What is a switching power supply for amps and

what is the advantage and disadvantage of using this vs. more conventional power supply, thanks
A switched power supply generates high frequency noise - lots of it. That's the major disadvantage. The advantage over a linear supply is that the overall size is much smaller, requires only a 5 volt starting power at the switch (instead of a line-voltage switch), rectifies the power down to low voltages and can provide much smoother dc.

A switched supply is mainly used for computers (the 5 volt push button start switch as opposed to the old 120 volt red switch), digital gear with standby mode and other similar electronics. The reason they generate noise is that they increase the mains frequency via diodes and transistors from 60 Hz to 20 kHz which enables the use of a smaller transformer and filtering capacitors. Because of the RF noise, I wouldn't want one of these in my amp or preamp - but I wouldn't rule it out because of it. As always, it's the sound not the specs.
Chord uses them in their amps.
NAD championed this approach way back in the late 70's and on. They use a power supply with two rails(one low voltage and one of higher voltage)off of a multi-tapped transformer. The first power supply carries continuous capabilities and the higher rail supplies voltage for instant peaks. You can make a smaller amp sound more powerful than its rating would initially indicate. The problem is in the switching. If it is engineered correctly, it works well even though the switch between supplies can be seen on a "O" scope. Since the switch is done at a higher output(usually the maximum continuous output point), the switching distortions are below an audible level. The drawback is the high rail only holds up for a brief period of time(as in miloseconds) Then the amp falls back on the low rail reducing the maximum output substantially. It does make a smallish amp sound fairly strong since a lot of output is only required during brief peaks. This allows the manufacturer to economically build an amp that has high peak capabilities without the added expense of a continuous power delivery.
For an example, the NAD 304 from the early nineties had a continuous rating of about 35 watts into 8 ohms 20-20k. It had a peak wattage output of 120 watts per channel at the same spec. This peak wattage was useful because the amp would hold it for almost 100 msec.
The amp sounded alot more powerful than its rating would suggest. Sterophile actually gave it a "Class C" rating and it retailed for $329. Quite a bargain considering it had excellent sonics and would drive most speakers fairly well.
Gs: The frequency that the power supply switches at is typically well above the audible range in an audio component. Most will be at 80 KHz or above from what i've seen. Some are quite a bit higher than that and / or are quite "dirty", producing strong harmonic output. This can result in not only a non-linear load being placed on the power line, but RFI being generated and fed back into the AC system.

Bigtee: What you described i.e. switching between two rails of different voltage potential was designed by Hitachi in the late 1970's and is called "Class H". Soundcraftsman used something similar and called it Class G. Don't ask me about the differences as i don't know what they are.

Kkm: Think of a "sine wave" aka "the hill and valley" waveform. Rather than having one continual line that swept up and then down and then back up again, draw the same thing using dashed lines rather than one steady flowing line. You now see the difference between a "Class A" sine wave and a switched sine wave. The switched supply actually pulses off and on at a pre-determined rate and adjusts the amount of signal / power as it needs to. As such, it is more efficient since it only supplies as much power as it needs to during the dashed lines and is not turned on during the gaps between those dashed lines. As such, efficiency is increased since less power is used and less power is wasted. Since less power is wasted, thermal losses are reduced and heatsinking can be minimized. Production costs are reduced, so the manufacturer thinks "this is good", etc...

The drawback to this approach is that it can't respond instantaneously to a signal change as it may be turned off ( the gaps between the dashed lines ) at that moment that the signal chance occurs. As such, the device would have to respond on the next cycle that it was pulsed on. In effect, this is kind of like trying to "counter steer" after you've already gone into a skid rather than staying in control of the situation to begin with by monitoring the situation 100% of the time. The end result is a lack of detail ( you've lost all the information that was included where the "gaps" in the dashed sine wave are ), lack of control ( the amp isn't "locked" onto the driver since it is being pulsed off and on all the time ), dynamics suffer ( can't respond in-between pulses, so there is a slight lag ), blackness of background suffers ( there is less of a difference between the amp being off or on ), etc...

On top of that, pulsed or switched supplies tend to produce their own types of distortion due to all of the switching off and on. This can result in increased RFI and other types of distortion. The RFI or pulsing can seep into other components tied into the same supply line, resulting in further degradation of their performance also.

Even with all of the technical disadvantages that they have compared to the "accuracy" and "control" that a Class A or a "richly biased" Class AB amp has, these devices are getting better and better and can sound amazingly good right now. As Gs mentioned, what you hear is more about the way that designs are implimented rather than the technology being used. Sean
All David Berning amps have used high speed switching power supplies since the early 1980s. Many people think that Berning amps are some of the best available.
The Nad amp(s) I refer to used two separate power supplies with separate filter caps and diode bridges. They used a solid state gate that brought the 2cd supply in as needed with its higher voltage. Whatever class it was I don't know but the thing worked. You received more apparent power out of a smaller package and you really didn't hear any audible effects. It wasn't a bad idea considering the sonics and price. Nad really advertised this thing.
Bigtee: Thanks for the additional info. I don't know how Hitachi or Soundcraftsman worked their designs, but i do remember the Hitachi being produced prior to NAD showing up on the market. Sounds like there were more than one company / engineer thinking the same thoughts. I guess we can all see who had the most success out of the three though : ) Sean