Apples and potatoes must be cooked differently to get best meals from each.
I am getting better result from pentodes with nested feedbacks than from best linear triodes. Well; if the sound quality from EL34 is the same as no-feedback paralleled 4P1L, still my EL34 amps have better bass control due to servo damping, and amplification of channels does not depend on particular tubes, i.e. no need to select and match so close to get the same loudness from both channels. It is defined by resistors.
I am getting better result from pentodes with nested feedbacks than from best linear triodes. Well; if the sound quality from EL34 is the same as no-feedback paralleled 4P1L, still my EL34 amps have better bass control due to servo damping, and amplification of channels does not depend on particular tubes, i.e. no need to select and match so close to get the same loudness from both channels. It is defined by resistors.
Member
Joined 2009
Paid Member
all my amps have almost 0 bass control, only feedback cures it, like a bandage on a bleeding wound 
, this is why I am building a new kt77 with cathode feedback (8-9 db) and apply another 8-9 db of GNF instead of the actual 13 db of GNF. The regulator should not let the voltage go down 0.1V at 10Vrms of output, which should take all the electrons that the tube can give, (AC heaters all the way, separate transformer lhahha0)....
i am still ignoring absolutely what is the servo damping....
defined by feedback resistors?
feedback equalize the gain between channels, for volume, we have to resort to black magic... (L pad averaging lol)
, this is why I am building a new kt77 with cathode feedback (8-9 db) and apply another 8-9 db of GNF instead of the actual 13 db of GNF. The regulator should not let the voltage go down 0.1V at 10Vrms of output, which should take all the electrons that the tube can give, (AC heaters all the way, separate transformer lhahha0)....i am still ignoring absolutely what is the servo damping....
defined by feedback resistors?
feedback equalize the gain between channels, for volume, we have to resort to black magic... (L pad averaging lol)
My earlier point. Good tube amplifiers don't require a lot of NFB because tubes, particularly triodes, are inherently more linear than other amplifying devices, i.e. semiconductors, beam pentodes etc. When you get into higher NFB, usually to correct for non-linearity in the amplifying devices themselves, then you run into a "slew" (pardon my pun) of other issues, including differences in device bandwidth, the ability to deliver current or charge into complex impedances downstream, etc.
There are usually more problems that can pop up than there are passive and active circuit components. In fact, there is a direct correlation between the number of components and the "Q" of the system, which, in the limit, is essentially, a complex bandpass filter with (hopefully) some voltage and / or current gain. High "Q" filters make lousy amplifiers for music signals.
I agree. And to the same end with other scaled-down situations - IME the speaker efficiency and active LF supplement can change things quite some.
Higher power SE, if wanted sure. At some point (and if you can find a good SE
P xfmr) PP triodes sound pretty good too (IME - better image width and depth at the expense of vividness and forward projection), and mitigate the need for higher B supply unless you really need more than 15W. HK
This is an excellent point, and difficult to argue against. Decent vacuum valve amplifiers are typically biased hot enough to remain in class A over almost all of their operating range. A good balance between a truly non-switching below clipping signal transfer and a switching-only-on-peak-levels transfer was worked out in early days, and really hasn't changed much since the classic era.
The seldom highlighted virtue of vacuum valves is that they run hot! Not a fashionable virtue, but very valuable for audio.
There is one other non-monotonic mechanism in conventional valve amplifiers. The output transformer's core has a permeability vs. signal curve (B/H curve) that passes through a large change of slope through the zero crossing region. Push-pull amplifiers' output valves, properly DC bias balanced, see this as a variation in the parasitic (primary) inductance. Low source impedances, triodes or feedback from the anodes, minimizes but can't banish this factor.
Single-ended designs bias the B/H curve somewhere away from the zero crossing, improving the low-level performance in that respect, but at other significant design costs. Everything on the Goddess's Good Green Earth is horses for courses.
All good fortune,
Chris
Once I used additional transformer after output transformer. But with very "fat" wire diameter. (No gap, 1:1, induction for max voltage p-p with respect to amplifier small power)
So the DC resistance of the secondary, from that added transformer, was much smaller than the tube OT. less than 0.5ohms. Somehow the tightness, low end control and low mid def. was significantly better?.
