RE: Kenpeter
"I hope Smoking-amp is happy to see his old theory proven."
I just read that in a tube book once, which was explaining the rollover
effect from internal Mu variation in the tube geometry factors. Seems reasonable, glad to see you have confirmed it.
RE: Elvee
For a P type C-Vac, I don't think you have to get the DC input levels matched up like a real PNP, just the AC phase correct could be useful. So a tube/Fet differential stage at the input could provide that. Current sensing off the plate or drain per triode or "anti-triode" simulation requirement (ie, pick one for the "pilot" device).
The "anti-triode" idea is a scheme where a triode and an "anti-triode" can be combined in P-P stage to get SE sound signature effect. The C-Vac approach can then be used to boost the two for more power or avoidance of the OT.
RE: 45
We all need to clarify what the objectives are. For some, the OT is an important part of the tube sound. For others it is an expense to get a good one. The C-Vac approach (or similar) can be used in more than one way. It could, for example, be used to boost the power of a small DHT to the level of a large power tube, while still using an OT if desired. It is another quite flexible tool in the tool kit, that's the way I see it.
If the OT is going to be left in, another version would be to just put the SS boosters on the secondary side of the OT.
That in turn gets close to another approach, where a complete (quality) SS amplifier has it's output summed (in series) with a small tube amp. The tube amp uses an OT with an unusually low output Z secondary (like say 0.1 Ohm) so that it can drive the same kind of currents as the SS amp (even though much less power is delivered). Feedback for the tube amp is taken from the summation voltage, so that the tube amp lays on a tube "veneer" onto the SS amp. to get what it wants. Ie, the feedback makes sure that the combined total looks like what the tube amp wants.
"I hope Smoking-amp is happy to see his old theory proven."
I just read that in a tube book once, which was explaining the rollover
effect from internal Mu variation in the tube geometry factors. Seems reasonable, glad to see you have confirmed it.
RE: Elvee
For a P type C-Vac, I don't think you have to get the DC input levels matched up like a real PNP, just the AC phase correct could be useful. So a tube/Fet differential stage at the input could provide that. Current sensing off the plate or drain per triode or "anti-triode" simulation requirement (ie, pick one for the "pilot" device).
The "anti-triode" idea is a scheme where a triode and an "anti-triode" can be combined in P-P stage to get SE sound signature effect. The C-Vac approach can then be used to boost the two for more power or avoidance of the OT.
RE: 45
We all need to clarify what the objectives are. For some, the OT is an important part of the tube sound. For others it is an expense to get a good one. The C-Vac approach (or similar) can be used in more than one way. It could, for example, be used to boost the power of a small DHT to the level of a large power tube, while still using an OT if desired. It is another quite flexible tool in the tool kit, that's the way I see it.
If the OT is going to be left in, another version would be to just put the SS boosters on the secondary side of the OT.
That in turn gets close to another approach, where a complete (quality) SS amplifier has it's output summed (in series) with a small tube amp. The tube amp uses an OT with an unusually low output Z secondary (like say 0.1 Ohm) so that it can drive the same kind of currents as the SS amp (even though much less power is delivered). Feedback for the tube amp is taken from the summation voltage, so that the tube amp lays on a tube "veneer" onto the SS amp. to get what it wants. Ie, the feedback makes sure that the combined total looks like what the tube amp wants.
Last edited:
If the objective is: I cannot afford a good traditional valve amp it's ok. You can boost the power from a small DHT if you wish I prefer a 211. Don't even need to buy expensive NOS 211's, Golden Dragon 211 is a perfect GE replica for instance. When I read that this is a solution to avoid OT's because the OT is a problem I disagree. The question is people that usually say OT's are a problem don't know what they are talking about as they have probably just read this argument several times in OTL or SS reviews....
Regarding the low Zout I simply don't care because it is not a problem a priori. Surely not for me. Low Zout amps are just one solution. They are the easy way because this solution has been adopted by industry and most speakers are designed accordingly. However low Zout amps are the industry standard mainly for cost reasons surely not because are the best. This is something that one should take into account before building a valve amp but often is not...
Last edited:
I am not sure if I have to laugh or I have to cry.
