Originally Posted by Curt Palme
MauneyM, interesting points. Not sure if I agree with all of them..
That's why I asked for comments!
How a modified wave at the AC input stage can REDUCE noise in an SMPS is beyond me. It's all converted to DC regardless.
OK - here goes. The ~100 kHz (an assumption) switching freq is high enough that it should cause little or no impact on audio signals. However, lower-frequency noise on the line could be inductively coupled through the power supply. Yes, the caps should get rid of it, but there IS a source impedance - nothing is perfect in practice. Get rid of voltage transients on the source, and you get rid of high-current spikes in the PS that can induce noise in anything nearby.
What could happen though is that the input capacitors in the AC->DC section might blow if they are rated a bit on the tight side.
Agreed, but we're generally talking about higher-end equipment, so I wouldn't expect this to occur.
You might get a larger DC voltage feeding the SMPS section, you might also have that SMPS run hotter and thus have a shorter lifespan.
I might actually expect the opposite, because with higher voltage comes lower current, given the same power requirement. This depends on the load characteristics, though, so........it's toss-up. [FWIW, I have seen a small SWPS that was designed and tested in a facility with a VERY soft source actually fail to start correctly when attached to a very stiff line. The design had been 'optimized' under a particular set of circumstances that turned out to be too far off the real-world averages. So, yes, input PQ can make a difference to a SWPS.]
The only section of any component that might benefit from cleaner AC power is a straight linear power supply section driving a speaker output stage. I'd say though that due to internal filtering of the capacitors of the linear power supply, very slight sags and surges of the AC input voltage are stabilized by the caps anyways.
It depends on the frequency and the total PS design. Remember that no cap in the real world is ideal - they all have ESR, and the source impedance from the low-voltage side of the transformer/rectifier comes into play. You can't block everything with a linear PS.
Once again, the virtues of something that I would consider a tweak have been waxed poetic about without any scientific evidence.
Agreed. I'm playing devil's advocate here, though, because I have seen some really bizarre real-world behavior solved by proper PQ mitigation. The real question is this: What was it that was wrong with the utility power feed that was causing the reduced performance to begin with?