Nobody is suggesting adding a 10K series screen resistor that not what the post indicated. I am suggesting adding an additional 10k resistor between screen and HT. That way if the switch breaks before makes the screen does not go briefly down to 0V. Its possible that the sudden disruption in current in the pramary of the OPT generates a large plate voltage due to the leakage inductance causing a breakdown in the KT88 although that's just a theory.
- The LED in the cathode of the driver is a good stabilizer (zener) BUT i tried it, did not retain it. A resistor can cope better with changing plate voltages. But its a design choice . . .
- I'm trying to figure out when there might be a reversal of phase in the feedback. In old days one could make a Bode plot. it should give a hint if that could be a reason for instability (high frequencies like 2-10 MHz do occor with instable feedback loops). I would at least ensure the wire connecting R12 is coax. The feedback loop limits the output to about 10x to 15x I guess.
- Then I think R10 is crazy low at 4700Ω. You should be able to increase it, say 100K-120K without harming the sound. Just my gut feeling. It will reduce the plate voltage hence strain on the plate of the driver. The design voltages are much too high there. This would cause flashing in the driver. There are only a few tubes capable of that high voltage (such as a EL84 driver in triode mode but then that has lower µ). The specs of the best tubes say 450V max in COLD state. ECC88 has max operating Vp=120V; ECC85 has max 300V.
- And I totally agree with baudouin, even though the tube can stand an open screen, changing states in hot mode in a running amplifier, is not wise. But there more design having that. But could it give rise to an abnormal state (my point 2??)
Not a great idea, it is in parallel with the load & in the cct all the time,
It will use some of the audio power. Don't do that! 👎
LED in the cathode is a poor choice, it is a non-linear device & increases stage distortion.
It also prevents the tube from adjusting to aging, that is what resister cathode bias does best by far,
Go back about 13 pages on this forum to July 20th,
you will find a thread that discusses the Blueglow amp in considerable detail.
Many issues are brought up, some might be useful. 🙂
you will find a thread that discusses the Blueglow amp in considerable detail.
Many issues are brought up, some might be useful. 🙂
Good points, all, thanks. I think for the time being I will keep it as triode, and maybe eventually add a terminal block or some other non-switch means of swapping back and forth between UL and triode. I'll leave the 10k resistor between pins 3 and 4 out of the plans for now.
Again, thanks to all for the help, pointers and advice. Eagerly awaiting a new set of tubes so I can get this back on my rack and playing music.
-Mike
Again, thanks to all for the help, pointers and advice. Eagerly awaiting a new set of tubes so I can get this back on my rack and playing music.
-Mike
Yes, 10K is much too high. You're kind of defeating the purpose of the triode configuration. I guess that was suggested as a possible solution to the tube failure, but it's not necessary. 100 ohms is sufficient.
Perhaps there is some confusion.
A Screen Grid Stopper resistor is one thing.
Often found in Beam Power / Pentode mode; Ultra Linear mode; Triode mode.
A Screen Grid Voltage Dropping resistor is another thing.
It is normally used when the tube is in Beam Power / Pentode mode.
Not normally used in Triode mode.
Not normally used in Ultra Linear mode.
Some Beam Power / Pentode mode circuits may use a 10k screen grid voltage dropping resistor. But in that case there often is a bypass capacitor from either the screen grid or screen grid stopper resistor, and the other end of the bypass capacitor connected to either the cathode, or to ground.
Just my opinions
A Screen Grid Stopper resistor is one thing.
Often found in Beam Power / Pentode mode; Ultra Linear mode; Triode mode.
A Screen Grid Voltage Dropping resistor is another thing.
It is normally used when the tube is in Beam Power / Pentode mode.
Not normally used in Triode mode.
Not normally used in Ultra Linear mode.
Some Beam Power / Pentode mode circuits may use a 10k screen grid voltage dropping resistor. But in that case there often is a bypass capacitor from either the screen grid or screen grid stopper resistor, and the other end of the bypass capacitor connected to either the cathode, or to ground.
Just my opinions
My idea is to connect string of power zeners (250-280 volts 5w) from screen to ground.Increase the power of 10k resistor at least in 5w. You can use bypass capacitor at least 10uf. This circuit in screen stabilize the voltage in screen with very good results. This circuit set your amplifier in beam power mode.You can set the tube current , if decrease self bias resistor . I use this circuit to my se tube kt88 amp. The schematic is almost the same Vanode-cathode is about 370v,the cathode resistor is about 200ohm and the tube current is about 94 ua! , the tubes biased very hot .The sound is very good to my taste.Try it ! screen voltage is 270 volts and B=400v
I think you may need to sketch that. I have used 100v zeners from the UL taps to the screens before. This gives more UL but you can have a higher HT and more output power without exceeding screen dissipation.
That also requires a recalc of the proper OPT impedance to get the optimum loading.
G2 voltage controls the entire plate family of curves.
I have a different hint, not in line with common thoughts about the screen.
- now you have a small resistor, a stopper. Good in a stable situation. Maybe not when pulling a switch and changing state.
- I saw the following suggestion: "So, what sensible people do is put a small-signal diode in series with the screen such that electrons can move -only- from the screen to the anode." The reasoning is that the screen in adverse situations will conduct in a state of super-conductance, i.e. in hundreds of mA; the output impedance of the screen can under adverse situation become negative. The diode prevents current to flow from screen to anode connection.
- Full background: search for an article 'Turning Pentodes & Beam Tetrodes into Triodes. Separating the DC conditions from the AC' by Bill. He additionally adds a zener to force the screen to be slightly lower than the anode.
Hundreds of Triode Strapped KT66 Williamson amplifiers did not have a problem with screen to plate current flow.
There were no diodes in the screen to plate connection of those amplifiers.
Use a diode from screen to plate if your brain says you must (but why do that?).
If you really need a diode there so you can not sleep at night, then you are almost certainly operating the output tube(s) beyond their maximum ratings.
More amplifiers and tweeters have been blown up when switching operating modes with the amplifier powered up (diode or no diode).
Turn the amplifier OFF, and wait till B+ discharges before changing modes from any one of these three modes to any other of these three modes . . .
Triode wired, Pentode/Beam Power, or Ultra Linear modes.
Go ahead, switch modes "Live", and your mileage may vary.
Just my opinions.
There were no diodes in the screen to plate connection of those amplifiers.
Use a diode from screen to plate if your brain says you must (but why do that?).
If you really need a diode there so you can not sleep at night, then you are almost certainly operating the output tube(s) beyond their maximum ratings.
More amplifiers and tweeters have been blown up when switching operating modes with the amplifier powered up (diode or no diode).
Turn the amplifier OFF, and wait till B+ discharges before changing modes from any one of these three modes to any other of these three modes . . .
Triode wired, Pentode/Beam Power, or Ultra Linear modes.
Go ahead, switch modes "Live", and your mileage may vary.
Just my opinions.