AD5682RBCPZ-RL7
Product Specifications
AD5682RBCPZ-RL7
Analog Devices — Analog & Mixed-Signal ICs
Product Overview
The AD5682RBCPZ-RL7 from Analog Devices is a high-performance analog / mixed-signal IC providing precision signal acquisition, conversion, and conditioning. Analog Devices (ADI) is a world leader in high-performance analog, mixed-signal, and DSP ICs for instrumentation, healthcare, automotive, and communications.
Superb Automation partners with leading brands including Qorvo, TI, ADI, XILINX, NVIDIA, Intel, ST, MACOM, Renesas, and Coilcraft to provide complete BOM and supply-chain solutions.
Key Features
- High-resolution data conversion with excellent SNR and THD
- Wide dynamic range and high linearity across operating conditions
- Integrated voltage reference with low temperature drift
- Flexible digital interface — SPI, I2C, or parallel options
- Industrial/automotive temperature range with guaranteed specs
- Low power consumption with power-down modes
Key Specifications
| Resolution | bits |
| Channels | — |
| Sample Rate | kSPS |
| Interface | — |
| Supply Voltage | V |
| Operating Temp | °C |
| Package | — |
| INL | LSB |
| SNR | dB |
| THD | dB |
Applications
AD5682RBCPZ-RL7: A 14 Bit Converter With the Reference Inside
What this part is
The AD5682R is a single channel 14 bit voltage output converter in a 2 mm by 2 mm eight lead LFCSP, with a 2.5 V reference inside the package. The reference drifts 2 ppm per degree C typical and 5 ppm maximum, and the output span can be set to the reference or to twice the reference, so a 5 V full scale comes from a 2.5 V internal reference without an external amplifier. The part runs from 2.7 V to 5.5 V and draws 1.2 mW at 3.3 V.
Putting the reference inside removes the two things that usually spoil a small converter: the external reference device and the trace that couples noise into it. The trade is that the reference is fixed at 2.5 V with a specified drift rather than being whatever the system already has, which matters when several converters have to track each other. A power-on reset writes the register to zero scale, and a power-down mode drops the current to 2 uA maximum at 5 V while keeping a software selectable output load.
Published figures that matter to a buyer
| Resolution | 14 bit |
| Channels | 1, voltage output, buffered |
| Internal reference | 2.5 V, 2 ppm per degree C typical, 5 ppm maximum |
| Output span | 2.5 V or 5 V, selectable |
| Relative accuracy | plus or minus 2 LSB as published on the product page |
| Total unadjusted error | 0.06 percent of full scale maximum |
| Offset error | plus or minus 1.5 mV maximum |
| Gain error | plus or minus 0.05 percent of full scale maximum |
| Output drive | 20 mA |
| Interface | three wire serial, clock up to 50 MHz |
| Supply | 2.7 V to 5.5 V, 1.62 V logic compatible |
| Power | 1.2 mW at 3.3 V, 2 uA in power down |
| Temperature range | minus 40 to plus 105 degrees C |
| Package | 2 mm by 2 mm eight lead LFCSP |
Where the headline numbers stop
Fourteen bits across a 2.5 V span is 152.6 uV per code, and across a 5 V span it is 305.2 uV, so the gain option changes the resolution as well as the range. The total unadjusted error of 0.06 percent of full scale on the 5 V span is 3 mV, which is 9.8 codes, so the converter is not accurate to its own resolution at zero and full scale even though the linearity is good. The offset error alone is 1.5 mV, or 4.9 codes, and it is the term a system can trim out most easily.
The reference drift is the term that cannot be trimmed. Two ppm per degree C over the 145 degree industrial span is 330 ppm, or 825 uV on a 2.5 V reference, which is 2.7 codes at the 5 V span and 5.4 codes at the 2.5 V span. Compared with a 0.06 percent total error of 9.8 codes, the drift is a meaningful share of the budget, so a system that calibrates at 25 degrees C and operates at 105 degrees C should expect the reference to move by more than one code.
Applications where it fits
- Digital gain and offset adjustment in a signal chain, where the converter sits in a feedback path.
- Programmable voltage sources and bias generators in process control equipment.
- Data acquisition systems that need a set point per channel with a small package count.
- Portable instruments where 1.2 mW of converter power is the difference between one battery life and another.
