when using a LPSU what else is important, other than ripple/noise?

[the_doc735]

when using a LPSU what else is important, other than ripple/noise?

 

...from what I have read on here, it seems that once you get below 10mV and down into uV & even nV, the ripple starts to become a little irrelevant? Even Sean Jacobs & Teddy Pardo admit that their ripple figures are so low that they can't be measured accurately! (Millivolt = one thousandth of a volt!).

A millivolt (1/1000 volt).

A microvolt (1/1000 millivolt).

A nanovolt (1/1000 microvolt).

 

So, "IF" that is agreed(?)   ...what else really matters? My SMPS can deliver 10mV. (Seasonic Prime Ultra ’80+’ Titanium 650 Watt ATX M-PSU).

 

e.g.   "broadband ultra-low impedance"?    ...etc.?    What else can the LPSU deilver, that the SMPS cannot? When in the audiophile context.

 

Thanks in advance!

[opus101]

Output impedance is important. Step response (to a sudden current demand) may well be important, depending on what its powering. Capacitance to mains is fairly important.

[PeterSt]

1 hour ago, the_doc735 said:

e.g.   "broadband ultra-low impedance"?    ...etc.?    What else can the LPSU deilver, that the SMPS cannot?

 

Nothing much really. So :

 

22 minutes ago, opus101 said:

Output impedance is important. Step response (to a sudden current demand) may well be important, depending on what its powering. Capacitance to mains is fairly important.

 

In all that the SMPS is better. This indeed means that these are the points of attention.

Btw, how the SMPS reacts to step response in combination with output impedance is a downside for the remainder of the system. So now make that LPS with continuous sufficient supply. Like a shunt continuously drawing 20 Amps or so.

 

Maybe not.

[opus101]

1 minute ago, PeterSt said:

 

In all that the SMPS is better.

 

Debatable I'd say - for example step response for the switching supply is going to be a function of its switching frequency. But with a linear regulator there's no switching to worry about so the bandwidth of the regulator can be freely chosen.

 

As regards capacitance to mains, the LPSU contains no common-mode noise aggressor where as the SMPSU does.

[PeterSt]

2 minutes ago, opus101 said:

But with a linear regulator there's no switching to worry about so the bandwidth of the regulator can be freely chosen.

 

IME (but never really measured anything but current draw) it is the supply system as a whole. Think : demand (in ATX) -> capacitor bank -> Regulator (12V) -> capacitor bank -> Rectifying -> Transformer (-> power cord -> mains socket). I'd say only the part between braces is equal for the LPS and the SMPS.

All is connected with wiring and all is subject to slugginess (and capacitance).

 

1. The trick is to have that LPS which is so fast that you can't see it on the current draw (spiking). Only if you try it, you'll know what I mean.

2. Compare this with the same "measurement" of the SMPS and see how it goes all over the place.

 

1) When not arranged for properly that too goes all over the place but more comfortable, so to speak. More smooth. However, it means that it will be too slow because no demand goes that slow (compare with the SMPS to see it happen).

Of course I am talking "during playback" and now also a whole operating system and playback software is involved. But it is like it is.

 

I find designing power supply systems an act of art.

[marce]

The impedance of the whole power delivery system is important with digital designs, that means adequate local decoupling capacitors, next to supply pins. If the power delivery system is not designed correctly it dose not matter what you put on the front end. Digital is all about sharp switching current spikes, so a local low impedance current supply is required, low impdeance SMD small case capacitors next to power pins, larger reservoir caps scattered near by, preferably fed by local LDO regulators and often local SMPS for higher current supplied. Clocks for ultimate accuracy should always have their own local supplies, as near as possible to the device.

[Superdad]

21 hours ago, opus101 said:

Output impedance is important. Step response (to a sudden current demand) may well be important, depending on what its powering.

 

Yup.

 

20 hours ago, marce said:

The impedance of the whole power delivery system is important with digital designs, that means adequate local decoupling capacitors, next to supply pins. If the power delivery system is not designed correctly it dose not matter what you put on the front end. Digital is all about sharp switching current spikes, so a local low impedance current supply is required, low impdeance SMD small case capacitors next to power pins, larger reservoir caps scattered near by, preferably fed by local LDO regulators and often local SMPS for higher current supplied. Clocks for ultimate accuracy should always have their own local supplies, as near as possible to the device.

