It is not so black-and-white as you make it out to be. Electrical and power engineering should be the guide here.
Removing Common-Mode and Differential-Mode Noise
It is difficult to know with certainty, i.e. without actual measurements, exactly how effective the capacitor is at removing common-mode and differential-mode noise.
Modern AC power is a mess and is full of hash. Much of the noise arises from motors (HVAC, air conditioner, dishwasher, washer/dryer, microwave, hair dryers, etc.), switching power supplies (computers, wall warts, appliances, etc.), and digital-logic switching (computer control in equipment). This may be in the home or in any home sharing the same transformer at the pole.
Such noise can stimulate parasitic tanks, typically form by the transformer's LC parasitics, to ring. It may also be amplified by a high-bandwidth amplifier, like a vacuum tube. So it is best removed.
The best filters for CM and DM noise are tuned tanks, combined with torroids for common-mode rejection. But the capacitor's reactance drops with frequency so it generally removes the higher-frequency noise. This is why it makes a tolerable common-mode filter which is why it is generally used.
X and Y Capacitors to Remove Noise
The X and Y capacitors are specifically designed for the type of filtering set forth in (1). I can't find my detailed posting about that and don't want to recreate it. Any of the manufacturer papers will describe this. Here's a summary:
The X-type capacitors are designed for line-to-neutral and fail as shorts, which is intended to blow the fuse or trip the circuit breaker. So no shock hazard. Such capacitors must never be used from line-to-chassis.
The Y-type safety capacitors are designed for line-to-ground, and do not leak. The design is to fail as an open, and the capacitor is not leaky. If a Y capacitor fails it ceases to remove noise, but does not create a safety hazard.
The XY-type is designed for either use and fails as an open.
I refer you to the manufacturer's datasheets for additional information.
Grounding is no Panacea
Grounding the chassis may create serious noise problems.
(1) It may create ground loops when equipment is plugged together using shielded cables.
(2) It connects the chassis to the house wiring which is a giant radio antenna.
(3) If any computerized equipment (audio or video) in the house is dumping digital noise to electrical ground this enters the amplifier and many solid-state or tube amplifiers have the bandwidth to amplify it.
A Surge-Protector or Varistor (MOV) Does NOT Remove Noise, Constitutes a Fire Hazard
A surge-protector does
not remove noise. It typically only contains a varistor as "protection", which it is not;
the MOV varistor is a fire hazard.
I'll post my previous comments on that.
A few words of caution which I learned the hard way: the MOV (Metal Oxide Varistor) may silently fail, offering no protection, and a failed MOV can start a fire and burn down your house. It is not a robust device, in fact it is sacrificial like a fuse, and it silently fails without warning or indication, unlike a fuse.
The underlying problem is inherent in the MOV construction: grains of metal oxide packed together, much like a sugar cube. The grains function as diodes, which is hardly surprising since the first dry diodes were fabricated from conductive metal oxides demonstrating rectification effects at the junction. (Many metal oxides exhibit such behavior, which is why a crystal radio may be constructed using a rusty high-carbon razor blade as a diode to rectify the AM signal.) Each of the oxide diodes has a forward-voltage drop across it, and the aggregate thus has a breakdown voltage to conduct all the way across the device, much like a zener diode.
The MOV has so many grains that the breakdown (conduction point) voltage can be high, typically arranged to be about sixty volts to hundreds of volts. Because the grains are jumbled together in no particular order the device will bi-directionally conduct. (A carbon-composite resistor has similar behavior, which is why it has peculiar effects, like a voltage-dependent coefficient of resistance. Why I always urge replacement with metal film. Bu that's an interesting side discussion which can be found elsewhere on AK, not the main point here.)
Each time an MOV suppresses a surge the oxide grains are heated and are slightly damaged, with the intra-grain junction resistance drops. This creates lower-resistance pathways through the jumble of grains. Which creates more heating, and thus more damage. Thermal runaway eventually occurs. The lower-resistance, and thus higher-current, pathways then merge and the MOV silently
shorts out, i.e. become a dead short across the AC line , typically hot to neutral or hot to ground, and it thus becomes exceedingly hot before the breaker in the socket strip trips. The manufacturer datasheets warn that a reduction in resistance of 10% is considered to be a failure. TEN PERCENT. Who regularly checks all of the MOVs in the house to verify proper function? Is this like a fire-alarm or CO detector which must be checked twice a year?
