I'm sorry c-J, you seem to have misunderstood my post. I place all these items under the same heading because they all have to do with main system power system integrity. Perhaps I can be more clear.
The inclusion of surge protection can have a direct impact on long term amplifier sound quality. Everyone knows that large voltage surges can wreak havoc; destroying equipment, burning out components, destroying tubes, etc. But far more insidious are the numerous small surges that never appear to do any damage but which can occur, in some systems, dozens of times a day. These repeated small surges have a cumulative effect of breaking down the insulation (and hence coil integrity) in mains transformers. This can lead to leakage, micro-arcing, and capacitive/inductive effects which actually generate high frequency noise in the transformer. The only way to combat this is appropriate surge protection. The fact that a transformer may have successfully passed a hi-pot test at the time of manufacture does not mean the repeated lower voltage surges cannot cause breakdown over time. A good rule of thumb is that the MOV in the amp should be no larger than about 150% of the rated primary voltage of the mains transformer. This helps to keep the main transformer "quiet" over it's entire design life.
Bypassing power switches with inductive loads is also noise related. The micro-arcing of unbypassed switches can cause carbon buildup on the switch contacts over time. This carbon buildup introduces a small resistance in the line. This new "carbon resistor" is a noise source. And with repeated switching over time, the noise will become big enough to affect the audio chain. This is a measurable phenomenon. It is why old switches in many unit are "noisy" (i.e slightly jiggling the power switch not only causes noise, but can change the overall noise level in the amp).
Now lets address this comment about fuses (please forgive me for getting off topic for a minute):
no fuse then your power cord must be rated to handle the full rated current of the circuit breaker on the mains.
Question: why the power cord "MUST" be "full rated current of the circuit breaker"???????
These is much confusion concerning the actual purpose of fuses and I want to be perfectly clear: The purpose of a fuse or circuit breaker is to protect wiring not loads.
This is a point which many people do not seem to understand. After almost 25 years of designing equipment for both FAA and US Military certification, I still run into engineers who don't understand this basic concept. The breaker or fuse protects the wiring not the load.
Lets take your example of the main power panel in a typical residence. As you said, the main panel may have a main breaker of 100A to 250A dependent on the size of the service panel. But if you add up all the branch circuits you will generally find that the sum total of branch circuits far exceeds the value of the main breaker (this is accepted practice under US electrical code). So with no faults, the branch circuits could easily over power the main breaker and cause it to trip. So what does the main breaker protect? The answer is that it's purpose is to prevent the wiring between the main panel and the distribution box from being overloaded (i.e. it is protecting the wiring). Now the branch circuits all have smaller values than the main breaker. In the US they are typically 15A or 20A, with larger values reserved for single point of use items (usually large appliances, electric water heaters, furnaces, etc.). For 15A circuits the branch circuit wiring must be at least 14AWG and for 20A circuits 12AWG. This is because smaller wiring would present excessive heating at the breaker values and present a fire risk. But what about the cords we plug into these circuits? These may be 16AWG, 18AWG, or even 20AWG. A fault in the load could easily cause excessive current draw in the lower rated power cord and still not trip the branch circuit breaker. Then our power cord would see excessive heating and itself become a fire risk. How do we protect against this? The answer is that we use a fuse or breaker in the load device that will trip before the current carrying capacity of the power cord is exceeded. Now obviously the cord needs to be rated to supply the current that the load will draw, but the limiting factor on the fuse is the capacity of the power cord, not the size of the load.
When I said "circuit breaker" in my post, the item to which I was referring was the branch breaker, not the main panel breaker. Now with the purpose of fuses clearly understood, we can reassess the comment about the EMI filters. The filter itself is a potential source of fault current on the cord (given a shorting failure mode). So if the fuse is before the filter but after the power cord, the fault will trip the fuse and prevent the power cord from overheating. But if the fuse is not there, then a fault in the EMI filter can cause excessive current to be drawn which may destroy the cord and present a fire hazard. The only way to protect against this eventuality is ensure that the power cord is rated for the full current of the branch
circuit breaker (generally 15A or 20A). In this way, if the fault current is less than the breaker value then there will be no excessive heating in the cord and no fire hazard will exist and if the fault current is larger that the breaker value, then the breaker will trip to protect the circuit and, again, no fire hazard will exist.
As an example, look at the this IEC entry module designed by Qualtek (http://www.alliedelec.com/search/productdetail.aspx?SKU=6894318
). If you look at the data sheet you will see that the fuse in the 880 series is indeed before the EMI filter, just as I explained above. But if you look farther down, you'll see that in the 882 series, the opposite is true. So what gives? The answer is that the 880 series is designed for equipment which will be using generalized power outlets (think desktop computers, lab equipment, etc.). But the 882 series is meant for equipment which is installed in a "configured power environment". The best example of this is a 19" equipment rack. Here all the equipment plugs into the power system supplied by the rack. And in this installation, all the power cords must be rated for the value of the breaker in the rack's power system.
Now, as to your final question about power strips. Yes, in general these contain an MOV but there are two limiting factors. First, their MOVs are generally rated at a very high voltage. These don't protect against the repeated small surges that can cumulatively damage power transformers over their lifetimes. If they were rated lower, then this would affect their reliability and seriously hurt sales. Some of the more expensive ones are better but still not great. The second factor is that in general these power strips contain some type of EMI filter. (One might say good, this is what we want.) However, the MOVs in these strips are generally upstream of the EMI filter which can significantly attenuate the peak levels of the surges. So they really only protect against surges generated up stream of the strip. Anything plugged into the strip which causes a surge (like a big power amp turning on) will affect all the other items installed in the strip. (Some may think that I'm "picking nits" but I just trying to explain my reasoning here.) So whereas I think that in general these power strips are a good thing to use, I do not rely on them when making design decisions.
Sorry for the long... long post, but I wanted to make sure that everyone understands my reasoning for including all the items from my post above.