 

Yup, yup.

 

And a big downside for SMPS is that the AC leakage currents are generally very high across a very wide bandwidth.  Examples (same SMPS, to 1Khz, to 50KHz, and to 1MHz):

[the_doc735]

I read somewhere on here that you want LPSU for one reason at least. The SMPS puts noise backwards into the mains and other devices and corrupts it all! If that is true then it's not just the SMPS's own device that suffers but the rest of the system as well?

I'm presuming this doesn't apply to LPSU's then? (i.e. SMPS infects the whole system not just the device it is connected to, whereas LPSU does not).

[Ultrarunner]

5 hours ago, the_doc735 said:

I read somewhere on here that you want LPSU for one reason at least. The SMPS puts noise backwards into the mains and other devices and corrupts it all! If that is true then it's not just the SMPS's own device that suffers but the rest of the system as well?

I'm presuming this doesn't apply to LPSU's then? (i.e. SMPS infects the whole system not just the device it is connected to, whereas LPSU does not).

Absolutely. Eliminating SMPSs anywhere in the home will improve sound quality. I’ve repaced most of mine. I’ve got a couple of Uptone JS-2s for important applications, and I’m currently looking for a reasonable cost replacement for our phone chargers. My system is on a dedicated line but the damage done by SMPSs on other lines is still noticeable. 

[bobfa]

I would think that almost anything these days has SMPS.  TV,s ??  Microwaves?? Desktop computers. Is that not a daunting problem, let alone expensive?  I do not leave phone chargers, or toasters plugged in, so I am not worried about them.  Oh, and laptop bricks are also SMPS.  

 

With the AC mains to my system being in metal conduit, I would think that a lot of the stray RF from this stuff would be shunted out.  I have not studied it that far.  I do know that the dedicated circuit helped the system.   My turntable and preamp both have SMPS that I really should merely unplug when I am not using them!

 

Where do we stop optimizing?

 

 

[the_doc735]

1 hour ago, bobfa said:

I would think that almost anything these days has SMPS.  TV,s ??  Microwaves?? Desktop computers. Is that not a daunting problem, let alone expensive?  I do not leave phone chargers, or toasters plugged in, so I am not worried about them.  Oh, and laptop bricks are also SMPS.  

 

With the AC mains to my system being in metal conduit, I would think that a lot of the stray RF from this stuff would be shunted out.  I have not studied it that far.  I do know that the dedicated circuit helped the system.   My turntable and preamp both have SMPS that I really should merely unplug when I am not using them!

 

Where do we stop optimizing?

 

 

A dedicated circuit with an isolation transformer right at the start/beginning where the power enters the house (i.e. a tap) bypassing everything including the electric meter and main fuse box? wink wink! (i.e. split the main cable into the property; one for hi-fi and one for everything else).

[Ralf11]

3 hours ago, Ultrarunner said:

Absolutely. Eliminating SMPSs anywhere in the home will improve sound quality. I’ve replaced most of mine. I’ve got a couple of Uptone JS-2s for important applications, and I’m currently looking for a reasonable cost replacement for our phone chargers. My system is on a dedicated line but the damage done by SMPSs on other lines is still noticeable. 

 

not exactly - EM noise attenuates with distance

 

two distances can be involved - [1] distance along the house AC wiring, which will be lower if the noise source is on a different circuit (hence the common advice to check your panel and put the stereo on a different column of breakers)

 

[2] radiated noise - this will go directly thru the air form the source; note also that you cannot blindly apply the inverse square law, as any attached wiring will operate as an antenna so you may be facing a line source (tho I usually expect radiated noise to be lower than xmitted noise for this situation)

 

I note the dedicated line but the methodology used to assess SQ issues is critical to any claims made 

 

emphasis added

[Ralf11]

3 hours ago, bobfa said:

I would think that almost anything these days has SMPS.  TV,s ??  Microwaves?? Desktop computers. Is that not a daunting problem, let alone expensive?  I do not leave phone chargers, or toasters plugged in, so I am not worried about them.  Oh, and laptop bricks are also SMPS.  

