I know that mains noise has an effect. Your thoughts about the amp confusing signal with noise are on the right track I think (perceptual confusion may be conceptually closer to nonlinear mixing than we realize). The noise in an amp will simply be mixed with the signal. There will be a nonlinear element to this mixing, which will cause extra frequencies to appear. This nonlinear element also causes shifts in operating point, though often too small to be technically significant.
I know that if you've optimized your amp for treble clearness and soundfield realism, filtering the mains can have an incredible effect. My amp has flat OLG to say 50KHz. So it deals extremely well with ultrasonic interference. This also means there are no or very few poles within the audio range; this means it is free of integrative effects which open the door for nonlinear mixing of signal with interference. Also it is not still "settling" within the audio range, it settles quickly. This can also be seen as there not being much energy storage in the audio range. Adding a beefy line filter, using an RC snubber on the rectifiers and damping the trafo secondary 500KHz resonance had a big impact on the treble imaging.
I've come to suspect that how EMI interacts with the amp determines much about the treble texture. In a slower amp with the OLG corner within the audio band, it theoretically could have effects down into the midrange or lower. The treble can be perceptually enhanced, but this appears to be always at the expense of clearness and imaging. It can have a big effect on emotional engagement. Manipulating the treble texture this way I think could be a valid way of tailoring the sound. I had a prototype once that made voices incredibly engaging, as though it did something to increase empathy. Once I cleaned up the treble, this was gone. The voices sounded great, but at the expense of everything else.
I've found that if you set compensation marginally, and the amp is fast, this makes treble clearer but can make it harsh as well. This is not the end, especially if you have not tried TMC or shunt compensation.
I've been working on my prototype for about a year, maybe longer. It still has no case, it is still sitting on my bench. I like it better than everything else I have so every time I wake up I turn the variac on, bring it up to voltage, and connect the speakers (haven't built the other channel yet, I've been listening in mono for a year!). I think Bigun would benefit from tinkering more with his prototypes. I have not seen him experiment with compensation much, or play with the circuit topology. I'd say 75% of the improvement in my amp's sound has come from the compensation and adjusting time constants of bootstraps and things. If I were him I would try shunt or TMC compensation, I think it would make a big difference, and TGM5 REALLY lends itself to it, topologically, because of the "CFB" and CFP drivers.
I know that if you've optimized your amp for treble clearness and soundfield realism, filtering the mains can have an incredible effect. My amp has flat OLG to say 50KHz. So it deals extremely well with ultrasonic interference. This also means there are no or very few poles within the audio range; this means it is free of integrative effects which open the door for nonlinear mixing of signal with interference. Also it is not still "settling" within the audio range, it settles quickly. This can also be seen as there not being much energy storage in the audio range. Adding a beefy line filter, using an RC snubber on the rectifiers and damping the trafo secondary 500KHz resonance had a big impact on the treble imaging.
I've come to suspect that how EMI interacts with the amp determines much about the treble texture. In a slower amp with the OLG corner within the audio band, it theoretically could have effects down into the midrange or lower. The treble can be perceptually enhanced, but this appears to be always at the expense of clearness and imaging. It can have a big effect on emotional engagement. Manipulating the treble texture this way I think could be a valid way of tailoring the sound. I had a prototype once that made voices incredibly engaging, as though it did something to increase empathy. Once I cleaned up the treble, this was gone. The voices sounded great, but at the expense of everything else.
I've found that if you set compensation marginally, and the amp is fast, this makes treble clearer but can make it harsh as well. This is not the end, especially if you have not tried TMC or shunt compensation.
I've been working on my prototype for about a year, maybe longer. It still has no case, it is still sitting on my bench. I like it better than everything else I have so every time I wake up I turn the variac on, bring it up to voltage, and connect the speakers (haven't built the other channel yet, I've been listening in mono for a year!). I think Bigun would benefit from tinkering more with his prototypes. I have not seen him experiment with compensation much, or play with the circuit topology. I'd say 75% of the improvement in my amp's sound has come from the compensation and adjusting time constants of bootstraps and things. If I were him I would try shunt or TMC compensation, I think it would make a big difference, and TGM5 REALLY lends itself to it, topologically, because of the "CFB" and CFP drivers.
Member
Joined 2009
Paid Member
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.