 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
76.90.100.230
In Reply to: RE: Are they ultralinear or triode? posted by mwhouston on October 09, 2024 at 22:30:39
That is not very linear. Are you using feedback?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Follow Ups:
The original ultralinear patent suggested an ideal location of the screen tap depending on the tube and other factors. To get around the patent, a number of manufacturers would misplace the tap slightly, resulting in less than ideal operation, which was then 'cleaned up' by feedback. This became a tradition and so most screen taps aren't optimized. But the patent is long expired, so if someone knew that and also knew what they are doing, it could be corrected.
My point here is we don't know if the tap is placed correctly and if not, how it errs from linear. Your chart shows one tap position only; quite likely its not the one in use.
please watch this
everyone
For example he states that PP amps will sound harsher on account of even ordered cancellation. This statement is false. He glossed over a bit in that no mention of the kind of non-linearity is mentioned (Quadratic vs Cubic in this case). In addition, he seems to not understand that the 3rd harmonic is treated much the same way as the 2nd by the ear. He compounds that error with false comments in various places in the video. He is correct though that you do want to minimize the 3rd an PP circuit.
In case there's any question about the 3rd being a problem to the ear, any properly functioning analog tape recorder produces a 3rd harmonic as its dominant harmonic distortion.
Ultralinear can be used in class A as well; he seems to gloss over this fact. The reason to use class A, triodes, ultralinear and other techniques is to minimize distortion because feedback isn't going to eliminate it- the more distortion the open loop circuit has, the more the closed loop circuit will have too. So there are class A ultralinear amps out there.
I'm not debunking this video- but I am saying its not 100% correct. For example you can build an ultralinear amplifier using only a pair of power tubes and a single 12AT7 or 12AU7. He is saying something else at 18:30 in the video...
At the same time there's a lot of really good information here.
"Ultralinear can be used in class A as well..."He never talked about UltraLinear (as in feeding the screen grid from a tap on the primary of the output transformer).
When he used the term 'ultralinear' he meant "Limited Class A" where the input signal is limited to keep the operation of the tube in the center of the most linear part of the transfer curve, staying clear of the non-linear cutoff and saturation regions. This is what he is talking about at 18:30 in the video.
Read my question and his answer from yesterday.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 10/14/24
He needs a different term! 'Ultralinear', ultra linear and ultra-linear are pretty well accepted to be the same thing.
BTW he didn't address the fact that for it to really be ultralinear, the taps have to be properly placed. The graph you showed earlier was of a 6L6 whose transformer designer was dodging the patent or working from a design that did (out of 'tradition').
"BTW he didn't address the fact that for it to really be ultralinear, the taps have to be properly placed."Somehow you are missing what I said and what he said.
He never talks about a circuit that has the screen grids of the output tubes connected to taps on the output transformer. What you and I (and most everyone else) calls "Ultra-Linear". No where in the video does he address that in any way.
When he uses the term ultra-linear he means "limited Class A" where the amplitude of the input signal to the amplifier is limited to prevent the output tubes from operating in the non-linear cutoff region and the non-linear saturation region of the dynamic operating curve. The tubes are only permitted to operate in the center (the most linear part, straightest part) of the curve. What I simply call "properly executed Class A".
You and I talked about this type of tube operation by phone years and years ago. Your remark to me was "My amplifiers are not that kind of Class A".
"The graph you showed earlier was of a 6L6 whose transformer designer was dodging the patent or working from a design that did (out of 'tradition')."
I said, and it is true, that it doesn't matter where you place the taps, a pentode or a beam power tube ran in ultralinear is less linear than that same tube ran in triode mode. Also none of them, no matter how you operate them, are as linear as a directed heated triode.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 10/15/24
" Also none of them, no matter how you operate them, are as linear as a directed heated triode."I am sure you have seen the anti-pentode, plate-to-grid traces on DIY Audio. There are a few iterations/variations, but for fun there was a before and after of a MOSFET. It went from horizontal drain lines to exceedingly evenly spaced verticals this way.
Add to that, not all DHT are created equal either. There is a fair variation there... :)
I would throw down a gauntlet, for a 10k a-a load, presented as a 5k SE, I get 500V delta plate on a positive grid swing, and a 450V delta upwards on a negative swing. A shallower load line would no doubt assist in improving that.
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Edits: 10/17/24
"I am sure you have seen the anti-pentode, plate-to-grid traces on DIY Audio."
