Test Runs
This repository contains data and analysis from a 18-day test run of an Apocalypse sensor kit in Mt Lavinia, Sri Lanka. This test aims to monitor the performance of the sensor kit in a real life setting. This will enable us to spot and fix or improve any issues or anomalies the setup might have.
- Duration: 18 days
- Location: Mt Lavinia, Sri Lanka
- Number of Sensor Setups: 1 (out of 3 total locations)
- Date Range: August 10, 2024 - August 28, 2024
The following environmental factors were measured:
- Temperature (°C)
- Humidity (%)
- Soil Moisture (%)
- Light Intensity (Lux)
-
Range: 0°C to 40°C
-
Daily Temperature Fluctuations: The temperature shows a clear pattern of daily fluctuations, with peaks and troughs occurring regularly. The peaks suggest daytime highs, while the troughs represent nighttime lows.
-
Consistent Pattern: The graph indicates a consistent pattern throughout the period, with no significant outliers or abrupt changes in temperature. This suggests stable environmental conditions over the monitored days.
-
Slight Temperature Dip Around 8/20/2024: There is a noticeable dip in the overall temperature trend around the middle of the graph, particularly around August 20, 2024. The area have been experiencing cloudy weather for few days in a row which could be the cause of a brief period of cooler weather during this time.
-
Gradual Increase in Daily Peaks: Towards the end of the graph, the daily peak temperatures appear to be gradually increasing, indicating a slight warming trend as the cloudy / rainy weather conditions came to an end.
-
Range: 0% to 100%
-
High Humidity Levels: The humidity remains consistently high throughout the period, typically ranging between 75% and 100%. This indicates a generally moist environment, which might be typical for the location being monitored.
-
Minor Daily Fluctuations: There are minor fluctuations in humidity levels each day, but the changes are relatively small compared to the overall high values, suggesting a stable and consistently humid environment.
-
Sudden Drop on 8/18/2024: There is a noticeable and abrupt drop in humidity around August 18, 2024, where it dips close to 0%. This is likely an outlier or error in the data collection rather than a true environmental change.
-
Occasional Dips in Humidity: Other brief and smaller dips in humidity are observed sporadically, which might represent brief changes in environmental conditions or potential sensor anomalies.
-
Slight Decrease Towards the End: Towards the latter part of the graph, there is a slight downward trend in the humidity peaks, indicating a minor decrease in overall humidity levels, but they still remain relatively high.
- Note: The microcontroller reads input from the soil moisture sensor as an analog voltage input. Afterwards this input is converter into a percentage using below formula.
100 - ((Vr - Wv)/(Dv-Wv)*100 where,
Vr - analog voltage reading
Wv - wet voltage reading
Dv - dry voltage reading
-
Range: 0% to 100%
-
Gradual Increase: Moisture percentage gradually rises from the beginning, indicating the starting of the rainy weather.
-
Peak Moisture: Peak moisture is show around August 18, 2024 where the area have being receiving rain for few days.
-
Gradual Decrease: Gradual decrease in moistures indicates the ending of the rainy days and the ground slowly drying up.
-
Occasional spice: Small occasions spikes may suggest the reading are noisy.
-
Range: 0 to 60,000 Lux
-
Distinct Daily Peaks: The graph shows distinct daily peaks in light intensity, with each peak representing periods of high light exposure, during daytime hours. The peaks are sharp, indicating a rapid increase and decrease in light intensity.
-
Maximum Intensity Variations: The peak values vary significantly, ranging up to 60,000 Lux. The intensity of these peaks fluctuates, suggesting variable light exposure across different days, possibly due to weather conditions or other environmental factors.
-
Low Baseline During Night: Between peaks, the light intensity drops close to zero, indicating nighttime periods with little to no light exposure.
-
Period of Reduced Peaks Around 8/18/2024: There is a noticeable reduction in peak intensity and frequency around August 18, 2024. The area have been experiencing cloudy weather for few days in a row which could be the cause of a brief period of cooler weather during this time.
