Yes, Gary is pretty clued up, but he still misses on a few things. What I like is that Jack had certain attitudes on what's essential for getting good sound, which are evolving -- it's very satisfying to see a next level of understanding take place ...
hate to tell you Frank, you just listen to the sound like everyone else, all else is make believe
First you let the amp roll off at 20 kHz, then you use a graphic equalizer to "correct" for that?
An odd approach, I think. In my view, it solves exactly nothing because the amp is still free to pahse shift. In fact, in my view, if anting it makes the rpoblem worse because an EQ is an additional layer of circuitry inserted in between, with all its properties thrown in.
Me, I would have let the amp go to at least 80 kHz at -3 dB, and I'd throw the EQ out. But that's just me.
you are trying to read too much into it.
first, the amp is not my design.
second, I'm not trying to correct it, it was just an experiment about the audibility of a 2.5 dB boost at 20kHz. the results were contrary to my expectation.
why the said amp has -2.5dB @20kHz is hard to say, thing is it sounds good to me.
I personally think that phase is over-rated.
many conveniently forget that we're listening to music from speakers which sometimes barely achieve a -3dB point at 25kHz which, on a minimum phase system (which speakers aren't but the approximation is good enough, IMO) and considering a 2nd order roll-off, translates to some 45 degrees of phase shift at 6-7 kHz.
and music comes from a mic of limited response, with its own phase shift. then the signal passes through electronics with limited FR. the phase-shift accumulates. and, dare I say, the fashionable non-linear phase reconstruction filters of today create some phase shift of their own in the DAC, the single place where we have the most control over without great expense 😀
but for some reason, people seem to care about it in the amp. may I speculate on why? because it's the most convenient place where they have a considerable degree of control over it 😀 that is, of course, eliminating the ones that are trying to sell you a very expensive amp 😀
looking at it from a philosophical stand-point, it kinda makes sense not to be very sensitive to phase in this world where the movement of the head by half an inch equals the period of a 20kHz sound wave. 4 degrees at 20kHz means 0.16 mm. in order to achieve less than 4 deg at 20kHz we'd need a corner frequency of about 1MHz I think.
Last edited:
Of course what one "hears" in one's head can be called make believe, the mind is interpreting the clues in the sound to work out what's going on acoustically. Only thing, some audio systems make it easier to follow the threads of sound, and form a mental sense of everything interesting that's happening.
Think of this "test": you walk up to a closed door, and on the other side you can hear a full blown choir letting rip. Do you think you would have any trouble, before you open the door, on deciding whether it's the real thing, or just a decent audio system 'pretending'?
wait a minute.
isn't he the guy making an amp based on class D modules, the most basic Hypex ones? 😀
http://www.6moons.com/audioreviews/genesis5/reference.html
The dullness often associated with equipment is not always frequency response .
Many years ago my brother and I designed an amp using EL 34's . The output transformer was a modified VOX AC 30 made by Avel Lindburg . Avel put some UL taps on for us . The thing of note was exceptional LF performance . The downside was a gentle roll off from 7.5 kHz .
In pentode 48 watts , UL 28 watts , triode 14 watts . Bias set at - 33 V .
Negative feedback was applied . Pentode a washout , UL was nasal . Triode was OK .
Finally no feedback triode . Powerful and real zing . To be truthful it was far more open and had more treble energy . The 7.5 kHz never a problem . One of the best amps I ever heard . After much thought I abounded it as I felt 7.5 kHz was noncommercial . It sounded disgusting in the other forms I would say . Now I own a spectrum analyzer I think it would have sounded better in it's other guises . A recent UL amp I made seems best that way .
Phono stages seem very sensitive subjectively to bandwidth . Piano espeically shows it ! That is as bizarre as that valve amp went in the other direction . My ears say it is about feedback working properly .
Has anyone noticed how op amps sound best at unity gain or above 20 ? Not surprising I guess as unity is so important . Try an op amp at gain 3 , often it sounds nothing like as open as gain of 1 . Same op amp at gain 50 potted down . Very open and a little soft . I suspect inverting the op amp would be better at gains of 3 ( Often a very useful gain ) . RIAA stages might help themselves in what they have to do . Often a high gain stage with passive 75 uS followed by a 3180/318 uS stage with gain circa 16 at 1 kHz ( 63 x 16 = 60 dB ) . If using 100 R loading ( Denon 103/R ? ) there is an opportunity to build a double inverting all active stage . If inverting there is less of an argument to use passive EQ .
The point about the gain of 3 op amp is the treble sounds rolled off , at gain 50 it doesn't . It just sounds sweet . At gain 1 it sounds like a better piece of wire . I don't expect many to agree with this , I probably don't even agree with it myself ( different circumstances and something else ) . That is not the point . The point is to think of feedback and how it changes things .
