Texas Instruments ADS1262 32-bit ADC evaluation, design tips (part 3)

Article 3 – Would be covering hints and tricks using such sensitive ADC/DACs (24+ bits)

  • Layout practice for precision long-scale ADC
  1. Make your assembly tidy and clean. Flux, alcohol residues and fingerprints will deteriorate the surface resistance of a PCB, which can cause big problems for high-impedance sensitive signals. A good ultrasonic bath, and decent bake at around 80°C will make your board clean and free of magical random performance issues. Keep your non-ESD gloves on, when working with precision analog hardware.
  1. Make design simple and without extra complexity. Add more blocks intro circuit only after they are tested and their impact well understood. Design modular!
  1. Stable reference goes a long way. Usual rule of thumb is to select voltage/current reference at least 3 times lower noise and better stability that analog front end to keep accuracy and resolution uncompromised.
  1. Don’t assume. Timing diagrams and datasheets aren’t always apply for every case, so avoid interpretation and test your digital bus interface signals to verify you are transferring what you think. This can save days, if not weeks for firmware debug.
  1. Do not abuse analog circuitry. Electrical stress conditions due to lack of ESD protection devices can upset sensitive circuits and flaw chip’s performance. Ensure your input signals are always within allowed range, at all times, including failure events.
  1. Consider actual and parasitic capacitance very carefully. Unwanted capacitance and capacitor types can ruin performance. High harmonic distortion, piezoelectric effects, flicker noise could be a result from capacitors with X7R or other high-k dielectric materials. Try C0G or film capacitors if your circuit not stable, as expected.
  1. Make an measurement/source error budgets. Account for temperature drift, voltage coefficients, power derating and thermal dissipation in PCB and hardware design. This will help to determine if precision design would live up to real world variations. Will your design still meet it’s spec after 90 days? 1 year? 5 years?
  • Examples from commercial products, such as 7.5-8.5 digit DMMs, calibrators, etc.
  • Seebeck effect (thermocouple EMFs due different metals on connectors, PCB, components, cables)
  • Thermal-aware design for layout
  • Bonus : Cryogenic test at 77K

Return back

Hamon resistor divider

VDC Ω Ω Ω Ω Ω Ω Ω 5.0000000000 mV 5.0000000000 mV 1.56638000 1.56638000 3.19207344 mA 3.19207344 mA 3.19207344 mA 0.00000 A 15.96037 mW 15.96037 mW 0.5000000000

Conrad R. Hoffman : An Easy to Build 0.1X and 0.01X Resistive Divider

Here we will study if it’s possible to operate commercial SD ADC at liquid nitrogen temperatures (77K or -196°C) and how such drastic temperature could affect it’s operation. This is not supported or adviced type of operation, provided for pure research and educational purpose. Same test rig was used to sample data at room temperature and cryocooled. Cooling was performed by full submersion into 30 liter vacuum Dewar with liquid nitrogen. Clock, reference voltage, power and input signal was supplied externally.

Author: Illya T.
Published: Nov. 24, 2015, 3:16 p.m.
Modified: Nov. 25, 2015, 11:48 a.m.