Would you happen to know if any of his writeups such as the one you mentioned are available online?
Are we afraid of solid state devices in our tube amplifiers?
I occasionally use a solid state current source as the plate load of a driver tube. It replaces the more usual plate load resistor.
Are we afraid of interstage transformers?
I have used interstage transformers. An interstage transformer has a low frequency limit (core, turns, DC current, peak AC swing), and a high frequency limit to impedance (distributed capacitance). The tube distorts at those lower load impedances at the frequency extremes. Interstage transformers have greater or lesser sensitivity to magnetic interference fields.
I still use resistors as plate loads.
Each tool has its function. But do not respond to every design with one tool, i.e. the hammer (otherwise everything begins to look like a nail).
I occasionally use a solid state current source as the plate load of a driver tube. It replaces the more usual plate load resistor.
Are we afraid of interstage transformers?
I have used interstage transformers. An interstage transformer has a low frequency limit (core, turns, DC current, peak AC swing), and a high frequency limit to impedance (distributed capacitance). The tube distorts at those lower load impedances at the frequency extremes. Interstage transformers have greater or lesser sensitivity to magnetic interference fields.
I still use resistors as plate loads.
Each tool has its function. But do not respond to every design with one tool, i.e. the hammer (otherwise everything begins to look like a nail).
Last edited:
There is always a law of diminishing returns. For a start, you could. say, reduce transformer distortion to zero - (that's possible if its entirely air-cored - but such a thing would the size of a house, or bigger), but there is always going to be a residual of distortion from other sources - directly in the tube transfer characteristics, power supply modulation, etc etc.
In a practical situation, one can always design a transformer to meet a spec. But the amount of iron and copper, and labour to build it, goes up faster beyond a certain point. You can always design a transformer to have half the distortion of another, while keeping are aspect, eg bandwidth, the same. You can start by increasing the turns. That increases winding self and inter capacitance, so you choose a bigger core so you can use less turns. That increases leakage reactance which also reduces bandwidth, so you go bigger again. So you need to reduce window utilization, on no that means a bigger core too. Now we are getting too much core loss, so we increase turns.....
Or we can start by increasing the core mass, which increases core loss, so we increase turns...
You claim of experience to the contrary is most likely observer bias, or Volkswagen Effect (see earlier post on this).
And that list isn't complete. I've got a few on hand that are not in the list.
Keit,
One factor of solid state amplifying devices is that as the voltage changes, things like stored charge change, And junction capacitance changes. Talk about slew rates, they are dynamically changing versus the voltage.
Tubes do not normally have changing capacitance versus voltage, at least not unless the electronic fields start pulling one element closer, or not as close to another.
We have done research on Electro-Phonics where the music notes are at a frequency which also happens to be at one of the mechanical resonances. The pulling rate activates the resonance. When that happens, you can stop the music and hear the mechanical resonance decay, sometimes quite slowly.
Micro-phonics is activated by sound.
Electro-phonics is activated by changing voltages (even if the tube is isolated from sound).
One factor of solid state amplifying devices is that as the voltage changes, things like stored charge change, And junction capacitance changes. Talk about slew rates, they are dynamically changing versus the voltage.
Tubes do not normally have changing capacitance versus voltage, at least not unless the electronic fields start pulling one element closer, or not as close to another.
We have done research on Electro-Phonics where the music notes are at a frequency which also happens to be at one of the mechanical resonances. The pulling rate activates the resonance. When that happens, you can stop the music and hear the mechanical resonance decay, sometimes quite slowly.
Micro-phonics is activated by sound.
Electro-phonics is activated by changing voltages (even if the tube is isolated from sound).
Last edited:
6A3sUMMER:
Actually, inter-electrode capacitances in tubes DO change with voltages applied.
Between the cathode and control grid is a space charge - an accumulation of electrons. Some electrons get thru the grid and end up in the anode (or screen) and make tube do useful things. But most electrons in the space charge are in a continuous process returning to the cathode from which they came, as a function of tangential velocity, initial radial velocity, and repulsion by the overall space charge and grid potential.
This means that there is a form of electrical conductor extending out from the cathode to close to but not touching the grid wires. That's a capacitor. And its capacitance depends on the grid voltage, because the more negative the grid, the further away stay the space charge electrons.