DF 96 said something which I will bend into dullness might be truth . I suspect it isn't as I don't hear it in real life . Good op amps at gain 1 perhaps are just right ?
Many years ago my brother and I designed an amp using EL 34's . The output transformer was a modified VOX AC 30 made by Avel Lindburg . Avel put some UL taps on for us . The thing of note was exceptional LF performance . The downside was a gentle roll off from 7.5 kHz .
In pentode 48 watts , UL 28 watts , triode 14 watts . Bias set at - 33 V .
Negative feedback was applied . Pentode a washout , UL was nasal . Triode was OK .
Finally no feedback triode . Powerful and real zing . To be truthful it was far more open and had more treble energy . The 7.5 kHz never a problem . One of the best amps I ever heard . After much thought I abounded it as I felt 7.5 kHz was noncommercial . It sounded disgusting in the other forms I would say . Now I own a spectrum analyzer I think it would have sounded better in it's other guises . A recent UL amp I made seems best that way .
Phono stages seem very sensitive subjectively to bandwidth . Piano espeically shows it ! That is as bizarre as that valve amp went in the other direction . My ears say it is about feedback working properly .
Has anyone noticed how op amps sound best at unity gain or above 20 ? Not surprising I guess as unity is so important . Try an op amp at gain 3 , often it sounds nothing like as open as gain of 1 . Same op amp at gain 50 potted down . Very open and a little soft . I suspect inverting the op amp would be better at gains of 3 ( Often a very useful gain ) . RIAA stages might help themselves in what they have to do . Often a high gain stage with passive 75 uS followed by a 3180/318 uS stage with gain circa 16 at 1 kHz ( 63 x 16 = 60 dB ) . If using 100 R loading ( Denon 103/R ? ) there is an opportunity to build a double inverting all active stage . If inverting there is less of an argument to use passive EQ .
The point about the gain of 3 op amp is the treble sounds rolled off , at gain 50 it doesn't . It just sounds sweet . At gain 1 it sounds like a better piece of wire . I don't expect many to agree with this , I probably don't even agree with it myself ( different circumstances and something else ) . That is not the point . The point is to think of feedback and how it changes things .
DF 96 said something which I will bend into dullness might be truth . I suspect it isn't as I don't hear it in real life . Good op amps at gain 1 perhaps are just right ?
I know that. there are other factors affecting perceived tonality besides FR.
as a matter of fact, the amp I'm talking about has more than decent mid-highs, that's part of what I like about it.
but in no way should that imply that flat FR to at least 20kHz is not a reasonable goal.
it'd be interesting to speculate what could make a manufacturer choose such a limited FR with a solid state amp.
Last edited:
You did, of course, adjust the feedback in each case to take account of the different forward gain of these three different modes of operation? And took account of any variation in g2 quiescent voltage? Or did you work on the basis of equal output impedance? No? Then you were not comparing like with like, so the conclusions tell you almost nothing generic about the modes themselves but merely that mode in that amp with that bias point and that amount of feedback.nigel pearson said:
People seem to do this all the time. A similar mistake is made about levels of feedback. Changing feedback will alter so many things about an amplifier that comparing like with like is almost never done, yet people almost always assume that it has been done so draw conclusions about feedback itself.
A 7.5kHz rolloff has more treble energy? Anecdotal evidence that subjective testing can be virtually meaningless. The simple act of owning a spectrum analyser could make the UL version sound better? You couldn't make it up!!
Yes jack is pretty serious too , even if he listens to VR speakers ...🙂
There are others too, Mike Lavinge as another, on a technical level Gary has them all and the best speakers system too ...
Regards
Thru correspondence , i liked what he has done to the Genesis line , pretty much what Nudell should have done years ago ...
Regards ...
Yes all was religiously applied . My brother alas is no longer alive . I would have loved him to be here to answer . Our intention was that pentode would be valid . Simon was a valve fanatic , often writing about TV repairs . All the pensioners locally had him repair their TV's . He smiled and said all were as reliable as his favourite ITT once sorted out . Here is the only mention I can find on him . Why I write here is you guys have to replace him .
30FL1 @ The National Valve Museum
It was the worse feedback amp I ever heard and without feedback one of the best amps . I have never met it's like before or after .
I wish I had a pair of Quad ESL 63's at the time to try the pentode mode . The basic measurements were not too bad . I would have tried some internal feedback perhaps ? Pentodes as a concept I rather like .
Two feedback amps I have like are Dynaco ST 70 ( EL 34 ) and Leak TL 15 + ( EL 84 ) .
30FL1 @ The National Valve Museum
It was the worse feedback amp I ever heard and without feedback one of the best amps . I have never met it's like before or after .
