 
|
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 |
152.131.10.193
The question I had in my initial posting was inspired by this schematic..
http://www.audiodesignguide.com/my/2a3_amp.html
Scroll down to where he has a choke loaded 6sn7 cap coupled to a 2A3.
What interested me is that there was no dropping resistor in the b+ rail..only the resistance of the anode choke is dropping B+.
I guess I wanted to know if it's the quality/manufacturer of the tube
tube that determines sound quality or it's operating point.
Could be both to a certain extent?
(Whoops..I asked another question)
Follow Ups:
Iczerman
Look again -
The choke loaded cap coupled tube is a 6SL7.
The 300V to the plate will fry the puppy.
6SL7 is rated at only 250V max.
I would not try to drive a 2A3
with a 44K output impedance.
Sound quality is determined by BOTH
make and operating point.
AFAIK the make does not change the
"sweet spot" of any given tube.
DanL
Every sheet I've seen says the 6SL7 is a 300V tube. My question: If the output Z of the 'SL7 doesn't impair the desired frequency response, is there still a problem?
--------------------------
Buy Chinese. Bury freedom.
"If the output Z of the 'SL7 doesn't impair the desired frequency response, is there still a problem?"No, I don't think so. But with that high of a drive impedance the Miller capacitance of whatever you are driving needs to be very low or there will be impaired FR.
edit, I should say phase. But if the phase is shifted then the response is down. It might not be down much but the phase will not be right at 20kHz unless the -3db point is at, at least, 200kHz.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 12/22/10 12/22/10 12/22/10
TK
My bad - I looked at TubeCAD
and they had the max at 250V.
The 6SL7 drives the 2A3 with an
F3 of 80KHz which means that the HF
will be affected from 8KHz and above.
While flat at 20KHz the phase isn't.
DanL
"The 6SL7 drives the 2A3 with an
F3 of 80KHz which means that the HF
will be affected from 8KHz and above.
While flat at 20KHz the phase isn't."
It will be a fine OPT that can do 80 KHz -3dB bandwidth, not even getting on to the speaker issues.
It really should be more of a concern about ratio'ed phase matching rather than absolute phase fidelity through 20 KHz. That's because there's more that can be done to match the phases than can be done to extend the amp's flat phase response. If the left and right out of a speaker pair is well phase matched to the listener's ear, where at 20 KHz left is -32.3 degrees and right is at -32.4 degrees, then it's a good bet that there will be good imaging because timing is nearly identical.
I would wager that left = -10 degrees and right = +10 degrees would be worse to the sound than -32.3 and -32.4 degrees respectively. Speaker designers who know what they're doing will not attempt the former as much as the latter, by driver parameter and crossover component value matching.
And there are those guys like Siegfried Linkwitz where left = right = almost 0 degrees, through electronic correction techniques. Better for some, too much op amp signal processing for others.
-Kurt
Kurt
As I have said before ...
I will not agree that because
other stages create phase errrors
then it is fine not to try to keep
the phase as flat as posible.
If good enough is fine by you OK.
But don't tout that it is the standard.
Why design an amp with less than perfect
response (on paper) because in reality
it will not even reach the design ideal.
The flatter the response the better.
The flatter the phase the better.
Reach for perfect not good enough.
DanL
Flatter or better in one area may not, or often is not, better overall.Whether you agree or not is irrelevant to the others making very good and valid points.
All it shows is that you do not agree. The others have acknowledged your point but have forwarded many more areas that are, for the most part, of greater importance.
So yeah yeah, we get your minor point and yes, that's what you want to concentrate on because you are special (by your own admission) but for many, the varied contributions by T, Kurt, TK and others are more relevant and insightful to real world music reproduction than your small point about HF bandwidth repeated over and over again.
Edits: 12/24/10
"others are more relevant and insightful to real world music reproduction than your small point about HF bandwidth repeated over and over again. "
But why can't we have it all.
Let's make sure the phase and FR is flat where we easily can (circuits) and work on the harder parts (speakers and transformers etc)?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Hi,
> Let's make sure the phase and FR is flat where we easily
> can (circuits) and work on the harder parts (speakers and
> transformers etc)?
That sounds, please forgive me the reference, exactly like:
"You blind guides, which strain at a gnat, and swallow a camel."