In beam tetrodes and the like, there is a space charge between screen and anode, so that anode-screen capacitance is a function of anode & screen voltages.
These effects are much smaller that the capacitance vs voltage changes in transistors, and is not normally a concern in tube amps. In any case the main tube variation is in the grid-cathode capacitance. In transistors, the main variation is in the more important collector-base capacitance. Other causes of distortion predominate in tube amps.
But it IS a concern in vacuum tube RF oscillators, which show drift in frequency with HT variation, and with heater voltage variation, as space charge density is a function of cathode emission efficiency.
Tektronix used to make a capacitance and inductance meter, the Type 130, which can easily resolve down to 0.05 pF. It was marketed partly on the basis that you could quantify the change in tubes capacities with change in applied voltages - something not normally included in datasheets (nor with transistors, either).
The space charge also suppresses shot noise. The noise generated in a tube is always a little bit greater in practice than what the classical theory predicts. A Bell Labs boffin in the 1950's showed that as both noise and capacitance depend on the space charge, you could more accurately predict what noise you would get in a circuit by measuring the change in grid-cathode capacitance - making Tektronix's nice meter even more useful. I have one, it dates to 1959 and its serial number is in the 7000-range. So they made quite a few after starting in 1954.
A bit of interesting trivia for you....
Actually, inter-electrode capacitances in tubes DO change with voltages applied.
Between the cathode and control grid is a space charge - an accumulation of electrons. Some electrons get thru the grid and end up in the anode (or screen) and make tube do useful things. But most electrons in the space charge are in a continuous process returning to the cathode from which they came, as a function of tangential velocity, initial radial velocity, and repulsion by the overall space charge and grid potential.
This means that there is a form of electrical conductor extending out from the cathode to close to but not touching the grid wires. That's a capacitor. And its capacitance depends on the grid voltage, because the more negative the grid, the further away stay the space charge electrons.
In beam tetrodes and the like, there is a space charge between screen and anode, so that anode-screen capacitance is a function of anode & screen voltages.
These effects are much smaller that the capacitance vs voltage changes in transistors, and is not normally a concern in tube amps. In any case the main tube variation is in the grid-cathode capacitance. In transistors, the main variation is in the more important collector-base capacitance. Other causes of distortion predominate in tube amps.
But it IS a concern in vacuum tube RF oscillators, which show drift in frequency with HT variation, and with heater voltage variation, as space charge density is a function of cathode emission efficiency.
Tektronix used to make a capacitance and inductance meter, the Type 130, which can easily resolve down to 0.05 pF. It was marketed partly on the basis that you could quantify the change in tubes capacities with change in applied voltages - something not normally included in datasheets (nor with transistors, either).
The space charge also suppresses shot noise. The noise generated in a tube is always a little bit greater in practice than what the classical theory predicts. A Bell Labs boffin in the 1950's showed that as both noise and capacitance depend on the space charge, you could more accurately predict what noise you would get in a circuit by measuring the change in grid-cathode capacitance - making Tektronix's nice meter even more useful. I have one, it dates to 1959 and its serial number is in the 7000-range. So they made quite a few after starting in 1954.
A bit of interesting trivia for you....
Last edited:
Keit, I am referring not necessarily to the size, but of the design.
C core, quad core, amorphous cobalt quad core, perfectly symmetrical computer windings of both primary and secondary.
All these differences have made a very distinct sound change, which is unmistakable.
I have not been able to measure in any way what makes the toroid etc transformer sound so much better...
it is not bandwidth, not power, not inter-modulation, or square waves, I believe it is their ability to convey music in a real way without interference, like the proverbial straight wire. I had no idea how much the OT contributed to sound before that...
So in terms of sound quality, ceteris paribus, in order of sound alteration generators, it could be resumed this way: feedback, OT, Tubes, and obviously Power supply, the resistors and capacitors in a properly designed wide bandwidth open loop with decent GNF are very small marginal (in the 1% ) contributors to sound characters of the whole system.
I cannot stand amps without feedback, nor too much, I can survive with cheap resistors and cheap caps everywhere in the amp, it is not so bad, and I can live without NOS tubes. A good OT is like good wine.
