Checking the balance and verifying the thermometer

Purpose

To find out how far each instrument sits from a reference you trust, to write it down, and to decide what to do about it — correct it arithmetically, replace it, or accept it and note it. An instrument with a known, recorded offset is more useful than one you have never checked, and it costs nothing to keep.

When to use it

At commissioning, and then at whatever interval you set for yourself. This course publishes no interval, because it has verified none for any domestic instrument. What it can name is the triggers: the balance has been moved, dropped, or has had its battery out; the thermometer has been dropped or has been in boiling water; a result surprised you; or a small mass is about to matter.

The third of the three commissioning checks — finding what your graduate actually delivers, by weighing the water it pours out — belongs beside these two and is taught in the same commissioning lab.

Before you start

You should already have read measurement and uncertainty, which is where the distinction between repeatability and trueness comes from and why only one of them needs a reference mass.

On the bench: the balance; a calibration mass if you have one; two trays of ice, crushed in a bag with a rolling pin; a beaker; a stirring rod; the thermometer, and the second one if you own both; the calibration log, which lives at the front of the notebook so that one instrument’s history is on one page.

Procedure

1. Stand the balance on a level, still surface, away from a window and a radiator. Switch it on and let it settle for a minute.
2. Zero it, then record the reading five minutes later without touching it. That is drift. Write it down even when it is 0.00, because “0.00 after 5 min” is a result.
3. Place the calibration mass and record the reading. Remove and replace it ten times, recording every reading. The spread of the ten is repeatability; the gap between their mean and the stated mass is trueness.
4. Repeat twice with the mass centred on the pan and twice near one edge. A large difference between the two positions is an instruction to centre everything you ever weigh.
5. Half-fill the beaker with crushed ice and add cold water until the ice is wet but still packed — a slurry, not ice floating in water.
6. Stir with the rod rather than with the thermometer. Immerse the thermometer to its normal depth, keep it clear of the sides and the bottom, and stir gently around it.
7. Wait for the reading to stop moving, which takes a minute or more, then record it. Repeat twice more, rebuilding the slurry each time.
8. If you own both a liquid-in-glass thermometer and a probe, read them against each other at the ice point and again at room temperature, and note whether they disagree by the same amount at both. A constant offset and a changing one are different faults.
9. Optionally, and only if you are content to handle boiling water, read the thermometer in a rolling boil, entering from the side rather than leaning over the vessel. Treat it as a check that the span is not badly wrong rather than as a calibration: the boiling point depends on atmospheric pressure and this course has verified no correction table.
10. Write an action against each instrument: correct arithmetically, replace, or accept and note.
11. Write the deviation on the instrument itself in permanent marker, so that the finding travels with the object rather than living only in a notebook you may not have open.
12. Write the calibration log lines.

How you know it worked

• Every reading is written down, including the ones that came out at zero.
• You can state your balance’s repeatability, and — if you had a mass — its trueness, and say which of the two the mass was needed for.
• The ice slurry was still packed at the end of the third trial, not floating in melt water.
• Each instrument carries its deviation on it, and the log carries the history rather than only the latest figure.
• There is an action beside each number rather than only a number.

If a step fails

Step 2, the reading will not settle. A draught, an unlevel surface or a warming instrument. Move it away from the window and the radiator, let it warm for five minutes, and re-zero.

Step 3, you have no calibration mass. You can still measure drift, repeatability and off-centre loading, which is three of the four findings. Trueness is the one that needs a reference you already trust; until you have one, write in the log that trueness is unmeasured and treat every concentration you calculate as relative rather than absolute — comparable with your own other batches, and not with a published formula. A coin is a serviceable repeatability object and a poor accuracy reference, and this course publishes no coin table, having verified no issuing authority’s figures or tolerances.

Step 3, the ten readings are identical to the last digit. Test whether the balance is re-measuring at all: switch it off and on between two readings, or add and remove a second small object each time. Note in the log that you had to. This course has verified no manufacturer’s statement about how any particular balance behaves, so treat that as a test to run rather than as a finding it has made for you.

Step 7, the ice-point reading drifts upwards while you watch. The slurry has melted into cold water, or the bulb is against the glass. Add crushed ice, stir, keep the bulb central, and take the reading only once it has been stable for thirty seconds.

Step 7, the ice point reads several degrees off and keeps doing so. That is a finding rather than a failure. Record it and correct or replace: one degree of processing temperature is worth about 9 per cent of development time on ILFORD’s published compensation chart, which is what makes a few tenths of a degree worth this much trouble.

Step 9, you would rather not handle boiling water. Do not do it. The ice point is the check that matters and this one is optional. If a glass thermometer does break in hot water, turn the heat off, leave the room until the steam clears, and lift the glass with a dustpan and a wet paper towel rather than with your fingers.

What to record

The calibration log, one page per instrument, at the front of the notebook: date, what was measured, the reference it was measured against, the reading, the deviation, and the action. Then the same deviation on the instrument itself. Ten entries on one page make a drifting thermometer visible, which is something no single check can do.