Honestly.
If Phase at high frequencies is such a concern, why not first fix it where the problems are gross and then worry about where they are subtle and not the other way around?
Or, to take refuge to the phraseology of the good book again, why argue about the straw of a few degree of phaseshift one way or the other in a high impedance SE Amp driver, when you ignore the rafter of huge phase shifts in Speakers, Microphones and others?
And why not remove that rafter first, after which you may be in a better position to tell your brother "But what about that 20 degrees phaseshift in your driver circuit?".
Happy holidays, a good new year and great tunes to all wishes
T
Sometimes I'd like to be the water
sometimes shallow, sometimes wild.
Born high in the mountains,
even the seas would be mine.
(Translated from the song "Aus der ferne" by City)
"If Phase at high frequencies is such a concern, why not first fix it where the problems are gross and then worry about where they are subtle and not the other way around?"
Again I say, why can't we do both?
"Or, to take refuge to the phraseology of the good book again, why argue about the straw of a few degree of phaseshift one way or the other in a high impedance SE Amp driver,"
I don't remember arguing about that.
" when you ignore the rafter of huge phase shifts in Speakers, Microphones and others?"
I don't remember ignoring them.
Merry Christmas
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Hi,> Again I say, why can't we do both?
We can. You might want to engineer some significant positive phaseshift at 20KHz into your speakers (and negative at 20Hz) to oppose that of bandwidth limited systems. You may be able to compensate the HF phase to flat at 20KHz with only three or four suitable all-pass circuits.
Or you can add one of the mystical "200KHz frequency response" Supertweeters (Hint 1, 150KHz frequency response demands < 0.1" radiating surface and or a phasegrid with < 0.1" spacing if using a compression driver. Hint 2, measurement microphones that are meaningful even at 100KHz are exceedingly rare.)
In this case maybe the best way to do it is likely in the digital domain and you can compensate both frequency and phase on a systemic level to "flat", compensate crossover phase-warp and do many other things.
Of course, then the whole debate about a little phase-shift in the Amp is moot again.
I know my (personal) speakers go up pretty flat to 50KHz (ring radiator Ribbon) on axis and reach above 25KHz off axis before crapping out and I have a very good idea what the phaseshift is as well (also at LF where with room assist sub 30Hz are reached). Since converting them to a 3-way series X-Over (used to be parallel with inverted mid) they even do credible square-waves (for a speaker anyway).
I have also measured my Amp's, including output transformers and they are are not really better than my Speakers and not due to a lack of bandwidth in the driver. I'd still not bet I could hear a 50KHz in this system with any reliability (remember, I participated in a DBT decades ago that illustrated that this can be audible under some conditions) and not just because age and abuse has attenuated some of my HF hearing.
So yes, you can do both. You even should do both. But I retain that the common way in engineering is to fix big problems first and to start worrying about small ones later. So fix your speakers to have flat phase-response at 20KHz and we talk again about Amplifiers.
Ciao T
PS,
I personally would probably build a DC coupled 2A3 amp using a triode wired 7721/D3a, which gives a lot of Mu (78), low output Z (around 3K) and ton's of transconductance (which is actually what is needed for "jump" or "live" factor) and incidentally around 700KHz -3dB point between driver and output tube (not that I am interested specifically in that high turnover, I want to the low driver Z to improve overload behaviour).
And I would probably use a 45 instead of a 2A3, but I have to try the new single plate 2A3's first, most of my "bad" 2A3 experiences came from double anode types, so I'll hang fire on that till I have time.
But as I said, that's just me and such a design would not claim universal goodness...
Sometimes I'd like to be the water
sometimes shallow, sometimes wild.
Born high in the mountains,
even the seas would be mine.
(Translated from the song "Aus der ferne" by City)
Edits: 12/25/10
"But as I said, that's just me and such a design would not claim universal goodness..."
Oh but I beat it would be good...
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Because we live in the real world and have to make a balance of choices.
If you want more bandwidth, then just go transistor or OTL. Those have more high frequency detail, but it might affect the tone in the midrange. YMMV. Certainly don't expect a SET amp to be within 1 degrees of phase flatness at 20 KHz. Try for better, but you might lose something else.
