4 data logger clock drift checks to perform when power generation is low

When a low power output is suspected, the first things usually considered are dirty panels, equipment failures, shading, wiring anomalies, or output control. However, if on-site data review finds no clear abnormalities, one often overlooked item to check is a time shift in the data logger. Even if the generation equipment itself is operating normally, if the recorded timestamps are off, it becomes difficult to correctly interpret the relationships among solar irradiance, temperature, power output, shutdown history, and alarm history. As a result, periods that are actually generating power may appear to be recorded at different times, generation during clear-sky conditions may appear low, or periods that are not stopped may be mistakenly identified as stoppages.

This article explains four checks to verify data logger time drift for practitioners investigating "low power generation". It is not simply about synchronizing clocks; from an operational perspective it organizes how to confirm the reliability of data when isolating causes of reduced generation and where to look to reduce misclassification.

Table of Contents

• Why data logger clock drift can appear as a decrease in power generation

• Verify that the current time on the monitoring screen matches the local time.

• Check the shift in the peak positions between solar irradiance and power generation.

• Confirm the time alignment between the stop history and the alarm history.

• Confirm the cutoff time between daily and monthly aggregations.

• Records to retain before correcting time discrepancies

• Summary

Why a Data Logger's Time Drift Can Appear as a Drop in Power Generation

When checking solar power generation, many operations staff rely on the numbers shown on monitoring screens and reports. By laying out daily generation, generation by time of day, output per power conditioner, solar irradiance, temperature, shutdown logs, and alarm histories, they can narrow down the cause to some extent. However, if the reference timestamps are misaligned, the same data can appear to mean something different.

For example, even though the generation peak actually occurs around noon, if the data logger's clock is one hour behind, the generation peak will appear shifted into the afternoon. If the clocks of the solar irradiance recording device and the power generation recording device are offset differently, the graph may look unnatural—showing low power when irradiance is high, or high power when irradiance is low. Looking only at this condition, one might mistakenly conclude panel soiling, shading, output curtailment, or equipment malfunction.

When investigating the causes of low power generation, it is important to confirm that the datasets being compared are aligned on the same time axis before suspecting equipment abnormalities. Even when looking at power output alone, time consistency is indispensable when making comparisons with the previous day, the same month in the previous year, solar irradiance, or weather conditions. If timestamps are shifted, you may end up evaluating morning data under afternoon conditions or including downtime in another day’s generation.

One point to pay particular attention to is that a time offset does not necessarily appear large. A deviation of a few minutes may have little impact on daily power generation, but it can cause problems when checking downtime records or alarm logs. Conversely, an offset of 30 minutes to several hours can disrupt the correlation between time-of-day generation graphs and solar irradiance, making it easy to misidentify the cause of reduced generation. If the offset crosses dates, the daily aggregation itself will no longer match reality, which can affect monthly reports and inspection records.

The time on a data logger, even if correctly set at installation, can drift due to improper time synchronization settings, changes in the communications environment, recovery after power outages, inaccuracies in the internal clock, or setting changes during maintenance. In the field, a time drift does not necessarily stand out as an alarm. Because graphs appear as usual on the monitoring screen and numerical values are still being collected, it can be difficult to notice from the appearance of the data alone.

When you receive a report of low power output, rather than immediately proceeding to equipment replacement or an on-site inspection, first check the time axis of the data to reduce unnecessary troubleshooting. A data logger time offset does not itself reduce the generation performance of the equipment. However, it can distort the data used to assess a decline in power output. In other words, it is not an abnormality of the power generation equipment but a misalignment in the foundation of the diagnosis.

Verify that the monitoring screen's current time matches the local time

The first thing to check is whether the current time displayed on the monitoring screen matches the actual local time at the site. This is a basic check, but it is sometimes put off during investigations into reduced power output. Even if the monitoring screen appears to be updating with the latest data, if its time display does not match the local time at the site, subsequent analysis results will be misaligned.

