With Crystal: set the crystal startup wait time to 1 second. It was 2
seconds before, and that seeemed too long.
With USB-Sync: scan for up to 1 second for the USB to be registered and
carry on with boot as soon as it it. Before, the code just waited for
500ms.
Side change: improve related comments.
Signed-off-by: robert-hh <robert@hammelrath.com>
It turned out that the result of calling ticks_us() was always either odd
or even, depending on some internal state during boot. So the us-counter
was set to a 2 MHz input and the result shifted by 1. The counting period
is still long enough, since internally a (now) 63 bit value is used for us.
By using the phase jitter between the DFLL48M clock and the FDPLL96M clock.
Even if both use the same reference source, they have a different jitter.
SysTick is driven by FDPLL96M, the us counter by DFLL48M. As a random
source, the us counter is read out on every SysTick and the value is used
to accumulate a simple multiply, add and xor register. According to tests
it creates about 30 bit random bit-flips per second. That mechanism will
pass quite a few RNG tests, has a suitable frequency distribution and
serves better than just the time after boot to seed the PRNG.
Allowing to increase the clock a little bit to 54Mhz. Not much of a gain,
but useful for generating a RNG entropy source from the jitter between
DFLL48M and FDPLL96M.
Methods implemented are:
- rtc.init(date)
- rtc.datetime([new_date])
- rtc.calibration(value)
The presence of this class can be controlled by MICROPY_PY_MACHINE_RTC. If
the RTC module is used, the time module uses the RTC as well.
For boards without a 32kHz crystal, using RTC makes no sense, since it will
then use the ULP32K oscillator, which is not precise at all. Therefore, it
will by default only be enabled for boards using a crystal, but can be
enabled in the respective mpconfigboard.h.
Which just sets the CPU clock to 200kHz and switches the peripheral clock
off. There are two modes:
machine.lightsleep(duration_ms)
and
machine.lightsleep()
In any mode any configured pin.irq() event will terminate the sleep.
Current consumption in lightsleep for some boards:
- 1.5 - 2.5 mA when supplied trough an active USB
(Seeed XIAO w/o power LED, Adafruit ItsyBitsy)
- 0.8 - 2 mA when supplied through Gnd/+5V (Vusb)
(Seeed XIAO w/o power LED, Adafruit ItsyBitsy)
- < 1 mA for SAMD51 when supplied trough a battery connector
(Sparkfun Thing SAMD51 plus)
Related change: move the calls to SysTick_Config() into set_cpu_freq(). It
is required after each CPU freq change to have ticks_ms run at the proper
rate.
The range is 1MHz - 48 MHz. Note that below 8 MHz there is no USB support.
The frequency will be set to an integer fraction of 48 MHz. And after
changing the frequency, the peripherals like PWM, UART, I2C, SPI have to be
reconfigured.
Current consumption e.g. of the Seeed Xiao board at 1 MHz is about 1.5 mA,
mostly caused by the on-board LED (green LED with 1k resistor at 3.3V).
Two new compile flags are:
MICROPY_HW_DFLL_USB_SYNC: Effective only if DFLL48 does not run from the
crystal. It will synchronize the DFLL48M clock with the USB's SOF pulse.
If no USB is connected, it will fall back to open loop mode. The DFLL48M
clock is then pretty precise, but with a higher clock jitter at SAMD51
devices.
MICROPY_HW_MCU_OSC32KULP: Effective only if the devics uses a crystal as
clock source. Run the MCU clock from the ULP 32kHz oszillator instead of
the crystal. This flag was added to cater for a interference problem of
the crystal and Neopixel/Debug pins at Adafruit FEATHER Mx boards, which
causes the board to crash. Drawback: ticks_ms() and time.time() vs. than
ticks_us() and the peripherals like PWM run at not synchronous clocks.
robert-hh
Damien George