-Kurt
Understood.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Thanks, Stephen, but I think it's about time we let Dan be with his own opinions. He seems to always take issue with me responding in any way to him; and others, too.
So just Merry Christmas!
-Kurt
Hi,
> While flat at 20KHz the phase isn't.
Hmmm. And how flat is the phase of your speakers HF Unit at 20KHz?
Indeed, have you ever considered what the excess phase at 20Hz & 20KHz is for a hypothetical, completely flat in-band frequency response, crossover-less speaker is, which has -3dB points at 10Hz & 40KHz and the theoretically ideal 2nd order out of band rolloff of an ideal mechanical system (not that any such speaker exists anywhere in actual space/time)?
In the end we deal with complex systems and complex interactions. While I personally do like to maximise bandwidth prior to a well designed output transformer and use wideband speakers, I hold few illusions that that a few tens degree excess phase-shift from my electronics will be terribly noticeable once the amplifier has left the test-bench and is attached to speakers with often multiples of 360 degrees EXCESS phaseshift across the 20Hz - 20KHz bandwidth.
Ciao T
PS, the DC coupled 6SL7 SRPP driven Amp from J.C. Morrison I mentioned in the earlier thread is this:
Here the full "Fi-Primer" reproduced, with several interesting and proven designs:
This is an online version of the classic "Fi Primer", by J.C. Morrison, published with the kind permission of J.C. himself.
Sometimes I'd like to be the water
sometimes shallow, sometimes wild.
Born high in the mountains,
even the seas would be mine.
(Translated from the song "Aus der ferne" by City)
T
As much as I respect you I have to disagree.
1) The higher the frequency the more our ears
are sensitive to phase deviation.
2) If we pile on phase/response deviation because
"well the speakers aren't perfect" then why
are we trying to get HiFi sound at all.
3) I can't stand metal domes because
the ~25KHz ringing gives me a headache
so my HF sensitivity is better than most.
DanL
Hi,
> As much as I respect you I have to disagree.
>
> 1) The higher the frequency the more our ears
> are sensitive to phase deviation.
The sensitivity of the human hearing towards phase deviation broadly parallels the ears overall sensitivity. General hearing acuity at very high frequencies is incredibly poor, due to the fundamental mechanics and "electronics" of the process.
I do not advise adding tons of extra phaseshift to an audio system, at the same time I refuse to be worried about 10 or 20 or even 30 Degrees at 20KHz, simply because I know the phase response of Microphones, Speakers and other devices. We are talking about very few percent added to the overall accumulated phase error.
> 2) If we pile on phase/response deviation because
> "well the speakers aren't perfect" then why
> are we trying to get HiFi sound at all.
If the speakers phase-error is 10 to 30 times that of the amplifier allow me to be somewhat blase' about the magnitude of the effect. I'd worry about other stuff than phase first.
In fact, if you listen to any Vinyl or analogue produced recordings on CD you should give up any hope that anything can sound good at all, given the phase-errors in these.
> 3) I can't stand metal domes because
> the ~25KHz ringing gives me a headache
> so my HF sensitivity is better than most.
Mine is not as good as it used to be, but pretty good.
I'm still trying to make mind up if the problems from ultrasonic resonances (yes, I observe these too - surprisingly few people do BTW) are down to bone-conduction or intermodulation in the non-linear air in ear-channel.
So, I agree that all else a higher f3 would be nice to have, but at the same I do not suggest anyone looses sleep, there are bigger fish to fry first.
Ciao T
PS, you did not comment on just how much phase-shift my rather idealised speaker would exhibit, nor on how much real speakers have.
Sometimes I'd like to be the water
sometimes shallow, sometimes wild.
Born high in the mountains,
even the seas would be mine.
(Translated from the song "Aus der ferne" by City)
T
> > I agree that all else a higher f3 would be nice to have, but at the same I do not suggest anyone looses sleep, there are bigger fish to fry first.
To me HF extension and non-resonance is
a top priority because of my sensitivity.
So to me there is no bigger fish to fry.
But that is just me I guess.
My ideal speaker is the ribbon but
I have yet to hear a speaker using
one that integrates it well with the
other drivers used in the speaker.