Two questions buyers ask
Can the internal reference be used by something else? The reference is inside the package and is not brought out as a general purpose supply, so it cannot bias another device. A system that needs a shared reference should use the version without the internal reference and a separate reference device, which is the AD5683 rather than this order code.
What does the part do at power-up? The power-on reset circuit writes the register to zero scale and leaves the output buffer in normal mode, so the output comes up at 0 V rather than floating. For a bias supply that must come up at zero that is the right behaviour, and for a loop that must come up at mid scale it is the wrong one.
Sourcing notes and next step
The AD5682RBCPZ-RL7 is the 14 bit version with the internal reference in the 2 mm LFCSP on a 13 inch reel. The same family covers 12 bit and 16 bit members with the same pinout, so a resolution change is a code change rather than a board change. The analog and mixed signal IC family page lists the converter line, the ADC and DAC cross reference page covers the substitution route across converter resolutions, and the BOM shortage recovery page sets out how a build is protected when one resolution option is short.
Fourteen Bits of Resolution and Eleven Bits of Accuracy
The AD5682RBCPZ-RL7 resolves 14 bits and its reference inside the package is the reason it can be used in 2 mm by 2 mm. The same reference is also the reason the part delivers fewer effective bits than its resolution suggests, and the arithmetic is short enough to do before a specification is written around it.
305 microvolts is the step, and the reference is larger than the step
Over a 5 V span, 14 bits is one step of 305.2 uV. The internal reference drifts 5 ppm per degree C maximum. Over a 100 degree ambient swing that is 500 ppm, which on the 2.5 V reference is 1.25 mV and on the doubled 5 V span is 2.5 mV. Divided by the 305.2 uV step, the drift alone is 8.2 LSB. The resolution is 14 bits and the drift is an 8 LSB term, so the two numbers cannot both be quoted on the same line as if they described the same thing.
Offset and gain add to the same budget
A 1 mV zero code error is 3.3 LSB at the 5 V span, and a 0.05 percent gain error is 2.5 mV, another 8.2 LSB. Adding the terms gives roughly 3.7 mV of worst case error, which is 12 LSB, so the part is good to about 12 bits over temperature before any calibration. Trimming the offset and the gain at 25 C removes the 3.3 LSB and the 8.2 LSB of gain and leaves the drift, which is the term that cannot be trimmed without a second measurement at temperature.
A smaller span does not improve the ratio
At the 2.5 V span the step is 152.6 uV and every error term halves except the reference drift, which stays at 1.25 mV in absolute terms and therefore doubles in LSB: 8.2 LSB at 5 V becomes 8.2 LSB at 2.5 V as well, because both the step and the drift scale with the span. The ratio of drift to resolution is set by the reference's ppm and not by the span, which is a useful thing to know when a specification asks for 12 bits on a 1 V span: the part is still 11 to 12 effective bits, and only an external reference with a lower drift moves that number.
| Term | Published figure | At the 5 V span | In LSB of 305.2 uV |
| Resolution | 14 bit | 305.2 uV per step | 1 |
| Zero code error | about 1 mV | 1 mV | 3.3 |
| Gain error | 0.05 percent | 2.5 mV | 8.2 |
| Reference drift, maximum | 5 ppm per degree C | 2.5 mV over 100 C | 8.2 |
| Combined, no calibration | not published as a sum | about 3.7 mV | about 12 |
Where the internal reference wins anyway
An external reference with 1 ppm per degree C drift would cut the drift term to 1.6 LSB, and it would also need a package, a bypass capacitor, a trace and an input that does not couple noise from the converter's own digital lines. In a 2 mm by 2 mm body the external reference is larger than the converter. That is the trade the part makes, and it is the right one when the measurement is ratiometric, when the same reference also sets the sensor excitation, or when the span is trimmed in production and only the drift remains.
The 1.2 mW is a real constraint on a battery board
At 3.3 V the part draws 1.2 mW, which is 0.36 mA, and the power-down mode drops to 2 uA maximum while keeping a selectable output load. A converter that is written once per second and powered down in between averages a few microamps, and the 2 uA figure is the one that decides whether the channel can stay on the battery rail permanently. The 2.7 V to 5.5 V supply range means the same order code works on a 3.3 V rail and on a 5 V rail without a change.
The family is listed on the analog and mixed-signal IC page, the footprint questions are on the pin to pin cross reference page, and a reel that has gone short is handled through the BOM shortage recovery route.
Request Quote — In Stock, 100% New & Original, Full Warranty