C core, quad core, amorphous cobalt quad core, perfectly symmetrical computer windings of both primary and secondary.
All these differences have made a very distinct sound change, which is unmistakable.
I have not been able to measure in any way what makes the toroid etc transformer sound so much better...
it is not bandwidth, not power, not inter-modulation, or square waves, I believe it is their ability to convey music in a real way without interference, like the proverbial straight wire. I had no idea how much the OT contributed to sound before that...
So in terms of sound quality, ceteris paribus, in order of sound alteration generators, it could be resumed this way: feedback, OT, Tubes, and obviously Power supply, the resistors and capacitors in a properly designed wide bandwidth open loop with decent GNF are very small marginal (in the 1% ) contributors to sound characters of the whole system.
I cannot stand amps without feedback, nor too much, I can survive with cheap resistors and cheap caps everywhere in the amp, it is not so bad, and I can live without NOS tubes. A good OT is like good wine.
Last edited:
Keit,
It is always time to learn. I just wish I could remember Some of the things I forgot. 1959 is the first time I made calibrated measurements: with a 55 milli-Ohm 0-1 mA meter.
It is always time to learn. I just wish I could remember Some of the things I forgot. 1959 is the first time I made calibrated measurements: with a 55 milli-Ohm 0-1 mA meter.
Please can you explain what it is? Is it da diode that blocks 'reverse' current like one would do on the screen grid? [I could not find a schema in a simple search on the net, it is so cluttered these days 😡 ]
From my experience with tube amps, i know i only heared one tube amp that sounded better without caps. That was a 6SN7/2A3 SET with transfo's on the input, interstage and output. The builder did wind his own trasfo's as he was not happy with what is availeble as interstage transfo's (limited in bandwith and coloring the sound he said) and that interstage is a big double C core with big copper wire and a Mumetal core. But he is an experienced electronic engineer with the right equipment and skills. Most can't do that. And if he had to sell those interstage transformers he build, it would cost close to 1K at least...
But for the rest of the transformer or DC coupled tube amps i heared, they are very coloured and often not very powerfull. Then a tube amp with wima MKP10 (looked down on by many audiophiles that are also into DIY) coupling caps sounds way better.
My tube amp is a stock Prima Luna Prologue 4 and i like it so it's fully stock. My pre is a modified (replaced tubes, caps and the potmeter for volume) chinese pre, where i replaced the cheap caps (i could not read the brand in chinese, but got a schematic in english) with Wima MKP10 and a Jantzen Crosscap i had laying around on the output. And that sounds very good now. Better than a homebrew preamp that is fully transformer coupled of a friend.
I think caps are still the best in most cases for coupling stages, and the discussion over wich cap is the best is very 'snake oil' colored. Use good quality filmcaps with a higher voltage rating than you need (double is very safe and my standard) and you're set. That can be mass produced Wima's or so, that does not matter much. You certainly don't need caps that cost a lot to do it right.
And caps in the PSU are rather essential i think. I rather have no ripple with some electrolites caps in the PSU than no caps and a lot of ripple. And an voltage regulator IC or even discrete also has an influence on the sound...
But for the rest of the transformer or DC coupled tube amps i heared, they are very coloured and often not very powerfull. Then a tube amp with wima MKP10 (looked down on by many audiophiles that are also into DIY) coupling caps sounds way better.
My tube amp is a stock Prima Luna Prologue 4 and i like it so it's fully stock. My pre is a modified (replaced tubes, caps and the potmeter for volume) chinese pre, where i replaced the cheap caps (i could not read the brand in chinese, but got a schematic in english) with Wima MKP10 and a Jantzen Crosscap i had laying around on the output. And that sounds very good now. Better than a homebrew preamp that is fully transformer coupled of a friend.
I think caps are still the best in most cases for coupling stages, and the discussion over wich cap is the best is very 'snake oil' colored. Use good quality filmcaps with a higher voltage rating than you need (double is very safe and my standard) and you're set. That can be mass produced Wima's or so, that does not matter much. You certainly don't need caps that cost a lot to do it right.