So after many trials I found a tweeter
that had the best highs I could afford.
The winner was the Dynaudio D-260
which I linked to the spec sheet below.
I have no idea as to the phase deviation
at 20KHz from this tweeter so
I didn't have a response to your query.
I don't know anyone who has measured
the acoustical phase response of
any speaker out there.
Electrical phase but not acoustical phase.
A cap is the ONLY part of my crossover
because I have found that inductors
and resistors rob too much music
when used in a driver crossover.
So phase deviation in my speaker
is minimal or as minimal as posible.
The cap is a 4uFd GE Mtr Run bypassed
with a .1uFd Vit Q for a crossover @~5KHz.
The midbass are Audax Aerogel drivers
again very damped from resonances.
DanL
"...phase deviation in my speaker is minimal or as minimal as posible.
The cap is a 4uFd GE Mtr Run bypassed with a .1uFd Vit Q for a crossover @~5KHz."
That cap produces about 45 degrees of shift in an area that's critical to vocal tonality and balance. I have to agree with Thorsten here; by comparison, 20-30 degrees of amplifier shift at 20kHz or above seems truly inconsequential. As for resonance, it seems unlikely that you're hearing the fundamental ultrasonic vibration. The resonance probably creates mechanical disturbances in the diaphragm that effect reproduction at much lower frequencies.
--------------------------
Buy Chinese. Bury freedom.
TK
That tweeter/cap output is in parallel with
the midbass who is operating with an inherant
inductively produced roll-off so phase cancelling
is also involved which will minimize phase shift.
Basic speaker crossover theory here.
40+ years ago many stores had an ultrasonic
security systems that drove me crazy.
As well as my brothers (genetics?).
My older brother had it the worst
because he was taller than the racks.
Believe what you want but I do hear
above 20KHz as I was also driven nuts
by KEF tweeters which had a cone breakup
that was at 22KHz that irratated me.
I couldn't understand how a soft dome
did this to me until I saw a graph
that showed a glitch at 22KHz.
DanL
Hi,I agree with you on ribbons/magnetostats.
There are reasons for the integration problems and they are moderatly trivial to resolve.
As for my point about bandwidth vs. phaseshift, an IDEAL transducer with a 40KHz -3dB Point (few exist BTW) AND without crossover (not even a capacitor highpass is allowed) will show an ideal 2nd order rolloff and so 45 degrees excess phase shift at 20Khz.
Obviously -3dB points at 20KHz are challenging for 1" dome tweeters, so they will usually be way over 90 degrees phaseshift at 20KHz over a reasonable range of the listening area. Add in the crossover and we accumulate more and more, for even the simplest designs.
In your case the Dynaudio tweeter (only on axis) has around 90 degree phaseshift at 25KHz, 30 degrees off axis it will be around 90 degrees phase shift at 12.5KHz and approaching 180 degrees at 20KHz. Again, crossover not accounted for.
Meanwhile a first order rolloff (like high z driver tube into miller capacitance) at 50KHz will add around 20 Degrees excess phase at 20KHz, which pales next to 90 Degrees and more from most practical speakers at that point.
If phaseshift bothers you so much, I suggest you fix it in your speakers first, where it is dramatic and gross, before criticising much lesser phaseshift levels in electronics...
Again, I agree in an ideal world we would not have them at all, in the real world it is all about finding appropriate compromises.
And in that context the excess phaseshift of a high impedance condemns it no more to unacceptability than that of speakers (or microphones, or output transformers or LP Cutting heads and LP-Pickups or tape heads [record and playback] to naim a few major phase response offenders).
Ciao T
Sometimes I'd like to be the water
sometimes shallow, sometimes wild.
Born high in the mountains,
even the seas would be mine.
(Translated from the song "Aus der ferne" by City)
Edits: 12/23/10
T
Just for the record I researched my Dynaudio tweeters
and they get less than 10 degrees phase shift at 20KHz.
Maybe that is what I heard that I liked about them.
And it was done by an independant reviewer not Dynaudio.
I didn't know for sure so I didn't want to say anything.
DanL
Hi,
> Just for the record I researched my Dynaudio tweeters
> and they get less than 10 degrees phase shift at 20KHz.
