SBAS652A May 2014 – August 2014 ADS7950-Q1 , ADS7951-Q1 , ADS7952-Q1 , ADS7953-Q1 , ADS7954-Q1 , ADS7956-Q1 , ADS7957-Q1 , ADS7958-Q1 , ADS7959-Q1 , ADS7960-Q1 , ADS7961-Q1
PRODUCTION DATA.
In general applications, when the internal multiplexer is updated, the previously converted channel charge is stored in the 15-pF internal input capacitance that disturbs the voltage at the newly selected channel. This disturbance is expected to settle to 1 LSB during sampling (acquisition) time to avoid degrading converter performance. The initial absolute disturbance error at the channel input must be less than 0.5 V to prevent source current saturation or slewing that causes significantly long settling times. Fortunately, significantly reducing disturbance error is easy to accomplish by simply placing a large enough capacitor at the input of each channel. Specifically, with a 150-pF capacitor, instantaneous charge distribution keeps disturbance error below 0.46 V because the internal input capacitance can only hold up to 75 pC (or 5 V × 15 pF). The remaining error must be corrected by the voltage source at each input, with impedance low enough to settle within 1 LSB. The following application examples explain the considerations for the input source impedance (RSOURCE).
This application is the most typical application, but requires the lowest RSOURCE for good performance. In this configuration, the 2xREF range allows larger source impedance than the 1xREF range because the 1xREF range LSB size is smaller, thus making it more sensitive to settling error.
The design is optimized to show the input source impedance (RSOURCE) between the 100 Ω to 10,000 Ω required to meet the 1-LSB settling at 12-bit, 10-bit, and 8-bit resolutions at different throughput in 1xREF (2.5-V) and 2xREF (5-V) input ranges.
Although the required input source impedance can be estimated assuming a 0.5-V initial error and exponential recovery during sampling (acquisition) time, this estimation over-simplifies the complex interaction between the converter and source, thus yielding inaccurate estimates. Thus, this design uses an iterative approach with the converter itself to provide reliable impedance values.
To determine the actual maximum source impedance for a particular resolution and sampling rate, two subsequent channels are set at least 95% of the full-scale range apart. With a 1xREF range and 2.5 Vref, the channel difference is at least 2.375 V. With 2xREF and 2.5 Vref, the difference is at least 4.75 V. With a source impedance between 100 Ω to 10,000 Ω, the conversion runs at a constant rate and a channel update is issued that captures the first couple samples after the update. This process is repeated at least 100 times to remove any noise and to show a clear settling error. The first sample after the channel update is then compared against the second one. If the first and second samples are more than 1 LSB apart, throughput rate is reduced until the settling error becomes 1 LSB, which then sets the maximum throughput for the selected impedance. The whole process is repeated for nine different impedances between 100 Ω to 10,000 Ω.
These curves show the RSOURCE for an unbuffered MXO.
The use of a buffer relaxes the RSOURCE requirements to an extent. Charge from the sample-and-hold capacitor no longer dominates as a residual charge from a previous channel. Although having good performance is possible with a larger impedance using the OPA192, the output capacitance of the MXO also holds the previous channel charge and cannot be isolated, which limits how large the input impedance can finally be for good performance. In this configuration, the 1xREF range allows slightly higher impedance because the OPA192
(20 V/µs) slews approximately 2.5 V in contrast to the 2xREF range that requires the OPA192 to slew approximately 5 V.
The design is optimized to show the input source impedance (RSOURCE) between the 100 Ω to 10,000 Ω required to meet a 1-LSB settling at 12-bit, 10-bit, and 8-bit resolutions at different throughput in 1xREF (2.5 V) and 2xREF (5 V) input ranges.
The design procedure is similar to the unbuffered-MXO application, but includes an operation amplifier in unity gain as a buffer. The most important parameter for multiplexer buffering is slew rate. The amplifier must finish slewing before the start of sampling (acquisition) to keep the buffer operating in small-signal mode during sampling (acquisition) time. Also, between the buffer output and converter input (INP), there must be a capacitor large enough to keep the buffer in small-signal operation during sampling (acquisition) time. Because 150 pF is large enough to protect the buffer form hold charge from internal capacitors, this value selected along with the lowest impedance that allows the op amp to remain stable.
The converter allows the MXO to settle approximately 600 ns before sampling. During this time, the buffer slews and then enters small-signal operation. For a 5-V step change, slew rate stays constant during the first 4 V. The last 1 V includes a transition from slewing and non-slewing. Thus, the buffer cannot be assumed to keep a constant slew during the 600 ns available for MXO settling. Assuming that the last 1-V slew is reduced to half is recommended. For this reason, slew is 10 V/µs or (5 Vref + 1 V) / 0.6 µs to account for the 1-V slow slew. The OPA192 has a 20-V/us slew, and is capable of driving 150 pF with more than a 50° phase margin with a 50-Ω or 100-Ω Riso, making the OPA192 an ideal selection for the ADS79xx-Q1 family of converters.
These curves show the RSOURCE for an OPA192 buffered MXO.