docs/library/pyb.ADC.rst: Document new features for ADCAll.
This commit is contained in:
parent
d6cf5c6749
commit
8f11d0b532
|
@ -14,10 +14,12 @@ class ADC -- analog to digital conversion
|
|||
val = adc.read() # read an analog value
|
||||
|
||||
adc = pyb.ADCAll(resolution) # create an ADCAll object
|
||||
adc = pyb.ADCAll(resolution, mask) # create an ADCAll object for selected analog channels
|
||||
val = adc.read_channel(channel) # read the given channel
|
||||
val = adc.read_core_temp() # read MCU temperature
|
||||
val = adc.read_core_vbat() # read MCU VBAT
|
||||
val = adc.read_core_vref() # read MCU VREF
|
||||
val = adc.read_vref() # read MCU supply voltage
|
||||
|
||||
|
||||
Constructors
|
||||
|
@ -81,27 +83,42 @@ The ADCAll Object
|
|||
|
||||
.. only:: port_pyboard
|
||||
|
||||
Instantiating this changes all ADC pins to analog inputs. The raw MCU temperature,
|
||||
Instantiating this changes all masked ADC pins to analog inputs. The preprocessed MCU temperature,
|
||||
VREF and VBAT data can be accessed on ADC channels 16, 17 and 18 respectively.
|
||||
Appropriate scaling will need to be applied. The temperature sensor on the chip
|
||||
has poor absolute accuracy and is suitable only for detecting temperature changes.
|
||||
Appropriate scaling is handled according to reference voltage used (usually 3.3V).
|
||||
The temperature sensor on the chip is factory calibrated and allows to read the die temperature
|
||||
to +/- 1 degree centigrade. Although this sounds pretty accurate, don't forget that the MCU's internal
|
||||
temperature is measured. Depending on processing loads and I/O subsystems active the die temperature
|
||||
may easily be tens of degrees above ambient temperature. On the other hand a pyboard woken up after a
|
||||
long standby period will show correct ambient temperature within limits mentioned above.
|
||||
|
||||
The ``ADCAll`` ``read_core_vbat()`` and ``read_core_vref()`` methods read
|
||||
the backup battery voltage and the (1.21V nominal) reference voltage using the
|
||||
3.3V supply as a reference. Assuming the ``ADCAll`` object has been Instantiated with
|
||||
``adc = pyb.ADCAll(12)`` the 3.3V supply voltage may be calculated:
|
||||
The ``ADCAll`` ``read_core_vbat()``, ``read_vref()`` and ``read_core_vref()`` methods read
|
||||
the backup battery voltage, reference voltage and the (1.21V nominal) reference voltage using the
|
||||
actual supply as a reference. All results are floating point numbers giving direct voltage values.
|
||||
|
||||
``v33 = 3.3 * 1.21 / adc.read_core_vref()``
|
||||
``read_core_vbat()`` returns the voltage of the backup battery. This voltage is also adjusted according
|
||||
to the actual supply voltage. To avoid analog input overload the battery voltage is measured
|
||||
via a voltage divider and scaled according to the divider value. To prevent excessive loads
|
||||
to the backup battery, the voltage divider is only active during ADC conversion.
|
||||
|
||||
If the 3.3V supply is correct the value of ``adc.read_core_vbat()`` will be
|
||||
valid. If the supply voltage can drop below 3.3V, for example in in battery
|
||||
powered systems with a discharging battery, the regulator will fail to preserve
|
||||
the 3.3V supply resulting in an incorrect reading. To produce a value which will
|
||||
remain valid under these circumstances use the following:
|
||||
``read_vref()`` is evaluated by measuring the internal voltage reference and backscale it using
|
||||
factory calibration value of the internal voltage reference. In most cases the reading would be close
|
||||
to 3.3V. If the pyboard is operated from a battery, the supply voltage may drop to values below 3.3V.
|
||||
The pyboard will still operate fine as long as the operating conditions are met. With proper settings
|
||||
of MCU clock, flash access speed and programming mode it is possible to run the pyboard down to
|
||||
2 V and still get useful ADC conversion.
|
||||
|
||||
``vback = adc.read_core_vbat() * 1.21 / adc.read_core_vref()``
|
||||
It is very important to make sure analog input voltages never exceed actual supply voltage.
|
||||
|
||||
It is possible to access these values without incurring the side effects of ``ADCAll``::
|
||||
Other analog input channels (0..15) will return unscaled integer values according to the selected
|
||||
precision.
|
||||
|
||||
To avoid unwanted activation of analog inputs (channel 0..15) a second prarmeter can be specified.
|
||||
This parameter is a binary pattern where each requested analog input has the corresponding bit set.
|
||||
The default value is 0xffffffff which means all analog inputs are active. If just the internal
|
||||
channels (16..18) are required, the mask value should be 0x70000.
|
||||
|
||||
It is possible to access channle 16..18 values without incurring the side effects of ``ADCAll``::
|
||||
|
||||
def adcread(chan): # 16 temp 17 vbat 18 vref
|
||||
assert chan >= 16 and chan <= 18, 'Invalid ADC channel'
|
||||
|
@ -140,4 +157,5 @@ The ADCAll Object
|
|||
def temperature():
|
||||
return 25 + 400 * (3.3 * adcread(16) / 4096 - 0.76)
|
||||
|
||||
|
||||
Note that this example is only valid for the F405 MCU and all values are not corrected by Vref and
|
||||
factory calibration data.
|
||||
|
|
Loading…
Reference in New Issue