USB connectors have 4 connections:
ground
+5 volts DC
Data+
Data-
The +5 VDC connection has 100 milliamps available. 100 ma = .1 amp. .1 amp X 5 volts = .5 Watts, AKA 500 milliwatts - 500 mw
some of that 500 mw of available power is used to perform other functions in the USB wireless adapter. leaving less than 500 mw available for making RF power.
The Chinese adapters that claim to produce 3800 milliwatts are producing more than 7 times the power they take in. A remarkable feat worthy of a Nobel prize.
generating RF power is not an energy efficient process. most of the power goes into making things hot, only about 25 percent goes out the antenna.
Wifi cards are all built around a very few IC chips. all manufacturers have access to those very few chips. you will note that no manufacturer claims their PCI wireless adapter generates 4 watts ( 4 times the FCC legal limit ) despite having access to 12 volts and more amperage. the power in excess of the standard 15 dbm has to come from somewhere. this somewhere would be an amplifier
An amplifier that can produce 4 watts at 2.4 ghz would be the size of a brick, and consume 16 watts. it would have vast heat sink fins, and it would be wise to use a fan to cool those fins
USB adapters that claim to put out more than 15 dbm, AKA 30 mw, are peddled by liars
this brings us to antennas. antennas work on principles based on the laws of physics. nobody is clever enough to break the laws of physics. a 5 dbi gain antenna is bigger than a 2 dbi gain antenna, a 15 dbi gain omni antenna is much bigger than either, and
anybody who claims more than 15 dbi gain from an omni antenna has discovered a new law of physics he isn't sharing
directional antennas get their gain by herding RF energy in one direction. these antennas obey the laws of physics:
you have to have a bigger reflector to get more gain. any 30 dbi gain antenna made by anyone must be bigger than any 24 dbi gain antenna.
anyone who claims more than 24 dbi gain from a dish less than 30 inches across is