Slide 1: Digital to Analogue Conversion
Digital to Analogue Conversion:
the mcp4275
Lecture 10
Uli Raich
UCC semester 2017/2018
Slide 2: Analogue versus digital
Up to now we only treated digital signals:
- on/off for the LEDs
- on/off to read the LED state
- Powering or not powering coils to generate magnetic fields in a stepping motor
However:
The world is mostly analogue:
- Temperatures are changing continuously and not in steps
- Pressure is an analogue value
- Distance, time, current, resistance … are all analogue values
Slide 3: Converting from digital to analogue
Since our computer is a digital device we must
- Convert digital values to analogue voltage levels
Digital to Analogue Conversion (DAC)
- … and we must convert external analogue values to digital
Analogue to Digital Conversion (ADC)
Slide 4: Digital to analogue conversion
A digital to analogue converter does not really convert into a continuous waveform
Since we have digital values as a base, there will be steps in the output waveform
The size of these steps depends on the resolution of the DAC
What is the smallest step a 12 bit DAC can produce on a 0..5V scale?
Slide 5: Can we smooth the output signal?
Yes, it is possible to smooth out these steps.
We need a low-pass filter, which filters out high frequencies
(the abrupt steps we have in the output signal) and lets pass only slower transitions.
When looking carefully at the output of our sine generator you will
also see these steps. In this case however they come from the
limited number of sine values (100) we calculate. To improve the resolution
we would have to increase the number of samples and the frequency
with which we send these values to the DAC.
Slide 6: How does a DAC work?
If you want to know more about DAC technology (and you should!)
have a look at this
excellent tutorial, from which I have copied the above illustration.
Slide 7: The MCP4725 12 bit DAC
As a demo device we bought the MCP4725 DAC
This is a 12 bit DAC which can be accesses by the
I2C bus
Here is its
data sheet
and here a photo of the device
Slide 8: The I2C bus
The
I2C bus is an industrial standard 2 wire bus using a data (SDA) and a clock (SCL) line.
Adding Vcc and Gnd we need a mere 4 wires to connect
a
I2C device to the Raspberry Pi cobbler
The
I2C bus was invented by Philips
Slide 9: Typical I2C connections
Slide 10: Open drain signals
Open drain signals mean that you can pull a line down
but you cannot set it to a high level.
If nobody pulls the line down, then it is at Vcc level,
pulled up by a pull-up register.
Like this the contention problem is solved where one device tries
to set a line to a high level, while another sets it to low,
thus creating a short circuit.
Slide 11: I2C Master and slave
The
I2C bus has at least one master (in our case the interface in the
Raspberry Pi’s ARM processor) and several slaves
We have the following
I2C slave devices:
- mcp4275 DAC
- bmp180 barometric pressure sensor
- pcf8581 8 bit ADC
- ads115 16 bit ADC
- at24c32 eeprom
- ds1307 real time clock
- mma845x accelerometer
- pcf8574 I/O expander used on the 2-line LCD display
Slide 12: I2C addressing
Since there can be several slaves on the bus there must
be a means of distinguishing them through addressing:
Every
I2C slave has a 7 bit address associated with it
Usually this address is determined by the manufacturer but
often there are address pins on the devices allowing the
user to have several devices of the same type on the bus
Slide 13: I2C buses and I2C addresses on the Raspberry Pi
The Raspberry Pi has 2
I2C buses with bus 1 being put onto the cobbler
You can find out the addresses of the
I2C slaves are currently
connected with the
i2cdetect command
Slide 14: Initiating an I2C transfer
Slide 15: Start stop conditions
The master starts a transfer by creating a
start condition:
- high to low transition on SDA while SCL is high
and ends the transfer with a
- stop condition: low to high transition on SDA while SCL is high4
Slide 16: Data transmission
Data are transmitted with the SDA line stable with SCL is high
Slide 17: A write cycle
A write cycle:
After the start condition the device address. The eighth bit, the R/W bit is kept low.
The second data byte is considered the register byte and the last one the data byte
The slave acknowledges each byte transfer.
Slide 18: The read cycle
The read cycle is a bit more complex:
First the master sends the slave address with R/W set to write
Then it sends the register information.
After that another address byte is sent, this time with RW set to read
And finally the slave sends the data. The master still sends the clock but
releases the SDA line allowing the slave to control it
Slide 19: I2C clock frequency
I2C bus supports different transfer speeds:
- Standard: 100 kbps
- Fast: 400 kbps
- High speed: 3.4 Mbps
pigpio seems to use the standard speed
while the Raspberry Pi can run
I2C transfers of up tp 1.66 Mbps
However: you need more low level libraries or direct register
access to accomplish these high speeds
Slide 20: MCP4725
Slide 21: Specs of the MCP4275
Slide 22: MCP4275 registers
Slide 23: MCP4275 I2C fast write
How does the
I2C write cycle look like in the MCP4275?
The MCP4275 fast write cycle write only the DAC register and not the EEPROM.
This is enough for what we want to do.
Slide 24: MCP4275 normal write cycle
Slide 25: Read back the DAC data
Slide 26: pigpio: i2c_open
Slide 27: I2C access initialization with pigpio
In addition to opening the library with pigpio_start we now also must
make a connection to the
I2C driver:
Slide 28: Accessing I2C with pigpio
Coming back to fast write mode, we must write 3 bytes:
- Address + R/W
- Register
- Data
The register byte contains also the high 4 bits on the 12 bit DAC data word
The first byte is created and written within the library extracting the I2C address
for the I2C initialization call
Slide 29: i2c_write_byte_data
The pigpio library provides a function
This means we have to split our 12 bit DAC data into 2 parts:
The highest 4 bits go into i2c_reg, the lower 8 bits into bVal
Slide 30: Writing the DAC value, an example
dacValue is a
short, while
reg and
value are
unsigned char
Slide 31: DAC performance
There are a number of criteria which characterize the performance of a DAC
- Its resolution or its least significant bit (lsb)
- Its settling time (maximum speed you can go)
- Its integral non-linearity or relative accuracy
- Its differential non-linearity
- Its offset error
- Its gain error ...
Slide 32: Relative accuracy
Slide 33: Differential non-linearity
Slide 34: Offset error
Slide 35: Gain error
Slide 36: Settling Time
--
Uli Raich - 2017-10-31
Comments