One sensor reading describes one recorded observation. A trend needs a sequence that can be compared meaningfully. When readings arrive as separate screenshots, differences in time, units or operating conditions can be mistaken for a change in the component itself.
Compare the same field on the same basis
Keep the field name, unit and machine identity with each value. Two numbers labelled only 'pressure' may come from different measuring points. A temperature value in one unit should not be compared directly with a value in another.
The definition also needs to stay consistent. A requested value, a measured value and a calculated value are different kinds of information. The guide to measured and calculated quantities explains why similar screen positions do not establish a common source.
If a field name or display configuration changed between records, flag that change before combining the values. A continuous-looking line can conceal a break in meaning. Keep the underlying observations available so the reviewer can see how the comparison was assembled.
Preserve observation time, not just message time
A photograph forwarded this morning may have been taken last week. Record when the reading was observed where that is known, and distinguish it from when the file was received. If the original time is unknown, do not invent it from the order of messages.
Uneven intervals also matter. Three observations collected minutes apart and another collected the next day do not form an evenly spaced series. A chart should preserve those gaps instead of giving every observation the same apparent time interval.
A missing value is not zero. Leave it marked as missing, unavailable or not recorded, according to the actual evidence. Replacing a gap with zero creates a change that nobody observed and may distort the impression of the whole record.
Keep working conditions beside the readings
An idle observation and one taken during a working task describe different conditions. Record the context that is genuinely known. Do not assume constant load or an unchanged mode simply because the operator did not mention a difference.
Context is not an invitation to recreate a condition. Do not operate the machine, change settings or repeat a warning event to improve a data series. Use records available through the approved process and follow the equipment's required response to any alert.
Where several people supply readings, keep their sources visible. One entry may be a direct observation, another a recollection and a third a value copied from a stored report. Treating them all as equally precise measurements can overstate what the record supports.
A smooth graph can still hide uncertainty
Connecting widely separated points does not prove what happened between them. An average can also conceal brief changes within the recorded period. Preserve any known sampling or averaging description rather than presenting a summary as a continuous direct measurement.
If that information is not available, say so. The purpose is to communicate the limits of the evidence, not to make the graph look complete. Keep the original values alongside any derived summary so another reviewer can follow the calculation.
Note any known component replacement between observations. Values collected before and after that event should retain the boundary, even if the displayed field name and unit stayed the same.
Use a trend to frame a question, not name a failed part
A repeated change can help the responsible person decide what needs assessment. It does not, by itself, prove that a sensor, display or controller must be replaced. The same visible result may require context that the record does not contain.
Describe the observation neutrally: which field changed, across which times and under which recorded conditions. Keep the proposed explanation in a separate note. This prevents an early guess from being repeated later as a confirmed finding.
For routine transfer between operators, use the display handover record guide. Retain longer-term observations in the equipment history. If a part is eventually specified, connect that decision to its own confirmation rather than treating the chart as the purchase reference.
