I wondered if Critic could more seriously address the technologies behind various forms of Switched Mode Power Supplies , rather than blaming every ill in a system on that technology lurking everywhere in the house.

According to the literature I found in the Devialet room in Munich, the Devialet produces LESS noise in the mains than a linear supply.

Hans
Of more concern to me is the residual A.C. from case to earth, of equipment using SMPS supplies, due to the obligatory small value, high voltage capacitor(s) between the primary and secondary sides of the SMPS. This can range between 80V AC and 120V AC with various brands of DVD players,or STBs for example.This can result in a nasty little bite, even when the player is not switched on, when you unplug a lead to an earthed preamp or power amp, or plug them back in. It's even worse with a preamp with multiple inputs,when there are several SMPS devices plugged in and you unplug or plug in the interconnects to the power amplifier. The effects are additive, and the "bite" can be quite painful.
If that isn't enough , the S/N performance of the source devices using SMPS is degraded. With my own preamp,
I also switch the earth side of the inputs to help reduce this degradation.It's not unusual with SMPS devices to rub a finger across the case of an SMPS powered device and feel a tingle or vibration.
Regards
Alex

This article may be relevant (though being technically pretty incompetent I'm not sure):

http://www.6moons.com/au...views/bluehorizon/1.html

Yes, I'd love to see something on this subject. It bothers me. SMPSs are far more efficient and so the way of the future. Eventually, the likes of Naim, Krell etc will all have to go down the SMPS route and in order to do so they'll need to develop ones that are immune to the extra noise and which still perform adequately, or even better than their current linear supplies.

Somewhat weirder, I have a friend who applied a linear power supply to his brand new Sinology 212 NAS, and this seemed to improve the general sound substantially. He then applied a linear supply to his gigabit switch and the sound improved again, if only by a small amount.

There is a crop of new Class D amps here in the states; the promised reductions in size and price seem to have disappeared. I am thinking of amps like Rowland, Levinson, etc. Perhaps a certain level of cost and size is necessary for top performance. Possibly this will be true of power supplies as well.

Everything produces noise. Linear power supplies are going to ripple at mains frequencies, yes, and if they have Schottky diodes in them which aren't perfectly "snubbed", then they're going to produce short bursts of RF as well.

I think, probably, that what makes all this quite hard is that we insist on buying bits of hifi that come in separate, separately powered boxes, and connecting them together with all sorts of various bits of electric string, where every input or output is connected, via this unknown string, to a port on another box with unknown characteristics.

Now compare with the all-in-one approach of Devialet (since they were mentioned above). He's got his digital inputs at one end, which he should be able to heavily filter and isolate himself from, and they've got the power amp output terminals at the other end. And in between he has complete control, and it even looks like he has access to a decent CNC machine and (in his low volume) I assume that he can machine out nice RF-tight sections within his case. It makes it a lot, lot, easier....

Martin didn't quite say that - I've just fetched the magazine from the loo :-). He said that Linn are advocating an extra switch in the line, so no X-over cable required, just more lengths of cable. Actually, I lie. He didn't write "extra", he just said "recommend ... switch is placed in the cable close to the player to reshape pulses..." (I've deleted some extra words). So that could imply a second switch, or it could just imply the one switch with a long CAT5 cable out to the NAS.

If you look at Linn's website, they don't say that; but I can believe that info that they've given to Martin verbally could well be different to what they've published - websites often lag current thought in companies by months or years! Mind you, you could interpret what they publish in that way. Here's their "entry level" system, which relies on a multi-port router (such as a BT Home Hub). Their "high-end" system just adds a switch to unload the router, as you'd expect.

I do remember looking at their diagram and thinking that the extra router was superfluous, but in any real-world network you do need at least one switch. I've got two in mine, and I'm not exactly a geek! But I don't buy Linn's reported argument about pulse reshaping, unless it can be explained better from an engineering perspective. From where I'm sat, either the pulses will arrive sufficiently square that they can be decoded - and they will, if the cable is in spec - or they won't, in which case you'll be flooded with retransmissions.

So, personally, I'm with Martin (and William of Occam!) on this. If you don't need an extra switch, then don't add one!

Hmm, my copy's at home...I could swear I read that removing the switch resulted in a 10% improvement...

I have a Linn Radikal motor with outboard S/M supply

but

I run it off a separate spur and it has a custom high dynamic range mains filter, and this does make a difference to sound quality

I gave a good review to the Akurate DS with on board S/M , either because or in spite of S/M

but this unit also befitted from careful mains filtering and also custom RFI filtered interconnects.

Linn do make relatively quiet S/M supplies but the plug top types and the rest are built to a price and are significant broadband noise transmitters both up the mains lead and radiated

The cheaper Class D amplifiers compound their S/M supply noise with their audio modulated output noise.

How do you switch ten amps peak at 500kHz and not make a broadband racket?

Martin Colloms

Few S/M supplies make noise in the audio range

their noise is in the radio range and corrupts the sound of electronics through complex modulation processes.

Martin C