Scanned Insulation Tests: Track Bias Age Separately for Every Channel

Measurement purpose

Verify that each scanned insulation observation represents the specified channel-specific bias duration and valid measurement window.

Specimens and conditions

Channel and specimen identity

Keep the physical insulation path, fixture position and specimen identity attached to each channel timeline.

Prior electrical state

Record prior bias, discharge, rest and any reconnection that changes the specimen's voltage history.

Equipment and records required

• Bias and switching observation: Use the actual connection diagram and suitable voltage/status observations to identify when each path receives and retains its defined bias.
• Timed current acquisition: Record aperture start/end, channel selection and range/settling state on a consistent clock within the approved electrical-safety arrangement.

Method sequence

1. Map switching behavior

   Identify which connections change during channel selection and whether each specimen remains continuously biased.

   Record: Bias/meter topology and channel event definitions.

2. Build individual timelines

   Record voltage-valid and acquisition events for each specimen rather than assigning only a global elapsed time.

   Record: Per-channel bias-age ledger.

3. Validate the schedule

   Compare the actual aperture and bias duration with the approved observation requirement.

   Record: Timing-valid result or explicitly incomplete acquisition.

Decision and uncertainty

A channel result is comparable at the specified bias age only when its actual voltage history and measurement window satisfy the method; global elapsed time alone cannot establish this.

Clock alignment, voltage-event uncertainty, current change with time and switching disturbance contribute separately from current-reading uncertainty.

The measurement owner defines observation timing and missed-window disposition; qualified personnel control energized switching, guarding and discharge.

Traceable outputs

Distinguish bias switching from meter switching

Draw the source, specimen paths, scanner contacts, current meter and guard connections. In one arrangement every specimen remains biased while the instrument selects a current path. In another, selection applies voltage to only one specimen at a time. These arrangements do not create equivalent electrical histories merely because they share a channel count and scan rate.

Changing the meter connection can also disturb a supposedly continuously biased path if the return or guard moves with it. Establish continuity of the intended voltage from the actual circuit and suitable observations. Do not infer uninterrupted bias from a source that remains switched on somewhere upstream.

Define the electrical event that starts each channel's age

For channel j, record the command time, actual voltage transition and the event selected by the measurement method as its time origin. A method may reference ramp start or entry into a defined voltage band; it must state which. The examples below use the latter for clarity without prescribing that convention universally.

Keep this event separate from channel selection and conversion completion. Software scheduling or relay delay can separate them by a meaningful interval. If voltage leaves the permitted state, record the interruption and apply the method's rule before treating a later reading as part of the same uninterrupted exposure.

Expose the error created by one global timestamp

Suppose three hypothetical specimens enter their defined voltage state at 1 s, 4 s and 7 s after the overall sequence starts. Observing all three at global time 10 s would give bias ages of 9 s, 6 s and 3 s, not 10 s for every specimen. Different time-dependent insulation currents could then appear to be a position effect.

For a required five-second observation age, their nominal observation times would instead be 6 s, 9 s and 12 s, subject to the actual aperture and method tolerance. The scheduling problem is therefore per-channel. A faster overall scan does not solve it if it continues to assign the wrong electrical age to each result.

Assign an interval rather than an unexplained instant

A current reading may integrate over an aperture or combine several samples. Record its start and end relative to the channel's bias event. A displayed timestamp can refer to the end of conversion or receipt by the host, neither of which necessarily equals the effective observation time.

If a reading averages from age 4.8 s to 5.2 s, calling it a five-second point is a method choice that must be justified against current variation and timing requirements. Avoid moving the aperture after viewing the results to obtain a favorable value. Keep any allowed observation tolerance explicit before acquisition begins.

Do not accumulate disconnected intervals into one exposure age

A scanner that energizes a specimen for 100 ms on each visit has not necessarily applied one uninterrupted 60 s bias because the overall sequence has run for a minute. Between visits, the specimen may float, discharge or retain an uncontrolled voltage. Its history must be described by the actual electrical states.

For a requirement based on continuous bias, redesign the acquisition arrangement or schedule so that the required exposure is genuinely supplied. If an intermittent-bias method is intentionally used, define it as such and validate the resulting quantity. Prior charging and relaxation can affect later observations, so the sum of visit durations is not automatically equivalent to one continuous interval.

Budget meter settling without resetting a verified bias history

In a genuinely continuously biased arrangement, selecting a meter channel can introduce acquisition settling while the specimen's bias age continues to advance. Keep the settling interval separate from the bias timer. Restarting the exposure clock every time the meter selects a channel would misdescribe this topology just as using one global timer misdescribes sequential energization.

Verify that the planned switch-to-read delay fits before the required observation window. If the meter cannot settle within that schedule, the result is a timing limitation of the arrangement. Do not hide it by reporting an unsettled value or relabeling a late reading as an on-time observation.

Relate clock uncertainty to the changing current

Near the specified observation time, a small timing uncertainty contributes approximately the local current slope multiplied by that timing uncertainty. For an independently assumed slope magnitude of 0.02 nA/s and a 0.5 s time-origin uncertainty, the corresponding first-order current contribution is 0.01 nA. This is separate from the instrument's current resolution.

Use a local slope supported by the relevant trace, not one fitted across a voltage ramp or range transition. A strongly curved transient can make a linear timing estimate inadequate. If the required window is missed, follow the predefined invalid-run or repeat rule rather than interpolating across an interruption that changed the specimen's history.

Challenge scan-order dependence with matched ages

Compare forward and reverse channel orders while preserving the required bias-age window and electrical conditioning for each specimen. If the apparent position difference changes when only the schedule changes, investigate timing, switching disturbance and instrument drift before assigning the difference to the ceramic or glass system.

The final record should allow another reviewer to reconstruct the age of every accepted observation. Retain missing windows and interrupted channels alongside successful ones, with the actual specimen identity and discharge history. This completes the multi-channel timing task; diagnosing dielectric charging mechanisms and choosing insulation acceptance limits remain distinct reviews.