I am not a fan they are simply better. Your hot air doesn't apply to music...
I just wrote a message to point out how.... intellectual .... was your comparison.
Last edited:
Don't wonder it is you. One one side you say it's not rocket science on the other you want to argue that OTL's are better....in principle without ANY evidence.
No you don't: they see the maximum Vgs of the MOS at most, and that will normally be <10V. Including a G-S zener of ~10-20V is anyway preferable for gate protection.
I like it a lot!Regarding the other feature Smoking-amp was asking for, might need explanation.
When normal devices operate push-pull, second harmonic distortion is cancelled,
and a smaller set of odd harmonics is created. Some like the cancellation of 2nd.
Some believe the 2nd sounds better than odd junk created by its cancellation.
So, he's asking can you optionally flip distortion profile of one triode multiplier,
such that the push-pull pair will distort indistinguishably from single ended?
Doing so in the context of an output transformer makes perfect sense, as PP
transformers tend to be less expensive per Watt, for those wanting SE sound.
Thus we purposely invert the distortion of one triode (folded cascode it) to an
anti-compliment for the other side. Giving a triode multiplication of 2x.
In context of an OTL totem pole output stage, such contortions are unnecessary.
You only need a single triode (or triode multiplier), let the other half follow it.
Topology that gives what Smoking wants is easier than Smoking suggests.
Take your output from the midpoint of the current set resistor, done...
When normal devices operate push-pull, second harmonic distortion is cancelled,
and a smaller set of odd harmonics is created. Some like the cancellation of 2nd.
Some believe the 2nd sounds better than odd junk created by its cancellation.
So, he's asking can you optionally flip distortion profile of one triode multiplier,
such that the push-pull pair will distort indistinguishably from single ended?
Doing so in the context of an output transformer makes perfect sense, as PP
transformers tend to be less expensive per Watt, for those wanting SE sound.
Thus we purposely invert the distortion of one triode (folded cascode it) to an
anti-compliment for the other side. Giving a triode multiplication of 2x.
In context of an OTL totem pole output stage, such contortions are unnecessary.
You only need a single triode (or triode multiplier), let the other half follow it.
Topology that gives what Smoking wants is easier than Smoking suggests.
Take your output from the midpoint of the current set resistor, done...
Last edited:
RE: Kenpeter
"Take your output from the midpoint of the current set resistor, done... "
I think you will need some inductor below that, or the sense resistor is going to get awfully hot.
Without the inductor though, I think one is back to the "anti-triode" fix to the bottom part of the totem pole stack (to get SE effect from a P-P stack).
"Take your output from the midpoint of the current set resistor, done... "
I think you will need some inductor below that, or the sense resistor is going to get awfully hot.
Without the inductor though, I think one is back to the "anti-triode" fix to the bottom part of the totem pole stack (to get SE effect from a P-P stack).
Inductor below, and sense resistors getting hot? I don't get it.
You must be imagining something else entirely. Certainly not
whatever I was was trying to describe.
I think we might be talking the same, in that last thing you said.
Except sense resistor are in the middle of the stack, not somehow
fixed to the bottom. You have drawn this on napkins before! So I
know you know how it works.
Plenty on the tube side have built SEPP with a split sense resistor
and it works (though the top half of the split sense is oftimes the
top cathode's own non-zero impdance). Elvee built one of my crazy
ideas with sense Schottkys in the middle. And that idea didn't work
out to be stable. But that was my fault, not his. Otherwise topology
is almost exactly the same. You guys have all seen it before.
I'm just saying if you want SE sound, you don't have to do anything
elaborate to bend one of the triode multipliers backwards. Just use
resistively split sense'd SEPP.
You must be imagining something else entirely. Certainly not
whatever I was was trying to describe.
I think we might be talking the same, in that last thing you said.
Except sense resistor are in the middle of the stack, not somehow
fixed to the bottom. You have drawn this on napkins before! So I
know you know how it works.
Plenty on the tube side have built SEPP with a split sense resistor
and it works (though the top half of the split sense is oftimes the
top cathode's own non-zero impdance). Elvee built one of my crazy
ideas with sense Schottkys in the middle. And that idea didn't work
out to be stable. But that was my fault, not his. Otherwise topology
is almost exactly the same. You guys have all seen it before.
