|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
69.108.66.175
Edit, I can't post but I can edit my posts.Ralph, when did I say you were lying?
If I did I apologize.
I don't think you are lying, I think that you just don't understand what Class A is all about.
If duty cycle is all there is to Class A then all Class A/B amplifiers are Class A because there is always some leakage current flowing.
Especially when using a tube like the 6as7g that is very non-linear near cutoff.
If all it takes is a little current flowing at all times to make it Class A then Classes of Operation lose their meaning and what's the point?
Why do you accept the first sentence of a definition but ignore the rest of the pages and graphs of that same definition?
I will never understand that.
I am sorry this causes such grief and I am sorry you can't just discuss this matter in a technical way without getting personal with me.
I am guessing this will never be resolved and that's sad.
BTW I have no problems with Class A2 operation except that it puts a lot of stress on the driver. Class A2 circuits can be made to have symmetrical waveforms just like Class A1.
Tre'
I had loaned out my 325 ohm Freed output transformers but I'm getting them back.I just tested, with my Triplett 3444 tube analyzer that tests for idle current, 26 6as7g's and came up with 8 that are very close.
With a 325 ohm load I use just 4 6as7g's per amp for about 23 watts real Class A.
By real Class A I mean that the two halves of the wave form for each triode are symmetrical.
I have run these amps with these output transformers before and they sound way better than any transformerless circlotron amp I have ever heard.
At about 15 watts output the wave form looks like this.
This is what 8 6as7g's in a circlotron circuit playing into 8 ohms looks like at about the same power level as above.
This is not what I call Class A.P.S. My next test will be to run 2 6336's per amp into the same load. The 6336 is a MUCH more linear tube.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/17/11 08/17/11 08/17/11 08/17/11 09/02/11 09/02/11Follow Ups:
Is there a specific technical reason why a non-OTL class A circlotron (with low Rp tubes like 6AS7) should sound any better than a more conventional class A push pull amplifier that uses the very same output tubes and driver circuit?
I completely agree with you that ALL commercial OTL tube amplifiers operate in class A/B - no question about this.
And if I were not a DIY I would certainly push hard the magic triplet of arguments 'class A-OTL-circlotron" (maybe with some flavour of no feedback) for selling my supposed brand new products.
Unfortunately, a practical 'class A-OTL-circlotron' (tube) amplifier would require special high current/low Rp output tubes that have never been invented (and never put into production).
So I suppose that, knowing the above, if I were a manufacturer willing to push the circlotron argument, a manufacturer who knows that another manufacturer has already used two 'OTL-circlotron' (if not 'class A-OTL-circlotron') magic arguments, my only choice would be to push the 'class A-circlotron' remaining arguments.
But in doing so, then there should be some other technical argument to support the hypothesis that 'class A-circlotron' is better than conventional class A push pull.
Otherwise any manufacturer of class A conventional push pull amplifiers could easily argue that 'class A-circlotron' (as well as OTL-circlotron) is nothing special at all.
Based on my past DIY experience, IMHO class A (PP) and class A/B OTL perform almost the same.
However, I personally prefer a class A/B OTL in case the higher output power of class A (PP) is not really necessary.
Best Regards
Luca
ecc230
"Is there a specific technical reason why a non-OTL class A circlotron (with low Rp tubes like 6AS7) should sound any better than a more conventional class A push pull amplifier that uses the very same output tubes and driver circuit?"
I haven't thought about it.
I built the OTL thinking it would be Class A [I guess I had been reading to many advertisements] and then realized that it's not Class A but it could be with a higher load impedance so instead of just taking it apart, I added an output transformer.
This makes it Class A and greatly lowers the harmonic distortion and improves the sound.
Tip of the Day, don't piss off the sponsors.
Tre'
Per Title.
P.S.: Ralph - you are not the only one Tre' has an opinion about.
"Isn't this the "OTL" Forum (as opposed to the "non-OTL" forum)?"
Yes it is, and yet there are posts about the ZERO-Autoformers being used and no one complains about those. Or maybe they do? :-)
BTW We all have opinions.....don't we?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Hi Bayleiler !
Sorry but what kind of the opinion Tre`have? , based on some limited and no real life virtual results from performed PC simulation ?, and based on the erroneous conclusions made on the basis of others' perceptions and misconceptions ? , he obviously can not clearly distinguish A1 mode of power tube operation from A2 mode of operation ,(Atmasphere Circlotron OTL Amps includes both class of output power tubes operation simultaneously ,and A2 Class of operation is by definition still A class of operation) .
In fact he does not understand the basic electronic behavior of a simple tube PP OPT Amplifier !, claiming that some AC power goes around through the set of two PP connected tubes ,which is actually impossible since the basic vacuum tubes physics that not allow , to tubes conduct opposite .
Actually I have no more illusion that sometime he come here on this OTL site with clear intention only to provoke some attention and to minimize the quality of the Mr.Karsten OTL Amps !
If his real reason for science? or is money?, I would say that though in the latter .
Best Regards
__
Enlightened Evolution-Astral Projection
This is one example of Class A2 where both halves of the output current waveform would be symmetrical.
I am at a loss to understand why Class A2 is being used as an excuse for non-symmetrical waveforms.
As for money, the only time [these days] that I make any money related to audio is when I fix a piece of equipment for someone.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
By real Class A I mean that the two halves of the wave form for each triode are symmetrical.
IOW, by 'real class A' you mean class A1.
BTW the actual waveforms we get off of an individual triode's cathode differs from the simulation in that it lacks the flat spot at the bottom of the curve unless the amp is driven into clipping.
I have run these amps with these output transformers before and they sound way better than any transformerless circlotron amp I have ever heard.
I've been castigated on the forums a number of times over the use of the word 'better' so these days I usually define it, however I'm pretty sure that you've not listened to any of the newer circuits out there in a while. You might want to avail yourself of that when you get a chance...
There is a very nice and rather long thread on diyaudio.com where some interesting ideas are presented. It is by no means the only thread where such occurs. For example there is another one about a hybrid Circlotron that uses an input coupling transformer for the output section, using MOSFETs for the output devices. We've built something similar in a tube embodiment, with a single-ended voltage amplifier and driver circuit. The simplicity offered works out very nicely for the guitar amp in which the circuit is used.
"By real Class A I mean that the two halves of the wave form for each triode are symmetrical.
IOW, by 'real class A' you mean class A1."
No, in a real Class A2 circuit the two halves of the wave form are symmetrical.
Non-symmetrical halves of the output waveform are indicative of Class A/B operation.
Even if the current never stops flowing completely in the 6as7g because it just that kind of tube, if there is no symmetry to the waveform then it's not Class A, not in my book.
Tremaine Audio-Cyclopedia 2nd Edition page 523 "The quiescent plate current is selected for a value in the most linear portion of the plate-current characteristic."
