get most of it right without me, although they won't get the entire result, it will be good.
I would refer people to Sound Practices articles on Loftin-White amps. If one studies those enough, he will invent ways to get the desired bandwidth for music.
---Dennis---
Edits: 06/27/12
What do you mean by that?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Here's my original post, and yes I was coached by both J and D as the project was made. And yes, there's what I'll call micro differences, some of which "may" have been improvements....
http://db.audioasylum.com/mhtml/m.html?forum=tubediy&n=153234&highlight=Satan+patiently&search_url=%2Fcgi%2Fsearch.mpl%3Fforum%3Dtubediy%26searchtext%3DSatan%2Bpatiently
I had not "yet" tried multiple by-pass caps, because I didn't have any at the time. I did however make an effort to minimize wire, see the pictures (much less than what Dennis does, I've seen the underside of his amp BTW).
I also had a much lower DCR power trans than Dennis uses, whatever that's worth, and I used dual 5R4GY's instead of one 5U4GB, which Jeff should like. I had custom made filament transformers with a tab for a CT that the R and Ck directly soldered to, from there to ground, (no wire).
Yes the voltage on the 2A3 was a little bit higher than what was "recommended", I didn't think that could "ruin" anything as the 2A3 was still operating UNDER max dissipation. I tore the amps apart 2 years ago or more. Before I did, I changed the driver to a CCS loaded lo mu triode,
instantly cleaner, better. I also changed the PS to "normal" which made other noise go away, but we now know that the original PS was "no good".
Jeff later felt happy to thrash my build, yet thankfully admits having misadvised.
See:
http://db.audioasylum.com/mhtml/m.html?forum=set&n=69636&highlight=grant+handley&r=&search_url=%2Fcgi%2Fsearch.mpl%3Fforum%3Dset%26searchtext%3Dgrant%2Bhandley
Hello GSH,You relied primarily upon me, and I was NOT correct in values and voltages in several areas, as my understanding of the circuit was not as good back then, as it would be today.
We both should disclose the fact that I told you to make the AC cord (ONLY) out of inexpensive Carol 300 VDC or 600 VDC 12 AWG wiring, and the audio out of TCSS and silver wire.
The TCSS never got installed. You wired the amp mostly with the inexpensive 12 AWG Carol, which has a "big wire" (very negative) thick and out-of-time sound to it.
Use of huge Carol 12 AWG hook up wire, in places other than the AC cord, guaranteed poor sound IMHO.
I also had the B+ supply filtering and shunt regulators mis-done, totally my fault back then, as far as obtaining low ripple. I made the same error with Slownlo, and even my own implementations in 2008-09.
You also never sent me a finished schematic of what you made, so that I could comment on voltages, and various obvious errors I would have spotted. IMHO, we BOTH made errors back then.
That a CCS improved things is not surprising at all to me, it was a band aid for numerous implementation and circuit errors. You gave up on the project, before fully informing me of exactly what you built and allowing me to suggest corrections. About two years later, in 2010, I told you quite precisely in what three areas that amp was "boogered", and it certainly was.
Its a shame you have never heard one of Dennis 2A3 amps, to know what performance level is possible. I have, that is what keeps me active.
Live and learn.
Cheers,
Jeff Medwin
Edits: 06/28/12
"The TCSS never got installed. You wired the amp mostly with the inexpensive 12 AWG Carol, which has a "big wire" (very negative) thick and out-of-time sound to it."
Yes, true BUT: the Carol 12g wire was only to the rectifier (yellow and red), and the ground (less than 6" total in one uncut piece from the HV CT
to the input, through the cap terminals. There's no other wire in the circuit, period, except for the lead to the OPT which was paralleled with silver wire. BTW we both know Dennis uses the stock red and yellow leads out of the stock Hammond power transformer to his 5U4GB. That sure isn't TCSS. So, the ONLY real wire "flaw" is the 6" ground, that actually only counts from C2 to the input, which is less than 3".
So the 3" ground wire of heavy copper, was "bad". OK
The difference in the op points is NOT large, as stated, still UNDER max dissipation for the 2A3, and the 7B4 was "close". I don't see how this could dramatically change anything. I don't think the small voltage differences could be "heard", in fact, they could easily blur between "perfect" and what you would call "way off" just with the line voltage changing. So I veto that.
The missing multiple by-pass caps is the largest "flaw" and I admitted that from the word go (see original post)
I also got tired of the hum and installed a Bourns pot, which fixed that.
I'm not really trying to defend or prove anything, just state the what is.
