There are three principal ways to power an amplifier:
1. a simple unregulated power supply consisting of transformer, rectifiers, and reservoir capacitors;
2. a linear regulated power supply;
3. a switch-mode power supply.
It is immediately obvious that the first and simplest option will be the most cost-effective, but at first glance it seems likely to compromise noise and ripple performance, and possibly interchannel crosstalk. It is therefore worthwhile to examine the pros and cons of each technology in a little more detail.
I am here dealing only with the main supply for the actual power amplifier rails. Many amplifiers now have some form of microcontroller to handle on/off switching by mains relays and other housekeeping functions; this is usually powered by a separate small standby transformer, which remains powered when the amplifier supply is switched off.
The design of this is straightforward – or at least it was until the introduction of new initiatives to limit the amount of standby power that a piece of equipment is allowed to consume. The International Energy Agency is urging a 1 W standby power limit for all energy-using products.
Simple Unregulated Power Supplies
Simple, reliable, and cheap (relatively speaking – the traditional copper and iron mains transformer will probably be the most expensive component in the amplifier).
No possibility of instability or HF interference from switch-mode frequencies.
The amplifier can deliver higher power on transient peaks, which is just what is required.
The power into 4 O will not be twice that into 8 O, because the supply voltage will fall with increased current demand. On the other hand, the amplifier will always deliver the maximum possible power it can.
Significant ripple is present on the DC output and so the PSRR of the amplifier will need careful attention; the problem is, however, not hard (if you read the second part of this chapter) and output hum levels below -100 dBu are easily attainable.
The mains transformer will be relatively heavy and bulky.
Transformer primary tappings must be changed for different countries and mains voltages.
The absence of switch-mode technology does not mean total silence as regards RF emissions. The bridge rectifier will generate bursts of RF at a 100 Hz repetition rate as the diodes turn off. This worsens with increasing current drawn.
Linear Regulated Power Supplies
A regulated supply-rail voltage means that the amplifier can be made to approximate more closely to a perfect voltage source, which would give twice the power into 4 O than it gives into 8 O. This is considered to have marketing advantages in some circles, though it is not clear why you would want to operate an amplifier on the verge of clipping. There are, however, still load-dependent losses in the output stage to consider. More on this later.
A regulated supply-rail voltage to a power amplifier gives absolutely consistent audio power output in the face of mains voltage variation.
Clipping behavior will be cleaner, as the clipped peaks of the output waveform are not modulated by the ripple on the supply rails. Having said that, if your amplifier is clipping regularly you might consider turning it down a bit.
Can be designed so that virtually no ripple is present on the DC output (in other words the ripple is below the white noise the regulator generates) allowing relaxation of amplifier supply-rail rejection requirements. However, you can only afford to be careless with the PSRR of the power amp if the regulators can maintain completely clean supply rails in the face of sudden current demands. If not, there will be interchannel crosstalk unless there is a separate regulator for each channel. This means four for a stereo amplifier, making the overall system very expensive.
The possibility exists of electronic shutdown in the event of an amplifier DC fault, so that an output relay can be dispensed with. However, this adds significant circuitry, and there is no guarantee that a failed output device will not cause a collateral failure in the regulators that leaves the speakers still in jeopardy.
Complex and therefore potentially less reliable. The overall amplifier system is at least twice as complicated. The much higher component count must reduce overall reliability, and getting it working in the first place will take longer and be more difficult. For example, consider the circuit put forward by John Linsley-Hood . To regulate the positive and negative rails for the output stage, this PSU uses 16 transistors and a good number of further parts; a further six transistors are used to regulate the supplies to the small-signal stages. It is without question more complex and more expensive than most power amplifiers.
If the power amplifier fails, due to an output device failure, then the regulator devices will probably also be destroyed, as protecting semiconductors with fuses is a very doubtful business; in fact it is virtually impossible. The old joke about the transistors protecting the fuse is not at all funny to power-amplifier designers, because this is very often precisely what happens. Electronic overload protection for the regulator sections is therefore essential to avert the possibility of a domino-effect failure, and this adds further complications as it will probably need to be some sort of foldback protection characteristic if the regulator transistors are to have a realistic prospect of survival.