Measured how and where? Have a look at 30 degrees off axis (which relates much more to what you hear than the on axis response.
Also, TBH, looking at Dynaudio's own frequency response (on axis) I see more than 10 Degrees phaseshift at 20KHz... If I have the time I'll work it out in detail later.
Ciao T
Sometimes I'd like to be the water
sometimes shallow, sometimes wild.
Born high in the mountains,
even the seas would be mine.
(Translated from the song "Aus der ferne" by City)
T
I could not find the original review that
had the measurement - even after a 200+ sites.
I found many references to that measurement,
on Ecoustics, AVaudio and TechHiFi sites
but the original eludes me.
The importance of the 30 degrees off axis
on what you hear is debateable at best.
But your assertion that I have 90 degrees
of phase shift at 20KHz is wrong.
I realize you believe what you say.
Let's agree to disagree.
DanL
Hi,> I could not find the original review that
> had the measurement - even after a 200+ sites.As you omitted to mention what review it was I did not find it either, but it took google and me around 20 seconds to locate this...
http://www.merlinmusic.com/vsm_wht.htm
This is Merlins own page on features of the VSM Speaker, which includes the Dynaudio Esotar, Dynaudio's best (two levels up from yours, but very much measuring the same).
It contains this graph:
This graph shows the phase of the whole speaker, as well as the frequency response, arguably 10 degrees off axis. It shows first of all quite substantial phase rotation (predictable) due to the crossover, however above around 5KHz we get the phase of the Dynaudio Esotar alone.
It shows small amounts of phase-errors around the -270 degrees phase of the tweeter, however at the top of the tweeter passband we see the phase rapidly changing, eyeballing (the graph is quite small) around 30 degrees at 20KHz and 90 degrees at 24KHz.
> But your assertion that I have 90 degrees
> of phase shift at 20KHz is wrong.You are correct. I was wrong, yo do not have 90 degrees phasesift DUE TO THE TWEETERS EXCESS PHASE ALONE (ignoring the crossover) at 20KHz.
You have 90 Degrees at 24KHz. I stand corrected.
Ciao T
Sometimes I'd like to be the water
sometimes shallow, sometimes wild.
Born high in the mountains,
even the seas would be mine.
(Translated from the song "Aus der ferne" by City)
Edits: 01/04/11 01/04/11
T
The Esotar is one level above my Esotec.
Just converted the graph to bmp and
I came up with just under 20 degrees
of phase deviation at 20KHz.
Also flat from 5K to 17.5KHz.
Also what I read was a Dynaudio bookshelf
speaker review not a Merlin speaker review.
One thing I will tell you the Dynaudio tweeter
loves a higher crossover point than the 2,150Hz
that the Merlin uses so bookshelf or 3 way
is the preferred use for the tweeter.
4-5KHz crossover will let the puppy shine.
DanL
Hi,
I gave up on soft dome tweeters ages ago.
The Dynaudio are better than many, but next to decent magentic planar (even the cheap ones in the various Infinity Speakers) or a Ribbon they just do not the slightest chance, be it in terms of impulse response, phase response at high frequencies or stored energy. Actually a simple CSD plot tells all.
Ciao T
Sometimes I'd like to be the water
sometimes shallow, sometimes wild.
Born high in the mountains,
even the seas would be mine.
(Translated from the song "Aus der ferne" by City)
T
I feel like I am argueing a given here.
You keep saying all things being equal
you would go for wide bandwidth.
They are equal at the designing point
which we are at here in Tube DIY Asylum.
Why design a HF phase shift at 20KHz
when you could design it without any?
Because other things before and after
have phase shifts is irrelevant.
If we shouldn't both with phase shifts
then we shouldn't bother with response
because speakers have aren't perfect.
We shouldn't worry about output impedance
and loadlines because speakers don't
have a flat impedance (except ribbons).
You could wave away with your hand almost
any design parameter with similar arguements.
I just do not see any credibility
to your arguement.
Every parameter should be optimized in a design
if it isn't then redesign until it is.
Good enough should not be good enough here.
I strive for excellance not good enough.
DanL
this is the main unquantifiable....
Yes phase shifts have their problems....
2
Minor phase shifts have less problems...