And caps in the PSU are rather essential i think. I rather have no ripple with some electrolites caps in the PSU than no caps and a lot of ripple. And an voltage regulator IC or even discrete also has an influence on the sound...
Triode-al,
I'm not terribly surprised you could not find an example, as it was rarely done back in the tube era. I've done it because there is a narrower range of new tubes these days and you need more options. I've seen it done elsewhere.
The fact is, though, its not often needed - simpler remedies against grid blocking can be done (eg the techniques used by GEC, Neville Thielle), that may not be a completely full cure, but it often doesn't need to be, in a commercial sense. It depends on how far in quality you really want to go.
It is a property of human hearing that gaps in sound lasting millisecs occurring one in a while (say a few seconds) are damn near inaudible to almost everybody. However, once you have "learnt", so to speak, you can pick it a lot more easily.
But if gaps in sound lasting miliseconds occur frequently (say several times a second), that is an entirely different story. It sounds dreadful to even the most cloth-eared. There is a continuum of audibility in between these extremes.
By clamp diode, I mean a diode placed between a fixed reference voltage (say a voltage divider off the HT, with a few microfards across it), and the anode of the preceding valve, thus limiting how far positive it can go.
You can just bias that valve to it's anode is at a pretty high voltage all the time and can't go much further (That's what GEC used to do in their professional grade amps, but then it will be well down on its grid curve and will distort more. This extra distortion is not much of a problem if the output stage is driven by push-pull driver tubes, as GEC did, eg in the 88-50).
Both these methods are a 100% cure for grid blocking.
Alternatively, you can add a diode(s) between the output stage grid(s), and earth, preventing them from going to far negative. This provides a fast discharge path for the grid-coupling capacitor, stopping grid blocking as soon as possible. It still happens, but its less audible. The grid conducts on positive peaks, and the (reverse-connected) diode conducts on negative peaks, balancing the thing.
Back in the tube area, most engineers didn't like these sorts of diode tricks - vacuum tube diodes cost money and they were another source of hum, as the cathode was not grounded. The common sorts of vacuum diode, such as 6AL5 or diode connected 12Ax7 don't have a high enough heater-cathode voltage rating. 6X4's were ok ratings-wise, but expensive and would cause a lot of hum. These aspects should not be an issue now.
I'm not terribly surprised you could not find an example, as it was rarely done back in the tube era. I've done it because there is a narrower range of new tubes these days and you need more options. I've seen it done elsewhere.
The fact is, though, its not often needed - simpler remedies against grid blocking can be done (eg the techniques used by GEC, Neville Thielle), that may not be a completely full cure, but it often doesn't need to be, in a commercial sense. It depends on how far in quality you really want to go.
It is a property of human hearing that gaps in sound lasting millisecs occurring one in a while (say a few seconds) are damn near inaudible to almost everybody. However, once you have "learnt", so to speak, you can pick it a lot more easily.
But if gaps in sound lasting miliseconds occur frequently (say several times a second), that is an entirely different story. It sounds dreadful to even the most cloth-eared. There is a continuum of audibility in between these extremes.
By clamp diode, I mean a diode placed between a fixed reference voltage (say a voltage divider off the HT, with a few microfards across it), and the anode of the preceding valve, thus limiting how far positive it can go.
You can just bias that valve to it's anode is at a pretty high voltage all the time and can't go much further (That's what GEC used to do in their professional grade amps, but then it will be well down on its grid curve and will distort more. This extra distortion is not much of a problem if the output stage is driven by push-pull driver tubes, as GEC did, eg in the 88-50).
Both these methods are a 100% cure for grid blocking.
Alternatively, you can add a diode(s) between the output stage grid(s), and earth, preventing them from going to far negative. This provides a fast discharge path for the grid-coupling capacitor, stopping grid blocking as soon as possible. It still happens, but its less audible. The grid conducts on positive peaks, and the (reverse-connected) diode conducts on negative peaks, balancing the thing.
Back in the tube area, most engineers didn't like these sorts of diode tricks - vacuum tube diodes cost money and they were another source of hum, as the cathode was not grounded. The common sorts of vacuum diode, such as 6AL5 or diode connected 12Ax7 don't have a high enough heater-cathode voltage rating. 6X4's were ok ratings-wise, but expensive and would cause a lot of hum. These aspects should not be an issue now.
