I'm wondering if it would be more efficient for me to replace several AC/DC adapters (linear, non-switching) with a regulated power supply? Two of the adapters get very warm and are rated 9 watts input and 3.6 watts (12 volts * 300 ma) output. So 40% efficiency doesn't seem that good and I don't need the heat.
Or are there good non-RFI producing switching adapters that you could recommend? How does their efficiency compare?
If by “a regulated power supply” you mean specifically a non-switching one as your second paragraph, then “a regulated power supply” and “an AC/DC adapter (linear, non-switching)” are exactly the same thing.
There are basically only three categories of Things You Plug Into The Wall And Get DC Out:
Unregulated linear power supplies. These contain a transformer, a rectifier, and smoothing capacitors. Their voltage output is significantly higher than rated when less than their rated load is applied, and their output has lots of “ripple” (a bit of 120 Hz AC on top of the DC output).
Regulated linear power supplies. These add a linear regulator circuit to the above, producing a stable, lower, output voltage. The difference in voltage, times the output current, is dissipated as heat; therefore they are less efficient than an unregulated supply, and most efficient when under maximum or minimum load.
“Laboratory” linear power supplies may add a current-limiting feature to this design. This is an addition to the control section but the power-handling components are the same.
Switch-mode power supplies. These produce a stable output voltage (or current, as in LED drivers) and do not dissipate heat, but they can produce RFI due to the high switching frequency. A well-designed switching supply will keep the noise inside the box by way of shielding and output filtering capacitors.
Any of these may come in “wall-wart”, “lump on the cable”, or desktop/benchtop form factors. This is determined more by their power rating (and hence their necessary physical size) than anything about their type of functionality, though a linear power supply will be larger and heavier than a switch-mode power supply of the same capabilities.
The “Class 2” mentioned in your question title is a safety standard: it says that the adapter is guaranteed not to produce too much voltage or current, so that the output will not pose the kinds of safety hazard AC line power does. This means both under normal operation (the maximum output current must not be too high) and possible faults (internal physical protection against the AC input jumping over to the output side).
Historically, an electronic device might use an external unregulated linear supply and place the linear regulator on the main circuit board instead. Modern devices may also have linear regulators or switching converters built in, in order to accept variable voltage or to produce multiple lower voltages for internal use. In this case, you could replace a switch-mode power supply with an unregulated linear one, as long as its open-circuit voltage is not too high for the device's input circuit components. This will reduce RFI but increase waste heat.
It is OK to replace a regulated linear power supply with a switch-mode power supply of the same output voltage and current rating, or vice versa. The tradeoff you are making is efficiency versus RFI.
It is not OK to use an unregulated power supply with an electronic device that does not have an internal regulator, as the varying input voltage may be too high unloaded or too low unloaded, causing damage or incorrect operation. It may be OK to use an unregulated power supply with a merely “electrical” low-voltage device having only lamps, motors, and such, but it should be chosen to match the current rating of the power supply to the actual load of the device.
A good answer already given; I am taking a slightly different approach. It seems to me that you have two concerns, first the efficiency of the power supplies, and second RFI.
So lets take efficiency in terms of $.
Assuming maximum use 24h/day, 1 year. Taking 10.42 cents per kWh (source). And lets take a "loss" of 5.4 Watts from your details; 9 Watt in, 3.6 Watt out.
Lets see what this is in a year:
0.0054 (kW) * 24 (h) * 365 (d) * 0.1042 = $ 4.93
Lets take a theoretical switch mode power supply with an efficiency of 90%, and output of 3.6 Watts, so the "loss" is 0.4 Watts, and calculate the same for a year:
0.0004 (kW) * 24 (h) * 365 (d) * 0.1042 = $ 0.37
As expected it is a big difference in annual cost: about four and a half dollars. However, the more efficient switch mode power supply have to be purchased, and probably/maybe EMI/RFI filtering is needed to clean up interference.
So now the approach of this answer is an economical one:
Taking into account purchase price for the replacement, possible purchase price for EMI/RFI-filters, and the difference of the running cost. What is it worth to you; running a low-RFI high-loss power supply, versus, a possibly-high-RFI low-loss power supply ?
And I purposely omitted "durability" in this answer. In my opinion linear power supplies have a longer life than switch mode power supplies. However I have no data to back up that claim. If that is true then in the economic answer you will have to figure in depreciation of both, and set this off over time/annual.
In my experience over the years, the lower cost switch mode power supplies are prone to failure and high RFI. In regards to RFI: I have had to put filter on both input as well as output of those cheap switch mode power supplies. However I have some more expensive ones, which are OK.
One experience I had was with mobile/cell phone chargers. Two exactly the same branded phones (which shall not be named, but which is in the news lately due to burning phones), with the original power supplies, both PSU exactly the same, same print on the casing, same type number and other indicators/writing... except one showed "made in China" the other showed "made in Taiwan"... The one "made in China" had no noticeable RFI (I did not do measurements, other then SDR-waterfall observations), the PSU which was "made in Taiwan" had RFI all over the place, to a point that when the phone was charging I could not operate on HF. So even when you purchase known-good-brand switch mode power supplies, you can get a bad one. The obligatory YMMV applies here.