Comparatively expensive, requiring at least two more power semiconductors, with associated control circuitry and over-current protection. These power devices in turn need heat-sinks and mounting hardware, checking for shorts in production, etc.
Transformer tappings must still be changed for different mains voltages.
IC voltage regulators are usually ruled out by the voltage and current requirements, so it must be a discrete design, and these are not simple to make bulletproof. Cannot usually be bought in as an OEM item, except at uneconomically high cost.
May show serious HF instability problems, either alone or in combination with the amplifiers powered. The regulator output impedance is likely to rise with frequency, and this can give rise to some really unpleasant sorts of HF instability. Some of my worst amplifier experiences have involved (very) conditional stability in such amplifiers.
The amplifier can no longer deliver higher power on transient peaks.
The overall power dissipation for a given output is considerably increased, due to the minimum voltage drop through the regulator system.
The response to transient current demands is likely to be slow, affecting slewing behavior.
This is initially only going to be of benefit for the big boys again!! While the smaller organisations will have to wait for an out the box equivalent.
The only difference is it has a multi-volt flux capacitor, apologies for the poor attempt to a bad joke!!
Regarding your comment:
"It's also worth noting that secondary (load) current does NOT increase flux density. Only the primary voltage affects flux density for a given design."
It's very refreshing to hear someone with strong links to the audio community expressing this simple but so frequently misunderstood truth. Time and again I wearily see reviewers talking with great approval of (needlessly) huge power transformers with high current ratings, used in order "to avoid saturation".
The problem doesn't seem to be confined to the "hi-fi" community either: one close friend who's a seasoned amplifier designer of some repute was surprised when I put him right on this point at a recent AES convention. The ignorance of transformer basics seems so widespread that I'd actually thought of submitting an AES tutorial paper called "The Misunderstood Transformer" (or something like that). I've held back to-date not least because although I've designed plenty of power transformers (for low-frequency and HF switched-mode use) I've less experience of detailed design for high quality audio, and there are other luminaries such as your good self (and dare I mention Brian Sowter, Per Lundahl...) who have greater knowledge here. Still seems like a good idea though.
Perhaps we could get some T-shirts made bearing the slogan "The Volts Determine The Flux" !! :-)
Although much is made of the "low" external magnetic field of toroids, they must be rigorously constructed to live up to that reputation. First, each winding must COMPLETELY cover the entire core. In most commercial units, the magnetic field emanates from the egress of wire leads, where core coverage has a gap. Second, virtually all power transformers are designed to be as cheap as possible. Less core and less copper are used in designs that operate very near magnetic saturation of the core. But the downside is that, as saturation is approached, radiated magnetic field drastically increases. It's also worth noting that secondary (load) current does NOT increase flux density. Only the primary voltage affects flux density for a given design. Therefore, a design on the verge of saturation at 115 VAC will likely be a problem at 130 VAC. To make any power transformer magnetically-friendly, simply operate it at reduced primary voltage (with reduced secondary voltages, of course). Another problem with toroids, because they don't have even the smallest air gaps in the magnetic path, is inrush current when they're first turned on. In large transformers, this can cause nuisance breaker tripping. I think the advantages of toroids are usually over-stated. -- Bill Whitlock, president & chief engineer, Jensen Transformers, Inc. www.jensen-transformers.com
While I agree that using a iron-core, 50/60 Hz transformer makes a lot of sense for audio amplifier design in the range of 10 -100W of output, I think that customers wish for smaller box, along with new requirements for standby power efficiency will push the use of switching power supplies more and more over time. These switching power supplies don't suffer the poor power factor of a bridge-rectified iron transformer with big filtering capacitor, plus they allow designer to use tricks like variable power rails to save power when the volume control position is low. These can also suppply the control voltage along with the power rail.
The initial investment in designing those may pay off for big manufacturers. This is obviously more difficult for a small shop to design and certify a line-level switching power supply for a power amp so there is a definite opportunity here for companies to design specialised off-the shelf unit for sale to the smaller audio components designers that can't afford to design them.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.