I wish I had a pair of Quad ESL 63's at the time to try the pentode mode . The basic measurements were not too bad . I would have tried some internal feedback perhaps ? Pentodes as a concept I rather like .
Two feedback amps I have like are Dynaco ST 70 ( EL 34 ) and Leak TL 15 + ( EL 84 ) .
I wont hold the hypex flaw against him ....🙂
If so , shouldn't feedback be adjusted for different loads ....?
Thing I like about Gary is that he went to buy some Genesis speakers only to find they had gone out of business . The store owner said " Gary why don't you buy the company " . From what I can tell it was like entering the church for Gary . As someone who has always been " in the church " I guess it wouldn't occur to me that hi fi isn't normal life . To Gary it did .
It makes no difference who actually limited the amp's output, you, I or someone else. The fact remains that it is too limited to cause problems.
Everything you say about mics and stuff is all right, but then we should ask ourselves: with all that going on, do we need further dterioration from the amp? Isn't it bad enough already? Will a phase shifted signal come out of my amp even more phase shifted?
Obviously, I personally think phase shifts should taken seriously - not as a fetish, but seriously nevertheless.
And this is not the domain of only very expensive amps. Harman Kardon and Sony have paid much attention to this, and you can hear it as clearer and more precise amps than most of the compeition in their price class. That said, I do admit both tend to cost somewhat over the average, but far from serious.
By way of example, my Citation amp, rated at 100 WRMS/8 Ohms is said to have a 20 kHz square wave tilt, resulting mostly from phase shift, is less than 4 degrees. Its open loop power bandwidth extends to just over 80 kHz into 8 Ohms, and its global NFB is just 12 dB. No surprise then that it hits 250 kHz quite eaisly.
Citation series is the upmarket part of their portfolio, but from personal experience (I also own a 6550 and a 680 integrated amp, much lower price wise, with very similar specifications, the 6550 uses 17 dB of global NFB), I can attest to the fact that things are hardly any different. Compared to their speed, most others are handicapped.
I wonder what you'd think if you compred them head to head to your amp? I mean, once the problem has been much reduced, and some reasonable slew rates (HK 680, Vitation both quoted as 80 V/uS) were thrown in for good measure?
And they achieve this by having closed loop bandwidths of just over 240 kHz; by 1 MHz, as Goldmund amps regularly do, I suspect that the phase shift at 20 kHz would be less than 2 degrees.
yes, you're right, hi-fi is about as little deterioration as possible so if you can make an amp having low phase shift in the band of interest, why not.
but at the same time I can't help but think of the accumulated phase shift, starting from the air produced by say the singer's mouth and ending in our living rooms 🙂
and, again, how about the low frequency phase shift, caused by the decoupling caps? I did some calculations a while back and it looks like it can be a real problem, with objectionable group delay extending to the low-mid region.
do you have any explanation/speculation about why 80V/us would correlate to good sound?
but at the same time I can't help but think of the accumulated phase shift, starting from the air produced by say the singer's mouth and ending in our living rooms 🙂
and, again, how about the low frequency phase shift, caused by the decoupling caps? I did some calculations a while back and it looks like it can be a real problem, with objectionable group delay extending to the low-mid region.
do you have any explanation/speculation about why 80V/us would correlate to good sound?
If we were to start with the harm already done to the sound, we might as well as give up right there and then, because no matter what we have, it will only add more degradation.
I prefer to think about adding as little degradation as possible to what has already been added on its way from the studio to my CD player.
Low frequency shift can be minimal with a little effort and thinking. While it is smearing, I believe it's far less damaging than high frequency shifts, which are much more easily heard. Few speakers reach down as low as 20 Hz, and besides, as far as I know, at 20 HZ one only has a few mammoth church organs and that's it. These days, I do beleieve frequency corners in the low region are at 3 Hz or lower, and if DC, then below 1 Hz. They are no longer relevant down there.
For those who still use decoupling caps, I note the are aware of group delay problems and tend to use really top quality caps, thus minimizing trouble.
To the best of my knowledge, no-one has established an iron clad connection between slew rate and quality sound, HOWEVER, only if one has gone above a certain point, 1 V/uS per every peak volt output, so SID, TID and TIM do not bedome a problem.
So, for a say 50 WRMS amp, minimum should be 28 V/uS, for a 100 WRMS amp 40 V/uS, and so forth. I mentioned it only because H/K likes it double, so the 680 integrated does 80 V/uS (yet needs 37 V/uS), the Citation needs 40 V/uS yet does 100 V/uS, etc. The point being that more can't hurt, while less can.
Also, in my experience, each and every superior sounding amp I have ever heard also satisfied this requirement without fail.
Does yours?
- Status
- Not open for further replies.