When checking, first look at the last retrieval time of the data currently displayed. It's important to confirm not just the time shown at the top right of the screen, but when the generated power and equipment status data were actually retrieved. The screen display time, data acquisition time, data storage time, and server-side reception time may be handled separately. In practice, confusing which time indicates what can lead to mistaking communication delays for time discrepancies.

In the case of a communication delay, the recorded timestamps of the data themselves are correct, but their reflection on the monitoring screen is delayed. On the other hand, in the case of a data logger time shift, the timestamps recorded are themselves offset from the local time. Both give the impression that "the latest data appears delayed," but their meanings are different. If it's a communication delay, the data may catch up later. If it's a time shift, the entire time axis of the accumulated data may remain shifted.

When verification can be done on site, directly compare the local time with the data logger unit's time. When verifying remotely, have the on-site person share the local time, current weather, and generation status, and cross-check these against the state shown on the monitoring screen. For example, if it is morning on site and generation is just starting to rise but the monitoring screen shows an output trend like early dawn, suspect a delay in data being updated or a time discrepancy. Conversely, if it is evening on site and generation is falling but the monitoring screen appears to show the midday peak continuing, you also need to check the time axis.

What to watch out for here is that finding a time offset does not mean you should immediately change the settings. You need to first record how much the time is off, when it started, and which data are affected. If you fix the time first, it can become difficult to determine how far back the historical data were affected by the offset. In investigations of reduced power generation, both correcting the cause and preserving the basis for your conclusions are important.

Also, care must be taken in how local time and reference time are handled. In normal domestic operations these are often managed according to local time, whereas some devices or aggregation systems may store timestamps using a different reference time and convert them when displaying. Even if the display looks correct, the times in exported data may be represented in a different form. When exporting to reports, CSVs, etc. for analysis, it is safer to confirm whether the times shown on the monitoring screen and the times in the exported data have the same meaning.

Determining that power generation is low is, in most cases, done by comparison — lower than yesterday, lower than a day with the same solar irradiance, lower than neighboring systems, or lower than the expected value. As the starting point for those comparisons, confirming that the current time is correct is an unglamorous but important step. If you skip this, even if subsequent detailed analyses look neat, you'll end up investigating causes based on a misaligned baseline.

Confirm the shift in peak positions between solar irradiance and power generation

Next, what we want to check is whether the peaks of the solar irradiance and power-generation graphs are in a natural relative position. In photovoltaic power generation, power output is strongly influenced by solar irradiance. Strictly speaking, air temperature, module temperature, azimuth, tilt, shading, equipment characteristics, output control, and so on also play a role, but as a general tendency on clear days, power output rises during the hours when solar irradiance is high. Therefore, if the peak of irradiance and the peak of power generation are greatly displaced, it can be an indication of a time shift.

For example, if the solar irradiance peak is around noon but the power generation peak appears late in the afternoon, the timestamps on the power data may be delayed. Conversely, if the power generation peak occurs in the morning while the irradiance peak is shown in the afternoon, the timestamps for the irradiance data may be shifted. When staff feel the power generation is low, many look only at the power graph, but checking its relationship with irradiance makes it easier to distinguish equipment anomalies from data anomalies.

For this check, it is practical to choose a sunny day with as stable weather as possible rather than a cloudy or rainy day. On days with frequent cloud passages, both solar irradiance and power output fluctuate repeatedly, making it difficult to determine whether deviations are due to time shifts or weather changes. On a sunny day, the peaks in the graph appear relatively smooth, making it easier to find shifts in the peak position and in the timing of the rise and fall.

However, you should not immediately conclude a time shift just because the peaks of solar irradiance and power output do not perfectly match. The timing of the generation peak can change depending on the orientation and tilt of the PV system. East-facing systems tend to show stronger output in the morning, while west-facing systems tend to show stronger output in the afternoon. Also, during periods when panel temperature rises, output may not increase much even when irradiance is high. Therefore, what you should look for is not a perfect match but whether an unnatural time lag persists.

