A balance is the simplest instrument in the laboratory and the one most often installed carelessly. It gets whatever bench space is left over, usually near a window or beside the centrifuge, and then people wonder why the last digit will not settle.
Almost every complaint we get about a balance turns out to be about its surroundings.
## Readability is not accuracy
The specification everybody compares is readability — the smallest increment the display shows. A balance reading to 0.001 g looks better than one reading to 0.01 g.
Readability is what it displays. Accuracy is how close that display is to the true mass, and repeatability is whether you get the same figure twice. A balance can show four decimal places and be wrong in the second.
Match readability to what you actually weigh. Weighing 100 mg of reagent to 1 per cent means you need to resolve 1 mg, so a balance reading to 0.001 g is right. Weighing 50 g of media powder does not need that resolution at all, and a more sensitive balance will simply be harder to use — more susceptible to draughts, slower to stabilise, and needing a draught shield you will find irritating.
The Aczet range covers the usual laboratory span: the CY 223 for analytical work needing milligram resolution, and the CG 602 or CG 302 for general preparation work where speed matters more than the third decimal.
## Where to put it
This is where the problems live.
On a solid surface. A dedicated stone or anti-vibration bench for analytical balances, not a bench somebody leans on. If a bench moves when you rest your elbow, it moves when the centrifuge runs.
Away from the centrifuge, the refrigerator compressor and anything else that vibrates. Away from the door, and away from any air conditioning vent or fan — moving air across the pan is the commonest cause of a display that will not settle.
Out of direct sunlight, and away from equipment radiating heat. A temperature gradient across the balance causes drift that looks exactly like a fault.
Level, checked with the built-in bubble rather than by eye, and rechecked after anyone moves it.
Give it time to stabilise after switching on. Analytical balances need warm-up, and a reading taken five minutes after power-on is not trustworthy.
## Calibration, internal and external
Many balances self-calibrate with an internal weight, either on demand or automatically when temperature shifts. That is useful and it is not sufficient on its own.
Internal calibration checks the balance against its own reference. If that reference has drifted, everything drifts with it and nothing looks wrong.
So verify externally with certified weights at a defined frequency, record the result, and act on it if it fails. For accreditation you need traceable weights with valid certificates, the verification record, and defined acceptance criteria. A balance calibrated annually by an external agency, plus your own periodic checks with certified weights between visits, is the usual arrangement.
Check at more than one point — a low mass near where you actually work and one nearer capacity. A balance can be correct at 100 g and off at 100 mg.
## Small habits that keep it honest
Weigh in the centre of the pan. Off-centre loading introduces corner-load error, which is worse on cheaper balances.
Handle weights with forceps, never fingers. Skin oils are enough to matter on an analytical balance, and fingerprints accumulate permanently.
Close the draught shield and wait for the stability indicator rather than reading the number as it settles.
Clean spills at once. Hygroscopic powders left on the pan absorb moisture and quietly change what you measure.
And keep the calibration record where the balance is, not in an office folder. The person using it should be able to see when it was last verified.