The Second Pass Prototype

A project log for A Halo For Lucy

It's not what you think.

Bud BennettBud Bennett 04/20/2019 at 19:090 Comments

A list of new features and/or changes after testing the first prototype:

  1. Only three VL53L0X distance sensors. Five sensors is confusing and not necessary. There will be dead zones because the outer two sensors will be 50° off of the front sensor. This was not a problem in the first prototype when two of the sensors were disabled.
  2. Exchanged the Arduino Feather for the Arduino ItsyBitsy -- its a lot smaller with nearly all of the same capabilities.
  3. Added a LIS3DH accelerometer for motion and tilt sensing. This will allow the system to mute when no motion is detected or the unit is tilted significantly off-axis (i.e. sleeping). It should also disable the forward sensor detection when Lucy is pointing her head at the ground.
  4. Added a battery charging function. 500mA when the USB is plugged into a 5V source. The charge indicator is a LED on the Itsy via the D12 pin.
  5. Added a separate PIC10F322 µC to monitor the battery voltage and put the system to sleep (<100µA) when the battery voltage drops below about 3.3V. The PIC will also monitor the accelerometer activity and put the system to sleep after TBD seconds of inactivity -- a battery saving measure.

The Second Pass Schematic:

The MCP73832 is the same as the MCP73831 but the STAT output is a pulldown only. The ItsyBitsy will present a weak pullup on its D12 input and only pull to 3.3V.

There is a separate 3.3V LDO regulator providing power to the accelerometer and the PIC. The accelerometer is configured to latch the INT1 pin high when motion is detected above a certain threshold. The latch is reset when the Itsy reads the INT1_SRC register every TBD seconds. If INT1 doesn't go high after TBD seconds then the PIC will disable the Itsy by pulling down the En pin, which turns off the Itsy's 3.3V regulator. This puts the system to sleep with a relatively low current drain. If the accelerometer detects motion again it will assert INT1 and the PIC will re-enable the Itsy. 

(Note: the connection from INT1 to D7 is not necessary since the Itsy can read the status of the accelerometer over the I2C interface, but I thought there might be some value to have be able to read the INT1 pin for debugging purposes. This brings up a sticky issue because the accelerometer and the Itsy are on two separate supplies. The Itsy's LDO is fed by two Schottky diodes to pick up power from either USB or BAT. When the battery voltage gets below about 3.5V the Itsy's supply voltage will fall below 3.3V because of the drop across the Schottky. But I measured the total voltage drop between BAT and the 3.3V supply as only 200mV, which should not cause a latchup problem.)

The PIC has one other function besides monitoring INT1. It also measures its supply voltage and will shutdown the Itsy if the implied battery voltage drops below TBD. This PIC has an on-board 8-bit ADC using VDD2 as a reference. The ADC input will be an internal FVR(Fixed Voltage Reference) set to 2.048V. If the measured ADC value rises above TBD, then the Itsy will be put into a sleep mode by pulling the En pin low. This should save the battery from imminent damage.

The Layout:

It's smaller by quite a bit -- only 1.575" x 1.156" (40mm x 29.5mm). The PIC10F322 is a SOT26 and two pins are unused! The battery and USB connection are at the top -- I'm a bit worried that the right speaker might interfere with the USB connector access. Hopefully that will be solved by clever packaging.

I'm waiting to see the results of a couple more training sessions before I order new PCBs.