If a time shift is suspected, confirm whether the same pattern appears across multiple days. If only one day shows a misalignment between irradiance and power generation peaks, other factors such as weather changes, partial shading, equipment shutdowns, or communication delays may be involved. However, if the same degree of time difference occurs consistently even on clear (sunny) days, it is worth checking the time settings of the data logger and related measurement devices. In particular, if the recording times of the pyranometer and the power output are managed by separate systems, you need to verify separately which one is misaligned.

Also, when judging low power generation by looking at power generated per unit of solar irradiance, the effect of time misalignment tends to become large. If the periods of high irradiance and high generation are shifted, you end up dividing or comparing data that do not actually correspond to each other. As a result, you may see unnatural evaluations such as “no generation despite clear skies” or “high generation efficiency despite low irradiance.” These are apparent anomalies caused not by equipment performance but by the compared datasets’ time axes being out of sync.

For practical on-site assessment, it's easiest to check three patterns: the morning rise after sunrise, the midday peak, and the evening decline. If you notice tendencies such as the start of morning power generation being clearly delayed relative to the increase in solar irradiance, the peak shifting by a consistent amount each day, or the evening power drop not matching the decrease in irradiance, proceed to check for time offsets. By looking at the overall shape of the graph as well as the detailed numbers, you can reduce missed anomalies.

Verify the time alignment of stop history and alarm history

When power generation is low, checking shutdown and alarm histories is essential. If records of power conditioner shutdowns, grid-side abnormalities, protection operations, communication errors, or measurement anomalies remain, it becomes easier to narrow down the cause of reduced generation. However, if the data logger's clock is out of sync here as well, there is a risk of misinterpreting the meaning of shutdowns and alarms.

When checking the stop history, confirm whether the time of the decline on the power generation graph and the occurrence time in the stop history are consistent. For example, if generation suddenly falls from 10:00 AM but the stop history is recorded in a different time period, it may appear at first glance that the shutdown and the drop in generation are unrelated. However, if the data logger or the log timestamps are shifted, they may actually indicate the same event. Conversely, if generation is normal during the period when an alarm was issued, you may be misattributing the alarm to a drop in generation.

In alarm history, not only the time of occurrence but also the recovery time is important. By checking whether power generation fell between the occurrence and recovery times and whether generation returned after recovery, you can assess the impact of the alarm. If there is a time offset, it may appear that power output had dropped before the alarm was issued or that output remains low even after recovery. If you determine the cause in this state, you may mistake a temporary stoppage for a persistent power generation fault.

Also, when a data logger is collecting information from multiple devices, the internal clocks of each device may not be synchronized. Even if the data logger’s clock is correct, if the historical timestamps on the connected devices are offset, the order of records on the monitoring screen can appear unnatural. When comparing multiple power conditioners, if occurrence times are recorded inconsistently even though they should be the same system event, you need to check each device’s clock management and the data acquisition mechanism.

When checking the time alignment between stoppage histories and alarm histories, it is easier to make a judgement if you check not only the generated power but also the AC-side output, the DC-side voltage and current, the communication status, and the total output of the entire facility. For example, whether only a single device’s alarm timestamps are shifted or the records for the whole facility are shifted changes which items should be suspected. If the facility’s total generated power and the outputs of individual devices disagree within the same time period, the aggregation-side timestamps and the data acquisition interval should also be checked.

In investigations into declines in power generation, it is common to treat the timestamps in the history as "the facts themselves." However, historical timestamps are also information recorded by the measurement system and depend on clock settings and synchronization status. Even if alarm names and shutdown details are correct, if the timestamps are off you can mistake the order of cause and effect. Whether the power output fell first or the alarm was triggered first, whether the system recovered after shutdown, or whether output did not return even after recovery — all of these judgments depend on the consistency of the time axis.

