With the symposium approaching and the prototype functioning as expected during the demo with the profs, we decided to do a second iteration of the GOAT glove. We wanted to use the perfboards we have to make a cleaner, compact, and modular device.
Planning
Before we could even start soldering, we needed to map out all the connections and see how we could improve upon our previous work with the prototype.
Glove Perfboard
To improve modularity, a perfboard for the glove was made. This housed input sockets for every sensor and actuator. The perfboard also had output sockets that would connect the sensors and actuators to a 90-degree perfboard that connects to the main perfboard on the battery mount. There is an output socket that groups all the flex sensors and another one that groups all the haptic motors. There are also input and output sockets for the Peltier module and thumbstick.
Critical Common Connections
One improvement we wanted to make was to reduce the number of excess connections we had on the prototype so that we could make room for more sensors and actuators. We also wanted to make the device modular, so we found a 16-pin 90-degree header and socket that is removable and could be used to house all the connections to the sensors and actuators on the main board. The 90-degree header makes it so that the plugs come out to the side rather than sprouting upwards from the main board. Initially, we thought that we would need 18 pins to connect all the sensors and actuators. However, we noticed we had excess ground connections; one for the flex sensors and 2 for the thumbstick. This led to us grouping all the grounds to reduce the total number of required pins down to 16, which matches our 16-pin 90-degree header.
The Main Perfboard
On the prototype, 2 perfboards contained all the headers for each component. On the new iteration, there is 1 large perfboard that contains all the headers for each component. The connections on the prototype were mostly improvised, so it was not very organized despite achieving the required functionality of the device. With this iteration, we wanted to improve the organization of the connections so that the components are packed closely together. As a result, the following figure was made to map out where wires above and below the perfboard would go when connecting the components to the appropriate pins on the microcontroller.
Execution
After completing the layout of all the components and sockets for each component, the soldering began for the 3 perfboards. During this time, an upgrade was made to some of our components for thermal management.
Soldering Results
After hours of soldering and checking continuity, the 3 perfboards for this new iteration had all the sockets, headers, and wires needed to connect the electrical components together. Below are the top and bottom views of the perfboards. Some of the connections from the mapping diagram were followed while most of them were changed to use wires that connect directly to components. There are many wires incorporated into the main perfboard to connect the appropriate pins to the correct sensors, actuators, and drivers.
Some pins from the microcontroller in the mapping diagram were changed in the final soldering stage because debugging results showed that these pins have special outputs that could affect the boot-up sequence of the ESP32. All the components except for the L9110H IC chip use 3.3 V as the input voltage. The L9110H IC chip uses the voltage of the USB breakout board, which is connected to the battery. Furthermore, all the components were verified such that there was no bridging between pins and they were tested; the thumbstick, Peltier module, flex sensor, and haptic motor could all operate and there were no issues with the L9110H IC chip and DRV2605L driver.
Next Steps
With the soldering complete, the connections verified, and the components working, there is a need for a case to enclose the main perfboard. The glove must also now accommodate the glove perfboard.
Kommentare