At milliohm levels the fixture is part of the measurement. Before arguing about the last digit, establish how repeatably your contacts touch the same spot, and treat the display as presentation rather than proof.
Low-resistance measurement splits into two questions that are easy to merge: what the instrument can convert, and how reproducibly the sample is presented to it. This guide stays on the second question, covering contacts, fixture behaviour and settling, and leaves the ADC resolution and range budgeting to the related milliohm meter accuracy-budget reference.
Why the fixture dominates
When the quantity being measured is very small, small mechanical differences between measurements can rival the signal. Whether the fixture uses a separate current path and sense path, how firmly the contacts press, and whether the same spot is used each time all shape the reading. Those are choices to document, not assumptions to leave implicit.
The converter side, as documented
The MCP3422 datasheet describes a differential ADC with two input channels, an internal reference and programmable gain. It also pairs 18-bit mode with 3.75 samples per second and offers higher data-rate options at lower resolutions. That pairing matters for procedure: the highest-resolution mode is the slowest, so settling and window length have to be chosen with the mode in mind. A differential input does not by itself guarantee milliohm accuracy.
A diagnostic view of variability
| Source of variation | Evaluation question | Practical check |
|---|---|---|
| Contact placement | Are the probes on the same spot each time? | Re-seat the sample and compare readings |
| Connection scheme | Does the fixture separate current and sense paths? | Document the terminal arrangement explicitly |
| ADC configuration | Which channel, gain and resolution are active? | Read back the programmed settings |
| Settling | Is the chosen mode fast enough to settle? | Allow for the resolution-versus-rate trade-off (18-bit pairs with 3.75 SPS) |
| Display | Does the display resolve the changes you care about? | Compare displayed digits against recorded values |
An ordered repeatability process
- De-energize and detach the component under test before making any connection.
- Fix the contact geometry and write down where the terminals go.
- Verify the ADC settings actually in use, including channel, gain and resolution.
- Allow settling consistent with the selected resolution mode.
- Repeat placement several times to gauge contact repeatability.
- Compare with an independent reference or a known standard before claiming accuracy.
- State the display limit separately from the measurement claim.
Reader questions
Why does changing the resolution mode change how long I wait?
Because the higher-resolution mode is slower: the datasheet pairs 18-bit operation with 3.75 samples per second and offers faster rates only at lower resolution.
Can I trust the last digit on the LCD?
The display shows what it is given. Whether the underlying value is accurate is a separate question that needs an independent comparison, so treat the digits as presentation.
What makes milliohm readings wander between attempts?
Contact repeatability is a leading candidate. Re-seating the sample and comparing results is the quickest way to expose how much the fixture contributes.
Does a differential ADC guarantee a milliohm-accurate result?
No. A differential input with an internal reference and programmable gain is a documented capability, not a calibration result.
Documentation
Scope note: this link documents one differential ADC datasheet only. No measurement or calibration was performed for this guide, and the current-source accuracy, circuit integrity, Kelvin contact behaviour and display ranges remain unverified; measure only detached, de-energized components.
