Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Sahib]

dcollins wrote on Fri, 20 April 2007 01:17

Larrchild wrote on Thu, 19 April 2007 15:24 Absolutely. I don't understand the mechanism completely, but as explained to me, the electrons that hang out between the grid and cathode create free degenerative NFB. Last time I posted this they made fun of me!  http://www.next-power.net/next-tube/articles/Stockman/Stockm an.pdf DC

Larry,

The paper talks about the electrons between the anode and grid. Whereas what you were told means the space charge.

What I don't understand from the paper is that why the author felt the necessity to model the triode as a pentode. Why not take it as triode, as it is and not clutter the article?

In terms of the claim, my knowledge in tubes are limited and I thought that the so called back action was only of significance  in tetrodes and that is why the secondary screen was introduced to form a pentode. Would the distance between the anode and the grid be sufficiently close to cause this back emf? Or would it not be that the grid would have to be more positive than the anode to attract electrons from it?

Regards,

Cemal

[bruno putzeys]

Hauled a tome off the bookshelf called "Grondslagen van de radiobuizentechniek" (Basics of Radio Tube Technology), 1946, Philips.
Space charge means the charge of the electrons that happen to be traversing the vacuum at any time, whether between k and g or between g and a. Space charge has the effect of repelling electrons trying to come off the cathode, thus limiting the anode current for any given anode voltage. It's responsible for the 3/2 power law.

[Sahib]

Thank you Bruno. What I meant was the charge area between the grid and the cathode.
Anyhow, I am making an innocent inquiry here. DC said that he was loughed at when he posted this article but what is the basis that he was loughed at? Would what the paper say not be described as feedback or would it just not happen in a triode?

Regards,

Cemal

[bruno putzeys]

Once something as technical as feedback (or the absence of it) becomes an article of faith, technical inroads into the matter are treated like evolution science by the religious: laughter at best, abuse at worst.

(Shame on me, going off topic like that)

[Sahib]

Huh.

It's a good job that the feedback is not mentioned in the Koran. Imagine those suicide bombers with triodes wrapped around them.

Regards,

Cemal

[Larrchild]

Bruno wrote:

Quote:

Once something as technical as feedback (or the absence of it) becomes an article of faith, technical inroads into the matter are treated like evolution science by the religious: laughter at best, abuse at worst.

Bruno, at least you understand that IC's were once single-grid creatures that crawled from the primordial DeForest and Fleming soup, to walk, upright, on their 8 little legs.

Some, your age think the universe of amplification was created in 1947 A.D.

That's why I like this place.

[maxdimario]

What's really funny, or sad, is that all the blabbing away is usually done by people who have never heard a low feedback class-a amplifier playing back a 50's vinyl..

THAT is the test which makes people understand over here at my place..

I have been screwing around with audio amps for decades now, and it was my highest priority as a kid to get as close to the music as possible..

people come over and listen to the records and their expression is that of pleasent surprise... it sounds like a performance.

you can, in my experience, only get this reaction from a low feedback tube amp setup in class-a.

triodes have inherent neg. feedback.. true... but it is local.

I prefer the sound of pentodes with neg. feedback from anode to grid sometimes.


Sahib, you will NOT get the answer at the local hi-fi store because low feedback amps are very expensive to make properly, with PSU parts being hyper-critical as well as other issues such as stability etc.


without high feedback ALL of the deficiencies of the various caps and transformers become evident..

STILL... the low feedback design when used in conjunction with anode chokes is the most lifelike of sound amplifiers.

it ALL depends on the quality of the active and passive components.

one reason why the 50's tube mics burn modern 'low distortion' mics is because they are a low-feedback, class-a design..

or why the fairchild, with all it's distortion, is great on vocals.. the only feedback is the internal feedback of the tubes, and the 4 tubes are wired as one stage with no coupling caps and an inductive load..

but what would our engineer friends do with the new technology then?

[Quince]

I'm guessing the reason some people like tubes is that the time constant of thermal memory distortion is way below the audio band, unlike with solid state (chip amps are the worst here).  Check out this AES paper: http://www.aes.org/e-lib/browse.cfm?elib=7497 (or I can email you if you don't have membership).

Someone tried to address this in solid state design, but I haven't tried to replicate his circuit and measure it: http://peufeu.free.fr/audio/memory/

[Larrchild]

The raw, uncorrected linearity of silicon semiconductors is, as a rule poorer than vacuum tubes. Thats why semiconductors make such darned good computers, lol.  On...offf


That's why people dig tubes. (sometimes)

But the stuff you attach to the tube usually has coloration, so that's what some dig.

[maxdimario]

Quince wrote on Mon, 30 April 2007 15:33

I'm guessing the reason some people like tubes is that the time constant of thermal memory distortion is way below the audio band, unlike with solid state (chip amps are the worst here).  Check out this AES paper: http://www.aes.org/e-lib/browse.cfm?elib=7497 (or I can email you if you don't have membership). Someone tried to address this in solid state design, but I haven't tried to replicate his circuit and measure it: http://peufeu.free.fr/audio/memory/

I think you just explained indirectly what I have been hearing for years.

thanks!

[maxdimario]

would a BIGGER active component have different thermal memory characteristics, because of the larger mass and surface area?

[Quince]

Larrchild wrote on Mon, 30 April 2007 11:37

But the stuff you attach to the tube usually has coloration, so that's what some dig.

I find I don't like even new output transformers, despite them supposed to be perfect.  Berning's switchmode output transformer is one great way to deal with this, and I've been thinking of doing a three-phase version that avoids the glitches of a two-phase version.

Hybrids can also be interesting, especially if you match nonlinearities in such a way as to decrease them overall.  An example is Kevin Gilmore's amp for electrostatic headphones like Stax.  The output stage has about 0.05% THD before feedback because the tube curves are nearly inverse in this operating point to the cathode driving MOSFETs: http://www.headwize.com/images5/gilmore4_1.png

[Quince]

maxdimario wrote on Mon, 30 April 2007 17:01

would a BIGGER active component have different thermal memory characteristics, because of the larger mass and surface area?

Yes.  I think another way to decrease the effect is run the parts at a higher idle bias.

Chip amps are the worst because you have the output stage in the same package as the input and VAS...

[Larrchild]

Quince

Quote:

Berning's switchmode output transformer is one great way to deal with this, and I've been thinking of doing a three-phase version that avoids the glitches of a two-phase version.

The load lines on that thing look Really good, don't they?
But is it a switching supply with 6L6's modulating it or 6L6's with a virtual output transformer on them, lol I can't decide. either way, it's clever.

[Quince]

I even have the large potted core ferrite transformers here.  Just have to find the time to build the thing...

[Edit:] I wanted to ask something actually.  I ended up building that electrostatic headphone amplifier, and I've been thinking of something.

Bob Cordell mentions somewhere that sufficient HF getting into an amplifier, either at the input, or from the output through the feedback loop, could potentially exceed slew limitations and cause TIM-related distortions.  So he recommends, in addition to the usual RC lowpass seen on some amps' inputs, to have some inductance at the output (even if not necessary for stability, so we're not talking about the usual amplifier output inductors here).  He suggested 1-3 uH air cores with RC Zobels to ground on either end.

But this is for power amps, whereas I've been wondering about how to size this filter for my headphone amps.  I have headphone amps both for regular headphones, and also the electrostatic one above, and especially for that one I'm most unsure of since the electrostatic headphones are essentially a capacitive load (roughly 94 pF for the Stax Omega 2).