I'm just saying if you want SE sound, you don't have to do anything
elaborate to bend one of the triode multipliers backwards. Just use
resistively split sense'd SEPP.
Last edited:
You want phase splitter function to be separate from anti-compliment function?
Go to the sand side and search SRJLH. Current steering phase splitter works in
parallel with SEPP totem. But in tubeland or CVAC land, I don't see advantage
to not just let SEPP totem handle both functions. Would be pointlessly difficult
to compute an antitriode function independantly of both the phase splitter and
the opposing Triode function. Draw me some scribbles if you think you could
separate those functions with reasonable simplicity and minimal sand.
Go to the sand side and search SRJLH. Current steering phase splitter works in
parallel with SEPP totem. But in tubeland or CVAC land, I don't see advantage
to not just let SEPP totem handle both functions. Would be pointlessly difficult
to compute an antitriode function independantly of both the phase splitter and
the opposing Triode function. Draw me some scribbles if you think you could
separate those functions with reasonable simplicity and minimal sand.
RE: Kenpeter
"Would be pointlessly difficult
to compute an antitriode function independantly of both the phase splitter and the opposing Triode function. Draw me some scribbles if you think you could separate those functions with reasonable simplicity and minimal sand."
-----
What I'm thinking of is to just use a "conventional" triode/Mosfet differential stage, to impliment the usual triode/anti-triode functions. These of course would operate in class A mode as required. Then two C-Vacs would multiply the current from each of these model elements to provide a strong totem pole output stage. The advantage would be if the C-Vacs can operate in class B mode instead, for efficiency. (some current offsets would be biased in, to remove the continuous class A current levels from the "models". ) So the C-Vacs would be using the same input triode for modeling, except one C-Vac would be emulating the "anti-" current part.
(Yeah, a separate input splitter and two identical triode models would be wasteful. So combine them as described above.)
"Would be pointlessly difficult
to compute an antitriode function independantly of both the phase splitter and the opposing Triode function. Draw me some scribbles if you think you could separate those functions with reasonable simplicity and minimal sand."
-----
What I'm thinking of is to just use a "conventional" triode/Mosfet differential stage, to impliment the usual triode/anti-triode functions. These of course would operate in class A mode as required. Then two C-Vacs would multiply the current from each of these model elements to provide a strong totem pole output stage. The advantage would be if the C-Vacs can operate in class B mode instead, for efficiency. (some current offsets would be biased in, to remove the continuous class A current levels from the "models". ) So the C-Vacs would be using the same input triode for modeling, except one C-Vac would be emulating the "anti-" current part.
(Yeah, a separate input splitter and two identical triode models would be wasteful. So combine them as described above.)
You do realize the top half drive signal has to swing the output voltage + drive?
The lower half only had to swing the drive. This sounds to me like you need a
bootstrap as-yet unmentioned in your plan. Yet you are going to use class A
differential pair of triode and mosfet to create some drive signals that provoke
totem CVACs to perform triode and anti-triode functions in class B. Somehow I
just can't picture how this Rube Goldberg contraption would gel. Give me some
specifics. Whip out the crumpled napkin and crayons or whatever, and this time
try to make it legible...
You do realize concertina already performs triode and anti-triode drive functions?
The lower half only had to swing the drive. This sounds to me like you need a
bootstrap as-yet unmentioned in your plan. Yet you are going to use class A
differential pair of triode and mosfet to create some drive signals that provoke
totem CVACs to perform triode and anti-triode functions in class B. Somehow I
just can't picture how this Rube Goldberg contraption would gel. Give me some
specifics. Whip out the crumpled napkin and crayons or whatever, and this time
try to make it legible...
You do realize concertina already performs triode and anti-triode drive functions?
Last edited:
I don't think this requires "Rube-Goldberg" technology. I know I've seen the standard diff. stage driver to a totem pole output stage using pentodes (with an output cap bootstrap back to one of the diff. stage loads) in Broskie's writeups somewhere. (no luck finding that diagram yet.)
