Any Diy PC Power Noise filters ?

[nigelwright7557]

A few notes on spread spectrum SMPS's,.

I use spread spectrum pulses on a flyback SMPS and the transformer plays some great tunes on power down.
It uses 4 different pulse widths depending power demand.

 

[marce]

Yes they can sing.
One advantage of multiphase is you can use cheeper and smaller magnetics, often off the shelf Wurth parts or similar.
Here is a link to another thread that has some relevant info in the links regarding keeping the noise down, filtering etc:
http://www.diyaudio.com/forums/power-supplies/193705-switcher-emc-design.html

 

[Eudyptes]

Other SoTM products that supposedly filter & solve/reduce a PC noise problem is the SoTM USB PCI/PCI-E cards.

SOtM - Ultimate High Performance Audio

SOtM - Ultimate High Performance Audio​

These things are more than twice the price of a decent sound card in the UK and for all I know are just more audiophile bs.

Is there any merit in these products? I'd particularly like to hear marce's and phofman's view.

 

[zibra]

What do you think about this product?:

APEXI Power Filter Card for Computer Audio Enhancement for PCI and PCI E Slots | eBay

 

[marce]

Not a lot. Filtering has to be done in the right place, or the noise is out there. And whilst SMPS docontribute some noise to a digital system, but the system switching also conributes a lot on noise. Also the supplies on PCI will more than likely be routed down to a voltage plane, so sittibng the filter in the middle of a plane wont do an awful lot.
The Sota cards, again may have some effect, but again I think they are taking advantage of the fear of noise, and as there marketing says...I have highlighted the relevant word, could.

For this purpose, tX-USBexp has noise filter circuit which could reduce the noises inside of PC affect to USB audio devices, and also use Ultralow noise voltage circuits and ultra-low-jitter oscillator circuit to reduce the noises causing the product itself.

Basicly if the noise reduction isn't built in to a digital design through proper design strategies then these are only like putting a sticking plaster on a gaping wound (IMO)

 

[JLOP]

Hi Marce

You are of course right, better design is superior to remedies. But how to select properly designed components for Pc based music centers ? No or few data is available on that issue on datasheet ? How to manage it practically ?

BR
Jean-Louis

 

[marce]

A thread here is discussing motherboards:
http://www.diyaudio.com/forums/pc-based/218853-best-motherboard-audio.html
I use logitech squeezboxes over wireless so have full galvonic isolation. PC components generaly will not publish noise figures, not a good marketing move, and also somewhat dependent on other parts of the system.

 

[crosswinter]

A great PCI HiFi

I found a great pci power noise filter "AXF-104 Pro"

An externally hosted image should be here but it was not working when we last tested it.

 

[qusp]

whats great about it? what exactly will a bunch of normal quality polymer electrolytic caps do about smoothing a HF signal? it will at the same time depending on what its connected to, reduce the effectiveness of the regulation, so may actually make noise worse

 

[magnoman]

Regarding spread spectrum SMPS for audio:
First does spread spectrum include phase modulation (as opposed to frequency)?
It would appear to me that given a fair comparison (same number of converters near their sweet spot etc) the total radiated energy is the same, thus its just it's characteristic that can be changed from narrow peak to modulated multiple peaks.
Although possibly measuring better wouldn't any kind of frequency modulation potentially introduce components moving throughout the audio range?
Wouldn't it be better to synchronize switching frequencies with the sampling systems, such as the DAC's ? At least then aliasing shows up at dc and one can even adjust phase to minimize this dc effect?

Thanks
-Antonio

 

[dmills]

The best approach is probably to design the audio card on the assumption that the power rails have all sorts of hash over the range of a few Hz to a few hundred Mhz, and design the card on board regulators and filtering to cope, not that hard.

Ferites and smallish smt caps go a long way at high frequency (Where regulators tend to have poor PSRR), I would be leery of adding large amounts of bulk capacitance to the rails on the PCI bus as that can bite you with regulator stability and also start up sequencing (Plus an add on card full of caps will be electrically too far from the load to help at more then a few Mhz).

Regards, Dan.

 

[magnoman]

I too am a big fan of ferrites but would advise being very careful if using any on the decoupling of an oscillator.
Thanks
-Antonio

 

[dmills]

Why? A little shunt R will kill the Q, and most SMT RFI ferrite becomes appreciably lossy at high frequency anyway by design, I routinely use them in power isolation for low phase noise crystal oscilators and seldom see a problem.

Granted a -2 mix or something would probably not be a great idea for power isolation at, but the more common high mu mixes used explicitly for RFI supression don't tend to have too much in the way of annoying self resonances below the point at which they become resistive.

Regards, Dan.

 

[magnoman]

Why? A little shunt R will kill the Q, and most SMT RFI ferrite becomes appreciably lossy at high frequency anyway by design, I routinely use them in power isolation for low phase noise crystal oscilators and seldom see a problem.

Granted a -2 mix or something would probably not be a great idea for power isolation at, but the more common high mu mixes used explicitly for RFI supression don't tend to have too much in the way of annoying self resonances below the point at which they become resistive.

Regards, Dan.

Dan,

I cautioned about using them with care, as you do.
My motivation was a lesson learned related to a switched turn on of the oscillator power. Oscillators dont start from their final resonant frequency so one does need to ensure there are no lower frequency resonances.

Thanks
-Antonio