Plain-language explanations of the measurements, sensors and numbers behind environmental monitoring.
An instrument will happily display a value anywhere across its full range, but the manufacturer only stands behind readings taken inside a narrower band, and outside that band there is no published statement of how wrong the number might be.
Read Article→The interval has to be short enough to resolve the fastest event that matters and long enough that the memory spans the whole deployment, and those two requirements are frequently in direct conflict.
Read Article→A live display answers only what the value is at the instant somebody is standing in front of it, while a record answers how long a condition lasted, how fast it moved, whether it repeats and whether two quantities moved together.
Read Article→A drive mode rolls off everything above a corner frequency so the fundamental reads cleanly, which is the right answer on a motor terminal and the wrong one whenever the high frequency content is the fault you are chasing.
Read Article→Low impedance mode puts a few kilohms across the input so that a capacitively coupled voltage collapses to almost nothing, while a genuine source barely notices the extra load.
Read Article→An average responding meter multiplies by a constant that is only correct for a sine wave, so on the switched output of a variable frequency drive it reports a figure that is arithmetically wrong rather than slightly optimistic.
Read Article→An averaging meter rectifies the signal, takes its mean and multiplies by the fixed factor 1.1107 that is correct only for an undistorted sine, so every other shape is reported with a predictable error.
Read Article→Multiplying the two sides of a grid produces a number that grows as the square of the thing buyers care about, so a fourfold pixel count is only a twofold improvement in how finely the scene is sampled.
Read Article→A detector reports one radiance weighted average for everything inside its cone, so a part smaller than that cone reads cold, and computing extra pixels from those averages cannot recover what was never sampled.
Read Article→The native array size appears in one of four places when it appears at all, and the method for extracting it works the same way on any instrument whose decisive hardware figure a listing would rather not print.
Read Article→A super resolution figure describes the pixel grid a thermal camera writes out, not the number of detector elements that produced it, and only some listings publish both numbers.
Read Article→Ingress protection is a voluntary claim that manufacturers publish when they have it, so silence in a listing is evidence of absence rather than a gap in the paperwork.
Read Article→