KEYSTONEIndustrial Services
Interactive trainer

pH measurement

A pH reading is really a millivolt measurement: a glass electrode makes about 59 mV per pH unit against a stable reference, pH 7 at zero. Six loops, four reads — grab sample, raw mV, reference junction, 2-buffer cal — including the pairs that look identical until you read the millivolts: a fouled reference versus an aged electrode, and a coated bulb versus a cracked one.

By Francisco Padilla Interactive Technician level

Before you start — job brief

A pH loop is a millivolt measurement, not a direct read. A glass measuring electrode develops about 59 millivolts per pH unit against a stable reference electrode, with pH 7 sitting at 0 mV — so pH 4 is about +177 mV and pH 10 about −177 mV. Everything that goes wrong shows up as a shift, a shrink, or a stall in that millivolt signal. The only truth you have is a fresh grab sample on a handheld meter: read the loop against it, not against yesterday.

  • The glass electrode makes ~59 mV/pH; pH 7 = 0 mV, below 7 positive, above 7 negative
  • A fouled reference junction is the #1 pH failure — it shifts the whole curve (an offset)
  • Read four things: indicated pH vs a grab sample, the raw mV, the reference junction, and a 2-buffer cal
  • Slope errors are right at 7 and wrong away from it; offset errors are wrong everywhere — including at 7

A pH reading is a millivolt measurement in disguise

A pH loop does not read pH directly. A thin glass membrane on the measuring electrode develops a small voltage that depends on the hydrogen-ion activity of the solution, and the analyzer measures that voltage against a second, stable reference electrode. The relationship is close to linear: about 59 millivolts per pH unit at 25 °C, pivoting around pH 7 at 0 mV. Below 7 the signal goes positive, above 7 it goes negative — so pH 4 sits near +177 mV and pH 10 near −177 mV. Every pH fault is really a fault in that millivolt signal: an offset that shifts it, a shrunken slope that flattens it, or a dead cell that collapses it. Read the raw mV alongside the pH and the electrode tells you which.

The reference junction fails first

The most common pH failure is not the glass — it is the reference. The reference electrode has to maintain a fixed potential through a small porous junction that lets its internal fill solution (usually KCl) make electrical contact with the process. Plug that junction with solids, coat it, or poison it with a reacting ion, and the reference potential drifts. Because the analyzer measures the glass against the reference, a drifting reference shifts the whole curve by an offset — the reading is wrong at pH 7 and wrong everywhere else, and it usually wanders. The tell is the reference check itself (no flow, dry or plugged junction) and the raw mV being wrong even on a near-neutral sample. The cure is to refresh the reference — refill the KCl, clear or replace the junction, and move to a double-junction cell if the process attacks it.

Slope, offset, and why pH 7 is special

A two-buffer calibration measures two things: the offset (where the curve crosses, ideally 0 mV at pH 7) and the slope (how many millivolts per pH unit, ideally about 59, reported as a percentage of that ideal). They fail differently. An offset error — a fouled reference, a bad zero — moves the reading by the same amount everywhere, including at pH 7. A slope error — an aged, worn glass membrane down at 76% — pivots around the pH 7 isopotential: the electrode still reads correctly at 7 and gets worse the further you move away from it. That is why an aged electrode can pass a one-point check at buffer 7 and still be a full unit off in a pH 4 or pH 10 process. Watch the slope at every calibration; a slope that keeps falling is a membrane wearing out, and no recalibration brings it back — replace it.

Field note — the two pairs that look alike

Two off-scale readings, two different faults. A fouled reference and an aged electrode both read wrong, and you tell them apart by asking where they are wrong: the reference offset is wrong even at pH 7 (check the mV at a near-neutral sample and the junction condition), while the aged electrode is right at 7 and wrong only away from it (check the calibrated slope). The other pair is two sluggish probes reading toward 7: a coated bulb still develops a real, finite millivolt signal — attenuated and slow, but alive — while a cracked bulb collapses to about 0 mV with low impedance and is completely dead. Finite-but-slow mV means clean it; near-zero mV at low impedance means replace it. In both pairs the raw mV, not the pH number, is what separates them.

Temperature belongs in the loop

The millivolts-per-pH slope is itself temperature-dependent — it rises with temperature — so the same electrode reads differently hot and cold. Automatic temperature compensation corrects for this, and like a slope error its effect is zero at pH 7 and grows away from it. Turn temp comp off (or feed it the wrong temperature) on a hot or cold process far from neutral, and the reading drifts by a predictable amount that looks exactly like an aged electrode — except the glass is fine. On a caustic line at 80 °C reading a real pH 10, temp comp off can cost most of a pH unit. It costs nothing to leave on with a temperature element in the sample, so leave it on and check it before you condemn a healthy electrode.

The discipline

Four reads name every fault in this trainer. Compare the indicated pH to a fresh grab sample — the only ground truth you have. Read the raw mV, sign and magnitude, against the ~59 mV/pH the electrode should make. Inspect the reference junction — flow and fill — because it fails first. And run a two-buffer calibration for slope and offset with temperature compensation on. Offset wrong everywhere is a reference; right-at-7-wrong-away is a slope, whether from an aged membrane or temp comp off; near-zero mV at low impedance is a cracked bulb; a finite but sluggish signal is a coated one. Read all four, compare each to what the physics says it should be, and the loop tells you exactly what it needs.