I do believe that the paralleled caps (if done whatever right is) could be an area of improvement for ANY CIRCUIT, which is why I brought it up in my other recent post. I give Dennis direct credit for having pursued this, although I don't have any direct experience with "what exactly" he has done, other than note that he said "it's expensive, and will cost more than the sum of the rest of the parts". With that, it's hard to get moving, expensively, into darkness, to maybe achieve light.
Furthermore, I think a direct coupled SE amp is a good sounding thing.
Perfecting it, is just an interesting set of possibilities.
I and many others may not prefer the 7B4-2A3 to what I or those others are preferring now, but overall improvements in PS, wire and caps are still valid for other circuits, so whatever real findings Dennis can show,
can likely "help" other designs in these areas. I wish this "tone" could be upheld in this discussion.
"There are wrong parts values and entire sections are left out. Voltages are also off."
So....If it's wrong tell us where and how.
This is what you and your friend told my friend to build. If it's wrong it's not the builders fault. He just built what he was told to build.
Entire sections? What do you mean by that?
You amps is a two stage DC 2a3 amp? Correct?
What sections are missing?
"No wonder you don't think it works-- I don't think THIS schematic does, either!"
So tell us what is wrong with it.
Put up or shut up!
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Correction! I don't tell anyone to build anything-- or completely how to build it.
Others do, and their contributions sometimes help, sometimes don't.
I, for one, admire them and appreciate their efforts. Remember, that these guys are working for nothing-- on other people's behalf. Now, that's pretty decent in my book, even if a few honest mistakes are made. (They will be-- SO WHAT? It's free- just use the GOOD parts of it, for God's Sake!).
That's life! Full of advice-- and only a few things actually work. Then, the real problem is-- as always-- how to APPLY some of that advice-- and maybe what to leave out also!
Again, that's life! Get used to it-- I happen to enjoy it and like to live on Planet Earth-- with all its faults!
As you should expect, I don't give out schematics or tell people exactly how they should build their own projects, since I'm part proprietary, and part understanding-- of the fact that every guy has the right to build anything any way he wants anyhow!
All we're doing-- at least that's me-- is allowing a peek into the THINKING PROCESS that can lead to excellence. You DO NOT get schematics, pre-designed kits, or any advice on how to build things-- except as general accurate, honest thinking processes that can-- if properly understood, lead to sheer excellence.
That's what you get for free from me. Nothing else. Be happy that it's available! Learn from it! Or-- don't. It's your call.
Developing an antagonistic attitude over it is childish.
---Dennis---
The overall concept of "modern" SE tube amplifiers is what is important, not the rote copying of a schematic.
I took a lot of information from Jeff Medwin aka Drlowmu and but I did not copy what he did exactly.
I adapted the overall "modern" approach to my circuit and came up with something useful and good sounding to my ears.
The detractors are not willing to buy the recommended parts, build a circuit or learn anything.
They want to publish poorly executed/incorrect Serious Stereo schematics and say that "modern" SE tube amps don't sound good.
The negative people just want to police the AA tube forums and try to bully other builders into silence.
dt 667
Don't you-- or others-- worry about it. They don't have anything that could work-- the mistakes are just too numerous to bother mentioning.
Their thing can't, and won't perform as they have it drawn up/perverted.
Now you know why I don't give out schematics-- I don't want them butchered.
---Dennis---
The only antagonism I have for you is because of the misinformation you spew to the detriment of the neophyte.Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 06/29/12
Sound Practices, Issue #5, Spring 1994, Page 3, "A Direct Coupled Single Triode Amplifier" by Ciro Marzio and Cristiano Jelasi".
Very nice article.
I'd suggest a modern power supply.
Jeff Medwin
Well, guys, what are you waiting for? Good Ole' Jeff has given you the goodies!Now, you can read the articles and learn WHY these Italians liked this approach better than anything else. Remember, they were listening to Italian Opera-- which is demanding musically.
Sure, I have perfected it. But, you cannot deny that their version of the concept did sound very realistic-- musically, even though it was operating in a narrow bandwidth-- measurement-wise..
Of course, that didn't really hurt it much.
Why is that? Even after rolloff of both highs and lows, the thing still responded well to micro-signal input. What happened is that although highs were rolled-off, THEY WERE STILL THERE!
That is FAR better than an amp that is measured flat-- say-- out to 35,000 HZ, but WILL NOT RESPOND to micro signals in frequency ranges even well below that, but must be POWER-DRIVEN to give even a passing nod to those energies.
