I’m hacking on adding Arduino Zero support to the Zephyr Project at the moment. I had a look at speeding up crc16_ccitt() but ran into a problem: there’s so many variants of CRC16-CCITT to choose from!
Zephyr uses CRC-16/AUG-CCITT which is equivalent to CRC-16/CCITT-FALSE with two zero bytes prepended. The effective seed is 0x1d0f and test vector output is 0xe5cc.
Mynewt uses CRC-16/XMODEM. The seed is 0 and test vector output is 0x31c3.
There is an issue when trying to program a SAMD21 like in the Arduino Zero using OpenOCD with a ST-Link v2 which gives errors like this:
** Programming Started ** auto erase enabled Info : SAMD MCU: SAMD21E18A (256KB Flash, 32KB RAM) Error: Failed to erase row containing 00000000 Error: SAMD: failed to erase sector 0 Error: failed erasing sectors 0 to 0 The issue is that NVMCTRL_CTRLA is a half-word register and the ST-Link v2 emulates the half word write using two single byte writes.
In follow up to my mysensors.org build, I’ve always wanted to design PCB and try my hand at SMD soldering. I normally use veroboard on a 0.1” pitch but the time to cut and solder wires makes for a slow build. Here’s the result:
An AT91SAMD21E18 (256 KiB flash, 32 pins, 0.8 mm pitch) USB, regulator, Cortex SWD debug header, LED, and pullups on I2C Footprint for a RFM69CW radio Adafruit Feather form factor Two layer PCB with 0603 passives The soldering turned out well.
It gets warm and humid here in the summer and feels too dry in the winter. To measure this I built up two MySensors Sensebender Micro boards and designed a 3D printed a case to go with them:
The holes in the case are to save on print time and let the air flow. The battery is far oversized and probably self discharges faster than the board draws current.
I ended up using a Raspberry Pi Zero with a RFM69HW as the gateway.
As part of adding MySensors based sensors around the house, I want to make an integrated version of the Sensebender Micro upgraded with a SAMD21 and a RFM69HCW radio. I’ve never had much success with making PCBs at home, so I thought I’d give it another try and record the experiments.
Try 1: HP M252dw, laminator @ 10 mm, 4 times through, Migros-brand T-Shirt transfer paper. Some transfer but lots left behind.