The voltage is now manageable - but what are your end objectives for the design ?
- output swing;
- gain;
- input impedance of the driven load
For lower supply voltages, the 6SN7 gets worse still, with a vertical load line.
You could replace all the triodes with a single bottle of E88CC, and still be able to run the anodes at 90V or even less - if it gives enough swing for your application. The gain will rise dramatically, and the distortion will drop. Instead of those FET current sources, unbypassed cathode resistors in the cathode will stabilise the anode currents, and offer a tradeoff of gain with distortion.
- output swing;
- gain;
- input impedance of the driven load
For lower supply voltages, the 6SN7 gets worse still, with a vertical load line.
You could replace all the triodes with a single bottle of E88CC, and still be able to run the anodes at 90V or even less - if it gives enough swing for your application. The gain will rise dramatically, and the distortion will drop. Instead of those FET current sources, unbypassed cathode resistors in the cathode will stabilise the anode currents, and offer a tradeoff of gain with distortion.
I'm definitely starting to see short comings of the 6SN7. I've seen the E88CC from Marcel's ValveDAC (I looked at a 6SN7 version).
Gainwise it needs to go from 447mV to ~±16V for the 6AS7s and to be honest they need a bit more than that. The loadwise AS7s aren't that bad but they need voltage swing with enough current.
Last edited:
Achieving a gain of 50 and a swing of ±20V from a supply of 100-150V is very easy using many different types of TV valves with a gm of 7.5mA/V of more, in shunt cascode.
For example: use an E180F or Russian 6Ж9П-E in triode at Ua=150V. Load resistor 10K, cathode resistor unbypassed. ZTX558 pinned by voltage regulator. Cascode DN2540 CCS fed from 170-190V unregulated. Even an exhausted 6Ж9П-E gives 11mA/V - so overall gain of 55 into 10K. The distortion at 40V p-p will be lower than you might imagine - very low!
For example: use an E180F or Russian 6Ж9П-E in triode at Ua=150V. Load resistor 10K, cathode resistor unbypassed. ZTX558 pinned by voltage regulator. Cascode DN2540 CCS fed from 170-190V unregulated. Even an exhausted 6Ж9П-E gives 11mA/V - so overall gain of 55 into 10K. The distortion at 40V p-p will be lower than you might imagine - very low!
I've gone back to the maths, and starting from first principles on the PNP side of things so I understand why - I needed to understand more about the capacitor, how Rload and how the frequency would change this.
Running E88CCs and ZTX558 in a shunt - I'm seeing 100Vp-p into the grid of the 6AS7 which is enough to create 4Vpeak at 140mA into 32ohms at full volume, so I have some wiggle room for down-tuning.
Running E88CCs and ZTX558 in a shunt - I'm seeing 100Vp-p into the grid of the 6AS7 which is enough to create 4Vpeak at 140mA into 32ohms at full volume, so I have some wiggle room for down-tuning.