However all the questions of how to drive a 2A3 remain....it needs a little gain...the 6SL7 has a little gain, and it is very linear...
I have built this one so this I can say 6Sl7 (rca round plate) DC into 2A3 (crappy Sovtek) sounds quite good....it does everything well in one stage except it's output Z is high..6SL7..
So from a philosphical, or just plain good empirical listening view the 6SL7 is such a good choice that the minuses don't overbalance the plusses....etc ok...??? it is not that I want to design in phase shift...it is just that the 6SL7 is agreat tube and the price is right (6 already in my tube box from the old days....)
So having said this the implimentation that you see, and handles all sorts of problems is to parallel the sections...hehe ( and listen to everyone go wild !!!)...this was the best sounding, it drives the 2A3 a tiny bit into A2...or really it is saturating way less near 0v..etc, and then of course we are at 160khz -3db now, and GLORY GLory...around 12 degrees phase shift....so now we have pleased most people....
Yes I have designed it in...hehe but it was really to handle the transition to A2 and overload where it really sounded better, and I diminished the phase shift to boot...ok....
Why??? because it sounded so good....Also 6SL7 is quite good in SRPP also, but parallel is better, and single is also really good....
So, my view: it is a matter of balance and judacious choice....
and of course the 6N1p is really excellent also...not quite the definition, but I can cheaply choke load it with the Hammond 156c, and now the impact and sweetness are so good that the little bit of lost definition I will probably except as well as a bit of gain: it needs 3v from the player....(not to mention the hot high choke loaded 12AT7 intermediate experiment which was also excellent as Mr WLEE of late posting will attest...anyway thanks for the Christmas card..and all the best to you...
Sincerely Happy New Year,
-3db
-3db
The 6N1P has a gain of almost 40.
With a plate choke you get near that.
Why the need for 3V from the player?
You should need less than 1.5V.
I assume diode or battery on the cathode.
If you have an unbypassed cathode resistor
then you will need more than the 156C.
DanL
dynamically so 3v....
and yes with the battery bias...it is really good...
onto the 2a3c's next....
Have we beaten the 6SL7 to death...and I acknowledge they were Sylvania round plates, not RCA's...tall chromed tops...
I love those tubes...
Somehow sometimes I don't love the 61np...but it is really good...
and it just takes one 156c....the 12at7 takes two to sound really goood..
so anyway to bed...good night.
-3db
Hi,
> I feel like I am argueing a given here.
Not quite...
> I strive for excellance not good enough.
Not quite. You strive for excellence where it contributes very little on a whole system level, as the small degree of difference in phase shift between for arguments sake a driver stage with 50KHz bandwidth into the output stage vs. one with 100KHz bandwidth (which still has phaseshift at 20KHz anyway) pales next to the phaseshift in your speakers and likely even your output transformers.
So you do not strive for excellence at all, where it matters, where gains are huge (like fixing the phaseshift in your speaker you can gain between 90 - 180 degrees improvement!), you only strive for it in a narrowly defined area that ranks low in relative impact on phase-response.
> You keep saying all things being equal
> you would go for wide bandwidth.
> They are equal at the designing point
> which we are at here in Tube DIY Asylum.
They clearly are not.
Using a driver valve with much lower anode impedance than the 6SL7 normally means we have much lower gain and now need add another stage. So it always is a tradeoff. I would probably choose the 50KHz lowpass from the 6SL7 over using two stages of 6SN7 for example.
The only option for "all else mostly equal" would be to use the 7721 as driver, which I would advocate over the 6SL7 in most cases, but which is not as easy to source and apply.
So, things never are truely caeteris paribus.
> Why design a HF phase shift at 20KHz
> when you could design it without any?
Because in exchange I gain something else I want in my design. Designing is about making the right compromises in the context of the design.
You may prefer trading off using many stages for wide bandwidth/low phaseshift, I may prefer trading some bandwidth/phaseshift for using fewer stages.
> Because other things before and after
> have phase shifts is irrelevant.
Hmmm. I disagree. It is the malaise of modern science to see too many things in isolation.
If you wish to make amplifiers that are used only on the testbench and conform to certain intellectual ideals, you are free to ignore all else.
If you wish to make amplifiers that work in a given system of recordings, speakers etc. et al. you ignore the behaviour of the rest of the chain at your peril.
