Glove PCB Assembly - Githubissues

digitaldanny / AudioEffectsGlove

A hand tracking glove that controls parameters of various audio effects.

1 stars 0 forks source link

Glove PCB Assembly #48

Closed digitaldanny closed 3 years ago

digitaldanny commented 3 years ago

Describe the solution you'd like Assemble the Glove PCB and note any issues with the PCB layout here.

digitaldanny commented 3 years ago

PCBs delayed until next Monday (3/29). Remainder of parts came in from Digikey this week, and I will assemble the PCB then.

digitaldanny commented 3 years ago

To program my device, I need to connect debug probe's Vcc and GND pins as the device power source too. Luckily, the BSTBCK1 header can be disconnected and the power can be provided to the DBPWR1 header's MSP_3V3 and GND. Otherwise, I will get a "JTAG Device not found" type error in CCS when programming.

digitaldanny commented 3 years ago

My LCD demo is playing at a much slower speed, and my HC-05 demo is not working on the PCB. I think there may be an issue with the clock source because both of these are affected by the SMCLK, which is the HFX clock divided by 4.

This StackExchange thread discusses the issue I am having. It looks like some external circuitry is required to get the crystal to 48 MHz. I am going to see if I can borrow one of the 48 MHz oscillators from the lab.

digitaldanny commented 3 years ago

Completed

Validated that 3.7V, 110 mA battery can power entire board through the 3.3V buck/boost converter.
Successfully programmed / debugged MSP432 on PCB using JTAG pins and debugger on MSP launchpad.
Validated flex sensor circuitry and LCD with oscilloscope and SW unit tests.

Todo:

Replace crystal with 48 MHz oscillator. This should fix the issues I am seeing with slow LCD drawing and infinite RX polling for uart communication with HC-05.
Validate Open Circuit Voltage reading and Runtime Current Draw sensing HW with SW unit tests.

digitaldanny commented 3 years ago

Switching out the crystal I purchased with the crystal from the lab has the same results. The driverlib example "cs_hfxt_start" to toggle P6.0 after 16000000 periods shows that there is exactly 1 second between toggles. This proves that the clock is running on 16 MHz instead of 48 MHz as intended.

Instead of spending time debugging this, I am going to make the quick switch to my UART and I2C configurations to work with 16 MHz. This should still be plenty of speed for my application.

digitaldanny commented 3 years ago

Fully assembled glove PCB.

digitaldanny commented 3 years ago

Tests for porting from 48 MHz to 16 MHz.

delayMs and delayUs timings.
UART baud rate 38400 for communicating with HC-05.
Make sure communication still works with MPU6050 over I2C.
Make sure communication still works with LCD over SPI.