I have not. I will take a look. Thanks.
When I said "no matter how you operate them" I was referring to the three ways that Ralph and I were talking about. Pentode, UL and triode.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
The plate-to-grid FB circuits are most certainly rigged as pentodes.I can quite agree with you on the idea that particular FB method makes something 'new' and out of your limits.
Just to get you started... :)Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Edits: 10/18/24
It's not out of my limits. It's just outside the context of the conversation I was having with Ralph.
Thanks for the link. It looks interesting.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
But the curves you presented earlier are indicative of the tap being poorly placed.
You might want to read that patent through.
Can you please show me the curves for a 6l6 in UL when the taps are properly chosen?Thank you.
Edit, I have read and read and the proper UL tap % is 43. So this is a set of plate curves of the GE 6l6gc with 43% screen taps.
This is what I posted earlier. The taps ARE at 43%.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 10/16/24
-
The article in Audio Engineering Nov. 1951 by Halfer and Keroes says 18.5% impedance (which is 43% voltage/turns).
https://www.worldradiohistory.com/Archive-All-Audio/Archive-Audio/50s/Audio-1951-Nov.pdf
The patent says 18.5%.
"18.5% being clearly an optimum value, for the particular tube for which the plot of Figure 6 was made."
https://patents.google.com/patent/US2710312A/en
As far as I know the Acrosound TO-300 was used in the original 6l6 ultralinear amplifier and it has the taps at 43%
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
The patent says 18.5%.
"18.5% being clearly an optimum value, for the particular tube for which the plot of Figure 6 was made."
https://patents.google.com/patent/US2710312A/enNote the actual text. It says that 18.5% is optimal for the particular tube.
From the patent you referenced in your last post:
[sic]
The winding i4 is provided with fewer turns than is the winding it), so that its impedance at any frequency is preferably about 18.5% of the impedance of winding is}. Although t's value is optimum not critical, and a satisfactory range of values may be taken to extend from l5% to 25%, for tubes of the 6126, or the 807 type. It is conceivable that a different percentage of loading might be required for tubes of different design parameters than those tested by us to date, but in all cases the value is so selected that extremely linear composite response occurs, and so that tube impedance is low, but output high. This means that the tetrode i. operates like a tetrode in respect to output level, like a triode in respect to internal impedance, and with a plate transfer characteristic intermediate that of a triode and tetrode, and more linear than either.I have highlighted in bold the relevant text.
Edits: 10/17/24
From the patent,
"2. The combination in accordance with claim 1 wherein said screen grid loading is substantially 18.5%, and wherein said vacuum tubes are equivalents of the 6L6 type. "
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
These words are intentional in patent language. When you get a patent, you have to leave wiggle room or get painted into a corner- you want it as broad as possible (I have a few patents so I've been through this).
These words are there for that reason as they allow for the tubes to not be 6L6s and the ratio not be exactly 18.5%.
The fact that the body of the patent expands on this (as in the case of the text of my prior post) is in keeping with how patents are written.
The bottom line is the 18.5% is an example and not cast in concrete. IOW the patent is saying you optimize the ratio for a particular tube such that it obtains the desired characteristic.
.
Have Fun and Enjoy the Music
"Still Working the Problem"
I don't know. A variable tap on an OPT would be needed so you could vary the screen tap to obtain the correct set of curves. Its likely not too far off from 18.5%.You have to look at the assignments of the patent to figure out what happened. Its a bit intriguing - it only shows Arco. This means that Hafler, one of the inventors, had to get around the patent when he went to work at Dynaco. I suspect that since he was there and Dynaco never got assignment of the patent, that everyone else did the same thing. That became tradition and most people (like Mr. Whiteside) forgot that fact.
So as a result most of the 'UL' OPTs out there are simply built incorrectly, all because none of them ever got assignment of the patent from Arco.
What this further means is that if you did the proper research using the patent which is now expired, that you could build a 6L6 UL circuit that is a lot more linear than most people think it could be. The question is was that done for the amp of this thread?
Edits: 10/17/24 10/17/24
.
Have Fun and Enjoy the Music
"Still Working the Problem"
Unless the word 'about' (as highlighted) means 'exact' in this article- I doubt it does.
Please keep in mind that inventors don't really like to share their work by revealing all the secret sauce.
So unless you or I do the work this will never be resolved.
I have the impression that you think that if the taps are in the right place a non-linear tube (like the 6l6) will magically become linear. But you can/won't show me the curves because "everyone does it wrong".