What you should pay particular attention to are histories recorded late at night or early in the morning. Since solar power generation produces no output at night, alarms or communication errors recorded at night are not necessarily directly related to daytime production declines. However, if the timestamps are significantly offset, events that actually occurred during the day may appear to have happened at night. Before deciding that a nighttime record is unrelated, it is important to check its time relationship with other data.

Confirm the cutoff time between daily and monthly aggregations

An easy-to-overlook data logger time offset is the boundary time between daily and monthly aggregations. When reports of low generation arise, daily or monthly generation totals are often used as the benchmark. However, if the data logger’s clock is off, part of a day’s generation may be recorded as belonging to the previous or next day. If the time offset occurs at the end or start of a month, it can also affect how the monthly generation appears.

For example, if the data logger's clock is several hours fast, power generation during time periods near the date boundary may be aggregated as the following day's data. Conversely, if the clock is behind, morning generation may be treated as belonging to the previous day. Even if the total daily generation does not change significantly, when viewed day by day some days may appear low and others high. If this is misinterpreted as uneven generation by the equipment, it can lead to unnecessary investigations into the cause.

In monthly aggregation, the boundary between the last day of the month and the first day of the next month is important. If a time offset causes part of the generation that spans months to be recorded in a different month, the monthly report figures may not match the on-site reality. In particular, when inspection reports, internal management, or variance checks against generation plans are based on monthly generation, you need to verify that the aggregation period boundaries are correct. What has been treated as a low-generation month may in fact be only because some generation was shifted into the neighboring month due to a boundary misalignment.

When checking daily totals, it's effective to verify whether the sum of the time-of-day data is consistent with the daily display. If the time-of-day graph shows generation but it doesn't appear to be reflected in the daily total, check the reference time and cutoff time used for aggregation. Because the monitoring screen automatically displays daily totals, it's important not to assume those values always represent the sum from local 0:00 to 24:00. Depending on the system or settings, the handling of save time, display time, and aggregation time may differ.

Also, you need to check the data acquisition interval. Depending on whether data is stored at 1-minute, 5-minute, 10-minute, 30-minute intervals, etc., the way timestamp misalignment appears will change. When data is stored at short intervals, it is easier to inspect misalignments in detail, but the data volume is large and you may mishandle it during aggregation. When data is stored at long intervals, the precise times of brief stoppages and recoveries may appear rounded. By understanding both timestamp misalignment and the aggregation interval, you can interpret daily and monthly declines in power generation more accurately.

When you find a day with low power generation, don't just look at that day alone—check the generation for the previous and following days as well. If the target day is unusually low while the day before or after is unnaturally high, it's worth considering not only weather or outages but also a misalignment of the date boundary. Of course, actual weather changes or output control can cause day-to-day differences, but by looking at time-of-day data you can confirm which day the generation has been allocated to.

Daily and monthly aggregates are easy-to-understand metrics often used in field reports. However, the more straightforward a metric is, the less visible its underlying aggregation conditions become. To safely conclude that power generation is low, it is important to check not only the daily or monthly figures but also the time range from which those figures were produced. Data logger clock drift is precisely a factor that makes that aggregation range hard to see.

Records to keep before correcting time drift

When you find a clock offset in a data logger, you might be tempted to correct it to the right time immediately. Of course, correcting the clock is necessary to maintain monitoring accuracy going forward. However, if you are investigating the cause of a drop in power generation, it is important to preserve the recorded data before making any correction. Although fixing the clock offset will make subsequent data accurate, information about the offset is needed to interpret past data.

First, what you should record is the date and time of the check and the amount of deviation at that moment. Record how many minutes or hours the data logger’s clock was ahead of or behind the correct local time. If possible, separately record the time shown on the monitoring screen, the data acquisition time, the data logger’s own clock time, and the on-site confirmation time. This makes it easier, when reviewing the data later, to decide which time to use as the reference for correction.