> If we shouldn't both with phase shifts
> then we shouldn't bother with response
> because speakers have aren't perfect.
Not quite. We need to understand the relative magnitudes of the problems to understand if certain amount of imperfection is likely going to be trouble or not. So just as I will not worry about a few degrees phaseshift at 20Hz or 20KHz I will not worry about halve a dB or maybe even a full db attenuation at these frequencies, next to the influences of room and speakers.
> We shouldn't worry about output impedance
> and loadlines because speakers don't
> have a flat impedance (except ribbons).
Correct in part. We must actually consider the real load-ellipses, produced by real speakers (and let's not forget output transformer prinamy inductance and parasitic shunt capacitance) rather then getting excessively gung ho about the resistive load-lines, unless our tube will be loaded resistively.
Or in other words we must come out of the ivory tower of idealism and reference our work to the grimy, stinky and annoying reality of the world around us.
> You could wave away with your hand almost
> any design parameter with similar arguements.
I do. All the time.
Any technical parameter of any design must be referenced to reality and the purpose to be meaningful.
A while back there was a debate here "if there was a perfect amplifier would you want it".
As long as absolute perfection remains unattainable we must compromise and make sure the compromises work in reality.
> Every parameter should be optimized in a design
> if it isn't then redesign until it is.
> Good enough should not be good enough here.
So, go and fix the gross phases-shift in your speakers, if you want any credibility in saying "good enough should not be good enough here".
In my books an amplifiers phase-shift is "good enough" if it does not contribute materially (more than 10%) of the total systemic phase-shift at the frequency extremes, a paradigm that is ridiculously easy to fulfil.
If you want to advance that a lesser contribution than 10% is needed for the amplifier phase-shift to remain essentially inaudible I am open to hear your arguments in support of such a reduced margin or in favour why my "good enough" is not actually "good enough".
Honestly, arguing about phase-shift in Amplifier design and ignoring that in speakers is like the parable in the good book about straining the gnat.
Ciao T
Sometimes I'd like to be the water
sometimes shallow, sometimes wild.
Born high in the mountains,
even the seas would be mine.
(Translated from the song "Aus der ferne" by City)
I understand that we shouldn't get too excited by a little phase shift when elsewhere in the system there are bigger fish to fry but...
I'm going to try to refute some of what you have been saying by using quotes from one of my favorite authors.
"If we look at the harmonic spectrum of a 300B loaded with a the usual loads of around 2k3 at 350V/80mA operating conditions, we have a harmonic spectrum dominated by a second harmonic and reaching the measurement limits of anything I have access to by the 5th harmonic.
........However, most of the driver-circuits used with a directly heated triode will have distortion of their own. Furthermore, if a directly heated triode is driven from a non-zero source impedance (output impedance of the driver) there will be distortion introduced. "
"In addition, a directly heated triode (and any similar triode) has a quite substantial Input Capacitance comprising mostly the milleramplified Anode to Grid Capacitance. The input Capacitance of a given Valve is (Cga + Cstray) * Mu + Cgk. This capacitance must be driven from somewhere, not only with respect to the frequency response, but also concerning the current draw at higher frequencies. For a 300B V`alve the input capacitance is usually around 70pF, a little more in reality due to stray capacitance's from the wiring or pcb. If we take 80pF Input Capacitance and a Bias of 70V for our above mentioned operating point we must be able to supply the current drawn by this capacitance up to at least 100kHz (usually a much higher frequency is strongly advisable) at full voltage swing. The impedance of the 80pF Capacitance at 100kHz is around 20kOhm, the peak current drawn by this capacitance at full signal is 3.5ma.
Hence our driver stage must supply this kind of current without problem, in fact it is required to supply even more. "
The 2a3 does not need as much drive swing but does have more Miller capacitance (76pF vs. 67pF) than a 300b.
I would think that the above quote would indicate that something with more current and lower plate resistance than a 6sl7 is needed to properly drive a 2a3. Regardless of what might be wrong with the rest of the system.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Tre,
You need to account for the lowpass behaviour in this.
Clearly, even so, we must operate the 6SL7 at quite a high current to avoid slewing, but it CAN be done.
