Looks like your going to 3d print that, proper job.
You could always revert to a more conventional temp sensor like a ds18b20, theres a high probability that the temp of the motor will be fairly even, its not that the motor is subject to rapidly changing dynamic loads or supply power / direction changes like a servo, its just whizzing a weight around.
Heres a wacky idea, use the yellow rpm feedback wire to tell the 'duino how fast the motor is turning, and use that to calc the vibtation level, and further use that to scale the pot command value, so that turning the pot gradually increases vibration over its rotation, instead of the last bit making the biggest chage.
Glad to be of help.