Mixing fans on a Y-splitter is quite OK as long as they both are same design - that is, both 3-pin OR both 4-pin.
It helps to understand what the mobo automatic control system does, and why. We talk all the time about "fan speed controllers". But the truth is the automatic control system used by a mobo is really a TEMPERATURE control system. (There normally are two - one for the CPU alone, and one for mobo temperature control, more commonly considered the case ventilation fan control.) In each case the control system has a temperature sensor (inside the CPU chip for one system, and somewhere on the mobo for the other), a target temperature depending on exactly what device is being monitored, and some control loop tuning parameters. The control loop constantly compares the actual measured temperature to the target. If measurement goes higher, it increases cooling rate. If temperature decreases, it reduces cooling rate. How? For 3-pin fans it actually just changes the voltage supplied to the fan on Pin #2 of the header. For 4-pin fans it changes the "% On" value of the PWM signal on Pin #4. In either case the result is that the fan speed changes, which is why we speak often of a "speed controller", and that changes the air flow rate and resulting cooling effect. BUT the real truth is the control system does NOT care at all about the voltage or the speed. It only cares about the TEMPERATURE it is trying to keep at its target. It will make whatever change in its control signal (voltage or PWM) is required to make the measured temperature come into line. In fact, although the mobo circuits can receive the fan motor's speed signal and measure that speed for your entertainment, the temperature control system does NOT pay any attention to that and couldn't care less!
So, if you use a Y-splitter to connect two different 3-pin fans in parallel on one SYS_FAN header, it does NOT matter whether they will run at the same speed or not. They will both receive the same DC voltage on Pin #2 and each will run at whatever speed that gives, and the mobo control system will not know or care what that is! All it cares about is whether the resulting air flow delivers the temperature it wants at the sensor.
An important thing about mixing fans is that BOTH of the fans on one splitter MUST be of the same design as the header is capable of controlling. That is, if the header operates in Voltaqe Control Mode, then BOTH fans connected to that header via a splitter should be of the 3-pin type. If the header is operating in PWM Mode, then BOTH of the fans must be of the 4-pin type, AND the splitter you use for this MUST be a 4-pin splitter.
A further general rule for 2 fans on a header via a splitter is that the total current they draw cannot exceed the max capacity of the header circuitry. The header can supply a DC voltage (12 VDC or less for Voltage Control Mode, or 12 VDC fixed for PWM Mode) up to a maximum current load of about 1.0 amp for almost all common header circuits. Now most computer case fans will draw less than 0.25 amps at max speed, so two would draw up to 0.5 amps. BUT at start-up for a couple of seconds they will draw at least twice that, so that establishes the limit of no more than 2 fans on one header.
So, if the control system does not pay any attention to the actual fan speeds, why is that measured? For three reasons:
(a) it's interesting for some users;
(b) some users will actually use that information to make judgements about the performance of their systems; and,
(c) the mobo circuits also check those speed signals for a very different purpose: failure detection. Quite separately from the matter of controlling temperature by manipulating fan speeds, the mobo also makes sure that each fan it monitors DOES send it a speed signal. If that signal fails, the mobo interprets that as a total failure of that fan and sends out a warning to the user of this failure. Some more sophisticated circuits alarm more than outright failure - they will alarm a speed that is below a certain minimum threshold. In certain cases (especially for the CPU cooling system) the BIOS may be programed in such a failure situation not to wait for the user to react, or even for the CPU measured internal temperature to rise rapidly. It may simply issue the warning and, after a short wait, shut the entire system down to prevent disastrous overheating. But all this failure detection system is separate from system temperature control via fan speeds.
Third-party "Fan Controller" modules are similar but much simpler. Very few of them do automatic control by comparison to actual temperatures measured by a sensor. None of them have any access to the temperature sensor built into the CPU chip, so they cannot use that info in any way. Most give the user the tools to set a voltaqe being fed to the fans (a few operate in PWM Mode and hence set the PWM signal instead). Some actually can display the speed of the fan if you use a 3-pin or 4-pin fan with it that has a speed signal to send back on Pin #3. In those cases you can tell what speed the fan is running. But then the missing link is: is that good enough, or too much, or too little? The USER is the brains of this type of control system. The USER has the power and responsibility to make the judgement about adequate cooling and change fan settings accordingly, and to keep on monitoring those factors as workload changes.