When a transformer oil result looks wrong, the sample is the first suspect — not the laboratory. ASTM D923 governs how insulating liquids are drawn from electrical apparatus and bulk containers so the bottle that reaches the bench actually represents the equipment it came from. It is the living North American counterpart to the IEC sampling track, and it underpins almost every downstream insulating-liquid test we report.
What it covers
D923 sets out how to take representative samples of new and in-service electrical insulating liquids — mineral oils, askarels, silicones, synthetic liquids and natural ester fluids — from transformers, circuit breakers, cables and other apparatus, as well as from drums, tank cars and tank trucks. It applies to liquids below a defined viscosity ceiling, which covers essentially every transformer and switchgear fluid in normal service.
The practice describes each sampling device and container in turn: glass bottles, metal cans, glass syringes, stainless steel cylinders, drum thieves, pressure-type samplers and the manifold used for cable feeders. It pairs those with step-by-step field procedures, flush requirements, container cleaning, photoprotection during transport, and the minimum information that must travel with every sample. Crucially, it also tells the crew what not to do — carrying explicit prohibitions on drawing samples from certain energised, low-volume apparatus where doing so would compromise safety or dielectric integrity.
For dissolved gas, water and particle work it hands off to the matching test methods, so D923 is the sampling reference those methods rely on rather than a duplicate of them.
Why it matters in practice
Most sampling errors are invisible until the data is trended. A container that lets gases escape or draws air in, a flush that was too short, or a sample taken in driving rain can all shift a result enough to trigger — or hide — an alarm. D923 codifies the choices that keep those errors out: which container suits which test, how much fluid to flush first, and how to seal and ship without losing the volatile species that drive a dissolved-gas diagnosis. Its container guidance is also where the familiar syringe-versus-bottle debate is settled in cautious, evidence-based terms rather than habit.
The older D923 companion practice for routine field sampling has been withdrawn without replacement; D923 is now the standing North American sampling reference, with no current alternative to fall back on.
How we use it
We treat D923 as the cross-check whenever a programme is written to North American conventions, when a laboratory or framework agreement cites ASTM methods, or when a client's downstream tests name D923-compliant sampling as a precondition. It sits alongside the IEC sampling standard in our reviews of field procedures, and where the two differ — container philosophy, cable-feeder method, label content — we reconcile them explicitly so a client under either regime gets a sample that is analytically defensible. When we audit a procedure or specify a kit, D923 is one of the documents the recommendation is anchored to.