 

With the AC mains to my system being in metal conduit, I would think that a lot of the stray RF from this stuff would be shunted out.  I have not studied it that far.  I do know that the dedicated circuit helped the system.   My turntable and preamp both have SMPS that I really should merely unplug when I am not using them!

 

Where do we stop optimizing?

 

 

 

We stop optimizing when SQ benefits are nil (which will be never for non-blind 'testing"), when the benefits of improved SQ are low relative to the effort or cost involved, and when something else (ACOUSTICS) yields far better SQ than the minor monkeying with electronics, esp. stuff not in the signal path.

[Ultrarunner]

57 minutes ago, Ralf11 said:

 

not exactly - EM noise attenuates with distance

 

two distances can be involved - [1] distance along the house AC wiring, which will be lower if the noise source is on a different circuit (hence the common advice to check your panel and put the stereo on a different column of breakers)

 

[2] radiated noise - this will go directly thru the air form the source; note also that you cannot blindly apply the inverse square law, as any attached wiring will operate as an antenna so you may be facing a line source (tho I usually expect radiated noise to be lower than xmitted noise for this situation)

 

I note the dedicated line but the methodology used to assess SQ issues is critical to any claims made 

 

emphasis added

Ok, I’ll modify that to “anywhere in my home”. We live in a two bedroom condo so the distances involved would be shorter than most. The electrical panel is centrally located. The dedicated line helped a lot but by the time I had it installed I was sensitive to the type of noise added by SMPSs.  I’ve confirmed it many times (sometimes wife my wife as listener) using blind testing.

 

Replacing our router’s SMPS was significant. I’ve also got the switch and Roku Ultra on a JS-2. But I agree that it’s impossible to be completely free of SMPSs, and the law of diminishing returns is at work. 

[Ralf11]

Good - don't wear out the WAF

[fas42]

5 hours ago, bobfa said:

 

Where do we stop optimizing?

 

 

 

When you temporarily remove something from the environment, and you can't hear it making a difference, then the audio system is robust enough to not worry. The down side is that the better a rig gets, the easier it is, normally, to hear the impact of everything ... unfortunately .

[marce]

On 10/14/2018 at 11:21 AM, the_doc735 said:

I read somewhere on here that you want LPSU for one reason at least. The SMPS puts noise backwards into the mains and other devices and corrupts it all! If that is true then it's not just the SMPS's own device that suffers but the rest of the system as well?

I'm presuming this doesn't apply to LPSU's then? (i.e. SMPS infects the whole system not just the device it is connected to, whereas LPSU does not).

Far to simple, it all depends...

[marce]

On 10/14/2018 at 5:26 PM, bobfa said:

I would think that almost anything these days has SMPS.  TV,s ??  Microwaves?? Desktop computers. Is that not a daunting problem, let alone expensive?  I do not leave phone chargers, or toasters plugged in, so I am not worried about them.  Oh, and laptop bricks are also SMPS.  

 

With the AC mains to my system being in metal conduit, I would think that a lot of the stray RF from this stuff would be shunted out.  I have not studied it that far.  I do know that the dedicated circuit helped the system.   My turntable and preamp both have SMPS that I really should merely unplug when I am not using them!

 

Where do we stop optimizing?

 

 

All medical kit, the measuring kit people use, the world is full of quiet SMPS's...

[the_doc735]

15 hours ago, marce said:

Far to simple, it all depends...

? please continue.......

[the_doc735]

15 hours ago, marce said:

All medical kit, the measuring kit people use, the world is full of quiet SMPS's...

examples please?

[PeterSt]

2 hours ago, the_doc735 said:

examples please?

 

[Ralf11]

15 minutes ago, PeterSt said:

 

 

⛑️

 

 

"kit" is Brit-speak for stuff

[Sonicularity]

I'm sitting here listening to music on my RME ADI-2 DAC that uses an SMPS.  What noise?

[Ralf11]

14 minutes ago, Sonicularity said:

...  What noise?

 

There is always noise.  It's like the 5th law of thermionicdynamics or sumthin...

 

 

my fav is Johnson noise

[PeterSt]

6 minutes ago, Ralf11 said:

my fav is Johnson noise

 

 

Alice is the new secretary.