Page 543 "What is a class A amplifier? An amplifier in which the grid-bias voltage is set to approximately one-half the cutoff voltage to obtain linear operation."
Page 604 "When operating as a single tube, class a, little distortion exists because the grid signal operates in the most linear portion of the dynamic characteristic."
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
"IOW, by 'real class A' you mean class A1."No, that's not what I mean at all.
In a Class A circuit using linear tubes both halves of the waveform look the same.
"BTW the actual waveforms we get off of an individual triode's cathode differs from the simulation in that it lacks the flat spot at the bottom of the curve unless the amp is driven into clipping."
If each triode is biased at 70ma idle, it can't continue to go down below 70ma. down from idle. (0ma.)
The graph shows exactly that.
For an M-60 to make more than about 5 watts the bottom half of the waveform of each triode will have the "flat spot at the bottom of the curve".
If the tube is idling at 70ma. and it is then driven (by an input sine wave) up to, lets' say 170ma. the top half of the wave form will have a shape and size that represented a 100ma. change to the positive.
The tube in the other bank, at the same time, is being driven by an input signal that is anti phase and that signal will try to drive the tube to a negative 30ma. But there is no such thing. The current in that tube will go down as far as it can and then stay there until the input signal drives it back up.
The "then stay there until" part is the "flat spot at the bottom of the curve".
That half of the wave form cannot have a shape and size that represented a 100ma. change to the negative. A tube cannot flow 30ma. of negative current.
Understand I'm not talking about the output waveform to the speakers.Edited to say the whole thing again in one sentence.
In any type of tube amplifier, push pull, single ended, Circlotron or whatever...If the positive half of the input signal pushes the tube to more than twice it's idle current the negative half of the input signal will cause a "flat spot at the bottom of the curve".
There is no way around this simple truth.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/23/11 08/23/11
There is no way around this simple truth.
Yes- unless you simulation does not reflect the actual circuit. I have to assume that you have a parameter wrong somewhere. In the old days the simulators themselves had obvious problems - like the model for the triode - but these days I tend to assume that they are reasonably correct.
We covered the A1/A2 thing ad nauseum in the past. We even came to some resolution on the matter- now you seem to have reversed yourself. Tre, I have to point out that when you come to this forum and make your post with the 'better sound' comment included, its obvious to everyone that you are looking for a fight. Since that is the roots of your intention, you will not command much credibility without admitting it.
"There is no way around this simple truth.Yes- unless you simulation does not reflect the actual circuit."
Are you trying to say that a tube can conduct in the negative?
This has nothing to do with simulations and it has nothing to do with me or my intentions.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/24/11
Yes- unless you simulation does not reflect the actual circuit."
Are you trying to say that a tube can conduct in the negative?
No, and no idea where you got that idea either.
You said that there was no "flat spot at the bottom of the curve" and maybe there isn't.But the lower half of the waveform would have to be truncated at anything above 5 watts for a M-60.
Majorly truncated at 60 watts.
Whether it's flat or rounded makes no real difference.
The main point is that it's not the same shape or size as the top half of the wave form, it can't be without the tube being able to conduct in the negative.
In a real Class A (1 or 2) circuit both halves of the wave form are the same (or very close to it).
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/24/11
Tre': So what term would you use to describe that particular assymetrical waveform above 5W output (because we know that an M60 makes more than 5W)? Is that waveform a hallmark of Class AB or B operation? I do not think so, but enlighten me. Thanks.
If the tubes don't completely cutoff some say that it's Class A.That's what the books say.
But if we read the rest of the chapter we find that there is more to Class A operation than that.
If we judge the class of operation by what is really happening I would have to say the amplifier is "acting" like a Class A/B amplifier.
Please understand that a Class A/B amplifier does not operate "like" a Class A amplifier up to a certain power level and then operate "like" a Class B amplifier.
Class A amplifiers are biased and loaded and driven in a certain way.
Class B amplifiers are biased and loaded and driven in a certain way different than a Class A amplifier.
Class A/B amplifiers are biased and loaded and driven in a certain way
that is different than either a Class A or a Class B amplifier.The three classes of amplifiers are distinct and not to be confused with one another.
What the "A" portion of a Class A/B amplifier has in common with a Class A amplifier is that the current can and does go down by the same amount (depending on the linearity of the tube used) that is goes up, giving a symmetrical wave form.
As the power level goes up the lower half of the wave form gets truncated.
The higher the power level the more (as a percentage of the total) truncated the lower half of the wave form gets.
This is exactly what happens in an M-60 above about 5 watts.
BTW most tubes are not linear near cutoff. The bias can go more and more negative but the current does not go down in a linear way.
This is called the cutoff region.
Once this behaviour starts the tube is said to be "in cutoff".
There are a lot of tube that "refuse" to cutoff completely without an extreme amount of negative bias.
Using one of these tubes does not make a Class A/B amplifier a Class A amplifier.
A properly designed and operated Class A (1 or 2) amplifier should not reach the cutoff region let alone total cutoff.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/24/11 08/24/11 08/25/11
Wrote the above probably at the same time you were posting your last two posts.
By the way, you guys have had this same discussion here within the past few years, and you used the very same figure to support your argument, which is fine.
" By the way, you guys have had this same discussion here within the past few years, and you used the very same figure to support your argument, which is fine."Yea, I did use that same figure.
That figure fits perfectly the full description given in all the books of what Class A operation is.
While it is true that a tube being operated in Class A does not reach cutoff that is not the end of the definition of Class A operation.
That figure is.
BTW Lynn Olsen refers to it as "deep" Class A. I and the books just call it Class A.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/24/11
This is from the MosFet Circlotron pdf from the link you gave.Micheal Rothacher understands Class A.
"How much bias current will we need? This amplifier will be conservatively rated for 20 Watts @ 8 Ohms in Class A, so our peak output current will be 2.24A (1.414 * SQRT(20/8). So, we’ll need to set the bias for each output MOSFET around one-half of this value, and we’ll round it up to 1.2A DC."
This is what can't be done in a Circlotron using 8 6as7g's driving 8 ohms above about 5 watts.
One cannot set the idle current at half the needed peak output current to make 60 watts into 8 ohms.
At the idle current normally used in a 8 tube Circlotron OTL the current can only be driven down (as much as up) to an output level of about 5 watts.
If driven to an output wattage above that the "flat spot at the bottom of the curve" starts to show up.The current in Micheal's output devices can be driven down as much as they are driven up, to the full rated 20 watts, with out creating that "flat spot at the bottom of the curve".
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/24/11
(sigh...) MOSFETs lack a grid current window...
You are confusing the Class A1 vs. Class A2 issue with the linear operation issue again.
Linear operation is indicative of Class A (1 or 2) operation.
Non-linear operation (truncated lower half wave form) is indicative of Class A/B or Class B operation.
Just because you can and do run the M-60 into A2 does not mean that the wave forms can get non-linear (non-symmetrical) while still claiming Class A operation.
