I believe it works differently. The speed signal from a fan is simply a train of 5 VDC pulses (2 per revolution). When they arrive at the mobo header the system counts the pulses in a fixed time period and calculates the RPM from that. If you send two or more pulse streams from different fans into that counting circuit, it receives twice as many pulses. IF they were all clearly separated, you'd get a reading that is the sum. BUT the two pulse trains are never exactly the same in frequency, so the pulses shift in and out of "sync". This means that some pairs of pulses overlap and get counted as only one, and this phenomenon keeps changing all the time. The result is that the actual reading produced varies widely, from very low to nearly twice as high as "normal". Of course, it you fed MORE than two fans' signals in, the high end of that range would get even weirder. Then what happens is that the speed monitoring part of the software gets some "speeds" that appear to be VERY slow and that triggers warnings about fan failure, possibly resulting in further self-protective actions. I remember when Splitters were first brought to market and the bad ones merely connected all lines together, resulting in exactly this situation - two or more fan speeds fed to a single mobo header. It caused no end of trouble for users who saw wildly changing speeds and occasional system shut-downs.