-
Increasing Light Levels Towards the End: Towards the latter part of the graph, the peaks appear to increase in both frequency and intensity, suggesting longer or more intense daylight exposure during these days.
The environment exhibits stable daily cycles of temperature, humidity, and light, with minor anomalies and a slight warming trend. This pattern reflects typical daily environmental dynamics, possibly influenced by local weather conditions and seasonal changes.
-
This particular sensor module did not have the necessary circuitry to perform battery level monitoring. Hence we were unable to extract accurate battery voltage fluctuation data with this module.
-
Humidity regularly reaching 100% may indicate poor airflow through the PVC stand pipe were the DHT11 sensor is placed in. This may cause more humidity to be trapped inside the stand pipe influencing the overall humidity value.
- Increase the number of holes in the PVC stand pipe.
- Introduce the battery monitor circuitry to this module as well.
Raw data and detailed logs are available in the data directory of this repository.
This repository contains data and analysis from a 18-day test run of an Apocalypse sensor kit in Piliyandala, Sri Lanka. This test aims to monitor the performance of the sensor kit in a real life setting. This will enable us to spot and fix or improve any issues or anomalies the setup might have.
- Duration: 18 days
- Location: Piliyandala, Sri Lanka
- Number of Sensor Setups: 2 (out of 3 total locations)
- Date Range: August 10, 2024 - August 28, 2024
The following environmental factors were measured:
- Temperature (°C)
- Humidity (%)
- Soil Moisture (%)
- Light Intensity (Lux)
- Battery Voltage (V)
-
Range: 0°C to 50°C
-
Initial High-Temperature Spike: At the beginning of the graph, there is a sharp temperature spike reaching close to 50°C. This is followed by a quick drop, suggesting either an unusual environmental condition or a sensor anomaly.
-
Extended Flatline Periods: There are several extended flatline periods where the temperature remains constant for long durations. These sections indicate periods when the sensor module was offline and not recording data.
-
Temperature Stabilization at Lower Values: Following the initial spike, the temperature stabilizes around 30°C to 35°C. This stability persists for some time before another flatline period, which again could be attributed to the sensor going offline.
-
Minor Temperature Fluctuations Between Offline Periods: When the sensor is active, the graph shows minor temperature fluctuations, generally ranging between 25°C and 35°C. These fluctuations suggest normal daily temperature variations, though they are less pronounced than in the previous location.
-
Notable Temperature Drops: Around August 18, 2024, there is a noticeable temperature drop, followed by another flatline period. This could be due to cooler weather conditions before the sensor going offline.
-
End Period Activity: Toward the end of the graph (near August 25, 2024), there are several sharp peaks and drops. This indicates that the sensor was active again, capturing temperature variations more frequently. The peaks suggest periods of higher temperatures, while the drop indicate cooler periods or nighttime.
-
Range: 0% to 100%
-
Extended Periods of High Humidity: For most of the middle section of the graph, the humidity remains consistently at or near 100%. This indicates a prolonged period of high humidity, possibly due to weather conditions like continuous rain or high moisture in the air. Flatline periods near 100% could also mean the sensor might have been affected by moisture saturation or has gone offline.
-
Sensor Offline Periods: There are noticeable flatline segments where the humidity remains constant at 100%. These periods likely correspond to the sensor module going offline, similar to the previous temperature graph, with no data being recorded during these times.
-Increased Activity Towards the End: Towards the end of the graph (near August 25, 2024), the humidity levels become more variable, dropping sharply at times to around 50% and below, and then climbing back up. This could indicate the sensor coming back online and recording actual humidity variations, it reflects changing environmental conditions, such as transitions from wet to dry weather.
- Note: The microcontroller reads input from the soil moisture sensor as an analog voltage input. Afterwards this input is converter into a percentage using below formula.
100 - ((Vr - Wv)/(Dv-Wv)*100 where,
Vr - analog voltage reading
Wv - wet voltage reading
Dv - dry voltage reading
-
Range: 0% to 100%
-
Near 100% flat line: Near 100% flat lines indicate the sensor reading zero voltage due to defect or bad connection.
