Substrate support device, cleaning device, device and method for calculating rotation speed of substrate, and machine learning device

A polishing device comprising:

• a plurality of rollers that holds an outer edge of a substrate;
• a rotation drive unit that rotates the substrate by rotationally driving the plurality of rollers;
• a cleaning member that comes into direct contact with the substrate and cleans the substrate;
• a cleaning liquid supply nozzle that supplies cleaning liquid onto the substrate;
• a housing that houses the plurality of rollers, the cleaning member, and the cleaning liquid supply nozzle;
• a vibration transmission mechanism that is installed in such a manner as to extend from any of the rollers or the rotation drive unit up to the housing and transmits vibrations to the housing, the vibrations occurring due to a notch or an orientation flat on the outer edge of the substrate hitting the roller;
• a detection sensor that is arranged outside the housing, detects at least one of sound, vibration, and strain occurring from the housing, and outputs a signal corresponding thereto;
• and
• a rotation speed calculation section that calculates a rotation speed of the substrate, based on the signal outputted from the detection sensor.

A device that, in a substrate support device, calculates a rotation speed of a substrate, the substrate support device including:

• a plurality of rollers that is arranged inside a housing and holds an outer edge of the substrate;
• and a rotation drive unit that rotates the substrate by rotationally driving the plurality of rollers, the device comprising:
• a vibration transmission mechanism that is installed in such a manner as to extend from any of the rollers or the rotation drive unit up to the housing and transmits vibrations to the housing, the vibrations occurring due to a notch or an orientation flat on the outer edge of the substrate hitting the roller;
• a detection sensor that is arranged outside the housing, detects at least one of sound, vibration, and strain occurring from the housing, and outputs a signal corresponding thereto;
• and
• a rotation speed calculation section that calculates the rotation speed of the substrate, based on the signal outputted from the detection sensor.

A method for calculating a rotation speed of a substrate for a substrate support device, the substrate support device including:

• a plurality of rollers that is arranged inside a housing and holds an outer edge of the substrate;
• and a rotation drive unit that rotates the substrate by rotationally driving the plurality of rollers, the method comprising:
• by a vibration transmission mechanism installed in such a manner as to extend from any of the rollers or the rotation drive unit up to the housing, transmitting vibrations to the housing, the vibrations occurring due to a notch or an orientation flat on the outer edge of the substrate hitting the roller;
• by a detection sensor arranged outside the housing, detecting at least one of sound, vibration, and strain occurring from the housing and outputting a signal corresponding thereto;
• and
• calculating the rotation speed of the substrate, based on the signal outputted from the detection sensor.

A machine learning device comprising:

• a data acquisition section that acquires, as input data, data obtained by a detection sensor based on at least one of sound, vibration, and strain that occur from a housing as a result of vibrations being transmitted to the housing via a vibration transmission mechanism, the vibrations occurring due to a notch or an orientation flat on an outer edge of a substrate hitting a roller when the substrate held at the outer edge thereof by the roller is rotationally driven inside the housing;
• a label acquisition section that acquires label data indicating a degree of rotation abnormality at a time of rotation of the substrate under a substrate rotation condition included in the input data;
• and
• a learning section that generates a trained model by performing supervised learning using the input data acquired by the input data acquisition section and the label data acquired by the label acquisition section.