[atmasphere]

I guess he removed a BUNCH of his posts then.

Not that I remember, anyway.

Its a simple fact that we could have sold a lot more amplifiers if we only used negative feedback. The reason that we don't is that its use causes the amp to violate a fundamental rule of human hearing: how we perceive the volume of a sound. Negative feedback can increase trace amounts of odd-ordered harmonic distortion (while overall otherwise decreasing THD). Its the trace amounts (specifically the 5th, 7th and 9th harmonics) that are used by the ear/brain system to figure out how loud a sound is.

So when these harmonics are altered even slightly, the amp will sound brighter and louder than it really is. This is one reason why two amps on a bench can measure flat but one is bright and the other is not.

Many speakers (not all, by any means) are designed to expect the amp to have some sort of voltage response at its output that is or approaches a constant voltage characteristic. With tubes this pretty well means that you will have to run feedback. The result will be unnatural brightness and unnatural loudness cues. IOW, with such speakers **it will never sound real**.

If the result will never sound real, why bother?

more on this topic: http://www.atma-sphere.com/Resources/Paradigms_in_Amplifier_Design.php

from Roger:

I note that the Atmasphere S-30 has an odd number of output tubes per channel. I have not seen a schematic of this amp and would be interested in looking at one. I suspect the odd tube may be the driver for the other parallel output tubes and Ralph may be going into positive grid voltages to get more power. Any comments?

We built the driver circuit in our amps to be able to push the power tubes into the class A2 (grid current) region. The 6AS7G remains quite linear in this portion of the curve and so allows us to do so. Conversely, the tube does not go into cutoff very easily either, so (as long as the amp is driving the right load) they are class A2 designs. There are those that claim that this is not 'true' or 'pure' class A, and such statements would be true if we were claiming that the amp was A1. But its not, and does trade off some linearity for more power. However, we also gain instantaneous overload recovery, with no issues of bias stability, both of which are problems in most OTL designs.

The S-30 has an odd number of tubes only because it worked out that way- as many of you know, OTLs have an economy of scale, the smaller you make them the less efficient they become. So the additional tube sections offered by the 5th 6AS7G were helpful.

[Roger A. Modjeski]

Not that I remember, anyway.

Its a simple fact that we could have sold a lot more amplifiers if we only used negative feedback. The reason that we don't is that its use causes the amp to violate a fundamental rule of human hearing: how we perceive the volume of a sound. Negative feedback can increase trace amounts of odd-ordered harmonic distortion (while overall otherwise decreasing THD). Its the trace amounts (specifically the 5th, 7th and 9th harmonics) that are used by the ear/brain system to figure out how loud a sound is.

So when these harmonics are altered even slightly, the amp will sound brighter and louder than it really is. This is one reason why two amps on a bench can measure flat but one is bright and the other is not.

Many speakers (not all, by any means) are designed to expect the amp to have some sort of voltage response at its output that is or approaches a constant voltage characteristic. With tubes this pretty well means that you will have to run feedback. The result will be unnatural brightness and unnatural loudness cues. IOW, with such speakers **it will never sound real**.

If the result will never sound real, why bother?

more on this topic: http://www.atma-sphere.com/Resources/Paradigms_in_Amplifier_Design.php

from Roger:
We built the driver circuit in our amps to be able to push the power tubes into the class A2 (grid current) region. The 6AS7G remains quite linear in this portion of the curve and so allows us to do so. Conversely, the tube does not go into cutoff very easily either, so (as long as the amp is driving the right load) they are class A2 designs. There are those that claim that this is not 'true' or 'pure' class A, and such statements would be true if we were claiming that the amp was A1. But its not, and does trade off some linearity for more power. However, we also gain instantaneous overload recovery, with no issues of bias stability, both of which are problems in most OTL designs.

The S-30 has an odd number of tubes only because it worked out that way- as many of you know, OTLs have an economy of scale, the smaller you make them the less efficient they become. So the additional tube sections offered by the 5th 6AS7G were helpful.

Ralph,

Thanks for your information. I have great respect for your work. Going into the positive grid area is a good idea and I have done it in some of my designs. Have you measured the peak output current using a tone burst? I am curious as to what you can get.

[atmasphere]

Hi Roger,

Thanks! Not to get too much into a mutual admiration society thing but that's the same with me

I've not messed with tone bursts especially. We *have* done things with DC voltage settings on the grids of the tubes though. The tubes can support so much current (for short periods, pardon the pun, but we are talking seconds, not milliseconds) that half of the power tubes in one of our amps can blow a power fuse that is a multiple of the correct rating quite easily without themselves being damaged. This is why having the correct fuse in place is important in our amps- having one that is over-rated can have the same results as a shorted output transistor in a solid state amplifier.

This should not be surprising though- an OTL has to drive the speakers directly and so can have some transistor-like abilities in that regard. Most people are quite surprised at how much current the output section can handle without damage- to put some numbers to that one section of the 6AS7 will handle 500ma peaks, far beyond the 120ma limit commonly seen in many spec sheets! Eric Barbour, who used to be with the Svetlana company's US office, observed this years ago and published the resulting curves on their 6AS7G datasheet.

[Retsel]

Roger was asking about the schematic for an Atmosphere amp.  This thread discusses an Atmosphere design.

http://www.diyaudio.com/forums/tubes-valves/161112-what-tubes-tube-amp.html

Retsel