I don't know what specific Yamaha keyboard you have. The first thing which comes to my mind is your factory-supplied two-prong power cable. If the keyboard only allows connection to two contacts, then using a 3-wire power cord is not going to provide any results. The ground lead of the power cable will function only as a very minimal shield to the power cable itself, in this case. I strongly doubt any benefit will be provided by it.
I want to ask you this question.... does the factory-supplied power cord have one prong wider than the other? This wider prong is the "neutral" lead, and I'm sure the keyboard would have a keyed power input connector, expecting neutral to be in the "right" place. It may be internally bypassed to the shielding connection inside the keyboard.
If the "hot" power lead is connected to neutral, then the internal shielding of the keyboard gets connected, in the radio-frequency sense, to the "hot" electrical connection. This is unintended and may cause this kind of a problem,
the connection to shields in the keyboard would likely be made using capacitors of small value which have a very high reactance at the 50 or 60 Hz power line frequency, allowing a quite-small current from the power line, while having a low reactance at the RF. Therefore, the keyboard would "work" ok, except that an unintended return path for spuriously-coupled RF airborne interference is being allowed.
All of what I just told you is based only upon speculation. I have no actual knowledge of the internal wiring of your keyboard. Only knowledge of how similar powering schemes are done.
It's not terribly uncommon to find electrical outlets in buildings and houses which have neutral on the wrong side of the connector, or silliness like shunting the neutral to ground in a local outlet, rather than making it return to the "common point" ground only at the distribution panel. This can play havoc with any signal source utilizing house power, where "passive" signal sources like guitars don't exhibit problems.
Ferrites and powdered iron "chokes" are indeed one way to reduce conducted RF through your audio cable.
The other suspicion I have is of a "ground loop" - where the external shield of your audio cable is able to carry currents other than as the signal return. It may be useful to you to insert a "direct box" in the path between your keyboard and your mixer as an experiment to see if using the "ground lift" of the direct box causes the RFI to stop or decrease.
This seems like a plausible characterization for your problem, as the behavior "matches" this cause.
There is another way to reduce RF conduction to your mixer input. Use bypass capacitors of small values to shunt the center conductor of your cable to the shield at the mixer end. This would be made more effective by the use of a series resistor in the center conductor lead, on the cable side of the capacitor.