Of course, the Miller-Effect eliminated amp will not drive a H.-EFF. speaker well at all, even though it measures well AT FULL POWER..
What happens here is that the rolled-off Italian design reproduces FAR more highs than the amp that has much more H.F. extension-- but not into a H.-EFF. speaker-- the extension to 35,000 HZ (as an example) is only on paper-- it is real enough if the amp is driven hard, but it DOES NOT occur in a H.-EFF. SPEAKER when the amp IS NOT driven hard..
In this case, amp measurements on paper don't prove power-response to SMALL signal input. The Italian's amp does respond-- at small signal levels, and that is the secret of it.
It can be rolled-off at both ends of the audio frequency extremes and yet deliver FAR more input of those rolled-off frequencies into an actual speaker-- than can the Miller Effect-Eliminated amp that measures perfectly, but will not respond to micro-signal input..
Now, I'm NOT interested in another mud-slinging contest over this. This is FREE, valuable information, TRE, MACH1, and GEO.
TAKE it or LEAVE it.
The Italians liked it and had fun with it. And, that wasn't even close to what is possible.
Oh, I KNOW-- several of you don't think that these are serious amplifiers. What did they use in the front-end-- a 12AX7?
Oh, but that can't work! It's that horrible Miller stuff! Yeah, sure. But it DOES work.
Have you ever asked WHY? I did.
---Dennis---
Edits: 06/28/12
"What happens here is that the rolled-off Italian design reproduces FAR more highs than the amp that has much more H.F. extension-- but not into a H.-EFF. speaker-- the extension to 35,000 HZ (as an example) is only on paper-- it is real enough if the amp is driven hard, but it DOES NOT occur in a H.-EFF. SPEAKER when the amp IS NOT driven hard.."
So, if it IS "driven hard" then are we back to equal (other than gain) ?
The main reason I left the hi mu driver, was I didn't need the gain, in fact a mu of 18 is still more than I need. This means that I'm using more than half of my available 2V of signal from the source, whether a CD player or a phono-pre, and, of course, ATTENUATING LESS. Is this "driven hard" ?
In addition, regardless of Rp or Miller issues, selected lo mu drivers can be measurably if not also audibly, lower distortion, because they are more linear, period. This is usually a good reason to use such an item, right?
I always thought hi mu triodes were for situations where hi-gain was needed, accepting the fact that there's a distortion "penalty" for such an "advantage". I think this is basically still true?
The real question is whether or not hi mu equates to higher low level sensitivity. "technically" is doesn't...
I'll admit, that once I made a direct coupled SE parafeed amp with a CCS loading a lo mu driver, I hadn't heard anything better, and was also amazed by the low level retrieval. The amp also had 2 less caps in the signal path...You may want to try it Dennis, use your front end if you prefer.
I neglected to mention something very important in this kind of amp-- it's critical.
When you're running a High-Mu, low current device as a driver-- that item must have unlimited plate power available, and it must have very tight voltage regulation.
If you attach a constant-current device-- either to the plate or the cathode of the driver stage, the performance should degrade because of the added phase problems initiated by the Band-Aid thus applied.
If, instead, the sound improves, and the amp has more power with the C/C applied, then that is proof of a power supply problem on the driver's plate supply.
The supply should have very tight voltage regulation-- you can't do this with any regulating device-- it will fail to closely enough follow music's changes. Instead, you use resistors for shunt-regulation. Ideally, you can run up to 40 ma., to ground, and use only 8 tenths of a milliamp of that for the driver plate current-- maybe slightly more-- one milliamp with a single section of 12AX7A-- which is recommended because so many excellent varieties are available today.
Sorry I neglected to mention this. If you've had problems, look here.
---Dennis---
Very intelligent and good questions-- all of them.Let's see if I can shed some light on them, not all of this may apply to your own situation, so don't consider it the only things you can do.
My approach to anything is how can I get it perfect... that may require changing more than the unit in question-- it may require designing a whole system to get everything as perfect as I can make it. So, again, this may not be gospel for your circumstances-- I'll do it for your discussion:
(1) I'm not sure that paragraph makes as much sense as it can-- let's take another look at it-- let's try and improve on my wording:
What I mean is that the Italian design- with its low-plate current driver stage responds better to small-signal inputs than an amp that has a higher plate-current driver installed for the purpose to overdrive Miller-Effect on the output tube grid.The low-plate current driver is easier to drive with input signals entering the amp than the higher plate current driver is.
This is an advantage where the amp is not pushed to full output-- as in the case of High Efficiency speaker..