And no, I would not myself be likely to design an amplifier like this.
However I do however feel that to dismiss any such design outright is not reasonable.
As said, there are many variables here and it is probably worthwhile trying to understand what goes on and why such choices where made, instead just calling them "bad".
Ciao T
Sometimes I'd like to be the water
sometimes shallow, sometimes wild.
Born high in the mountains,
even the seas would be mine.
(Translated from the song "Aus der ferne" by City)
"As said, there are many variables here and it is probably worthwhile trying to understand what goes on and why such choices where made, instead just calling them "bad"."
Fair enough and if it wasn't being shoved down our throats as the only and best way to drive a 2a3.....
***I could look up dozens of quotes from Dennis where he seems to be claiming to have re-invented physics and claimimg to have understanding "way beyond" that of mortal man..... :-) ***
.....I may have just left it alone.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Hi,
> Fair enough and if it wasn't being shoved down our throats
> as the only and best way to drive a 2a3.....
I do not remember shoving the 6SL7 down anyones throat.
There are always some fringe people who will weave great mysteries about simple design choices, mainly because they would never like to admit that their choices are compromises (like everyone elses of course).
As for serious stereo, I know people (with decent ears and long experience to add) who have Amp's build on these principles and they seem to like the results with speakers similar to those promoted by serious stereo, so there is some merit here.
I think we are better off ignoring the mumbo jumbo and trying to understand what is done and why, than pouring scorn, but that's just me and WTFDIK...
Ciao T
PS, a quote from John Curl:
"Condemnation without Examination is Prejudice"
Sometimes I'd like to be the water
sometimes shallow, sometimes wild.
Born high in the mountains,
even the seas would be mine.
(Translated from the song "Aus der ferne" by City)
"I do not remember shoving the 6SL7 down anyones throat."
No, not you. Dennis is the one doing the shoving. I'm sorry for not being clearer on that.
"I think we are better off ignoring the mumbo jumbo and trying to understand what is done and why, than pouring scorn, but that's just me and WTFDIK..."
I'm am sure that is good advice. Advice that I may or may not be able to take.
But thanks.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
"As said, there are many variables here and it is probably worthwhile trying to understand what goes on and why such choices where made, instead just calling them "bad"."
I think this a very important point. Many many times around here, people concentrate on one fine detail of a design and completely ignore the whole; the major part that makes that design good or even great. They may even ignore the same error of greater magnitude elsewhere.
Because it is very very easy to pick on one detail and claim it is wrong in someone else and they have it right.
It is much harder to appreciate the whole without experiencing the total result because that takes a much greater effort. There may even be parts of the whole that are outside their understanding but that would take an admission of failure so it goes unobserved or unacknowledged.
Music is a whole body experience so it baffles me why some don't take that into the appraisal of other peoples designs and contributions.
Great contribution T :-)
cheers,
Stephen
T
W O W
We have formulated design philosophies
and priorities from experience and
experimentation and came to completely
different conclusions across the board.
Let's just agree to disagree.
Merry Christmas, Friend.
DanL
How far off is it?
Stephen
> > > How far off is it?
I don't know what you mean.
As is the 20KHz will be about 25° off.
You need lower than an 18K driver
to get the 20KHz in phase.
DanL
So of no great importance then in the real world.
Thanks.
DanMy tube manual lists the 6SL7 at 300V max:)
No matter though, why would anyone use one to drive a 2A3?
Now, a choke loaded 6SN7 or better yet a 7193, that could sound
very nice.Cal
Edits: 12/22/10
Cal,
Someone would use a 6SL7 because it has the more overall gain, to have a lively sounding , FUN amp, with a fantastic jump factor at varying listening levels, which a 6SN7 would never do, and the 6SN7 two-stage DC amp would be pedestrian at best, a boring and dead sounding amp, like most amps out there. FYI, a mu of 100 12AX7, properly implemented, half a section, will beat em both, be even more fun to hear.
I've heard and built it, others have too.
Jeff Medwin
"FYI, a mu of 100 12AX7, properly implemented, half a section, will beat em both, be even more fun to hear."
If you like the 'AX7, you should try the EF86. Wow! :)
--------------------------
Buy Chinese. Bury freedom.
Post a Followup:
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
