The problem of solving voltage Uh-k has two solutions, raising the potential against the cathode or floated filament voltage, I do not think an argument on this topic is useful, I prefer that someone familiar with the simulation procedure of this scheme to certify if it is correct or not, thanks
Yup. The grid might already draw current at idle with no AC signal, but the tube certainly will run into grid current once the AC signal exceeds 0.17 V, driving the grid to 0.0 V. You'll want to bias the input tube such that the grid always stays nicely negative to stay away from grid current, even at peak AC input.
In this operating point the ECC83 anode current about 2.5mA, which -for me- is "not healthy".
If you see datasheet, the (mu, gm, Rp) curves "ending" at about 1.7mA.
I usually use these tubes around 1mA.
The GE datasheet suggest at 180V anode voltage 100k grid leak resistor (and bypassed cathode resistor).
As many post pointed, this extreme low bias unusable, if you want 100Vpp anode swing.
I recommend to raise cathode resistor over 1k, and use bypass capacitor .... in the power tube cathode too.
Without bypassing, the lower -3dB point crawls up.
BTW, I'm not a fan of 12AX7 as VAS tube, at large swing these tubes distortion more, than another good linearity tubes.
Sample simulation (apart from upper tube Vfk):
If you see datasheet, the (mu, gm, Rp) curves "ending" at about 1.7mA.
I usually use these tubes around 1mA.
The GE datasheet suggest at 180V anode voltage 100k grid leak resistor (and bypassed cathode resistor).
As many post pointed, this extreme low bias unusable, if you want 100Vpp anode swing.
I recommend to raise cathode resistor over 1k, and use bypass capacitor .... in the power tube cathode too.
Without bypassing, the lower -3dB point crawls up.
BTW, I'm not a fan of 12AX7 as VAS tube, at large swing these tubes distortion more, than another good linearity tubes.
Sample simulation (apart from upper tube Vfk):
Popa, the cathode-heater voltage under full signal and bw cannot be simulated without knowing the actual heatercircuit to ground capacitance.
Let me explain, at dc the "floating" heater is coupled to the cathode trough the cathode heater leakage resistance resisistance.
This resistance can be from as low as 4M to just about anything.
The tube datasheet recommends a Rkf of max 20k.
Despite of that, many cathode followers with a grounded heater winding use cathode resistors of 100k and more without problems.
Therefore i thinck it is safe to assume that resistor values higher than 20k are ok, but i do not thinck that it would be smart to leave the heater circuit just floating without any cathode heater bypass resistor at all.
Looking at it, what happens when there is a high signal voltage at low frequency, things arent much different from dc, the heater circuit can follow the voltage swing of the cathode and Ufk voltage limit is hardly an issue.
But at the high frequency end of the audio band you have to consider the capacitances.
The often unknown, but smallish, cathode-heater capacitance forms a capacitive voltage divider with the serial connected, many times larger, transformer heaterwinding capacitance. One could shunt this capacitance with a cap that would provide a safe voltage division but this would still leave the notoriously unstable leakage resistance unattended.
What you should do, is bypassing the highly unstably cathode-heater leakage resistance with a resistor and form a <20kHz low pass filter with the heater circuit to ground capacitance.
To do that, i do the following:
First I short all transformer windings and ground all trnsformer parts except of the heater winding.
Next i measure the capacitance between the heater winding and ground.
Then i divide 8000 by this capacitance to get the max resistance value that assures that the max actually occuring voltage between cathode and heater is cut in half at all frequencies up to 20kHz.
Offcourse this 20kHz lowpass filter is additional load to your driver, but at least it provides safe and stable operation that practically doubles the Ukf voltage limit, in your case to 120Vp up to 20kHz.
Let me explain, at dc the "floating" heater is coupled to the cathode trough the cathode heater leakage resistance resisistance.
This resistance can be from as low as 4M to just about anything.
The tube datasheet recommends a Rkf of max 20k.
Despite of that, many cathode followers with a grounded heater winding use cathode resistors of 100k and more without problems.
Therefore i thinck it is safe to assume that resistor values higher than 20k are ok, but i do not thinck that it would be smart to leave the heater circuit just floating without any cathode heater bypass resistor at all.
Looking at it, what happens when there is a high signal voltage at low frequency, things arent much different from dc, the heater circuit can follow the voltage swing of the cathode and Ufk voltage limit is hardly an issue.
But at the high frequency end of the audio band you have to consider the capacitances.
The often unknown, but smallish, cathode-heater capacitance forms a capacitive voltage divider with the serial connected, many times larger, transformer heaterwinding capacitance. One could shunt this capacitance with a cap that would provide a safe voltage division but this would still leave the notoriously unstable leakage resistance unattended.
What you should do, is bypassing the highly unstably cathode-heater leakage resistance with a resistor and form a <20kHz low pass filter with the heater circuit to ground capacitance.
To do that, i do the following:
First I short all transformer windings and ground all trnsformer parts except of the heater winding.
Next i measure the capacitance between the heater winding and ground.
Then i divide 8000 by this capacitance to get the max resistance value that assures that the max actually occuring voltage between cathode and heater is cut in half at all frequencies up to 20kHz.
Offcourse this 20kHz lowpass filter is additional load to your driver, but at least it provides safe and stable operation that practically doubles the Ukf voltage limit, in your case to 120Vp up to 20kHz.
Last edited:
Thanks @euro21, excellent simulation for the driver, the results almost coincide. I have all the driver and the final stage on the test board, so I can easily make changes, so I changed the cathode resistor of ecc83 with a 1.5 k one, but if it is decoupled with a capacitor, the amplifier gain of the amplifier is too high (350mV/4w out) there is still an error in your diagram, the connection to the grid of PL36 is made from the cathode of E180F and not from the anode of ECC83 (lower THD in the OT secondary) I have chosen the primary impedance of the OT practically and the result was 3.8k for the minimum THD, but the fact that you did the simulation with 3.3k is logical because the program requires a real OT model
star_border
star_border
Yes, I agree with this noted limitation.
There is however a solution if the OP insists on using ECC83. A mosfet follower could be used. A mosfet with very low Crss is advised for ECC83, and the ZVN0545a is a known good choice - its best to run it below the 700mW limit, and protect the gate with a 15-18v zener.
Note: This solution is also very inexpensive, requires no heater current and takes up very little space.
@gorgon 53, your reasoning about the h-k capacitive coupling is correct and logical, the measured resistance for the E180F is greater than 99Mohm and the h-k capacitance <30pF, however, feeding the filament in DC reduces the risk of accidental breakdown and I can say that several amplifiers made by me with this type of driver it works without problems