-
Gradual Increase: Moisture percentage gradually rises from the beginning, indicating the starting of the rainy weather.
-
Peak Moisture: Peak moisture is show around August 18, 2024 where the area have being receiving rain for few days.
-
Gradual Decrease: Gradual decrease in moistures indicates the ground slowly drying up.
-
Occasional spice: Small occasions spikes may suggest the reading are noisy.
-
Range: 0 to 60,000 Lux
-
Periods of No Data: Following the initial fluctuations, there is a periods where light intensity is near zero (until around August 18, 2024). This indicates the sensor module was offline or in a low-light environment (e.g., night-time or cloudy weather).
-
Brief Peaks and Rapid Decays: Around mid-August (around August 18, 2024), the graph shows brief peaks in light intensity followed by rapid drops back to near-zero levels. These sudden peaks may capturing moments of bright light, such as direct sunlight breaking through clouds, followed by rapid end of light exposure.
-
Gradual Increase and Spikes Towards the End: Towards the end of the data collection period (near August 25, 2024), the graph shows a more gradual increase in light intensity, followed by multiple sharp spikes that reach higher values. This suggests that the sensor was active and capturing more consistent light exposure patterns, possibly indicating longer daylight hours, periods of sunlight, or varying light conditions.
-
Sensor Offline or Shaded Periods: Several flatline segments at or near zero lux indicate potential periods when the sensor was offline, malfunctioning, or in a shaded conditions (e.g., during nighttime or shaded areas).
-
Initial Voltage Levels: The battery starts at around 4V, which is consistent with a fully charged single-cell LiPo battery. There are small fluctuations in the voltage readings at the start, which are normal for battery-powered devices due to changes in load or temperature.
-
Gradual Decline Over Time: A slight, consistent decrease in voltage can be observed from August 11 to around August 18, indicating typical battery discharge over time. The decrease is relatively steady, suggesting a consistent power draw from the module and no significant spikes in power consumption.
-
Dipping below operating voltage: After August 18, it can be a gradual decline down to 3.50V which is the lowest voltage the sensor module can operate. This indicates insufficient recharging of the batteries. This could be caused by poor light conditions such as cloudy weather or shading.
-
Increase in Voltage: Around August 25, there is a noticeable increase in voltage, returning close to 4V. This suggests that the solar charging is happening at a higher efficiency.
-
Consistent Post-Increase Behavior: After the voltage increase, the battery maintains a relatively stable level, slightly fluctuating around 3.8V to 4V.
Location 2 demonstrates a high humidity environment with generally stable temperatures but variable light conditions. There are periods of missing data across all sensors due to known offline status, indicating challenges in sensor reliability or environmental interference.
-
Two major sensor downtimes were recorded.
-
Upon diagnosis, it was revealed that the first major downtime (from 12th to 16th) was caused by a loose LoRa module which prevented the sensor module from sending data. It was fixed by manual intervention.
-
The second downtime was caused by a battery drain. This could have happed due to a combination of few reasons.
- Higher power consumption of the module due to a build defect.
- The area experiencing a few cloudy days in a raw (from 16th to 21st) where the solar panels did not generate much power.
- Nearby trees regularly casting shadows on the solar panel.
After the light conditions improved, the sensor node came back online without other any manual intervention.
- Problems with loose components can be avoided by careful handling.
- It is recommended to redo the wiring of the sensor module circuit if its consuming higher current (more than 30uA) when its in sleep mode.
- Make sure to remove any tree branches or other objects that might cast shadows over the solar panel.
Raw data and detailed logs are available in the data directory of this repository.
This repository contains data and analysis from a 18-day test run of an Apocalypse sensor kit in Pannipitiya, Sri Lanka. This test aims to monitor the performance of the sensor kit in a real life setting. This will enable us to spot and fix or improve any issues or anomalies the setup might have.
