Choosing DMM Integration Time for Thick Film Resistance Tests

Measurement purpose

Determine which supported acquisition settings produce a valid resistance result within the required noise and timing constraints for the specified thick-film test cycle.

Specimens and conditions

Stable check artifact

Use an identified resistor at controlled temperature to compare acquisition settings without introducing intentional product drift.

Representative contact cycle

Include the real connection or scanner transition and the actual specimen support. A continuously connected artifact alone does not establish post-switch settling.

Equipment and records required

• DMM acquisition controls: Identify NPLC or aperture options, line-frequency selection, autozero, filtering, trigger delay and any mode-specific accuracy conditions.
• Timing observation: Capture trigger and result times and the settings used. If a faster diagnostic instrument is employed, define its bandwidth and loading separately.

Method sequence

1. Characterize stationary noise

   Compare supported apertures on an unchanged check artifact while retaining individual readings and timestamps.

   Record: Mean, scatter, range, excitation and integration settings.

2. Test the connection sequence

   Repeat controlled contact or channel transitions with varied pre-acquisition delay, then compare readings at a fixed aperture.

   Record: Transition timing, delay and residual movement.

3. Choose the reported cycle

   Select aperture and delay jointly, verify throughput and retain the diagnostic response to interruptions.

   Record: Approved timing diagram and conditions requiring renewed evaluation.

Decision and uncertainty

The chosen settings must meet the defined resistance uncertainty and valid-settling criteria without concealing required transient information. A quiet display alone is not sufficient.

Line rejection depends on acquisition architecture and frequency alignment. Serial correlation, drift and thermal response limit the benefit of repeated averaging.

The measurement-method owner approves aperture and timing; the process owner defines which transient events must remain detectable.

Traceable outputs

Separate aperture, waiting time and reading interval

The aperture is the interval over which the converter observes the input for one measurement. Trigger delay is a wait before that interval. The complete reading interval can also contain switching, range selection, autozero and communication. These times do different jobs and should not be represented by one field called measurement speed.

For example, waiting after a probe lands can allow an electrical transient to decay before acquisition. Increasing aperture instead may include the transient in the reported average. The two sequences can have similar total duration yet produce different results on a printed resistor, especially immediately after trimming or fixture movement.

Translate power-line cycles into seconds

Where the instrument supports integration expressed in power-line cycles, the nominal aperture is the selected cycle count divided by line frequency. One cycle corresponds to 20 milliseconds at 50 hertz and approximately 16.67 milliseconds at 60 hertz. Ten cycles correspond to 200 and approximately 166.7 milliseconds respectively.

These are aperture durations, not guaranteed readings per second. Confirm the detected or selected frequency and the actual instrument mode. A system moved between electrical environments should not retain a timing assumption merely because its software still requests the same NPLC number.

Understand what integer-cycle rejection can remove

An ideal rectangular average of an integer number of cycles of a sinusoidal interference signal has zero contribution from that signal. This explains the usefulness of line-related integration. Real rejection depends on frequency error, waveform content, analog coupling and the instrument's converter and filtering architecture.

Do not claim rejection of all electrical noise. A moving contact, changing thermal offset or intermittent force connection is not a stationary line-frequency sinusoid. Longer integration can reduce its visible peak without fixing the physical problem. Keep the distinction between rejecting a known interference pattern and averaging an unwanted event.

Use a delay study to identify settling bias

Hold the aperture fixed and vary the delay after a repeatable contact or scanner event. Plot the resulting resistance against delay while keeping excitation and specimen temperature consistent. A systematic movement with delay suggests that the measurement begins before the intended state is reached.

Then hold a sufficiently justified delay and compare apertures. This two-step study is easier to interpret than changing both settings together. If a long aperture merely approaches the late-time value because it includes more settled data, it has not demonstrated that the early part of the aperture was valid for the prescribed measurement.

Select the next comparison from the observed behavior

Keep mean movement separate from scatter. An aperture can reduce random scatter while shifting the mean if the specimen changes during the observation window. Compare both with the required uncertainty and physical measurement state.

Do not replace event detection with a smooth mean

A short interruption can have a small effect on a long average while being unacceptable for a sensor contact or in-process measurement. Determine whether the specification concerns stable resistance, interruptions, or both. If both matter, one acquisition setting may not serve both purposes.

A faster diagnostic record needs its own valid bandwidth and sampling definition. It is not enough to log the slow display more often. Conversely, multiple readings created by a digital moving average may share most of their input samples, so their apparent count does not represent independent observations.

Build throughput from the complete station sequence

An illustrative cycle might include 30 milliseconds for contact settling, 20 milliseconds of aperture and 15 milliseconds of control and transfer overhead. With those assumed nonoverlapping intervals, the cycle is 65 milliseconds, or about 15.4 cycles per second. Dividing one second by the 20-millisecond aperture alone would incorrectly predict 50.

Real instruments may overlap some tasks, perform occasional autozero operations or change range. Measure the distribution of actual cycle times and preserve exceptional cycles. A throughput target that requires skipping validity checks or thermal settling changes the measurement method, not merely its productivity.

Transfer settings with the measurement requirement

Record instrument identity, range, current, NPLC or aperture, frequency setting, delay, filtering, autozero and trigger behavior in the controlled method. Include the evidence that the selected settings still observe the required resistor state. A settings screenshot without the physical contact sequence is incomplete.

Recheck after changes to scanners, firmware, test current or trim-to-measure timing. A new setting can be appropriate, but its equivalence should be demonstrated rather than assumed from matching display digits. For quotation and test planning, provide the required stable resistance result and any separate interruption criterion so both can be addressed honestly.