Tremaine Audio-Cyclopedia 2nd Edition page 523 "The quiescent plate current is selected for a value in the most linear portion of the plate-current characteristic."
Page 543 "What is a class A amplifier? An amplifier in which the grid-bias voltage is set to approximately one-half the cutoff voltage to obtain linear operation."
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
You are confusing the Class A1 vs. Class A2 issue with the linear operation issue again.
I'm calling you out on the bullshit (again)- I have no such confusion nor ever did.
The diagram you show is for Class A1. I've seen it many times.
Tre, its obvious to me that I offended you sometime a long time ago as you have had an ax to grind for the last ten years or more. I don't know what I did you offend you, and you apparently are not man enough to talk about it. But one thing is sure- it has nothing to do with audio- it has to do with some human interaction of which I have no idea.
I've apologized for whatever it was in the past, as I have had no intention of offending anybody, but you persist like a rash that won't go away. Again, the persistence has nothing to do with electronics. Its something personal. At this point I don't see the point of arguing with you further as you have also persisted in logical fallacies like the Strawman in an attempt to make me wrong. At this point you have also seemed to have a short memory as well, obviously on purpose, going back through the archives will sort that comment out.
At this point you now owe me an apology. Plain and simple. Otherwise, get the hell out of here and don't come back.
Now you should know as everyone else does that I am simply not known for comments like this in over 20 years of the Internet. Its not that I am mad, its that you are being an ass.
"Its not that I am mad, its that you are being an ass."
I have re-read my posts to you in this thread.
I do not see where I have been an ass.
I have only talked about electronics except one instance where I claimed that you were confusing two different points.
I don't think that rises to the level of "being an ass".
You are, of course, welcome to your opinion.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Maybe someone could do an actual experiment, hooking a scope up to an M-60 running at high power but below clipping, and get an actual plot of the current flowing through the tubes? This would be quite instructive, and it could eliminate the debate about whether the simulation is accurate or not.
Chris
Because of the nature of the circlotron circuit, no matter where you hook up the scope you will be seeing the combined waveform from each bank of the push pull. The whole circle is part of a "feedback loop" if you will.Even though each bank's waveform is asymmetrical the combined waveform looks fine. This is normal for Class A/B amplifiers and for Class B amplifiers as well.
To look at the waveform of one bank will not be easy and to look at the waveform at a given power level will be harder.
If you pull one bank of tubes you will be able to see the asymmetrical waveform but you will have changed everything in terms of the feedback and loading.
I have no problem at all understanding that a tube idling at 60ma. can only reduce it's current by 60ma.
If the tube is driven to a current more than twice it's idle current then the lower waveform will be truncated.
If the current is increased to 200ma, that 1/2 current waveform will be of a size that represents 140ma. and will be much larger than the following 1/2 current waveform that can only be a 60ma. current waveform, at best.
200ma. per triode, 8 triodes per bank, one bank is cutoff or close to it.
(however much current is flowing in the other bank will subtract from the output power)200ma times 8 is 1.6 amps peak.
1.6 amps peak across an 8 ohms speaker is 10.24 watts RMS.
With do respect to all parties, I don't really need to do an actual experiment.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/26/11
I am assuming one would insert a very low value resistor in series with the cathode of one of the tubes, and put the scope across that. This would give a direct view of the current flowing through that tube.
I agree that the outcome seems inevitable, since there simply isn't room for anything else to occur. I just feel that one clear trace showing the waveform of the current flowing in a single tube would put a stop to the endless debating about whether it is or isn't class A.
What seems a shame in all this is that it detracts from the much more important point that it is a beautiful amplifier that apparently sounds superb.
Chris
The amp already has the resistors mentioned. So its been easy enough over the years to measure this,
A2 trades greater power for linearity. If the tubes were to cut off without the amp clipping, the result would be a class AB2 amplifier. This is explained in the FAQ section of the website and has been there for over a year.
"I am assuming one would insert a very low value resistor in series with the cathode of one of the tubes, and put the scope across that. This would give a direct view of the current flowing through that tube."I'm not sure it would. I think AC current flows in a circle that includes all the tubes. I could be wrong about that.
"I agree that the outcome seems inevitable, since there simply isn't room for anything else to occur. I just feel that one clear trace showing the waveform of the current flowing in a single tube would put a stop to the endless debating about whether it is or isn't class A."
I couldn't agree more.
"What seems a shame in all this is that it detracts from the much more important point that it is a beautiful amplifier that apparently sounds superb. "
Yes it is a shame and the amplifier does sound great. It's just not Class A in the true sense.
Transcendent Sound, makers of OTL amplifiers, makes it clear that their amplifiers (and most other OTLs) are Class A/B.
The paper from Transcendent, while using some language that I don't totally agree with, makes the same points.
A Class A amplifier needs to idle at 1/2 the total needed peak current [that's 1/2 per bank. If you add the idle current of the two banks the amplifier needs to idle at the total needed peak current].
It takes 3.86 amps peaks to make 60 watts into 8 ohms so there needs to be 1.93amp idle current per bank or 241ma. per triode.
That just can't be done with a 6as7g.
Thanks for your input.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/27/11 09/20/11
Hi Tre,
I don't know why you are trying to stir the pot about what Ralph has said about his amplifiers. I went back in the archives about this subject and it seems that Ralph has always made it very clear about what he is doing. You're implications by bringing up this subject every year or so is that Ralph is misleading us about something and by doing so, you are trying to impune his credibility and reputation. I have know Ralph for 20 years and he has never mis-lead me in any way. On the contrary----if anything, he always takes the conservative approach.
What I find amusing--is that when Ralph calls you on it. You simply respond that it's all about the science. I think that if there is any deception going on---it's on your part by your extremely weak ( it's all about the science ) response.
I must admit that I'm an avid Atma-Sphere user because the equipment just sounds exceptional in all areas. The icing on the cake is that the support that Ralph and his staff give is even BETTER !!
I have the money to just about buy anything I want and over the years I have gone through allot of equipment to come to a point where by comparison----the Atma equipment just sounds better and is even more reliable.
My only assumption is that you are one of those guys that just want some attention and like to get it by posting on this forum. But by doing so it's at the expense of one hell of a nice gentleman that has devoted his life to his public and his company.
I also think you owe Ralph an apology but I really don't think you are man enough to do it. We will just get some weak response about it ain't personal and it's about the science.
Looking at the Atma-Sphere website pages for the S-30, M-60, MA-1, MA-2 and MA-3 the text in all cases says "Class A operation".But that's not true.
Tremaine Audio-Cyclopedia 2nd Edition
Page 543 "What is a class A amplifier? An amplifier in which the grid-bias voltage is set to approximately one-half the cutoff voltage to obtain linear operation."
Page 604 "When operating as a single tube, class a, little distortion exists because the grid signal operates in the most linear portion of the dynamic characteristic."
