Having just built a LM3886 chip amp and finding it too clean, I'm looking to add some tube warmness. I don't need any gain in a preamp, but a buffer with a cathode follower would probably be too clean. I need a grounded cathode stage. I whipped this up:
It seems to work on the bench, 2V P-P in, 2V P-P out. However, I have a very janky benchtop SMPS, and there's a lot of noise on the scope so I can't really measure anything. Is this thing a good idea, or an abomination? I would think the amount of feedback would squash a lot of the harmonic distortion. Should I go with a voltage divider at the input and AC couple the cathode follower without the feedback? Thanks, all.
It seems to work on the bench, 2V P-P in, 2V P-P out. However, I have a very janky benchtop SMPS, and there's a lot of noise on the scope so I can't really measure anything. Is this thing a good idea, or an abomination? I would think the amount of feedback would squash a lot of the harmonic distortion. Should I go with a voltage divider at the input and AC couple the cathode follower without the feedback? Thanks, all.
That cathode follower can't really drive the 1k feedback resistor. And there's no series input resistor for the feedback to work with.
Remove that 1k, and add a 20k audio taper volume control at the output and set it for overall unity gain (roughly center of rotation).
Remove that 1k, and add a 20k audio taper volume control at the output and set it for overall unity gain (roughly center of rotation).
Oh, I had a 22k resistor to the ground at the output. I figured the load would be 22k||1.001M as the grid resistor would be part of the load?
The low impedance source shunts the 1M, so the cathode follower is loaded by a total of ( 1k + Rsource ).
Even if you add a series input resistor to form a proper feedback network, the 1k would still severely load
the CF because a virtual ground is formed at the grid.
Even if you add a series input resistor to form a proper feedback network, the 1k would still severely load
the CF because a virtual ground is formed at the grid.
John Broskie has an article for a "Harmonic Restorer" (https://www.tubecad.com/2013/06/blog0265.htm). He uses a grounded-grid triode followed by a cathode follower. He pushes the grounded-grid triode to high voltage swings to increase the second harmonic and then reduces the excess voltage with a resistor voltage divider.
I would suggest that the problem with a Class AB chip amp is not that it's too clean, but rather that it has significant (audible) low level ("crossover") distortion, which is very audible, but seldom measured. This can't be fixed by adding high level distortion. Maybe for a guitar pedal, but not for music reproduction. There are no "good distortions".
All good fortune,
Chris
All good fortune,
Chris
Too kinky for me😛, though I have a bunch of 6AM4 and 6AN4 tubes that were made for grounded grid operation. Gain's too high though.
Looks like I was trying to reinvent the wheel.
So volume attenuation at the input defeats the purpose as the smaller signal at the grid induces a smaller voltage swing, and thus, less distortion. Thanks everyone for your help!
1.6% THD, almost all second order. Super quiet, no buzz or hum.
