1. R7 and R8 are different, because there is no high impedance current source in that stage's cathodes (22k cathode resistor causes different gain per side). But just set R7 = R8, and there will be 2nd harmonic in that stage. Unfortunately, that 2nd harmonic will be covered up by the negative feedback which corrects that.
2. Triode wiring the outputs will reduce the output power, but also reduce the amplifier's open loop gain. Then the negative feedback can be removed altogether. Want less 2nd harmonic, leave R7 and R8 unequal as in the original circuit. Want more 2nd harmonic, then set R7 = R8.
3. The 100 Ohm resistors should be connected directly on the tube socket pin 5 of the EL34 tube grids. They do not belong on pin 1 and pin 6 of the 12AU7. Correct that, and the 100 Ohms will be real grid stoppers.
4. Triode mode for the EL34 tube is actually easier on them, than Ultra Linear. Just look at the family of screen current curves for an EL34 in the data sheet. Remember, the EL34 Screen is just fine wire. Beam Power Forming plates are thicker and more massive elements than screen wires.
Triode mode:
A. When the control grid goes far negative, the screen current goes down; the tube screen and plate both come near cutoff.
B. When the control grid goes Less negative, the plate current and screen current increase, but the plate and screen voltage are equal and low voltage. So, the plate draws much more than the screen current.
Ultra Linear mode:
A. When the control grid goes far negative, the screen current goes down; the tube screen and plate come near cutoff. And, relatively, the plate voltage has risen 100%, while the screen voltage has only risen 40%.
B. When the control grid goes Less negative, the plate current and screen current increase, but because the the plate voltage relatively went down by 100% versus the screen voltage only going down by 40%, the screen voltage is much higher than the plate voltage. Now the screen current is very much higher, because its higher voltage draws more relative to the plate which is at a much lower voltage than the screen.
Just my Opinion.
The reason R16a is only 1k Ohms, is that it has to drive the capacitance of the shielded cable.
If you are going to reduce the negative feedback, just change R16 and the C12 that is across it. For example, double R16, and cut C12 in half.
As to my post # 23 and #24, the results you get will largely depend on: the quality and the primary impedance of the output transformer.
2. Triode wiring the outputs will reduce the output power, but also reduce the amplifier's open loop gain. Then the negative feedback can be removed altogether. Want less 2nd harmonic, leave R7 and R8 unequal as in the original circuit. Want more 2nd harmonic, then set R7 = R8.
3. The 100 Ohm resistors should be connected directly on the tube socket pin 5 of the EL34 tube grids. They do not belong on pin 1 and pin 6 of the 12AU7. Correct that, and the 100 Ohms will be real grid stoppers.
4. Triode mode for the EL34 tube is actually easier on them, than Ultra Linear. Just look at the family of screen current curves for an EL34 in the data sheet. Remember, the EL34 Screen is just fine wire. Beam Power Forming plates are thicker and more massive elements than screen wires.
Triode mode:
A. When the control grid goes far negative, the screen current goes down; the tube screen and plate both come near cutoff.
B. When the control grid goes Less negative, the plate current and screen current increase, but the plate and screen voltage are equal and low voltage. So, the plate draws much more than the screen current.
Ultra Linear mode:
A. When the control grid goes far negative, the screen current goes down; the tube screen and plate come near cutoff. And, relatively, the plate voltage has risen 100%, while the screen voltage has only risen 40%.
B. When the control grid goes Less negative, the plate current and screen current increase, but because the the plate voltage relatively went down by 100% versus the screen voltage only going down by 40%, the screen voltage is much higher than the plate voltage. Now the screen current is very much higher, because its higher voltage draws more relative to the plate which is at a much lower voltage than the screen.
Just my Opinion.
The reason R16a is only 1k Ohms, is that it has to drive the capacitance of the shielded cable.
If you are going to reduce the negative feedback, just change R16 and the C12 that is across it. For example, double R16, and cut C12 in half.
As to my post # 23 and #24, the results you get will largely depend on: the quality and the primary impedance of the output transformer.
It already has "minimal feedback".
A quick read of the open-loop gain gives 55*8*0.4 or 179.
R16 R2 limit the gain to 152.
20% excess gain is very little NFB indeed.
Roughly like some classic Fender guitar amps. Damping Factor a little better due to UL connection, but roughly "2", not the >10 we like for High Fidelity.
Snip the 1k at the 8 tap. It will be very slightly louder, very slightly more "colored". Depending on your speaker, the change of DF may be audible (more bass boom) or not. It is the kind of thing a guitar amplifier builder calls "voicing".
G-amps sometimes have a back panel knob to vary excess gain from 4 down to zero, to tweak to taste for a given speaker or venue.
> Why does R7 and R8 have different values? isn't that supposed to be a LTP?
It is driven single-input and uses low-gain tubes. And the tail is not very "long". It is sure to have unbalance unless those resistors are different. This part might even give good balance.
@6A3sUMMER, analog_sa, PRR: Thank you very much for educational replies...
