Datasheet

MCP4901/4911/4921
DS22248A-page 32  2010 Microchip Technology Inc.
6.7 Designing a Double-Precision
DAC
Example 6-5 illustrates how to design a single-supply
voltage output capable of up to 24-bit resolution by
using 12-bit DACs. This design is simply a voltage
divider with a buffered output.
As an example, if a similar application to the one
developed in Section 6.5.1 “Design Example:
Design a bipolar dac using example 6-3 with 12-bit
MCP4912 or MCP4922” required a resolution of 1 µV
instead of 1 mV and a range of 0V to 4.1V, then 12-bit
resolution would not be adequate.
1. Calculate the resolution needed:
4.1V/1 µV = 4.1x 10
6
. Since 2
22
= 4.2 x 10
6
,
22-bit resolution is desired. Since DNL = ±0.75
LSB, this design can be done with the MCP4921
or MCP4922.
2. Since the DAC
B
‘s V
OUTB
has a resolution of
1 mV, its output only needs to be “pulled” 1/1000
to meet the 1 µV target. Dividing V
OUTA
by 1000
would allow the application to compensate for
DAC
B
’s DNL error.
3. If R
2
is 100, then R
1
needs to be 100 k.
4. The resulting transfer function is not perfectly
linear, as shown in the equation of Example 6-5.
EXAMPLE 6-5: SIMPLE, DOUBLE PRECISION DAC WITH MCP4921 OR MCP4922.
V
REF
DAC
B
V
DD
R
2
DAC
A
V
DD
R
1
DAC
A
(Fine Adjust)
DAC
B
(Course Adjust)
SPI
3
R
1
>> R
2
V
O
V
OUTA
R
2
V
OUTB
R
1
+
R
1
R
2
+
-----------------------------------------------------
=
G = Gain selection (1x or 2x)
D = Digital value of DAC (0-4096)
0.1 µF
V
CC
+
V
CC
–
V
OUTA
V
REFA
G
A
D
A
2
12
-------=
V
OUTB
V
REFB
G
B
D
B
2
12
-------=
V
OUTA
V
OUTB
V
O