Design Notes

The DACx3004 12-Bit and 10-Bit, Ultra-Low-Power, Quad Voltage and Current Output Smart DACs With Auto-Detected I2C, PMBus™, or SPI data sheet recommends using a 100-nF decoupling capacitor for the VDD pin and a 1.5-µF or greater bypass capacitor for the CAP pin. The CAP pin is connected to the internal LDO. Place these capacitors close to the device pins.
When the external reference is not used, the VREF pin should be connected to VDD through a pullup resistor.
The example circuit shows two methods for controlling the LED current. Current can be set via an R_SET resistor and varying the gate voltage of an external MOSFET with the DACx3004 output, or the LED current can be set using the current output mode of the DACx3004.
1. To adjust the LED current with the external MOSFET, select an R_SET resistor and vary the gate voltage with the DAC output. R_SET is calculated by:
  $R_{S E T} = \frac{V_{S E T}}{I_{L E D}}$
  If the DAC output voltage range is chosen to be 0 V to 2.5 V, and the required LED current range is 0 mA to 20 mA, R_SET is calculated to be:
  $R_{S E T} = \frac{2.5 V}{20 m A} = 125 Ω$
  The DAC codes can be calculated by:
  $C o d e = \frac{V_{D A C}}{V_{R E F}} \times 1024$
  In this design, the internal reference is powered down to limit the power consumption. This configuration compensates the gate-source voltage drop caused by temperature, drain current, and aging of the MOSFET. Assuming a typical gate-source voltage of 1.2 V and a power supply headroom of 200 mV, the VDD for the DAC must be a minimum of (2.5 V + 1.2 V + 200 mV) = 3.9 V. If a 5-V VDD is used as the reference, the high and low DAC values for the 10-bit DAC53004 become:
  $C o d e = \frac{2.5 V}{5 V} \times 1024 = 512 d$
  $C o d e = \frac{0 V}{5 V} \times 1024 = 0 d$
  where
  - d = decimal
2. The DAC can be used in current output mode to drive the LED directly with up to 250 µA. With the ±250 µA range selected, the 8-bit current DAC53004 codes are calculated by:
  $C o d e = \frac{(I_{D A C} - I_{M I N}) \times 256}{I_{M A X} - I_{M I N}}$
  The high and low DAC53004 codes become:
  $C o d e = \frac{(250 µ A + 250 µ A) \times 256}{250 µ A + 250 µ A} = 256 d$
  $C o d e = \frac{(0 µ A + 250 µ A) \times 256}{250 µ A + 250 µ A} = 128 d$
  256 decimal (256 d) is rounded down to 255 d to give a high value of 248.04 µA.
The power consumption of the DACx3004 will vary based on the configuration used, and the power down settings. The power consumption is given by:
$P = (V D D \times {I D D}_{S L E E P}) + \sum_{x = 0}^{N - 1} (V D D \times {I D D}_{X})$
where
- IDD_SLEEP is the quiescent current for the device in sleep mode
- N is the number of channels powered on
- IDD_X is the quiescent current per channel powered on
1. One DAC channel is powered on in voltage output mode in the voltage output configuration with the external MOSFET. The quiescent current in voltage output mode is 35 µA typical per channel. The quiescent current of the DAC in sleep mode is 21 µA maximum. With a VDD of 5 V, the power consumption equation becomes:
  $P = (5 V \times 21 µ A) + (5 V \times 35 µ A) = 280 µ W$
  This calculation does not include the load current sourced from VCC though R_SET.
2. One DAC channel is powered on in current output mode in the current output configuration. With a current output range of 0 µA to 250 µA, the quiescent current is 18 µA typical per channel. In this configuration, the load current sourced by the DAC output channel also needs to be added. The power consumption equation becomes:
  $P = (5 V \times 21 µ A) + (5 V \times (18 µ A + 250 µ A)) = 1.445 m W$
3. All of the DAC channels are powered down and the device quiescent current is 3 µA maximum in deep-sleep mode. The power consumption equation becomes:
  $P = (5 V \times 3 µ A) = 15 µ W$
The slew rate between the high and low DAC codes can be programmed if these two values are stored in the MARGIN-HIGH and MARGIN-LOW DAC registers. The slew time is determined by the settings in the SLEW-RATE and CODE-STEP fields in the DAC-X-FUNC-CONFIG register. The slew time is given by:
$S l e w T i m e = \frac{(M A R G I N_H I G H_C O D E - M A R G I N_L O W_C O D E + 1)}{C O D E_S T E P} \times S L E W_R A T E$
If the CODE-STEP is set to 1 LSB, and the SLEW-RATE is set to 4 µs/step, the slew time for the voltage configuration becomes:
$S l e w T i m e = \frac{(512 - 0 + 1)}{1} \times 4 µ s = 2.05 m s$
The slew time for the current output configuration becomes:
$S l e w T i m e = \frac{(255 - 128 + 1)}{1} \times 4 µ s = 512 µ s$
The GPIO pin is used as an input to enter and exit deep-sleep mode. A falling edge on the GPIO pin puts the device in deep-sleep mode. The LDO takes approximately 550 μs to switch off and the device remains in deep-sleep mode as long as the GPIO input is low. A rising edge brings the device out of deep-sleep mode. The digital circuitry and the LDO take approximately 550 μs to switch on. The register settings to enable the GPIO for this function are described in the Register Settings section.
The DACx3004 can be programmed with the initial register settings described in the Register Settings section using I²C or SPI. The initial register settings can be saved in the NVM by writing a 1 to the NVM-PROG field of the COMMON-TRIGGER register. After programming the NVM, the device loads all registers with the values stored in the NVM after a reset or a power cycle.