Next, check when the time offset may have started. Even if you cannot determine the exact start point, at minimum record the last day that is believed to have been normal and the first day when the offset was detected. By looking at past graphs to see when the peaks of solar irradiance and generation began to shift, and when any anomalous changes in daily aggregates appeared, you can narrow down the scope of the impact. Once the affected period is identified, you can add a note to any evaluation of reduced generation for that period.

We also record the details of any time correction work. We document who corrected the time, when, and by what method; whether, after the correction, the current time and the data acquisition time are consistent; and whether the post-correction power generation graph shows any unnatural gaps or duplications. If the correction set the clock backward, data for the same time band may appear duplicated. If the correction set the clock forward, blank periods may appear in the data. This is not necessarily an equipment failure, but because it may look abnormal to someone reviewing the data later, it must be recorded.

When preparing reports on declines in power generation, treat data affected by time shifts and the corrected data separately. Even when adjusting historical data for evaluation, it is safer to avoid deleting the original data or overwriting it without leaving the rationale. In practice, keeping the original data, the reasoning behind corrections, and the post-correction evaluation separate and explainable helps reduce misunderstandings among stakeholders.

The important point here is not to dismiss a time offset as merely a configuration error. If there are daily reports, monthly reports, inspection records, anomaly reports, or power generation comparison materials created during the period when the time offset occurred, the figures in those documents may also have been affected. It may not be necessary to make major corrections to everything, but knowing which documents might have been influenced by the time offset will make it easier to explain later.

Furthermore, from the perspective of preventing recurrence, including time checks in regular inspection items is also effective. There is no need to perform detailed analysis each time, but simply confirming that the monitoring screen's current time, the timestamp of the latest data acquisition, the peak positions of irradiance and power output, and the boundaries of the daily aggregation show nothing unusual can lead to early detection. Rather than checking only after you notice low power output, observing the integrity of the time axis during normal operation makes it easier to isolate issues when an anomaly occurs.

The correction of time discrepancies must be carried out using appropriate procedures in accordance with the specifications of the equipment and systems. For work involving electrical equipment or communication devices, it is important to confirm on-site safety rules and management authority and, if necessary, have a specialist handle the task. The more urgently you want to resolve a drop in power output, the more likely you are to hastily change settings, but by recording actions before making corrections you can protect later accountability and data quality.

Summary

When power generation is low, a data logger's clock drift is not as conspicuous a cause as equipment failure. However, because it can distort the very data used to judge a drop in power generation, it is an item you should check in the early stages of a cause investigation. Before suspecting panel soiling, shading, equipment shutdown, output control, wiring abnormalities, etc., first check whether the time axes of the data being compared are aligned to reduce misjudgments.

Points to check are whether the current time on the monitoring screen matches the local time, whether the peaks of solar irradiance and power generation have a natural positional relationship, whether the timestamps in the shutdown history and alarm history are consistent with the power generation graph, and whether there are any discrepancies in the boundary times between the daily and monthly aggregates. By checking these four items, it becomes easier to separate causes of low power generation into equipment-related issues and data-related issues.

When a time offset is detected, it is important not only to correct it immediately but also to record the amount of offset before correction, the period affected, the data that was checked, and the details of the correction. When evaluating historical data, clearly identify which periods were affected by the time offset and apply corrections or annotations as necessary. This makes explanations in inspection reports and monthly management easier and helps prevent misunderstandings among stakeholders.

The issue of low power output is not simply whether the numbers are low, but how those numbers are interpreted. Checking for time drift in the data logger is the entry point to correctly assessing the condition of the generation equipment. If on-site generation status, solar irradiance, shutdown history, and aggregated values are aligned on the same time axis, isolating the cause becomes much easier.

If you want to confirm a drop in power generation more efficiently, it can be effective to establish operations that integrate on-site inspections and data organization. If daily power generation, on-site conditions, inspection records, photos, and work histories are kept in a format that is easy to cross-check later, you can reduce misjudgments caused by time discrepancies or missing records. Aligning the data’s time axis and preserving the evidential basis of the records is fundamental to safely isolating the causes of low power generation.