None of the output tubes in Ralph's amplifiers operate this way.
What do you think I should do? What would you do?
Way back in the days of the ASOG forum this question came up and was never answered, at least not to my satisfaction.
Ask Kurt Strain about this issue. Paraphrasing Kurt "Tubes almost never cutoff completely. If a tiny leakage current is all it takes to call an amp Class A then all Class A/B amplifiers are Class A."
I like Ralph and I wish him the best.
My main focus here is to educate people.
I think it is Ralph who owes an apology to everyone who has read his website that says his amps are Class A, when they are not.
The link is to a very short paper I wrote years ago.
P.S. Challenging a claim or theory, in a scientific way, is a normal part of the scientific process.
Defending a claim or theory, in a scientific way, is a normal part of the scientific process.
But Ralph doesn't do that. He changes the subject. He tries to make it about me. ETC.
What would you have me do? I think I stayed on subject and just argued the science without getting personal. What else can I do?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/26/11 08/26/11 08/26/11 08/26/11
Okay, here's everything I know and stated better than I can. John Broskie is right. He wrote that when he also went in and explored the claims. I'll back up John on his essay.But I have an additional comment to John's article. When you use SPICE modeling, it is a small signal condition applied for rp, ip, and vgk calculations. It does large signal harmonic distortion calculations based on 6AS7 triode plate curve modeling, but not for these other parameters as curves for them are not provided in the model.
Therefore, the output impedance calculations are for small signals only and that will operate in the zone of pure class A operation, about a couple watts maximum output. That means that actual output impedance for more signal level than that is significantly higher.
-Kurt
Edits: 08/30/11
but unfortunately that URL only contains the first page of Broskie's treatise on the circlotron. I read the complete piece once upon a time, and I must say I could not/did not follow his argument about circlotron = totem pole (the Futterman topology). I would like to re-read it and try again.
Paul Speltz once made a case that is the exact opposite of what you state. His conclusion was that as power output from the circlotron increases, its output Z goes down. I printed out his post, possibly from the "Tweaks" secion of the old ASOG website, and I saved a hard copy, but I don't have an e-version to post here. I do recall that his analysis was based on his results in driving a signal backwards into the output of his own M60 amplifiers and observing how the signal could "push the output stage around" at different amplitudes. Maybe I can scan it and post it, or maybe Paul is out there somewhere.
But anyway, Tre' has not been making any claims as regards output Z. He is talking about Class A operation, or not.
It is possible to resistively inject a small signal to extract the small signal output impedance (with some error) as that will not disrupt everything.It is not possible to obtain impedance information by messing up the output stage's normal operation by jamming into it a very large signal. The fact that it takes it is a good thing. But that's ruining the amp's own efforts and it is as if, for an analogy, in particle physics he has the worst heaviest probe to measure position while that completely ruins any knowledge of it's velocity.
But this whole thing about all the old claims would never have been an issue if the claims were 100% in line with classic definition, accurate study, and world class measurement techniques including having guardband.
Ralph would never yield to those who already knew what they had in an M60 amp, and so that is a problem with the people who know test and measurement. It makes us want to scream that someone gets to claim something others will not due to better adherence to old definitions and careful factual argument.
There was no personal issue until Ralph saw people who don't like his sloppy methods as some kind of troublemaker. No, they do want accurate information spoken. John Broskie wrote that sarcastic article after losing it when encountered with no yielding to the facts of electrical engineering.
But we CAN as consumers brush it all aside and not give a hoot! The amp is a fantastic performer. It's real measure is how it sounds, and how reliable it is. I can recommend M60 amps like the one I had.
Edits: 08/30/11
Thanks, Kurt.
I really am not qualified to argue Paul's thesis, pro or con. Your post suggests to me that you read his original contribution on this subject in the Tweaks section of the old ASOG website. Is that the case? If not, please don't assume that I accurately described his rationale or his method. That would be unfair to Paul. Anyway, I am a scientist, albeit a biological one, every work day of my life, so in audio, which is "just" a hobby, I feel it is sometimes OK to adopt the Alfred E Newman philosophy: "What? Me worry?"
But just to satisfy my other self, can you describe the rationale that leads you to say the output impedance of a circlotron goes UP as the power output goes up?
To be fair to Paul, this is several years old, and I have no idea whether or not it reflects his current thinking. ("Current" being the operative word.)
Paul did a fine job in his article you showed us. Basically his 11 ohm result is what I also got. And he also explains that this number is for one condition and that it moves from there.The output impedance change is directly related to the lack of class A operation going on. At 0.6 amp quiescent bias, the amp operates class A to 1.44 watts into 8 ohms (0.6 amp peak current into 8 ohms). During this time, there are two tube "banks" (phases) driving the output impedance, even if Paul is not fully convinced.
When at max power, 60 watts into 8 ohms, the peak current has to be 2.74 amps for one phase and 0 for the other. When operating in this outside zone, only one tube bank contributes and the output impedance now comes from the cathode follower impedance of one phase bank of parallel 6AS7s. I don't have that formula right now, but it's easy to look up.
The 6AS7 is not fully linear, and so when one bank cuts off then the Zout is moving with the changing plate curve point for it. And this might help it or hurt it, depending. But for sure it has left the help of the other phase.
So the output impedance slides continuously over many changing values instantaneously during a full sinewave signal cycle.
There is instantaneous and average full cycle output impedance. Most people measure average on meters, few will do the complete job and measure the full plot of instantaneous output impedance over phase and over varying loads. This is the problem you get into when you try to get one number with highly assuming assumptions about the measurement that is not fully understood. Small signal offers a decent complete number but large signal is nonlinear in many ways, especially in this case, and there is no real good measurement technique for that, although Paul tried hard.
I hope this helps explain things.
Edits: 08/31/11 08/31/11
Isn't that approximated by the term 1/Gm? If so or if not, how to combine the two terms (the foregoing plus the output Z of the bank of output tubes that is not in cut-off) to calculate a net output Z for the amplifier?
1/gm is another small signal parameter. gm changes when you move the operating point's position from one place to another. Thus the output impedance must change.
The problem is that most want a single number in ohms for the average, not a graph of the output impedance shifting with phase of an inputted sinewave. The same goes for output power. People want the average power outputted over the full cycle, not a plot of instantaneous power. We know it's a number with large signal distortion involved.
The issue here is: can we calculate and then measure the average integrated impedance over the full cycle at every power level, knowing the larger signals are more distorted? Math can calculate almost anything once you get the idea of a good method.
It's really quite a complex effort, and it would need a DC element input for the large signal Rout=Vout/Iout when stopped there, and then add small signal wiggling sinewaves for the small signal zout at that position. Then move around 360 degrees in "DC" steps measuring all small signal output impedances at those positions. Then integrate for a mean value and check the standard deviation while you're at it. This is then the mean Zout with a now known distorted deviation.
how many manufacturers state in print that their amplifiers are "Class A", when their products do not in fact operate in Class A up to full power output? The answer, I think, is nearly ALL of them. One company makes a big deal out of the fact that their hideously overpriced linestage operates in "Class A" and uses "no feedback". I hope you are an equal opportunity gadfly.
