[Levi]

I think that an adept response would be a cheaper route.

So if I BRIBE a power company employee to install a dedicated transformer to chop power down to 60 vac coming in to my house, I can have balanced 120 vac circuits everywhere in the house?     I wonder what it will take........

[Occam]

TNT,

Thanks for the description of the NuForce ps. I hadn't realized that the Ref9 was a dual stage PFC controlled swicher. Is the Ref8 different other than its rail cap size?

.....Ref8 has about 2500uF of energy storage at 48V, or less than 1/10 (E = 0.5*C*V*V) .
......
So the combination of PFC+SMPS+CMC+DMC+CAPS+Speaker feedback (with control loop gain-bandwidth > 500kHz) assures that some line noise don't get propagated to the speaker.

Are you stating that increasing energy storage implicitly increases filtering efficacy by that same amount? As a first order appoximation, we might make that statement with regards to ripple, but I've never (till now) seen that extended to externally sourced noise. We are not dealing with perfect capacitors and they have inductance, resistance, etc.... and increasing rail capacitance does not reduce line noise by the same amount. Otherwise, we could just keep adding capacitance and be done with it. From my, and others experience, adding rail capacitance is not a substitute for powerconditing. I'm also assuming that the Ref8 also has those CM and differential chokes like the Ref9. I'd think their main purpose is to keep internally generated noise from propogating into the mains for regulatory compliance.

Certainly, any ps assures us that some line noise doen't get propogated to the speaker outputs. [and obviously, for that RFI extending into the FM band, its source isn't line noise but inherent to the specific implementation by NuForce]. But that simply begs the question. The question is whether the Nuforce Ref9 would benefit from additional powerconditioning in a noisey mains environment. I (and others) already know the Ref8 benefits substantially from a BPT conditioner in the Nasty Apple. With all due respect for your calculations (I'm a mathematician by training and vocation), the answer to this question is only answerable empirically, either by listening and/or actual measurements.

[larry99]

increasing rail capacitance does not reduce line noise by the same amount. Otherwise, we could just keep adding capacitance and be done with it. From my, and others experience, adding rail capacitance is not a substitute for powerconditing.

I have wondered about that. At some point I would expect problems since the peak currents through the diode bridge would get very high to pass the required energy to recharge the capacitor banks. More capacitors result in less voltage sag between each charging pulse from the 120Hz (rectified 60Hz) charging source. Since there is less sag there is less time that the supply voltage from the transformer secondary is high enough to forward bias the diodes. Less time but the same energy requirements results in higher peak currents. These current transients can generate all sorts of electro-magnetic emissions that could get into circuits within the device in question or into other equipment on the same supply line.

Has any one ever tried a string of car batteries to run a power amp?

[Occam]

Larry,

Good points. Your reference was specifically to linear supplies so that the point at which the rectifier conducts, and stops conducting, influences the shape of the pulse. And to some extent, the inductive leakage of the associated power transformer might have a strong influence. With a bridge rectifier in a linear supply, we are conducting on some portion of the |sine| and the larger the capacitance, the sharper the transition. I guess I should have added the inevitable 'all things being equal', and to make those things actually equal, one would pad to compensate for the lowering of esr (ignore the esl, multiple inductances in parallel, or assume we're not adding multiple of capacitors, but just substituting a bigger cap with equivalent inductance) so that we could see only the effect of increased capacitance which would be grossly wrt noise, a lowering of pole frequency of the formed filter.
But in the context of a switching supply, assuming some form of PWM, we already know shape of the rectified pulse, prior to any inductive or capacitive massaging its a rectangular shape, and its the spectral content of those sharply defined edges that is so frigg'n problematic in switching supplies.
So I'm digging myself deeper and deeper into this hole of power supply design in which I can't claim any particular expertise. My point is simply that first order appoximations and off the cuff calculations are a poor substitute for actual empirical verification. All I can say is that I've yet to meet a component that didn't benefit from the proper power conditioning.

Has any one ever tried a string of car batteries to run a power amp?

Actually, Gordy has build a stereo UCD 400 poweramp powered by dual 48v rails of gell cell batteries