Some manufacturers increase the inter-grain resistance to make the MOV into a high-value resistor so when it eventually fails it does not burst into flame. This reduces the ability of the MOV to suppress transients.
Some years ago I had an MOV fail—to the best of my knowledge I had no suffered a lightning strike or similar overvoltage condition—and saw the smoke emerging from the strip. I will not use an MOV as it likely voids one's fire insurance. I regard them as dangerous and unsafe, and removed them from my socket strips. I have opened socket strips and seen MOVs discolored from the heat.
This is a well-known problem. Yet the FTC has not banned them. Go figure.
Manufacturers reduce the risk of MOV fires by adding current-limiting fuses, thermal cutouts, and even encase the MOV in cement to limit temperature rise. I believe it was Monster which specifically advertised that it encased all MOVs in a cement block. All of this ignores the fact that the MOV is continually failing and may not offer any protection when a significant over-voltage condition occurs. Such events are, of course, rare, but the point is protection; when the event occurs the MOV will likely not offer protection. The longer the MOV is in service, the less likely it is to properly function.
The replacement protection device, which is far safer, is the solid-state Transient-Voltage-Suppression (TVS) diode also known as the "transil" or by the older name of "thyrector". See:
https://en.wikipedia.org/wiki/Transient-voltage-suppression_diode
As the TVS is a silicon device, functioning much like a zener by conducting at a certain voltage, it does not have the same issue with gradual degradation. The TVS cannot handle much current, but it handles more than the MOV, so the surge it must suppress must be of short duration and within its capabilities. This typically is the case.
Here are a few links I had previously located and posted to demonstrate why the MOV is dangerous and should never be used:
The burning issue of metal oxide varistors and how to stop thermal runaway
https://www.rs-online.com/designspa...ide-varistors-and-how-to-stop-thermal-runaway
After prolonged use they can degrade and the material that makes up the disc can start a leakage current due to internal damage after a surge event, this will create a higher leakage current over time and before you know it we have a thermal runaway event.
This can result in a fire, a quick look online at MOV Fires in a search engine will show the extent to which MOV failure has led to house fires and the finger normally gets pointed at the surge protection socket strip. These socket strips can be over loaded and that is a fire risk but the MOV design if fitted badly can be a cause of fires also.
...Wow, that's alot to take in
But the above raises an important point, MOVs are very popular in appliances and equipment whether that appliance device is fixed or portable.
Looking inside many different appliances I find MOVs soldered or strapped across the main terminals or connections without a fuse or thermal disconnection device of any kind.
All MOVs have the same reaction to over stressed events and those that do not have a protective device associated with them are a fire risk.
SURGE SUPPRESSOR FIRES by Jim Pharr, Fire Marshall, Gaston County, North Carolina
http://www.esdjournal.com/techpapr/Pharr/INVESTIGATING SURGE SUPPRESSOR FIRES.doc
Recent fires involving multiple outlet devices toted as surge suppressors raised attention at the Gaston County Fire Marshal's office primarily when one such fire occurred in a fire station. Investigation of a fire that started behind a desk in an office revealed the ignition source was a surge suppressor. Immediately prior to the fire workers were attempting to connect an emergency generator to the firehouse's electrical system to provide an alternative electrical source. Workers noted problems with the system and later, after the fire, discovered the generator was shipped from the factory with a loose neutral (a condition that causes voltage differences between legs of a 120 volt electrical system). Fire fighters noted fluctuations in their radio and other electronics thus started to disconnect all electronics from the system. In the office area they discovered a small fire burning behind the desk. A portable fire extinguisher was used to suppress the flames and an investigation was initiated.
Within that firehouse, three separate surge suppressors were recovered and examined. Each had failed, the one caught on fire, another suppressor ceased working, while the third continued working but later was found to have failed internally. These findings, coupled with suspicion of suppressor involvement in other fires, prompted in-depth examination of possible reasons.
The latter article (the .doc file) contains photos of failed MOVs and analysis. Impressive.