I have never read the Atma-sphere website in detail. I did not even know that there is a claim to full Class A operation. (Does it really say that? Or does it say "Class A", without specifics?) I really don't give a flying big one what you call the mode of operation, and I have long been aware of the phenomenon that limits Class A power, which you pointed out here for the second or third time in the last several years. Still, Atma-sphere amplifiers are the best sounding tube OTLs I have heard during my 40 years in this hobby.
The world we live in is full of much more consequential false claims that need to be debunked.
"Still, Atma-sphere amplifiers are the best sounding tube OTLs I have heard during my 40 years in this hobby."
I have never disputed that and I have a track record of stating that the Atma-Sphere OTL is a great sounding amplifier.
"The world we live in is full of much more consequential false claims that need to be debunked."
Very true.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
My main focus here is to educate people.
Here is the deal,this is Ralph OTL site!If your are not going to say good thing about his amps dont post.
I think it is Ralph who owes an apology to everyone who has read his website that says his amps are Class A, when they are not.
Goodluck,the Amps are Class A to Ralph an others that pay Class A Money!
I think it funny that Most who buy Ralphs amp an preamp say there the best,For $20k what would you say there fair?
Maybe
But Ralph doesn't do that. He changes the subject. He tries to make it about me. ETC.
What would you have me do? I think I stayed on subject and just argued the science without getting personal. What else can I do?
He,s The OTL king, Well he thinks so.
Tre' Thanks for the info
And now we can clearly see that your agenda has nothing to do with science.
Duke
Me being a dealer makes you leery?? It gets worse... I'm a manufacturer too.
.
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/28/11 09/02/11
Here, let me show you why I made the assumption:Tre', 8-26-11: "My main focus here is to educate people."
GFL, 8-27-11: "My main focus here is to educate people."
* * *
Tre', 8-26-11: "I think it is Ralph who owes an apology to everyone who has read his website that says his amps are Class A, when they are not."
GFL, 8-27-11: "I think it is Ralph who owes an apology to everyone who has read his website that says his amps are Class A, when they are not."
* * *
Tre', 8-26-11: "But Ralph doesn't do that. He changes the subject. He tries to make it about me. ETC."
GFL, 8-27-11: "But Ralph doesn't do that. He changes the subject. He tries to make it about me. ETC."
* * *
Tre', 8-26-11: "What would you have me do? I think I stayed on subject and just argued the science without getting personal. What else can I do?
GLF, 8-27-11: "What would you have me do? I think I stayed on subject and just argued the science without getting personal. What else can I do?
* * *
Now that I look back on it, maybe GFL really is someone else, and he was quoting you without giving any indication that's what he was doing.
You were pretty pissed when I accused you of duplicity, when it may well have been a misunderstanding on my part. Hmmmm. Isn't that what you're doing to Ralph? Except that you've made a years-long crusade out of it.
Duke
Me being a dealer makes you leery?? It gets worse... I'm a manufacturer too.
Edits: 08/28/11
Notice that in every case the duplicated sentence was authored first by Tre' and then by GFL. (That was pretty good detective work, incidentally.) I believe that GFL is just stealing Tre's phrases, because he has nothing else to say. Whereas Tre' is in earnest. I don't agree with Tre's approach, but I believe he is sincere.
You have to scroll down a little.
Since this is explained on the site and has been so for some time (and also as it has been in the past), I take umbrage to Tre's comments that I am lying. At best he seems to not allow that class A2 is class A while I maintain the A1 and A2 are both class A.
The problem that I have is if I can't call it class A2, then what is it?? In an AB amplifier, whether AB1 or AB2, the tubes go into cutoff at some point. Class C, D, G and H do not apply. Class A2 is about all there is.
Thanks for the link, Ralph.
It makes no sense to me that Tre' would crusade so vehemently over whether or not "Class A2" can legitimately be called "Class A" (and duh, it's obviously a type of "Class A"). That is certainly NOT worth calling someone a liar over!
So I do not think Tre's vendetta is about amplifiers or even about precise definitions of terms; rather, I think it's about a personal agenda. My guess is that this crusade is just Tre's ego trip, rather than him actually having some personal grievance against you and your company.
Duke
Me being a dealer makes you leery?? It gets worse... I'm a manufacturer too.
The personal agenda thing had not occurred to me and it makes more sense...
I get your point.Has there been a (technical) misunderstanding on my part?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/28/11
As i sead, Dont F++k with King Ralph,This is an has been A DEAD Site!
An Most here Like Lew an others have nothing to add,just the same thing over an over.I love My OTLs That it.
There was at one time Vary good info here on OTLs but looks like the King did not like it so he just put a end to it. Just like NOW.With help from others that have nothing to add to what your are saying.
Just let the site go back to The same crap, talking about how great the Kings Amps an preamps are,Theres no more to say but you can bet there well be more crap!That All thay Got.
The Amps Are not an never well be CLASS A.But thin what is True Class A
.I no what it is not,but your telling us what is,Thanks
PS. If Ralph was a true OTL KING he would have his on Site!Must be $$$
Would you like to step outside, GFL?
Lew are you a Bully?
"The diagram you show is for Class A1."
What in that diagram says it's A1 only?
Zero bias tubes like 805's require positive grid voltages and grid current but can still have a dynamic curve much like the one shown and still have both halves of the wave form be symmetrical.
I know you think this is something personal but it's not.
It's just about science and physics.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
4 tubes with the transformer gives me 14 watts. Which is more than I need.
2 tube give me 3.5 watts. That's all I really need. I think I'll try that.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Can We see your schematic, i would like to setup my Circlotron with output tranfourmers, i have a pr of M60 but i live in Fl an it to hot here to run all the tubes, thanks
The output transformer primary just hooks up where the speaker goes and the speaker hooks up to the secondary of the transformer.The same idea as Paul's Zero Autoformers only taken to an extreme.
http://www.zeroimpedance.com/Zero-New.html
Paul's autoformer is great, it just does not go far enough to load the tubes for real Class A operation. The 3 X hook up will make your 8 ohm speaker look like a 32 ohm load for the amplifier.
I hate to use the term extreme. It's not really. Loading 8 6as7g's in a Circlotron (Atma-Sphere M-60) circuit with 172 ohms is not extreme at all.
It's just proper loading for real Class A operation. It yields 42 Class A watts.If you remove 4 tubes per amp and load with 325 ohm you get 14 watts all Class A with very low distortion.
You will get less power, much less harmonic distortion and much better damping factor with a transformer.
I think a transformer winder like Heyboer could make any impedance you want.