My position is still, no matter where you put the taps you can't turn a sow's ear (non-linear tube like the 6l6) into a silk purse (like a 2a3 that is very linear to start with).
The other way to resolve this is to find out where the taps are on the to-300 transformer. If they are at 43% then I am right.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
here is f langford smith of rdh4 fame doing the work along with a compilation.
period correct articles.
So it doesn't seem like there is a "magic" spot where a pentode in UL turns into the "most linear amplifying device ever made". That is how Ralph made it sound when he was telling me that the 6l6 UL plate curves I posted "are indicative of the tap being poorly placed".
I figured as much when Ralph went from being sure that 43% was wrong to saying "I don't know" when asked what the proper % is.
I get weary of people talking in circles.
Thank you for your clear, straight up input. It is very refreshing and very much appreciated.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 10/18/24
So it doesn't seem like there is a "magic" spot where a pentode in UL turns into the "most linear amplifying device ever made". That is how Ralph made it sound when he was telling me that the 6l6 UL plate curves I posted "are indicative of the tap being poorly placed".
I figured as much when Ralph went from being sure that 43% was wrong to saying "I don't know" when asked what the proper % is.
I was merely telling you what the patent says. I've had no interest in building a UL 6L6 based amp so don't know the actual tap. No talking in circles here; the problem has been that in your posts you've conflated words like 'about' and 'exact' to be the same thing.
Its essentially trolling.
I didn't 'make it sound' like anything. I was simply informing you of what the patent actually says. I've not done anything beyond that; please desist with the personal attacks.
"please desist with the personal attacks. "
Nothing personal Ralph.
After mwhouston answered your question saying the amps were operating in UL I asked him if he was using feedback after showing him how non-linear a 6l6 in UL is.
You then posted this, "The original ultralinear patent suggested an ideal location of the screen tap depending on the tube and other factors. To get around the patent, a number of manufacturers would misplace the tap slightly, resulting in less than ideal operation, which was then 'cleaned up' by feedback. This became a tradition and so most screen taps aren't optimized. But the patent is long expired, so if someone knew that and also knew what they are doing, it could be corrected.
My point here is we don't know if the tap is placed correctly and if not, how it errs from linear. Your chart shows one tap position only; quite likely its not the one in use."
Did you intend to imply that a 6l6 in UL could have plate curves with the grid lines evenly spaced along a proper load line and be usable without feedback if only the screen taps were in the right place?
That is what I thought you were saying.
When you said "....misplace the tap slightly, resulting in less than ideal operation, which was then 'cleaned up' by feedback." did you mean to say that placing the tap in the right stop would allow "ideal operation" and there would be nothing to "clean up" with feedback?
That is what I thought you said.
All I am saying is that, no matter where the UL taps are placed, the plate curves will not become linear enough. A 6l6, in UL, single ended, will still need to be "cleaned up" with feedback.
Thank you.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Did you intend to imply that a 6l6 in UL could have plate curves with the grid lines evenly spaced along a proper load line and be usable without feedback if only the screen taps were in the right place?
That is what I thought you were saying.
When you said "....misplace the tap slightly, resulting in less than ideal operation, which was then 'cleaned up' by feedback." did you mean to say that placing the tap in the right stop would allow "ideal operation" and there would be nothing to "clean up" with feedback?
That is what I thought you said.
All I am saying is that, no matter where the UL taps are placed, the plate curves will not become linear enough. A 6l6, in UL, single ended, will still need to be "cleaned up" with feedback.
That is indeed what I was saying up to your last paragraph above. If you look at the article by Langford-Smith about UL operation, you see the distortion quite a bit lower than even a triode (see pages 57 and 58). How that works exactly when single ended I can't say with precision, but having built such amps (although not using a 6L6) they seemed to sound good without feedback so I suspect the principle works with single-ended in a manner as described by the patent.
So what I am saying is you might be pleasantly surprised to find out how good a 6L6 can actually sound when operated UL single-ended, in particular if the tap is optimized. You might try it and see!
I have tried it, but it was necessary to mentally
discard a lot of linearity theory-- not to eliminate that,
but instead to explore the whole "could be" SE Pentode
amplifier as a new set of sonic opportunities in itself-.
I adopted an attitude where one imagines what will
sound the most accurate, instead of relying too much
on what calculates as the most "accurate".