Should one wish to push the output tube to full output, then the higher plate current driver will get much more input from the source components, and we will enter the area where it can perform better. It will also have a wider bandwidth because it is now overdriving the Miller Effect on the output tube grid.
We run into a severe fidelity problem when we DO NOT push this high plate current driver amp to full output. In that case, the hefty input tube plate current acts as an incoming signal blocking device, rendering the amp hugely less transparent and dynamic. As you probably already know, amps have to be power-sized to their speakers.
A much more honest appraisal would be-- the amp DRIVER STAGE must be sized to the SPEAKER.
-----This is far more important than matching the output stage to the speaker-----!
The thing that I notice most when using tube amps is that the power from the amp with the low-plate current driver, while delivering less total watts available from that amp-- puts out audio power that is infinitely more powerful and potent-- and has far greater ability to drive the speaker AS LONG AS that speaker doesn't need more power yet.
If the speaker does, you have to have a larger driver stage. In such a case, you give up a LOT of fidelity-- you would be much better off to change the speaker out and opt for more efficiency.
(2) If the amp is driven hard with the high-current driver stage, more input is being put onto its grid. This overcomes some of the limitations that the high-current driver exhibits under lower drive conditions. Is this performance as good or as potent as running the low-current driver stage into a speaker that is efficient enough to place that driver stage into its correct operating range?
NO! NOTHING will outperform the amp with the High-Mu, low-current driver stage-- IF the speaker loads the amp properly-- and never overloads the amp.
(3) "I didn't need the gain". The above discussion is for a 2-stage amp, so the high-gain input stage is needed. The overall gain of such an amp-- using a 2A3 output tube-- is in the range of 19-20 db. This is discussed in Jim Smith's book "get better sound". In there, you can read what Jim considers ideal gain for an amp. That is the range we have here..
Typically, with this kind of amp, we'll use a CD player, computer Sound Card, or a Phono stage that outputs-- ideally-- 3.2 to 3.6 volts Single Ended. In such a case, no preamp stages are needed, the 2-stage amplifier is directly driven by the source component. Volume control is accomplished by a Ladder Attenuator, or an L-Pad Attenuator.
Series and transformer or autoformer type attenuators are not used because series types cannot maintain an ideal load on the source regardless of volume setting.
Transformer or Autoformer types are OK for good midrange and voice performance, but fall short of maximum transparency-- compared to what we can get by eliminating them. Another thing that MUST be eliminated, in any really good system is all active preamp stages except for the Phono stage-- which is, unfortunately, necessary.
(4) Don't let linearity charts or other data fool you into thinking that they always apply to all situations under all musically-driven conditions. These can certainly mislead, although you will often catch me looking at them, and trying to find the more linear areas-- IF everything else is ALSO OK. IF NOT, then a workable balance must be thought-out across the entire amp's operating range and conditions.
Remember this: all devices, be they tubes, solid-state devices, caps, resistors, transformers and inductors-- have ranges where they operate well and ARE NOT being stressed-- thermally, signal-input, or power dissipation-wise. When parts are not ever stressed, they are far more dynamic and transparent, and MUCH more musical. With some tubes, this WILL NOT be the most "linear" portion of the tube curve. SO WHAT?
Basically, getting too hyped-up over theoretical "linearity" or "low distortion" can cause one to make overall engineering mistakes in amp design that will rob one of the musically linear, distortion-free presentation that you really want----- IN THE SPEAKER.
(5) High-Mu triodes are used where they can eliminate an extra gain stage. An extra gain stage-- regardless of intent-- is always a sonic disaster if it can be avoided. The more parts and gain stages you have, the slower your amp is, and the less transparent it is.
Anytime you can eliminate a coupling cap-- do it! NO cap exists that can be accurately put in series with an audio signal.
---Dennis---
Edits: 06/28/12
"What happens here is that the rolled-off Italian design reproduces FAR more highs than the amp that has much more H.F. extension-- but only on paper."BS
"TAKE it or LEAVE IT."I'll leave it and argue that it is a flawed design.
"Oh, but that can't work! It's that horrible Miller stuff! Yeah, sure. But it DOES work.
Have you ever asked WHY?"
It doesn't work so I have no reason to ask why it works.
You still haven't answered my question.
What do you mean when you say, referring to the schematic I posted, "the proper connections into circuits are in the wrong places"?
The plate of the 7b4 is connected to the last cap in the PS filter with a plate resistor. Just like in your amplifier.