Talk to Douglas (PakProtector) over in the Tube DIY forum about dealing with Heyboer.
The Freed transformer I'm using came from a Krohn-Hite amplifier with push pull parallel 6550's.
It has four identical primaries that I have wired in parallel and two identical secondaries that are wired in series. That's how I get to a 325 ohm primary with my 13 ohm speakers.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/19/11 08/19/11 08/19/11
Interesting thing that Experience Electronics Germany 140W - A/AB Class , 6xEL34 Circlotron tube based Power Amp use the same 330 ohm Rkk primary OPT ( AP 634/2 ) impedance as yours rewired OPT transformer for yours 4x6as7g based OTL Circlotron Amp !
BTW , Circlotron cathode coupled OPT usually (and strictly ) need special ``split bobbin``coil configuration for EI typ of OPT to obtain the best Audio (electrical transfer) performance .
If you want to avoid your OTL Circlotron Amps to run in deep region of A2 class of operation but without to use impedance matching OPT , it is my personal opinion that it is better way to recone the single speaker with 330 ohm bobbin coil , or to use some line array loudspeakers configuration , with serial connected speakers of standard 8-16 Ohm impedance finally resulting in yours favorite non standard final high Z load impedance for yours OTL Cirlotron power Amps , running than only in A1 class of operation .
Best Regards
__
Enlightened Evolution-Astral Projection
Edits: 08/20/11
"Circlotron cathode coupled OPT usually (and strictly ) need special ``split bobbin``coil configuration for EI typ of OPT to obtain the best Audio (electrical transfer) performance ."I don't understand why a split bobbin would be needed.
I would think the best OPT for this would be a high nickel, parafeed (no gap) single ended transformer.
The only reason, I can see, to run into A2 would be to get more power.
I have 96db speakers and don't play that loud so power is not an issue.
Thanks for the link.
Stephens Trusonic had an OTL amp way back. They built a 500 ohm speaker for it.
The problem is a 500 ohm VC gets heavy.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/20/11
Reason for use the OPT ``split bobbin `` coil configuration for Circlotron output power stage is to get the perfect primary load symmetry & balance for both Circlotron PP output power stage phase , since the Circlotron topology use the whole OPT primary coil , not only the half of the OPT primary coil which is normal for classic based OPT tube PP Amp .
That`s mean equal OPT parasite capacitance & stray inductance for both half of the output Circlotron PP power stage & more important is to get minimum and Equal parasite capacitance between primary and secondary Circlotron OPT coils .
Usually there is no any DC flowing thru the Circlotron OPT primary coil ,so there is no appearance of switching transients ! , (accidentaly same as for SE parafeed configuration ), and central primary coil earth tap it is necessary only to get the ground reference point for driver stage unit & power stage bias , since the Atmasphere OTL Amps use two equal 600R resistors from each ``floathing `` speaker terminals conected to Amp star ground point for that , that is the reason why yours OPT CT don`t have to be grouded at all.
The best iron core for Circlotron output power stage it is not critical as for standard SE or PP based Amps ,and it is physically much smaller than ordinary PP OPT for the same projected output power , but in the same time it is with much wider both high & low Audio frequency ends response .
For low step up/down ratio request quality toroids finish the job 100 % , but for higher step up/down Circlotron OPT ratio the double C core iron with middle inserted split bobbin coil is far the best solution !
- Hi impedance VC is not heavier than low impedance VC ,since the high impedance coil is made of very tiny isolated Cu. wire ,this type of speaker it is ``excited`` with average relative high voltage signals from OTL Amp but with relative low average signals peak current for the same distributed power .
Friend of mine have reconed 12``/ 15W speaker with 2500 ohm VC !!!,``excited direct via condenser from 6l6 SET anode (parafeed), Sonic difference is huge compared with normal OPT based 6l6 SET ,and thats excellent system sound it is the normal appearance for pure tube OTL Audio !!! .
Best Regards
__
Enlightened Evolution-Astral Projection
Edits: 08/20/11
Back in the early 70s, I had a double pair of KLH9 ESLs wired in series and driven by a large home-made Futterman amplifier, using eight 6LF6s per monoblock. One wonders why I ever moved on from that, the point being that the series arrangement made for a nice load on the Futtermans.
Because Your Nuts! like most here at AsyLum.But you Are a Diyner so i say your ok.An you no what thats worth.....
I guess I'm not nuts in this sense, since I have hardly changed my setup since the eighties---NYAL OTL3 + Stax F83's. Someone suggested I should run the panels in series instead of parallel, which would be possible but not trivial, so I didn't do it. The only change I have made is to get some subwoofers, designed by Roger Modjeski, for which I am very grateful. On the other hand, I am nuts about Patricia Barber, having flown to Chico three times to see her at the Green Mill. I do admit to some terror about the future of my amps---I have some extra tubes, but who will bias them correctly when the time comes? BTW my NYAL amps are, I suppose, class B and have lots of feedback, but I like the way they sound and they don't run that hot. These are good facts, I think.
Thanks for all the great information.One thing...
"since the Circlotron topology use the whole OPT primary coil , not only the half of the OPT primary coil which is normal for classic based OPT tube PP Amp ."
In a normal Class A PP amplifier the whole OPT primary is used by each tube.
Only in Class B is the output primary use one half by one tube and other half by the other tube.
"The best iron core for Circlotron output power stage it is not critical as for standard SE or PP based Amps"
This is not true at all.
If there was concerns like DC being present, etc. then compromises are made.
When there is no DC present then there is no reason for compromises.
The best transformer has no air gap in the core and has a core made of the highest permeability allowable for the power level intended.
The signal source (when it's pure AC as in this case) is only part of the equation as to it's impedance.
The transformer is not going to know if it's being feed from a PP circlotron or single ended parafeed or differential coming from a single ended to balanced bridging transformer or differential output opamp or whatever.....
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/20/11 08/20/11
``In a normal Class A PP amplifier the whole OPT primary is used by each tube , Only in Class B is the output primary use one half by one tube and other half by the other tube.``
- Nope , this is not true at all ! Regardless to class of operation with the single source PSU connected to CT of classic PP Amp OPT each tube always use only one half of the transformer primary coil !
About OPT again :
Each type of tube Audio OPT it is one big compromise in itself !, even Circlotron OPT it is .
Some Amps topology just allow less OPT performance compromises , some Amps topology require more OPT performance compromises .
So the best Amps OPT it is actually No OPT at all , but just OTL !
Best Regards
__
Enlightened Evolution-Astral Projection
"- Nope , this is not true at all ! Regardless to class of operation with the single source PSU connected to CT of classic PP Amp OPT each tube always use only one half of the transformer primary coil ! "That is only true in the DC.
Let's take a pair of 300b output tubes with a common cathode bias resistor.
The AC signal path will only include the power supply (CT of OPT) and the common cathode resistor for the imbalance.
