Response time
The IEC 60846-1:2009 §8.9.2 response time evaluation — measuring t₉₀ after a step change in dose rate, the limit that tightens at high dose rate, and the mandatory check 60 seconds later.
Last reviewed
A dose rate indication lags the actual field. The instrument averages over an interval to reduce the statistical variation of the counting process, and that averaging both stabilises the indication and delays it. Response time quantifies the delay.
The practical consequence: after entering a field, the indication continues to reflect the previous field for the duration of that delay. A thirty-second response time means the instrument reports the new field only after the operator has been in it for thirty seconds.
Measurement procedure
Apply a step change in dose rate — as close to instantaneous as your setup allows, within about a second — between two levels that differ by at least a factor of ten. Use a shutter or a source transfer mechanism; a progressive approach to the source does not produce a step.
Then measure t₉₀: the time from the step until the indication reaches 90 % of the way from its initial value Gᵢ to its final steady value G_f.
The 90 % point is:
Entering Gᵢ and G_f into the tool computes that threshold in advance, so the target indication is known before the measurement rather than derived afterwards from a recording.
Measure in both directions. Rising and falling response times frequently differ, depending on how the averaging is implemented. An instrument that responds quickly to an increase and slowly to a decrease continues to indicate an elevated dose rate after the operator has left the field, which reduces confidence in the indication in routine use.
Dose-rate-dependent limit
The limit is not a single value:
- for final dose rates between about 1 µSv/h and 10 mSv/h, the limit is the value you set — ten seconds by default;
- at 10 mSv/h and above, the limit drops to two seconds.
The basis is that at high dose rate the event rate is high enough that a long averaging window is not required to control statistical variation, while the consequence of a delayed indication is greater. The tool applies the bracket per row automatically from the G_f you entered, so a test set that spans the threshold gets judged correctly without you having to remember which rows are which.
The 60 s settling requirement
t₉₀ alone is not enough for a verdict. Sixty seconds after the step, the indication must have settled to within a tolerance of the final value — ten percent by default.
This requirement detects a failure that t₉₀ alone does not: an instrument that reaches 90 % within three seconds and then continues to drift. Its t₉₀ is compliant and its indication is not settled. Without the 60 s check it would be recorded as a pass.
A row with no 60-second entry gets no verdict. The tool withholds it rather than judging on t₉₀ alone. This is a case where an incomplete measurement and a passed measurement must not look alike in a report.
Reporting requirements
The step is part of the result. Record Gᵢ, G_f and the method used to produce the step: a transition that took four seconds to establish makes a two-second limit unmeasurable, and a reviewer must be able to see that from the report.
Where the instrument has a selectable time constant or a fast/slow mode, record the setting used, and test both where practical. A response time measured in one mode does not characterise the instrument as delivered in the other.
Interpretation
t₉₀ inversely proportional to dose rate
Expected behaviour for an instrument whose averaging window adapts to the counting statistics: at low dose rate the required number of events takes longer to accumulate. This is not a fault.
t₉₀ long and constant across dose rates
Indicates a fixed filter rather than an adaptive one. At high dose rate it contributes delay without reducing statistical variation. Worth recording even when the row passes, since it is frequently adjustable in firmware.