WEBVTT

1
00:00:00.000 --> 00:00:02.000
For a separate worked example,

2
00:00:02.120 --> 00:00:06.963
assume a fictional process owner has
approved temperature indication at 20,

3
00:00:07.083 --> 00:00:14.208
40 and 60 °C with error no greater than
±0.5 °C.

4
00:00:14.328 --> 00:00:16.328
These are teaching assumptions,

5
00:00:16.448 --> 00:00:20.800
not general GMP tolerances or real project
approvals.

6
00:00:20.920 --> 00:00:25.037
Requirement R-03 concerns indication
accuracy.

7
00:00:25.157 --> 00:00:28.275
It does not establish product temperature
uniformity,

8
00:00:28.395 --> 00:00:32.555
mixing performance or an acceptable
manufacturing process range.

9
00:00:33.175 --> 00:00:39.660
The supplier lists three component error
bounds of ±0.2 °C each:

10
00:00:39.780 --> 00:00:41.935
sensor, transmitter and display.

11
00:00:42.055 --> 00:00:44.526
If they can all act in the same direction,

12
00:00:44.646 --> 00:00:48.400
the simple worst-case sum is 0.6 °C.

13
00:00:48.520 --> 00:00:51.848
That exceeds the worked requirement of
0.5.

14
00:00:51.968 --> 00:00:55.086
Calling each component accurate would miss
the chain.

15
00:00:55.206 --> 00:00:57.915
This conservative teaching calculation is

16
00:00:58.035 --> 00:01:01.321
not a complete uncertainty budget or an
automatic

17
00:01:01.441 --> 00:01:03.851
rejection of every possible design.

18
00:01:03.971 --> 00:01:06.126
Other combination methods exist,

19
00:01:06.246 --> 00:01:08.422
but their assumptions, correlations

20
00:01:08.542 --> 00:01:13.619
and intended decision must be justified by
appropriate metrology expertise.

21
00:01:14.239 --> 00:01:17.695
The row identifies R-03, the criterion,

22
00:01:17.815 --> 00:01:22.253
sensor schedule S-07 revision B and the
calculation.

23
00:01:22.373 --> 00:01:24.741
Its decision is not demonstrated.

24
00:01:24.861 --> 00:01:28.338
The action is to propose a justified
measurement-chain

25
00:01:28.458 --> 00:01:32.213
design or revise the requirement through
authorized change,

26
00:01:32.333 --> 00:01:34.658
not quietly relax the criterion.

27
00:01:34.778 --> 00:01:38.426
Installation checks must confirm the
selected components.

28
00:01:38.546 --> 00:01:40.743
Operational testing must challenge

29
00:01:40.863 --> 00:01:43.569
the assembled chain under the approved
method.

30
00:01:43.689 --> 00:01:47.679
DQ plans this evidence; it does not invent
its results.

31
00:01:48.299 --> 00:01:51.605
Suppose the supplier proposes lower-error
components.

32
00:01:51.725 --> 00:01:55.459
Review their applicable conditions and
recalculate the chain;

33
00:01:55.579 --> 00:01:59.248
do not close the row merely because a new
datasheet arrived.

34
00:01:59.368 --> 00:02:02.141
If the revised design supports the
requirement,

35
00:02:02.261 --> 00:02:05.013
record the rationale and reviewer
decision.

36
00:02:05.133 --> 00:02:07.133
Keep the later installation

37
00:02:07.253 --> 00:02:10.432
and operational checks in the verification
plan.

38
00:02:10.552 --> 00:02:12.963
Distinguish design evidence adequate

39
00:02:13.083 --> 00:02:15.536
from installed performance demonstrated.

40
00:02:15.656 --> 00:02:19.624
The workbook contains a filled row and a
blank row for practice.