If the tubes were perfectly linear and perfectly matched there would be no AC traveling out of the CT of the output transformer or across the common cathode resistor at all.
Take a look at the differential (forced balanced) and parallel feed differential examples in the link below.
I have ran my Class A1 PP 300b amp with and without a cathode cap and hear no difference at all. Not at the levels I play them.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/20/11 08/20/11 08/21/11
Hi Tre` !Don`t want to abuse this respected OTL Asylum forum with OPT staff but ,Yours attached schematic from Mr.Lynn Olson tutorial on 2004 European Triode Festival presentation it is incorrect and can lead the reader to false conclusion that each tube use the whole OPT primary coil , or Raa load !
Here is my explanation :
Each Basic PP Amp is formed from two equal basic SE Amp joined together ,and need at least double separate SE OPT Ra primary coil load for normal operation ( 2x SE Ra = PP Raa ) .
Best Regards
__
Enlightened Evolution-Astral Projection
Edits: 08/22/11
You show loop one with positive polarity sinewave input giving no inversion on the plate output. That is an error.
You show loop two with negative polarity sinewave input giving a proper inversion on the plate output. This is correct.
So you diagrammed a cancellation instead of a double boost in output current.
One tube operating and one cut-off except for DC will have one tube energizing the whole primary through mutual coupling to the other polarity as in class AB operation. With both helping, the load is easier as in class A operation. Both work. They are however, in different operating points.
Mr. Kurt.s without my sincere intention to belittle or insult you !!! , but a You sure in that what you claim ?
Found that the present level of your ignorance and misinterpretation of basic phenomena inside of a simple tube PP Amp electronic circuit represented in my posted schematic it is incredible !
Best Regards
PS: Now I can clearly see the reasons why Mr. Karsten sometimes loses self-control and react emotionally when he has to deal with some ignorant people.
__
Enlightened Evolution-Astral Projection
This forum is your place to go to get all your ideas fucked up in your brain. Enjoy it.
Perfect!Let's look at your schematic.
If the tubes are linear then when the current increases in one tube, the current will decrease, by the same amount, in the other tube.
So the AC current flowing through the Power supply/common cathode node will be equal and opposite.
Equal but opposite current is the same as no current at all and the AC signal path will bypass those nodes and include the whole primary winding and the both tubes only.
If there is no net AC current through the cathode resistor/power supply node then we are left with one AC current loop just as Lynn shows.
The current loops you are showing are for the DC only.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/22/11
If we galvanic separate both B+ supply for both separate SE Amps joined together in one PP Amp via two separate equal OPT primary coil we should expect that yours `` AC power path flow `` well be interrupted , and this Amp well be out of function ,unable to operate , but no !, this separate B + Amp well work with the same performance as Amp from my first attached schematic regardless to class of operation , still having two separate DC current loops .
Only real AC power appears and flow on the OPT secondary coil via load(speaker),on the primary side of OPT AC power appears but only virtually ,being generated from pure PP DC current flow through the both primary loops.Best Regards
__
Enlightened Evolution-Astral Projection
Edits: 08/22/11
"Amp well work with the same performance as Amp from my first attached schematic regardless to class of operation , still having two separate DC current loops "That is not true for (at least) 2 very obvious reasons.
Now each tube is operating into 1/4 Raa instead of 1/2 Raa.
The cathode bypass caps are now 100% in the signal path and so are the power supplies.
This is not how a Class A PP amplifier works. In a Class A PP amplifier the power supply and the cathode resistor are only in the AC signal path for the imbalance. A small % of the total.
I can prove all of this with an oscilloscope.
Calculate the amount of DC current change that should flow from one tube being modulated at a given amount. Measure the actual amount flowing through an un-bypassed cathode resistor in a Class A PP amplifier.
The difference is the amount that the other tube is changing in the opposite direction.
The amount of DC current change that is measured is a measure of the imbalance in the circuit.
The "missing" current change is the balanced current that is flowing through the tubes and the whole primary of the OPT to the exclusion of the PS and cathode resistor.
You should read and understand what Lynn Olson has to say. He knows what he's talking about.Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/22/11 08/22/11
My next schematic describe one of the PP Amp type which Really use only half Raa of conventional PP Amp OPT primary coil , it`s anode coupled Circlotron , again two separate SE Amp joined together , sharing common Ra load and still having two separate DC loops , but now with opposite DC peak currents loops direction .``` You should read and understand what Lynn Olson has to say. He knows what he's talking about.```
Found that respected Mr. Lynn Olson interpretation , plot and fable of certain type of electronic circuit in some particular cases have no real life basis , and sometime it is just his personal free view with very speculative nature .
Think that this is my last post here related to non OTL staff , since this respected forum it is primary dedicated to OTL staff .
Best Regards
__
Enlightened Evolution-Astral Projection
Edits: 08/22/11
Instead to loosing my energy arguing , I collect the parts for my new Atmasphere based OTL Circlotron pure tube DIY Amps ,this time with projected modest 2x30W /32 ohm / A class , to play some Crystal Clear Music in my home ! 24/7 !
Mr. Karsten many thanks again for the excellent sounding power Amp design ever and for helping DIY !!!
Best Regards to all
__
Enlightened Evolution-Astral Projection
Install two resistors and remove the cathode bypass capacitor as shown.Drive the output stage with 1kHz.
Measure the AC voltage across each added resistor. Calculate the current.
Now measure the AC voltage across the cathode resistor. Calculate the current.
What happened to the missing current?
That missing current is not flowing through the common cathode resistor and it is also not flowing through the power supply.
It's flowing in a circle that does not include the common cathode resistor or the PS.
It's flowing through the whole primary of the OPT and the tubes and the added resistors.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/22/11
"Each Basic PP Amp is formed from two equal basic SE Amp joined together ,and need at least double separate SE OPT Ra primary coil load for normal operation ( 2x SE Ra = PP Raa ) ."It is true that a Class A PP circuit requires a primary 2 times the impedance that is needed for each tube.
In a Class A PP circuit both tubes are using the whole winding at the same time and the effect is that each output tube is seeing 1/2 Raa.
But in a Class B amplifier 4 times is needed. Each half of the primary winding is only 1/4 of the impedance of the whole.
This is a problem with Class A/B amplifiers. The load shifts as soon as one tube is no longer changing in an equal but opposite way.
As long as both tube are contributing in an equal way the load seen by each output tube is 1/2 Raa. When one tube cuts off (or reaches the non-linear "cutoff region") the load seen by the other tube is shifted to 1/4 Raa.
The same is true with a Class A/B OTL amplifier.
When one bank of tubes reach the non-linear cutoff region and stop contributing equally, the output impedance of the amplifier goes up and that would have the same negative effect as a load change in a transformer coupled amplifier.