I have put that word "accurate" in quotation marks because the
human ear is decidedly not linear, but even more interesting is
the fact that no two humans hear even remotely alike anyway.
Building a great audio amplifier is an art, but
included in that must also be total reliability and
long component life, including the tubes.
I am currrently amazed at what the "premium" Hammond SEA
series of SE output transformers can do-- did their engineer
just get the taps (for ultralinear) right-- maybe it's
a special model that they got right_deliberately, or
is this one item just a piece of good luck? Or do we
have some excellent metals and winding techniques here,
or is it just what they say-- a "premium" piece? Who knows?
I do know that this one model works musically, like the
best transformers out there.
One could argue over that for decades, or he could
instead learn how to use the results to his sonic advantage.
That one can build a GREAT SE amplifier using a 6L6GC,
run in transformer tapped G2 mode is a fact as of this writing.
As many of you probably know, I used DHTs for many a
year, and won more than a few audio shows with those.
They are certainly low distortion and linear.
I just don't listen to DHTs anymore. The pentode SE
simply trashes them in the ways that count the most--
having more of the musical artifacts fully there and
fully fleshed-out during the actual enjoyment of the music.
These are conservatively run Zero feedback SE Pentode amplifiers.
Personal preference? Fidelity trumps theory and
calculation? It's not that simple-- it's more like
having ALL of the musical expressions there ALL of
the time. In other words, it's a better WORKHORSE.
Do any of you have good explanations for this? I
would love to hear those.
Thanks, Guys!
--Dennis--
-
I will have to look into that. What I have read says that a 0% tap is pentode and a 100% tap is triode. In triode tubes are more linear than in pentode and UL is operation somewhere between pentode and triode so I always assumed that the linearity of the tube in UL would be somewhere in between pentode and triode.
If what you are saying is true, then why did Hafler include feedback here?
Push pull cancels the even harmonics but single ended doesn't cancel anything. I would think feedback would be required more in a single ended amp than in a push pull amp?
One other thing. Langford-Smith also said this.
"distortion quite a bit lower than even a triode"
So maybe he was talking about the amplifier as a whole (including GNF) not the linearity of the tube (in UL) itself?
I say that because at 43% the tube itself is not very linear at all so even if 43% is a little off the tube should still be very linear in UL if it is the tube itself that is "quite a bit lower than even a triode" when used in UL. Again, he said the tapping point is not at all critical but the 43% tap UL 6l6 is not even close to the triode 6l6 in terms of linearity.
Can you understand my logic?
Thanks.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Actually working with the circuit and its variables is likely to bring a greater understanding.
Without feedback the distortion of the amp rises with the output power. So if you want the amp to be neutral, the usable amount of power goes down. With SETs this is typically about 20-25% of full power at clipping. This is why Hafler used feedback.
The other thing to understand about feedback is what happens to the output impedance of the amp. When EV and MacIntosh were introducing the Voltage rules for loudspeaker drive, the get the desired Voltage source effect out of the amplifier, feedback was a requirement for any tube amp. So if you want your tube amp to be compatible with speakers designed under the Voltage rules (these days that's about 99% of speakers made) then you had to have feedback. This would allow the amp to cut back its power into impedance peaks, such as the one in most box speakers at low frequencies, in order to obtain flat FR.
The trick with feedback is proper application. Outside of the audio world its known as 'control theory' and the rules of use are well understood.
Its often not applied properly in audio. Crowhurst and Baxandall both pointed to problems associated with linearity issues at the feedback node but neither really offered a good solution, although it was just right there all along. The trick is to wrap it around the amp in the same manner as seen in an opamp, since any amplifier is a primitive opamp anyway. IOW it should be applied to a grid rather than a cathode so as to prevent the feedback signal from being distorted by the tube.
I happen to not agree with Langford Smith as the UL patent says something else- I go with that. Langford Smith was referring to a zero feedback circuit; look at the rest of the article as he discloses his test procedure.
h
So I take it you have never done it and everything you say on the matter is just conjecture. Have you ever seen a set of plate curves for a tube in UL when the taps were in the right place? If you have will you please show it to me.
Do you think the Acrosound TO-300 transformer has the proper % taps for a 6l6?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
-
.
Have Fun and Enjoy the Music
"Still Working the Problem"
according to f langford smith.
here
El Blanco-- Thanks for the link.
Too bad Blumlein left us so young.
Good stuff. Thanks again.