- Duration: 18 days
- Location: Pannipitiya, Sri Lanka
- Number of Sensor Setups: 3 (out of 3 total locations)
- Date Range: August 10, 2024 - August 28, 2024
The following environmental factors were measured:
- Temperature (°C)
- Humidity (%)
- Soil Moisture (%)
- Light Intensity (Lux)
- Battery Voltage (V)
-
Range: 0°C to 50°C
-
Initial Temperature Variation: The graph starts around 25°C to 30°C and shows a period of fluctuations, peaking close to 35°C to 40°C multiple times. These fluctuations could be due to changing environmental conditions, such as day-night cycles or variable weather.
-
Sudden Temperature Drop: A significant drop to nearly 0°C is observed around August 14-15, 2024. This drop is likely an anomaly, possibly due to a sensor error, malfunction, or a brief power loss.
-
Gradual Increase After Anomaly: Following the anomaly, the temperature readings stabilize around 25°C to 30°C.
-
Flat Lines After Anomaly: It shows a period of not data from 15th to 19th and again from 20th to 28th. This could be due to sensor module going offline.
-
Range: 0% to 100%
-
Initial High Humidity: The humidity starts at around 50-60% and quickly rises to over 90%. The early part of the graph shows some fluctuations, but humidity mostly stays high, close to 100% for several days until August 18, 2024.
-
Sudden Spikes and Drops: Around August 19-20, 2024, there are sharp spikes and drops in humidity levels, indicating sudden changes. Humidity briefly spikes close to 100% and then drops sharply below 50%. These abrupt changes may be due to sensor noise, a temporary malfunction, or rapid environmental changes.
-
Flat Lines After Anomaly: It shows a period of not data from 15th to 19th and again from 20th to 28th. This could be due to sensor module going offline.
- Note: The microcontroller reads input from the soil moisture sensor as an analog voltage input. Afterwards this input is converter into a percentage using below formula.
100 - ((Vr - Wv)/(Dv-Wv)*100 where,
Vr - analog voltage reading
Wv - wet voltage reading
Dv - dry voltage reading
-
Range: 0% to 100%
-
Unpredictable pattern: The soil moisture sensor reading show almost random variations that does not seem to follow a patter, this could indicate of a faulting soil moisture sensor.
-
Range: 0 to 60,000 Lux
-
Initial Light Intensity Fluctuations: The light intensity shows significant fluctuation in the range from 0 to 20,000 lux in the first few days, This indicates that the sensor was not receiving any direct sunlight at any time of the day.
-
Sharp Peak Followed by Low Intensity: A sharp peak around August 11, 2024, reaching nearly 15,000 lux, may indicate a sudden exposure to bright light, such as direct sunlight or an artificial light source.
-
Near Zero Flat Lines: Near zero flat lines suggests that no readings were taken during this period indicating a sensor downtime.
-
Gradual decrease with Minor Fluctuations: From August 10 to August 15, 2024, the battery voltage decreases gradually overtime, fluctuating slightly around 3.7V to 4.0V. This indicates an unusual rate of battery depletion with not enough charging.
-
Sudden Drop to Zero: On August 13, 2024, there is a sudden drop in voltage to 0V, which then quickly returns to the previous stable voltage level. This drop suggests a potential temporary disconnection or power failure.
-
Down time: The graph shows the battery voltage dipping down to around 3.4V by 18th, and the sensor has gone offline due to insufficient power. The graph shows a the battery again draining from 4.1V to 3.3V in a single day. This may indicate a faulty battery.
Location 3 is an environment with low light intensity conditions and with higher temperatures. There are periods of missing data across all the readings due to sensor downtimes caused by the faulty battery.
- It is evident in the light intensity graph that the sensor module was sitting in a place where it does not receive direct sunlight for its solar panels to generate much power. But a properly built sensor module should be able to operate for a long time (more than 2 months) in these conditions.
- The soil moisture sensor seems to be giving inaccurate readings. It is likely a faulty component.
- This particular device seems to have a bad battery which cannot hold its charge for long.
- Replacing the battery and the charging circuit is a proper fix for the power issue.
- Replacing the soil moistures sensor with a tested new sensor.
Raw data and detailed logs are available in the data directory of this repository.