Yes, I am using the same transformer both to power B + and the AC heaters also for room reasons.
Furthermore, as can be seen from # 67, the psu consists of a voltage doubler + gyrator, so the first could be the cause of this.
I will first try to power the heaters in regulated DC, then I will see...Thanks
In effect RCA aren't short...if I had balanced ones things would have been different maybe as for "mains pickup",
I didn't know it was necessary to put a loop breaker in the amp too, I didn't think about it...I'll try your suggestion and see. Thanks triode_al
Thank you, cogsncogs and rongon! I'm quite surprised that your results are much worse that those for the much simpler circuits before.
Best regards!
I think the higher THD numbers are a consequence of less NFB employed in that circuit, The author states that the circuit yields 40x (32dB) of open loop gain, and is claimed to be unity gain stable. The circuit as shown yields about 8X gain (18dB). Apply more NFB and the THD will come down commensurate with the reduction of gain.
The circuit in post #235 yields about 2X (6dB) gain.
I think the THD is that low because of distortion cancellation between U1 and U2. In real life the THD probably won't be that low, but you know... it could happen. I used an Ayumi-method model made by cogsncogs (thanks!), but I also cross-checked using a model by Adrian Immler, and the THD and gain were very close between the two. One thing I noticed is that the gain into a 10k load doesn't drop that much, but the THD almost quadruples to 0.015% at 1V rms. 12AU7 has rp of about 7k to 10k ohms, so it is not capable of driving a 10k ohm very well. I'd prefer to use something like a 6DJ8 or 6N6P for the cathode follower if this circuit needs to drive 5k or 10k ohm loads.
IMO octal varijant of ECC82 as 6SN7 will drive that impedance much easier , or eventually even 6SL7 .
That might be your opinion, but the actual facts don't agree, Plate resistance of 6SN7 is very close to that of 12AU7. Even in a cathode follower, you want the load to be at least 2X the rp of the driving triode,
Plate resistance of both 12AU7 and 6SN7 is about 8k to 10k ohms (varies depending on operating point chosen).
Plate resistance of 6N6P or 6DJ8 is about 3k ohms (more or less, depending on operating point chosen).
6SL7 would be by far worse than 12AU7. Plate resistance of about 50k ohms, which is far too high.
Last edited:
Not necessarily. It is important to remember that output impedance and drive capability are entirely different things. The drive cababilty depends on how much current the output can drive into a load, and in all class A circuits like this, that depends entirely on the quiescent current.
Cheers
Ian
There's a good passage about this in MerlinB's "Designing High Fidelity Valve Preamps" book. To paraphrase, it basically says that a good rule of thumb is to design your cathode follower like it was a common cathode stage, so the load to be driven should be 5 to 10 times the triode's plate resistance, not the calculated Zout of the cathode follower.
I.e., since a 12AU7 has rp of about 8k ohms, figure it can drive a 40k ohm load pretty well (5x the triode's rp), or even a 20k ohm load (2.5*rp) but it will load down badly into a 10k ohm load (only 1.25x the triode's rp). The CF's gain will drop into a very heavy load as well.
LTspice simulations bear this out. A 12AU7 cathode follower that has something like 0.03% THD for 1V rms into a 50k ohm load will see its THD rise to close to 0.1% into a heavier 10k ohm load, and the gain will be lower too.
A 6SL7 or 12AX7 as a CF will have a Zout of perhaps 2k ohms, which you would think would be low enough, but it will still load down very badly into a 10k ohm load. A 6SL7 has rp of about 55k ohms at Ip = 1mA.
http://www.williamsonic.com/TubeDI/6SL7GT.pdf
I.e., since a 12AU7 has rp of about 8k ohms, figure it can drive a 40k ohm load pretty well (5x the triode's rp), or even a 20k ohm load (2.5*rp) but it will load down badly into a 10k ohm load (only 1.25x the triode's rp). The CF's gain will drop into a very heavy load as well.
LTspice simulations bear this out. A 12AU7 cathode follower that has something like 0.03% THD for 1V rms into a 50k ohm load will see its THD rise to close to 0.1% into a heavier 10k ohm load, and the gain will be lower too.
A 6SL7 or 12AX7 as a CF will have a Zout of perhaps 2k ohms, which you would think would be low enough, but it will still load down very badly into a 10k ohm load. A 6SL7 has rp of about 55k ohms at Ip = 1mA.
http://www.williamsonic.com/TubeDI/6SL7GT.pdf