-Dennis-
he will respond thoroughly when he has time.
dont send him any attachments...-he just deletes them.
has tons of videos on his channel but nothing for 10 months.
inductive load line series/output xfrms series is very good.
View YouTube Video
I assume that when he uses the term "ultra-linear" he is meaning that the input signal is limited so as to not drive the tube into it's non-linear regions (near cutoff and near saturation)? Having nothing to do with an ultra-linear output transformer (as developed by David Hafler and Herb Keroes who started Acrosound ) having screen grid taps so that a pentode ( or beam power tube) is operated some where in between pentode and triode?
I think his words and this graph prove that. Notice the absence of "ultra-linear' operation.
Edit, I received answer to my question about his use of the term "ultra-linear" and I am correct.
That aside, he is saying the exact same thing I have been saying. He does a better job than I do.
By the way (and just for fun) I love that when he says "we need a true triode" he holds up a 6b4g. That's what I listen to! :-)
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 10/11/24 10/11/24 10/11/24 10/12/24
sorry.
no, he doesnt mean the H&K UL.
just short stroking the output tube to minimize distortion.
BTW, its been a long time , but the only UL amp I ever heard that sounded right had those Acro TO-300 in them.
Those are not particularly evenly spaced. Draw a circle centered on the idle point... :)
I have one on the boards not that is... :) from idle, +40V on the grid gives a delta V anode of 500V, and 40 vots negative changes V anode by 450. That is pretty fine IMO...:)
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
I think Ive posted this before.
turns everything upside down
another way of looking at a dynamic curve.
In 2D, once you pick your operating point and plot your load line, you can then plot a dynamic curve (transfer function curve).
Here's a tube that is not very linear, with an operating point (and load line) that are chosen poorly making things worse.
The non-linearity of the tube, coupled with poorly chosen operating points, cause one half of the wave form to be exaggerated and the other to be truncated.
If the tube's operation were linear the ellipse would be symmetrical. Because of the distortion the ellipse is egg shaped.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
but its super cool.
I have that Preisman book and it always winds up being low and hot is best.
and not necessarily the flatness of dynamic line.
I've seen operating points described like this many times over the years and always found it vague and confusing. I'm never even sure if those who use these terms even mean the same thing.
Does "low and hot" mean lower voltage and higher current or does it mean lower current and higher voltage?
yes. lower voltage higher current.
A beam power tube or a power pentode is not very linear connected as a pentode, or UL (no matter where the taps are) or as a triode.They are all very non-linear compared to a direct heated triode (45, 2a3, 300b, 211 or 845).
Here is a 6l6 triode connected. It is more linear than when used in UL (no matter where you put the taps for the UL) but still no where near as linear as a DHT.
and here's a 300b direct heated triode.
For those who might not understand what I'm saying, pick you operating point, plot your load line and then see how much the plate voltage changes for a given decrease of the bias voltage (less negative on the grid caused the current through the tube to increase so the plate voltage is reduced) and then see how much the plate voltage changes for that same amount of increase of the grid bias (more negative grid bias caused the current through the tube to decrease so the plate voltage increases). In case you don't know, the music signal simply modifies the grid bias and the result is a change in the current flowing through the tube. The first half of a input wave form is positive causing the current through the tube to increase and the second half of the input wave form is negative causing the current flowing through the tube to decrease.
With a linear tube, with a proper operating point and a proper load line, the plate voltage change be symmetrical (the plate voltage will increase and decrease equally for each half of the wave form). If it doesn't, then distortion is the result. It all has to do with the spacing between grid voltage line following the load line. When they are equally spaced then the output wave form is symmetrical and the HD is low.
Below you can how one half of the output wave form will be much bigger than the other half. What you are looking at is harmonic distortion caused by the grid lines not being equally spaced along the load line. Now pick an operating point of the 300b plate curves above and plot a load line and see what a difference evenly spaced grid lines make.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 10/11/24
The 6L6 is a particularly good example of this. It needs a judicious amount of feedback to produce reasonable linearity, even in UL and triode modes. The same is true of the EL34. There are exceptions though. The 6V6 offers very good linearity in triode mode, and is useful in SE with no feedback at all (assuming damping factor is not an issue).
It was an early one, and if you wish a wee bit lower output Z, apply a bit of plate-to-grid( in pentode mode no less) sufficient to knock its mu down to 3.5 and *THEN let us compare it to a 300B.
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Tell us more! Do you have a schematic with component values?
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2024, Audio Asylum, all rights reserved ]
