Yes you are correct, although I would think whatever preamp it is probably has an output impedance far lower than 10k. What I meant was the voltage output of the preamp. With your 6n6p biased where it is, your preamp would need to have about 2.8Vrms output, 8v peak to peak. To get the swing you actually need, 3.5vrms, or 10v peak to peak. Most preamps these days are over 2v, but if the stage was looking for 2.8Vrms to begin with, then it's something to be aware of. Usually power amps are set up to achieve full power with something more like 1.5Vrms. What is the preamp? That'll solve that question. And after that, what do you think about some of these circuits that have been shared so far? I see some nice solutions 🙂
Loren
I have not read all comments in detail, so sorry if I suggest something that already have been discussed, but I'm curious to know why you don't bypass the 400ohm cathode resistor of the output tube with a capacitor?
You will have more gain and lower output impedance from your KT88 with a bypass cap of 100-470uF over the 400ohm resistor.
Which also will reduce the necessary amplitude from the driver stage to drive the grid of your KT88.
I assume you have put in the usual 100 ohm screen grid resistor to run your KT88 in triode mode, even if you don't have it in your schematic.
You will have more gain and lower output impedance from your KT88 with a bypass cap of 100-470uF over the 400ohm resistor.
Which also will reduce the necessary amplitude from the driver stage to drive the grid of your KT88.
I assume you have put in the usual 100 ohm screen grid resistor to run your KT88 in triode mode, even if you don't have it in your schematic.
Last edited:
Thank you all for advices.
I replace first filter cap 3.3uF with 22uF and now B+ is around 200V. Swing at Kt88 grid is 120Vpp after which driver clipping occours (sin. tops are cuted=no B+). Interesting is that with 120Vpp=+/- 60V at kt88s grid (biased at -40V) still isn't enough to push output swing to clipping. I only get 18Vpp and OT secundary with 8ohm resistor load (P=5,1W). Kt88 graph shows that output power for that configuration should be 8W (22-23Vpp). Amp sounds wonderfully on low sensitivity Falcon Q7.
Now I must add 1K gate stopper first, but will also try to bypass kt88's Rk with 200uF. In my older project with tubes I found that bypassing Rk increase gain and lower output Z, but also degrade sound quality. Maybe better qualitiy cap is needed, not ordinary electrolytic.
I replace first filter cap 3.3uF with 22uF and now B+ is around 200V. Swing at Kt88 grid is 120Vpp after which driver clipping occours (sin. tops are cuted=no B+). Interesting is that with 120Vpp=+/- 60V at kt88s grid (biased at -40V) still isn't enough to push output swing to clipping. I only get 18Vpp and OT secundary with 8ohm resistor load (P=5,1W). Kt88 graph shows that output power for that configuration should be 8W (22-23Vpp). Amp sounds wonderfully on low sensitivity Falcon Q7.
Now I must add 1K gate stopper first, but will also try to bypass kt88's Rk with 200uF. In my older project with tubes I found that bypassing Rk increase gain and lower output Z, but also degrade sound quality. Maybe better qualitiy cap is needed, not ordinary electrolytic.
Wth reference to the schematic in Post # 1 . . .
A bypass cap is needed across the 400 Ohm KT88 cathode reistor.
1. As flex2 said, the gain will be too low without the bypass cap.
2. Another important reason to have a bypass cap across the 400 Ohm resistor, Rk . . .
The damping factor will suffer if Rk is not bypassed.
KT88 u (mu) = 8 (Triode wired)
KT88 Gm = 11,500 uMhos
Plate impedance, rp, is u / Gm = 700 Ohms (only if the cathode is bypassed)
Damping factor is 3,000 Ohms / 700 Ohms = 4.3 Good for a non-negative fed back design.
But the plate impedance is increased by u x Rk
8 x 400 = 3,200 Ohms.
The new plate impedance is 700 Ohms + 3,200 Ohms = 3,900 Ohms.
The new damping factor is 3,000 / 3,900 = 0.77 very poor.
Connect the bypass cap across Rk, 400 Ohms . . . and get the damping factor back to 4.3, where it should be.
With the bypass cap in place, you get:
More gain
A Decent damping factor
And, did anybody already mention it . . .
More Power Out
Degeneration (un-bypassed cathode) in some stages can be good; but is not always good in an output stage.
Note: The above principles apply to Triode output tubes, Triode wired pentode output tubes, and Triode wired Beam Power output tubes.
A bypass cap is needed across the 400 Ohm KT88 cathode reistor.
1. As flex2 said, the gain will be too low without the bypass cap.
2. Another important reason to have a bypass cap across the 400 Ohm resistor, Rk . . .
The damping factor will suffer if Rk is not bypassed.
KT88 u (mu) = 8 (Triode wired)
KT88 Gm = 11,500 uMhos
Plate impedance, rp, is u / Gm = 700 Ohms (only if the cathode is bypassed)
Damping factor is 3,000 Ohms / 700 Ohms = 4.3 Good for a non-negative fed back design.
But the plate impedance is increased by u x Rk
8 x 400 = 3,200 Ohms.
The new plate impedance is 700 Ohms + 3,200 Ohms = 3,900 Ohms.
The new damping factor is 3,000 / 3,900 = 0.77 very poor.
Connect the bypass cap across Rk, 400 Ohms . . . and get the damping factor back to 4.3, where it should be.
With the bypass cap in place, you get:
More gain
A Decent damping factor
And, did anybody already mention it . . .
More Power Out
Degeneration (un-bypassed cathode) in some stages can be good; but is not always good in an output stage.
Note: The above principles apply to Triode output tubes, Triode wired pentode output tubes, and Triode wired Beam Power output tubes.
Last edited:
In case that I modify biasing with constant voltage -40V to grid and remove Rk, numbers for dumping factor are same as bypassed Rk?
Yes,
Damping factor is the same.
At 20Hz Signal frequency, a 400 Ohm Rk and a 1000 uF cap, look like 10 Ohms to the cathode, just like a 10 Ohm DC current sense resistor.
The point is that that degeneration subtracts from the grid to cathode signal, by allowing the cathode to move with signal voltage.
When using fixed bias, I suggest that Each cathode be returned to ground through 10 Ohm reistors. A DMM can measure the voltage across the 10 Ohm resistor, and the current is the Voltage / 10 Ohms.
Example, 0.5V / 10 = 0.050 Amps (50mA).
Note: That 50mA is the sum total of the plate current and the screen current.
-40V fixed bias will not always give the exact current you want, you need some adjustment of the -40V.
Being able to easily measure voltage, at a very low voltage from ground, is safe.
And, anytime you want, you can measure, then calculate the current of each output tube. Nice to know if all is well, with new tubes, and as the months pass by.
Damping factor is the same.
At 20Hz Signal frequency, a 400 Ohm Rk and a 1000 uF cap, look like 10 Ohms to the cathode, just like a 10 Ohm DC current sense resistor.
The point is that that degeneration subtracts from the grid to cathode signal, by allowing the cathode to move with signal voltage.
When using fixed bias, I suggest that Each cathode be returned to ground through 10 Ohm reistors. A DMM can measure the voltage across the 10 Ohm resistor, and the current is the Voltage / 10 Ohms.
Example, 0.5V / 10 = 0.050 Amps (50mA).
Note: That 50mA is the sum total of the plate current and the screen current.
-40V fixed bias will not always give the exact current you want, you need some adjustment of the -40V.
Being able to easily measure voltage, at a very low voltage from ground, is safe.
And, anytime you want, you can measure, then calculate the current of each output tube. Nice to know if all is well, with new tubes, and as the months pass by.