The cathode of the 7b4 is connected to ground through a bypassed cathode resistor. Just like in your amp.
The grid of the 7b4 is connected to ground through a grid resistor. Just like in your amplifier.
The plate of the driver is directly connected to the grid of the output tube. Just like in your amp.
The plate of the output tube is connected directly to the output transformer. Just like in your amp.
The output transformer is feed from the second C of the PS filter. Just like in your amplifier.
The cathode of output tube is connected to ground through a bypassed cathode resistor using the CT of the filament transformer. Just like in your amplifier.
So tell us Dennis, what connections are in the wrong places?
You say some awful silly stuff some times.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 06/28/12
Saved me some trouble.
Dan
Dan,
No need to build it to hear it, because I am sure - you will not do it right.
Why don't you come to RMAF this October, and hear it at its best - over a few days? Much easier, much cheaper.
You won't ever say (or think) a thing about Miller or filtering, non-issues, when you hear it properly, .... like on the last day of the show.
Cheers,
Jeff Medwin
Well, I've done all I can!
The showdown is-- as usual-- at Denver's RMAF 2012 in October.
No one "wins" it-- nobody expects to-- it's a Hotel after all, and setup time is severely limited.
One CAN, however, spot great equipment in several of the demo rooms at these shows.
All of the exhibitors at these shows are doing it for you. Oh yes, if he gets LUCKY-- one may get to sell something once in a great while.
Denver in October! It's fun, and it's nice. One might accidentally find something he likes in there.....
In any case, there's no harm done, now is there?
---Dennis---
Post simple answers to simple questions on a whiteboard in hotel room at the show..... The system can be a 'static' setup as the sound of the amp is not what is leading to these ridiculous threads and you know that.......
Geo.,
People come to the show to meet the Manufacturers, but most of all, to see and hear the equipment, as best as they can. Posts up here mean zilch, audio performance is what really counts the most.
Jeff Medwin
How mnay times does some one have to say that this stupid thread is not a criticisim of Dennis' amps. You can't be that dense.
'You can't be that dense.'
I beg to differ.
I'd be willing to bet that to most of us, it is ALL ABOUT the sound of the amps we build and listen to.
It is all about the sound.When things are done right, the sound will be better.
It might not be recognized as being better.
One needs to know what real instruments sound like to start with before they can judge if a system is doing a good job at reproducing.
Leave the Miller not fully driven.....leave the cathodes not fully bypassed and the system will not be doing a good job at reproducing.
Will some people like or even love the way that sounds?
You betcha!
Let me be clear.
If God Almighty Himself said that a system and a recording was perfect, not everyone would like what they hear.
I have made changes in my system, over the years, that were clear, definite improvements in the technical behaviour of the circuit.
When listening for the first time (or even the second time) I didn't recognize it as an improvement.
Given time it became clear that that piano DID sound more like a real piano.
That Martin acoustic guitar did sound more like a Martin acoustic guitar.
Being a recording engineer, I know what these instrument sound like in the real acoustic world.
The thing is, I had become accustomed to the flawed sound of the previous circuit.
This is something that we all need to guard against constantly.
Get to know the sound of real acoustic instrument (without mics or amps or loudspeakers). They are the only real point of reference we have.
Whenever I hear someone say a system sounds "good" I take that with a large grain of salt.
There are plenty of people that think 1970's Japanese receivers and Bose speaker sound "good".
I just had one at my house not 2 hours ago. He was impressed with my system but he should have been blown away. The bottom line....he doesn't have a clue.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 06/26/12 06/26/12 06/27/12
you have no idea of the origins or this stupid thread. probably started two years ago. move on.
I've been here much longer than 2 years.
What do you believe the effect would be if you and a few others would follow your advice and "move on"????
I believe these type threads wouldn't exist and the forum would be much more hospitable to us "lemmings" looking for advise!!!
Agree.
Additionally, Geo (1) has never heard Dennis's amps, (like about ALL the criticizers I see up here), and (2) he seeks to conveniently dismiss the MOST obvious evidence - that which proves he stands on a weak foundation.
Yes, its ALL about the sound. Who cares if it uses a banana peel for the finals!! Results count.
Jeff Medwin
I have not heard the amps therefore I can't critiize the sound of it. You are right. Dig through these post and point out where I criticized his amp....nimcompoop.
It matters not what you said or didn't say, you stand accused by the grand imperial nincompoop and you now have to PROVE your innocence.
Who was is who said, "I may agree with what you say but I will fight to tbe death against your right to say it" ?
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: