Zero Error, Twenty Blocks and Twenty-Four Hours
Twenty Fitter trainees are finishing a file-and-measure job: a rectangular block hand-finished to 30.00 ± 0.05 mm thickness. The section owns two 25–50 mm micrometers with 0.01 mm graduations. Before marking, you check the one in use against its 25 mm setting standard: with the anvil closed on the standard, it reads 25.03 mm. Trainee A submits readings for three blocks: 30.06, 30.09 and 30.02 mm, each taken as a single reading. When you re-measure the second block yourself, three times, you get 30.07, 30.05 and 30.09 mm. The second micrometer is in the tool room, and the tool room reopens after the assessment. The assessment is in 24 hours and all twenty blocks must be booked as accepted, reworked or rejected before then.
1. Correct trainee A's readings and decide the status of each of her three blocks. Show the reasoning.
Read the worked answer
The zero error. With the standard in place, the micrometer reads 25.03 mm instead of 25.00 mm. It reads 0.03 mm high, so this is a positive zero error of +0.03 mm. Every reading includes that 0.03 mm, therefore:
Corrected size = observed reading − 0.03 mm
Trainee A's blocks. Tolerance limits: 29.95 to 30.05 mm.
- Block 1: 30.06 − 0.03 = 30.03 mm → inside limits → accept.
- Block 2: 30.09 − 0.03 = 30.06 mm → 0.01 mm above the upper limit → rework, remove 0.01 mm by draw filing and re-measure. Do not scrap: 0.01 mm is well within what a file and a check pass can recover on a 30 mm block.
- Block 3: 30.02 − 0.03 = 29.99 mm → inside limits → accept.
The trap in this question. Trainee A's raw readings look like two rejects and one accept. The corrected set is two accepts and one rework. An instructor who books against raw readings will send good blocks back for rework and, on the block at 29.99, may let a trainee take another cut and put it under the lower limit. That is how a zero error turns into a scrap part.
Sign discipline to teach. Positive zero error means the instrument reads high, so subtract. Negative zero error means it reads low, so add. Write the rule under the sign the instrument actually shows, and check it against the setting standard every time the micrometer is picked up - not once a week.
2. You re-measured block 2 three times and got 30.07, 30.05 and 30.09 mm. What does that tell you, and what do you do about it in 24 hours?
Read the worked answer
What it tells you. The spread is 30.09 − 30.05 = 0.04 mm, on a part whose total tolerance band is only 0.10 mm. Measurement variation is eating 40% of the tolerance. The usual rule of thumb is that gauge or measurement uncertainty should be no more than about 10% of the tolerance band; 0.04 mm against 0.10 mm is four times that. The measurement system is not capable of sorting these blocks safely.
Consequence. Every part sitting within about 0.02 mm of a limit - and block 2, at a corrected 30.06, is one of them - has an undecided status. Booking it either way is a coin toss recorded as a decision.
Actions in 24 hours, in order.
- Sort the tool, not the parts. Use the known zero error as a correction (+0.03, subtract) if the micrometer has no adjustment; if it has a zeroing spanner, reset it and re-verify against the standard. Record the correction on a card at the bench either way.
- Attack the technique, which is the real cause of the 0.04 mm spread. Three separate faults produce a spread like that: tightening on the thimble instead of the ratchet, dirty or burred anvil faces, and measuring a warm part straight off the file. Brief all twenty trainees for ten minutes: clean the anvils with paper, use the ratchet until it slips three times, take three readings and record the mean, let the block cool.
- Re-measure only the near-limit parts. Anything outside 29.97–30.03 mm corrected gets three fresh readings and a mean. Parts comfortably inside need no rework of the record.
- Book borderline parts as 'conforming, verified by re-measurement', with the readings attached. Do not book a coin toss as an accept.
What you do not do is wait for the second micrometer. The variation is in the method, and a second instrument used the same way will give the same spread.
3. Write the three teaching points you would put on the board, and say how you would test that they have been understood rather than memorised.
Read the worked answer
The three points.
- Check the zero before you trust the number. Every time: close on the standard, read it, write the error on the job card. Positive error - subtract. Negative error - add.
- The ratchet is not decoration. Three slips of the ratchet, clean anvils, part at room temperature. A micrometer tightened on the thimble measures the trainee's grip strength, not the part.
- One reading is not a measurement. Three readings and a mean, and if the spread is wider than a fifth of the tolerance, stop and find the cause before you book anything.
Testing understanding, not memory. Do not ask them to recite the three points; they will.
Set a practical check instead. Hand each trainee an unknown block and the micrometer with its known +0.03 error, already mounted. Ask for a written result in three lines: the error and its sign, the three readings, and the corrected mean with a pass or fail decision. Then swap blocks with a neighbour and repeat.
What tells you it has landed:
- They write the correction with the right sign without being prompted.
- Their three readings agree within 0.01 mm.
- Their accept/reject decision matches yours on a block you have deliberately set 0.01 mm inside the limit.
What tells you it has not: they get the arithmetic right but their three readings still spread by 0.03 mm. That is memory, not skill - and it is exactly the condition that produced the 0.04 mm spread on block 2 in the first place. Send them back to the ratchet, not to the formula.