The rectangular chip is the brains of the operation, unfortunately it has absolutely no markings, so I have no idea what sort of brain it is. The control setup is quite interesting in that it is neither straight PWM nor a buck controller type voltage regulator (note the lack of inductors; no inductor, no buck regulator). In the interest of figuring out what method it does use I hooked my oscilloscope up to the output with a fan hooked up as a load, the following picture is a graph of the voltage delivered to the fan, each line vertically is two volts and each line horizontally is 10 milliseconds.
For starters we see that this is pulsing the voltage at a very low frequency, around 14 Hz or so. Beyond that it looks like the control chip feeds some voltage to a transistor that fills a capacitor and that capacitor in turn feeds voltage and current to a second transistor that controls the fan itself, as the capacitor drains out the second transistor lets less current/voltage through and the fan doesn’t get as much. It’s a rather interesting method and much cheaper to design and build than a buck regulator, while still capable of delivering the full 12 V when set to full, unlike a linear regulator setup.
The only real issue with this setup is that some fans will make a noise every time the voltage pulses. When set to full speed the controller doesn’t pulse the output voltage, so of course the noise goes away entirely. Both medium and low speeds require the voltage to be pulsed and hence can cause fans to make noise. Of the included fans only the rear fan makes noise; the top fan, the front fan and the two lower HDD bay fans make no noise on medium or low speed settings beyond airflow noise.
Other than that noise from the rear fans, all the fans are very quiet on low, and still quite quiet on medium. On high they ramp up and make more noise, but nothing offensive and no mechanical noises. Unfortunately one of the two SMD chips that are paired up for each channel is marked only with A79T, which I cannot find any data on at all. The other chip’s markings decode to a fairly standard NPN transistor with 0.2 A maximum capacity. Because I don’t know what the other chip is exactly, the above description of the fan controller is my best guess, rather than known fact.
The fan controller is rated for a maximum of one Amp per channel, not one Amp per connector! Each channel has at least two connectors, and one of them has three connectors. Be sure to add up the power draw of the fans on each channel to avoid letting the magic smoke out of one or more parts of the fan controller; unlike the controller that comes with the Cooler Master Silent Pro Hybrid 1300w PSU there does not appear to be any over-current-protection on this controller.