Linear vs Switched mode PSU

[RothwellAudio]

I've noticed this "PSRR" term being mentioned a lot. What does it mean (how is it applied to a circuit), what are its advantages, and what relevance does it have, in terms of the 'linear vs SMPS' argument?

PSRR stands for power supply rejection ratio. It's a measure of how well a circuit shrugs off power supply variations. However, a single figure can be a bit misleading. A circuit could have a very good PSRR at 50Hz but a much less impressive PSRR at 1kHz or 20kHz.

"The PSU of most power amplifiers is an unregulated linear supply and truly awful but you don't hear of people complaining about that!"

Perhaps because they hadn't thought of it before, thus done anything about it? That might change, however, now that you've mentioned it!

People have been doing things about it for a long time. The obvious solution is to use a regulated power supply but it has to be very big and expensive to compete with a simple unregulated supply.
An alternative approach which I have used myself is to split the "output" part of the power amp from the "input" part. The output needs a beefy supply, so leave it with a simple unregulated supply. The input part doesn't draw anywhere near as much current so it's feasible to use a regulated supply for that.
Some designers say that there's no need for any regulation at all because a well designed power amp (with good PSRR) doesn't need it.

[Marco]

Hi Andrew,

PSRR stands for power supply rejection ratio. It's a measure of how well a circuit shrugs off power supply variations. However, a single figure can be a bit misleading. A circuit could have a very good PSRR at 50Hz but a much less impressive PSRR at 1kHz or 20kHz.

Ah, I see. Thanks for that. I've heard of power supply rejection ratio, but didn't 'clock it' immediately from its abbreviation. Like I said before though, isn't it better/would make more sense, to create less to 'reject' in the first place and give the amplifier an easier time, than the opposite, and it attempting to fight the problem later?

'Source first', and all that!

People have been doing things about it for a long time. The obvious solution is to use a regulated power supply but it has to be very big and expensive to compete with a simple unregulated supply.

Ah, gotcha! However, the "people" I suspect Jez was referring to were consumers, not audio designers. I suspect that the former would've 'complained' more, and perhaps done something about it, if they'd realised it was an issue

For example, now that I know that, I'd focus very heavily on how the PSU has been implemented on any commercially produced amplifier I was considering buying, which is something I wouldn't necessarily have done before.

An alternative approach which I have used myself is to split the "output" part of the power amp from the "input" part. The output needs a beefy supply, so leave it with a simple unregulated supply. The input part doesn't draw anywhere near as much current so it's feasible to use a regulated supply for that.
Some designers say that there's no need for any regulation at all because a well designed power amp (with good PSRR) doesn't need it.

Thanks for the info. All interesting stuff!

Marco.

[RothwellAudio]

Like I said before though, isn't it better/would make more sense, to create less to 'reject' in the first place and give the amplifier an easier time, than the opposite, and it attempting to fight the problem later?
'Source first', and all that!

No, that isn't the way I would go about it. As an example, a simple emitter follower needs to have a bias voltage at its base. The classic textbook circuit shows the bias voltage derived from the power supply via a simple two-resistor potential divider. That's fine, but any modulation of the power supply then gets injected into the circuit's input - not a great idea. BTW, that modulation of the power supply can be caused by the circuit itself rather than any external influences. With a bit more thought it's possible to derive a bias voltage that won't be so easily affected by power supply modulations, so why not do it? To me, trying to create a perfect power supply to compensate for the lack of thought when deriving the bias voltage is "attempting to fight the problem later".

[Marco]

Fair enough, I'm not qualified to argue the technical points. My point however, is very simple: if a problem doesn't exist in the first place, it's a helluva lot easier to 'cure'!

A top-quality regulated linear PSU = the creation of a cleaner noise environment for a hi-fi system than any SMPS, yes? Therefore, why would someone like me, with no lack of funds at his disposal and a 'no compromise' approach to audio, use anything less?

Marco.

[LC1979]

Interesting thread, to the extent that I can follow it!

Can I conclude then that the benefit of a linear supply in audible terms is very likely dependent on the design of the audio equipment it is powering? If so, what kinds of devices are likely to benefit most from a linear or other kind of well designed (ie expensive, presumably?)
PSU?

There is also the issue of noise going back into the mains and having an audible effect. But in a typical household that will have all manner of noise from various SMPSs and other sources of noise, would a single "quiet" supply really make any difference if you couldn't somehow isolate all your hifi from the rest of the mains supply?

[Bonky]

"I reckon the bad rep of SMPS comes from when poor SMPS are used with equipment with poor PSRR. In which cases replacing the poor SMPS with a decent linear regulated PSU will indeed give a big improvement! We need to be comparing apples with apples here!"

So, I have 3 amps, some wall-wart PSs, some items on a linear PS, the usual house stuff (TVs,'fridges etc). How do I know which item is injecting the most 'noise'? How do I know which amp, if any, has the best PSRR? It's no good saying 'use your ears' as I'm 69 and think the sound I have from my 'best' system is perfectly OK, (but am always willing to try sensible improvements -ie those that have some sort of rational explanation). I'll try the AM radio technique later.

Does this seem better:
Has anyone bought or hired one of these:
http://www.emfields-solutions.com/detectors/detekta.asp

Thanks,

Richard

[Starterman]

To be honest I'm not sure why we are even debating this. Just read any of the links on this search:
https://www.google.co.uk/#q=using+sm...e+applications

They all say the same things, but to keep it simple just read the bit under "Choose Linear Regulators with Low-Noise and High PSRR Characteristics" here:

http://www.mouser.co.uk/applications...er-management/

[Marco]

There is also the issue of noise going back into the mains and having an audible effect. But in a typical household that will have all manner of noise from various SMPSs and other sources of noise, would a single "quiet" supply really make any difference if you couldn't somehow isolate all your hifi from the rest of the mains supply?

In many of my relevant experiences, yes, especially on a source component where, after all, the music originates that eventually emits from your speakers. In that respect, you can't fix something upstream that's broken, further down the chain, once it's too late to fix...

But then I'm of the school of thought that anyone who owns the property that they live in and is *truly* serious about extracting maximum performance from their hi-fi system, should be powering it from a dedicated ring main, thus isolating it (as far possible) from the noise created by all of one's household appliances.

Anything less, and you're simply 'playing at it'!

Marco.

[Marco]

So, I have 3 amps, some wall-wart PSs, some items on a linear PS, the usual house stuff (TVs,'fridges etc). How do I know which item is injecting the most 'noise'?

Best to keep things simple, Richard, so assume that any SMPS that isn't supplied with high-quality audio equipment is shit, so keep the components in your system well away from it/them.

A good way of doing that (although not as good as having a dedicated ring main) is only to have the main components for your system plugged into the distribution block powering your system, and the rest of your stuff plugged into separate wall sockets elsewhere (or vice versa depending on your room and set up).

Never, for example, have a SKY box or mobile phone charger plugged into the same distribution block as a phono stage, or anything that's similarly sensitive to interference. That's why many folks use two separate mains blocks: one for 'clean' components, and one for the 'dirty' stuff - and never the two shall meet...

Honestly, if you try that one simple thing (which is essentially free), you should hear a significant improvement in the sound of your system. It's just basic 'good housekeeping'

Marco.

[anthonyTD]

You just cant get away from the fact that due to the way SMPS work, they generate quite a lot of noise, and its noise that is very difficult to control' and deal with in a satisfactory manner, on the other hand, equipment being run from such supplies needent sound bad, but its the affect this equipment has on anything else around it that needs to be taken into account, this is the main reason I will not use any SMPs in my Audio system.
A...