Is there some other way to modify this Push Pull design towards SET sound (harmonic signature which is dominated by even harmonics)?
I'm comparing Primaluna to Almarro 318b, which is pure SET... and it is much more full sounding, and also vocals are more upfront.
One way seems to be resistors R7, R8... is there anything other than this?
Is there some other way to modify this Push Pull design towards SET sound (harmonic signature which is dominated by even harmonics)?
I'm comparing Primaluna to Almarro 318b, which is pure SET... and it is much more full sounding, and also vocals are more upfront.
One way seems to be resistors R7, R8... is there anything other than this?
Last edited:
Does not your amp able to switch between triode & UL mode ? My Dialogue 7 has this function but for my ML speakers I prefer UL mode. The last option if you cannot mode is to try different power tube. My set of Psvanne tii does move vocals forward a little & it's got a fuller sound & better bass but because of my room I acoustics, I prefer the sound of GL KT88.
Cheers
Cheers
> modify this Push Pull design towards SET sound
Short R208. It becomes semi-SE. (If you do this wrong, it will blow-up.)
This too is sometimes done in guitar amps. Put a pot on one of the output tube grids so it works, or half-works, or doesn't work, while maintaining DC current. Gives you something to play with.
Comparing this mish-mosh with a simple 300b SE is very unfair. The small but non-zero NFB actually *raises* distortion and intermodulation, "clouding" complex sounds.
Short R208. It becomes semi-SE. (If you do this wrong, it will blow-up.)
This too is sometimes done in guitar amps. Put a pot on one of the output tube grids so it works, or half-works, or doesn't work, while maintaining DC current. Gives you something to play with.
Comparing this mish-mosh with a simple 300b SE is very unfair. The small but non-zero NFB actually *raises* distortion and intermodulation, "clouding" complex sounds.
PRR,
Shorting R208 will reflect back an extremely low impedance to the bottom 1/2 of the phase splitter. The gain of the top half of the phase splitter will increase, and its distortion will go up.
If you short R208, OK, but then also be sure to open C203 at the grid end.
That will keep the phase splitter distortion relatively constant, and its gain relatively constant.
And, as I said before, put the 100 Ohm resistors where they belong, on the output tubes grids, not the other side of the coupling caps.
Shorting R208 will reflect back an extremely low impedance to the bottom 1/2 of the phase splitter. The gain of the top half of the phase splitter will increase, and its distortion will go up.
If you short R208, OK, but then also be sure to open C203 at the grid end.
That will keep the phase splitter distortion relatively constant, and its gain relatively constant.
And, as I said before, put the 100 Ohm resistors where they belong, on the output tubes grids, not the other side of the coupling caps.
Again, the 1k negative feedback resistor is chosen to drive the long shielded cable that drives the 150k negative feedback resistor.
1k and the shielded cable is a low pass filter. At what frequency? Well, what is the capacitance of the shielded cable?
Also, the series 1k resistor means that the capacitance of the shielded cable does not react with the leakage reactance of the output transformer.
Just my opinion, OK?
1k and the shielded cable is a low pass filter. At what frequency? Well, what is the capacitance of the shielded cable?
Also, the series 1k resistor means that the capacitance of the shielded cable does not react with the leakage reactance of the output transformer.
Just my opinion, OK?
This is something I'm planning to do... exchange many of resistors (R2, R3, R7, R8, R14, R15) for Allen Bradleys, which I'm using also in my DAC and in my opinion sounds much more natural than metal film (which is installed in Primaluna).
kamilrakyta,
Connecting this amplifier's output transformer's secondary signal to the 12AX7 parallel grids will cause the amp to become an oscillator (it is actually positive feedback when applied to those input grids).
Even if you re-connect the output transformer secondary in reverse phase, then although that will now be correct to apply to the 12AX7 grids, it will lower the input impedance at the input connector.
And with that connection, now you can not apply that same signal to the 12AX7 cathode circuit, it will be positive feedback.
The application of a particular phase of 'negative feedback signal' to a grid is the opposite of applying the same negative feed back signal to the cathode of that same stage.
You are thinking of a different schematic where negative feedback that connects to a grid, of a Long Tailed Phase inverter.
The 2 grids of a LTP phase splitter have opposite phase.
I hope that explains it.
Connecting this amplifier's output transformer's secondary signal to the 12AX7 parallel grids will cause the amp to become an oscillator (it is actually positive feedback when applied to those input grids).
Even if you re-connect the output transformer secondary in reverse phase, then although that will now be correct to apply to the 12AX7 grids, it will lower the input impedance at the input connector.
And with that connection, now you can not apply that same signal to the 12AX7 cathode circuit, it will be positive feedback.
The application of a particular phase of 'negative feedback signal' to a grid is the opposite of applying the same negative feed back signal to the cathode of that same stage.
You are thinking of a different schematic where negative feedback that connects to a grid, of a Long Tailed Phase inverter.
The 2 grids of a LTP phase splitter have opposite phase.
I hope that explains it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.