I've used Toroidy for my 6p14p-ev baby huey, and they are very good.
Do you play guitar too?
You guys are above my league for experience when it comes to more than 2 stages; high power; and combined with negative feedback.
I only do low power, 2 stage single ended or 2 stage push pull.
I do not always use negative feedback.
The only negative feedback types I like to use are Ultra Linear; and from the output tube Plate to the driver cathode.
Decades ago, I used the 16 Ohm tap for output cathode feedback (with the primary leads reversed). I no longer want to take the feedback signal from the secondary.
When I tried Schade feedback, I did not really like the tradeoffs. I probably did not play with it enough.
Presently, I only use Beam Power / Pentode outputs in either Ultra Linear, or Triode wired; and Triodes (Indirectly Heated Triodes, IDHT).
I am considering using DHTs (Directly Heated Triodes) again, this time with DC powered filaments; specifically using the 45 or 2A3.
I would say that there are so many topologies that work very well when the design pays careful attention to all the details.
I usually remember to write out the 3 letter and 4 letter acronyms, for the newbees.
Happy designing, Happy building, and Happy listening to all of you.
I only do low power, 2 stage single ended or 2 stage push pull.
I do not always use negative feedback.
The only negative feedback types I like to use are Ultra Linear; and from the output tube Plate to the driver cathode.
Decades ago, I used the 16 Ohm tap for output cathode feedback (with the primary leads reversed). I no longer want to take the feedback signal from the secondary.
When I tried Schade feedback, I did not really like the tradeoffs. I probably did not play with it enough.
Presently, I only use Beam Power / Pentode outputs in either Ultra Linear, or Triode wired; and Triodes (Indirectly Heated Triodes, IDHT).
I am considering using DHTs (Directly Heated Triodes) again, this time with DC powered filaments; specifically using the 45 or 2A3.
I would say that there are so many topologies that work very well when the design pays careful attention to all the details.
I usually remember to write out the 3 letter and 4 letter acronyms, for the newbees.
Happy designing, Happy building, and Happy listening to all of you.
Last edited:
Thanks, I will play on that. Looking at a fast simulation, it doubles the THD.
It will for sure be a compromise, the amp is "on the edge" on many sides.
J113 seems a good solution, I've found others in the SS section liking it alot (EG here: Complementary diff input with JFET and BJT cascode ). It would be great to have a single component with two balanced jfets inside, like LMK389 or U440, with characteristics like the J113 or better.
May I ask you some suggestions on output tube plate to driver cathode feedback? I've not used it yet. Thanks!
Thank you! I wish the same to you!
zintolo,
You probably are familiar with Schade feedback, from the output tube plate to the driver tube plate.
Yes?
If you have self bias in the driver tube cathode that has a bypass cap across the self bias resistor . . .
Then disconnect that bias network from ground; and insert a resistor in series from the bias network to ground.
Then connect a negative feedback resistor from the ouput tube plate, and a DC blocking cap in series, to the junction of the self bias network and the extra resistor in the cathode circuit that is going to ground.
There was an excellent thread using a 6550 or KT88 that used multiple feedback loops.
One of those loops was output tube plate to driver tube cathode.
I think that "Mona" was one of several contributors to that thread.
Note:
1. If you use a cap in series with the negative feedback resistor, the gain goes back up at low frequencies, due to less negative feedback.
2. If you use only a resistor, the DC current from the output tube plate to the driver tube cathode will raise the bias voltage of the driver tube, unless you reduce the resistance of the original self bias network of resistor and bypass cap.
I hope that explains it.
You probably are familiar with Schade feedback, from the output tube plate to the driver tube plate.
Yes?
If you have self bias in the driver tube cathode that has a bypass cap across the self bias resistor . . .
Then disconnect that bias network from ground; and insert a resistor in series from the bias network to ground.
Then connect a negative feedback resistor from the ouput tube plate, and a DC blocking cap in series, to the junction of the self bias network and the extra resistor in the cathode circuit that is going to ground.
There was an excellent thread using a 6550 or KT88 that used multiple feedback loops.
One of those loops was output tube plate to driver tube cathode.
I think that "Mona" was one of several contributors to that thread.
Note:
1. If you use a cap in series with the negative feedback resistor, the gain goes back up at low frequencies, due to less negative feedback.
2. If you use only a resistor, the DC current from the output tube plate to the driver tube cathode will raise the bias voltage of the driver tube, unless you reduce the resistance of the original self bias network of resistor and bypass cap.
I hope that explains it.
Last edited:
One thing about most of those types of feedback, is that they are local, not global.
Easy for me and my limited experience and math power to make it work, and easy to make it stable.
Nothing on the edge of oscillations, bad square waves, etc.
Just a personal preference.
I use Beam Power tubes in Ultra Linear.
I use Beam Power tubes Triode wired.
I use Indirectly Heated Triodes.
In the future, I hope to use 45 tubes, or 2A3 tubes, with DC powered filaments (I have the tubes, no time, etc.).
I mostly use these amplifiers near-field, with the speaker no more than 2-1/2 feet from my one ear (the other ear no longer works).
Since the speaker is so near, I do not need much output power.
Easy for me and my limited experience and math power to make it work, and easy to make it stable.
Nothing on the edge of oscillations, bad square waves, etc.
Just a personal preference.
I use Beam Power tubes in Ultra Linear.
I use Beam Power tubes Triode wired.
I use Indirectly Heated Triodes.
In the future, I hope to use 45 tubes, or 2A3 tubes, with DC powered filaments (I have the tubes, no time, etc.).
I mostly use these amplifiers near-field, with the speaker no more than 2-1/2 feet from my one ear (the other ear no longer works).
Since the speaker is so near, I do not need much output power.
Last edited:
baudouin0,
I have to admit, I used to dis-like the idea of Ultra Linear.
Now, I really like Ultra Linear.
I have experimented with 20%, 40%, 50%, and 80% taps.
Of course, for a given tube type, Real Ultra Linear only occurs at a specific Tap %.
But some of the benefits can be partially realized, even if the tap is at a different % than what is optimum.
And, I have also used "Ultra Linear" at the 100% Tap (Triode Wired).
. . . Just another way to think of "Triode Wired".
I have to admit, I used to dis-like the idea of Ultra Linear.
Now, I really like Ultra Linear.
I have experimented with 20%, 40%, 50%, and 80% taps.
Of course, for a given tube type, Real Ultra Linear only occurs at a specific Tap %.
But some of the benefits can be partially realized, even if the tap is at a different % than what is optimum.
And, I have also used "Ultra Linear" at the 100% Tap (Triode Wired).
. . . Just another way to think of "Triode Wired".
Last edited:
The cathode tap only gives a few dB of feedback, if it is only a Common, 4, 16 Ohm secondary.
A separate cathode feedback winding can give more dB of feedback.
But with every dB of feedback, you need a dB more signal from the driver.
That is why so many Unity gain cathode feedback circuits will use positive feedback (there is a resistor from the output tube plate to its driver tube plate).
A separate cathode feedback winding can give more dB of feedback.
But with every dB of feedback, you need a dB more signal from the driver.
That is why so many Unity gain cathode feedback circuits will use positive feedback (there is a resistor from the output tube plate to its driver tube plate).
I've already used a pair of matched LSK170 and a U440. The LSK389 is already on the list ( http://www.linearsystems.com/lsdata/datasheets/201122 - LSK389 Datasheet Rev A14 2018 06 14.pdf ).