So the DC resistance of the secondary, from that added transformer, was much smaller than the tube OT. less than 0.5ohms. Somehow the tightness, low end control and low mid def. was significantly better?
.
Ah, so you are talking about low-level zero-crossing magnetic effects. I remember a thread years ago where SY tried to measure these effects in a real push-pull amp and was unsuccessful. Ever since then, I have wondered if the effects are so small as to be insignificant. Are you aware of any test data that shows how significant the effect is in the real world? It seems like with a sensitive test setup, it would be easy to observe the phenomenon in an amplifier but I've never seen such test results. Do you know of any? I'd love to see them.
I've also heard people attribute increased detail in SE amps to the lack of a magnetic zero-crossing. That never made sense to me, as I expect that the "details" would be riding on top of larger, lower frequency waves and wouldn't just exist at the zero-crossing region. It seemed more likely to me that euphonic distortions might just be a source of perceived extra detail.
I will tell you that the SE amplifier I built on the bench sounded very, very nice.
Zero crossing is an irrelevant issue. A lot theory but I have yet to find someone who can detect anything. The only real case where something is visible on the scope is when using a heavily gapped transformer in PP. The B_H curve has a sort of bubble around zero but it's a "stationary thing" and its influence on performance is...zero. In pathological cases (I mean, people who get very biased by theories) one can use special core materials with very narrow loop and high permeability where this is basically invisible even on the scope.
Member
Joined 2009
Paid Member
SY failed to measure it, but I am sure you can measure something, it could be a reduced bandwidth, which I suspect,
Many factors are biasing the results here:
1. bias changes B+ voltage,
2. Bias changes output resistance
It could be that the unpleasant transition to class AB is caused by the reduced B+
Class AB needs a good deal of voltage...
and unregulated supplies, like choke loaded, split choke cathodes etc... (to be discussed)
p.s. I find the actual detail, low sounds less in S.E.T....
Many factors are biasing the results here:
1. bias changes B+ voltage,
2. Bias changes output resistance
It could be that the unpleasant transition to class AB is caused by the reduced B+
Class AB needs a good deal of voltage...
and unregulated supplies, like choke loaded, split choke cathodes etc... (to be discussed)
p.s. I find the actual detail, low sounds less in S.E.T....
It's difficult to measure things that happen below the overall noise floor, and some could argue that anything below the noise floor can't matter. Might be right, but in the interest of good conservative engineering we shouldn't jump to conclusions. OTOH we wouldn't want the "perfect" to be the enemy of the good.
Iron core effects at low currents appear as a variation in slope (permeability) of the B/H curve through the zero-crossing region (although I just did a Google search which showed NO curves with this included, so it's apparently not well known any more) and permeability only matters in its effect on primary inductance.
Primary inductance, in turn, only matters because it appears as a parasitic load to the output valves. At lower signal levels, presumably, reduction in primary inductance wouldn't be so important.
In solid-state (output-transformer-less) amplifiers these small signal effects are getting more attention these days, and valve amplifiers are relegated to snake oil status, but it doesn't have to be so. (Unless Jean-Luc P. says to make it so!)
All good fortune,
Chris
Reduced bandwidth? How? I get well above 100 KHz with standard EI laminations. So I don't think so. SY failed to measure it because it is not even second order effect. It's really tiny and irrelevant. Something is only visible at low frequency. At high frequency I really doubt you will get any zero crossing...a small DC bias which is likely present in most PP amps will ensure that actually zero crossing never happens at mid-high frequencies in normal operation. The AC induction is too small.
Member
Joined 2009
Paid Member
I am not sure about this. What is into the pot is the (output) voltage as function of B, turns, frequency and core area. However at mid-high frequency the impact of the core is smaller and smaller even for high output.The AC induction is inversely proportional to frequency. As 1KHz is 50 times bigger than 20Hz and max B is calculated to not reach saturation @ 20Hz it means that B will be 50 times smaller. At 20KHz it will be 1000 times smaller.....without considering that musical signal at high frequency usually does not require more than a couple of watts.....
All this after considering that zero crossing is a small effect on its own....hence irrelevant really.
Member
Joined 2009
Paid Member
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.