So:
Take the original C-Vac stage and remove the input model tube, so that we have just the Q1 base input. Stack two of these up in totem pole configuration (as the original thread's setup).
Then for the input stage we have a differential stage using a triode on one side and a FET on the other. With a CCS tail. Input signal to the triode grid. A voltage reference to the FET gate.
The triode side gets it's load resistor (100 Ohms ) bootstrapped from the output. (a big cap from the output, at center of the C-Vacs, to the top of the 100 Ohm load, and then a 5K from there to B+ )
The FET side of the diff stage gets a 5K Ohm load resistor to B+.
The top C-Vac's input comes from the Triode plate ( a coupling cap over to Q1 base, with no 100 Ohm internal current sampling resistor here, since it is already in the bootstrapped load.)
The bottom C-Vac's input comes from the FET drain (a coupling Cap here too, to the Q1 input). This C-Vac retains its internal 100 Ohm sampling resistor to B- (or gnd, whatever).
That's it.
So:
Take the original C-Vac stage and remove the input model tube, so that we have just the Q1 base input. Stack two of these up in totem pole configuration (as the original thread's setup).
Then for the input stage we have a differential stage using a triode on one side and a FET on the other. With a CCS tail. Input signal to the triode grid. A voltage reference to the FET gate.
The triode side gets it's load resistor (100 Ohms ) bootstrapped from the output. (a big cap from the output, at center of the C-Vacs, to the top of the 100 Ohm load, and then a 5K from there to B+ )
The FET side of the diff stage gets a 5K Ohm load resistor to B+.
The top C-Vac's input comes from the Triode plate ( a coupling cap over to Q1 base, with no 100 Ohm internal current sampling resistor here, since it is already in the bootstrapped load.)
The bottom C-Vac's input comes from the FET drain (a coupling Cap here too, to the Q1 input). This C-Vac retains its internal 100 Ohm sampling resistor to B- (or gnd, whatever).
That's it.
Last edited:
"If the triode sees an inconsistent share of load, its not gonna act
much like a triode. How thats gonna work driving class B?"
---------------------
I see your point, but I think it is not as all important as you think. Lets assume the outputs are also operating in class A, with constant current gain from the class A tube model(s).
Now if we start removing some of the constant (class A) shoot-thru current from the output stage (approaching class aB), the current which actually drives the load remains the same at any given point in a cycle. And the tube still sees the same varying plate voltage. Does the tube even notice?
Since the current gain is varying now though, the load Z appears to be changing for the tube. (With highest loading during crossover) But if the current gain is sufficiently high to begin with, then the dynamic loading on the tube plate can be insignificant compared to its plate resistance at all times.
So we need a high current gain, or equivalently, the output Z from the SS drivers needs to be very low compared to the load Z. And we need to stay in the class A part of the output's class aB op. mode a bit more, so that the tube doesn't pick up the full load during crossover.
Some type of dynamic bias scheme could also be conjured up which would help to maintain the impedance seen by the tube more constant, by increasing the cross-thru current near crossover. But that would be in the optimisation category.
much like a triode. How thats gonna work driving class B?"
---------------------
I see your point, but I think it is not as all important as you think. Lets assume the outputs are also operating in class A, with constant current gain from the class A tube model(s).
Now if we start removing some of the constant (class A) shoot-thru current from the output stage (approaching class aB), the current which actually drives the load remains the same at any given point in a cycle. And the tube still sees the same varying plate voltage. Does the tube even notice?
Since the current gain is varying now though, the load Z appears to be changing for the tube. (With highest loading during crossover) But if the current gain is sufficiently high to begin with, then the dynamic loading on the tube plate can be insignificant compared to its plate resistance at all times.
So we need a high current gain, or equivalently, the output Z from the SS drivers needs to be very low compared to the load Z. And we need to stay in the class A part of the output's class aB op. mode a bit more, so that the tube doesn't pick up the full load during crossover.
Some type of dynamic bias scheme could also be conjured up which would help to maintain the impedance seen by the tube more constant, by increasing the cross-thru current near crossover. But that would be in the optimisation category.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.