Last edited:
Another name for the Clamp Diode would be the DC Restorer often used in vacuum tube pulse circuitry.🙂
Aeroflex bought Marconi T&M in the 2000s. I don't recall the TIM TS being part of the product mix in the comm group at that point. But their 2026 RF Synthesizer was capable of generating three RF signals simultaneously, all locked to a common timebase. Real useful when testing for IM products. Locked TB is a requirement, independent generators don't do the test properly.
Aeroflex bought Marconi T&M in the 2000s. I don't recall the TIM TS being part of the product mix in the comm group at that point. But their 2026 RF Synthesizer was capable of generating three RF signals simultaneously, all locked to a common timebase. Real useful when testing for IM products. Locked TB is a requirement, independent generators don't do the test properly.
Prejudices can be extremely harmful.
Usually people who fall to them tend to neglect other aspects.
The key to achieving high performance and sound is to try looking from the most POV as possible.
Hi Waxx, Belgium is the birthplace of good sound 🙂
i believe you about this amplifier. The more mainstream transformers the more risk of a dull sound.
I thought GNF and other types of feedback like cathode or UL could reverse the transformer coloration of the sound but it is not doing a perfect just as I can easily hear the differences with the same transformer specifications...
When I had a lot of spare time I read the whole Turner Audio (Australia) website. I calculated the parameters of his schematic and draw all the load curves, input and output coupling bandwidth etc, operation points. He build his own C cores and balanced loaded chokes.
I almost bought transformers from him, and I would.
Mr. Turner is rational, no bs and I communicated with him when I started modifying some amplifiers, I would send him my revised schematics and he would advise me and revise, he never wanted any money in return.
His website is also a real treasure of easy to build schematics for serious diyers.
i believe you about this amplifier. The more mainstream transformers the more risk of a dull sound.
I thought GNF and other types of feedback like cathode or UL could reverse the transformer coloration of the sound but it is not doing a perfect just as I can easily hear the differences with the same transformer specifications...
When I had a lot of spare time I read the whole Turner Audio (Australia) website. I calculated the parameters of his schematic and draw all the load curves, input and output coupling bandwidth etc, operation points. He build his own C cores and balanced loaded chokes.
I almost bought transformers from him, and I would.
Mr. Turner is rational, no bs and I communicated with him when I started modifying some amplifiers, I would send him my revised schematics and he would advise me and revise, he never wanted any money in return.
His website is also a real treasure of easy to build schematics for serious diyers.
triode_al mentioned using a diode (in series?) with the Screen. And that is what I have seen on some schematics.
Is the screen going positive, with too few electrons?
Is the screen is getting negative (an excess of electrons)?
And why would a Beam Power tube, or Pentode tube have excess electrons on the screen in the first place? I thought the purpose of Beam Forming Plates, and Suppressor grids was to prevent secondary electrons from the plate going to the screen.
If the Beam Forming Plates and Suppressor grids are not working, then is it possible . . . That the tube is bad, so replace it? That the tube is being operated in conditions (voltages, currents, dissipation, etc.) that are outside of the range the tube was intended for?
Please, I would like an explanation of why use a diode there?
Thanks!
Is the screen going positive, with too few electrons?
Is the screen is getting negative (an excess of electrons)?
And why would a Beam Power tube, or Pentode tube have excess electrons on the screen in the first place? I thought the purpose of Beam Forming Plates, and Suppressor grids was to prevent secondary electrons from the plate going to the screen.
If the Beam Forming Plates and Suppressor grids are not working, then is it possible . . . That the tube is bad, so replace it? That the tube is being operated in conditions (voltages, currents, dissipation, etc.) that are outside of the range the tube was intended for?
Please, I would like an explanation of why use a diode there?
Thanks!
Last edited:
Wavebourn,
I am confused. You said me too. Then you seemed to want to know why I want an explanation. Did you mean you want someone to explain? Or, in my post # 177, did I fairly well answer my own question?
Thanks!
I am confused. You said me too. Then you seemed to want to know why I want an explanation. Did you mean you want someone to explain? Or, in my post # 177, did I fairly well answer my own question?
Thanks!