In a Class B amplifier each tube always only uses one half of the primary winding and the two output tubes take turns. Each tube sees it's 1/4 of the total primary impedance and that doesn't shift.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 08/22/11
" Posted by Damir _the real one on December 2, 2004 at 09:40:51
In Reply to: I'm confused... did everybody miss this class posted by Gordon Rankin on December 2, 2004 at 07:03:25:
Well, I mathematicaly described 30:1 turns ratio SE OPT with CT secondary, see below. In short, you can connect one 4 Ohms speaker on the half of the sec., one 16 Ohms speaker across the whole sec., or two 8 Ohms speakers across both halves. In all cases, our tube on the primary will "see" the same reflected impedance and "gave" the same power. Well, let`s reverse the thing:
1.) We have, say 300 turns CT primary (150+150 turns), and 10 turns secondary. Turns ratio from whole primary to secondary is 30:1, or from halves of the primary 15:1.
2.) Let`s connect the speaker of 8 Ohms across the secondary. Then our reflected impedance to the whole primary is (30^2)*8 = 7200 Ohms.
Reflected impedance across the one or another half of the primary is (15^2)*8=1800 Ohms, or 1/4 of the whole primary.
3.) Let`s connect the tube, or better - for simplicity AC generator that gives 100Vrms no matter of load. We can connect the generator through the whole primary, Raa=7200 Ohms. Our 100 Vrms will give the power Ppr`=U^2/Raa=100^2/7200=1,3889W. That power is "transferred" to the secondary load. Usec is Upr/turns ratio, or Usec`=Upr/30=3,333Vrms, and power is Psec=Usec`^2/Rsp=3,33^2/8=1,3889W.
We can connect our generator across only the half primary, our generator "see" Raa/4=1800 Ohms, and gives power Ppr``=100^2/1800=5,555W. On the speaker - Usp``=100/15=6,667Vrms, Psp``=6,667^2/8=5,555W.
4.) Now, we feel like goin` PP, and connect two identical generators with 100Vrms output, but in antiphase. Connection - one wire from the each generator on the one side of the primary, then remaining two wires connect together and grounding (the same like with tubes). We can now say that we have two generators in series, in antiphase, connected on the primary, and each one gives 100Vrms.
When we measured AC voltage across our primary, we measured 200Vrms!
Now, Ppr```=200^2/7200=5,555W, and our Usec```=200/30=6,667Vrms, and Psp```=6,667^2/8=5,555W.
5.) The question is - what load each tube/generator in PP A class (both tubes working) "see"?
The Law of power conservation - Ppr=Psec. Every tube/generator gives half of the total voltage and half of the total power on the primary (and transferred to the secondary load).
Usec```^2/Rsp = Ugen1^2/Rload1 + Ugen2^2/Rload2
We know that Ugen1=Ugen2=100Vrms, we know Usec```^2/Rsp=5,555W, and we know that Rload1 = Rload2 = Rload1,2 = unknown resistances of the one and another halves of the primary in PP case.
Then we have: 5,555 = 2*(100^2)/Rload1,2 and
Rload1,2 = 20000/5,5555 = 3600 Ohms (half of the total Raa=7200 Ohms).
We can express this with turn ratios, the result is the same :
1/Rsec = 1/(Rload1*(Nsec/Nhalf pr1.)^2)+1/(Rload2*(Nsec/Nhalf pr2)^2)
1/8=1/(Rload1,2*(10/150)^2) + 1/(Rload1,2*(10/150)^2)
and from that Rload1,2 = 3600 Ohms.
6.) Summary and conclusion:
Reflected secondary load across the whole primary is the square of the turns ratio times Rsec, or in this case Raa=7200 Ohms. This load (7200 Ohms) "see" the tube connected across the whole primary.
When we connect the tube across the only half of primary, our tube "see" Raa/4, or in this case, 1800 Ohms.
When we connect two tubes in PP A class, each tube "see" half of the primary resistance, Raa/2, or in this case 3600 Ohms. "
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Banat
Don`t want to abuse this respected OTL Asylum forum with OPT staff
Dont even think about it, go on with info.Thanks for your time
I see sead the Blind man!No i get it an i have tryed the ZerO not for me but Lew i think uses them?
I think that "BAT" is setup diff,more like a two tube Kt88 pushpull output
i would like to see there Schematic,but that for your time
"An Keep Working the Problem"I can Dig that
The BAT VK-60 is a very similar circuit using a differential cascade voltage amplifier and a Circlotron output, driving an autoformer. The original prototype that Victor K. showed in several locations actually used an Atma-Sphere Z-Music autoformer that belonged to his partner, Steve Bednarski.
Thanks for your time RaLph own the BAT
Personally I've never had a chance to see the Victor K. - BAT VK60 - Circlotron circuit design .
But regardless to that I want to say next : With inserting of autoformer in Circlotron output power stage circuit that Amp immediately stop to be the pure DC Amp , + take in account some negative Sonic sides of autoformer iron .
Contrary to that situation yours Atmasphere Circlotron Hi Fi OTL tube Amps use the pure DC Amp output power stage design configuration ! , but Without using of any DC servo loops (usually traced via GNFB loop(s) ) to preserve OTL Amps output power stage DC stability !
That`s extremely rare !, even on the world of SS Hi Fi Amps designs .Best Regards
__
Enlightened Evolution-Astral Projection
Edits: 08/23/11
As you know, Tre', this is not "news". The assymetrical wave form out of the OTL circlotron has been described before. To me, the surprise is that you like the sound WITH the transformers better than without them. In the commercial market, this is called a BAT amplifier. How does your version differ from BAT's, except in that you may not use a toroid, and they use 6C33Cs, not 6AS7s? Thanks.
"As you know, Tre', this is not "news"."
I know, I was just "sharing".
"The assymetrical wave form out of the OTL circlotron has been described before. To me, the surprise is that you like the sound WITH the transformers better than without them."
That asymmetrical wave form is the cause of harmonic distortion but because this is a push pull amplifier the wave forms from each bank will be combined and most of that harmonic distortion will be "covered up".
But it won't all be covered up. If we start with clean wave forms from each bank of output tubes then there is no high harmonic distortion to cover up in the first place.
The output transformer gives the output tubes a higher impedance load line (more horizontal), that's why there is so much less harmonic distortion.
The output transformer I use is one of the best ever made and whatever harm it's doing is way less than the harm from the harmonic distortion caused by a vertical load line.
"In the commercial market, this is called a BAT amplifier. How does your version differ from BAT's, except in that you may not use a toroid, and they use 6C33Cs, not 6AS7s? Thanks."
I would have to see the BAT schematic to answer your question.
The 6C33C is not very linear at all and I would not use it.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I enjoy learning from you. Thanks.
I have never heard a BAT amplifier, either, but by all accounts they make less "magic" than do Atma-sphere amplifiers, which is why I was surprised to learn that you prefer the sound of your amps when they are transformer-coupled, harmonic distortion notwithstanding.
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: