The following is a list of possible problems that have been compiled from past experiences, mainly from the Atmospheric Radiation Measurement (ARM) research facilities and the experiences from developing and operating a solar measurement station at our respective institutions.
Some key points before addressing possible equipment failures:
Proper maintenance is the final step in quality control of measurements. After a measurement is recorded, only data quality assessment is possible.
A situation with no data is often better than one that produces bad data. Unless bad data is identified and flagged, it can destroy confidence in and value of the set data.
The first section will cover observed problems, and possible causes are noted. Next the corrective maintenance actions will be listed. Because all ARM and UO solar monitoring network sites use Campbell data loggers for solar and infrared measurements, comments on data loggers will be specific for Campbell data loggers, but the gist applies to all data loggers.
An open circuit can be the result of a damaged signal cable, loose signal cable conductor connections at the data logger input terminal, improperly seated connector at the radiometer body, or broken thermopile winding (typically due to lightning strike).
The data logger system program may also have been changed or lost.
The data logger internal battery voltage may be below nominal limits.
Loss of proper electrical grounding and shielding of the low-level direct current (DC) signal cable may cause erratic voltage readings.
The radiometers have a time-constant on the order of 1 sec to changes in irradiance (voltage) due to clouds. Electrical noise will generally have a higher frequency than this natural variation.
Data logger system has failed.
The data logger system program may have been changed or lost
The data logger internal battery voltage may be below nominal limits
Other possible causes include
Moisture inside the radiometer case or signal cable connector
An improperly seated connector at the radiometer body
Loose signal cable connections at the data logger input terminal
Cold solder joint inside the signal cable connector
The GHI pyranometer can, with unique reflections from properly positioned towering cumulus clouds, exceed the solar constant for periods of less than tens of minutes.
Moisture inside the pyranometer dome or the pyrheliometer window can form small lenses that could increase the radiation on the radiometer detector.
Half-melted frost or snow on the dome can reflect sunlight onto the sensor receiver.
More often, high signals are the result of an open circuit causing the data logger to over range (6999), a ground loop introducing a spurious voltage, or an incorrect radiometer calibration factor or logger program.
Signal has been shorted by damaged cable.
Corrosion at signal cable connections has caused increased resistance.
Wrong calibration factor or application associated with instrument.
Moisture inside the pyranometer dome or behind the pyrheliometer window.
This is most easily determined from Equation 2.1. In addition to a bad radiometer calibration factors (Cf), possible causes are
DNI low
Incorrect tracker alignment
Dirty water, ice on NIP window
Nearby obstructions (trees, poles, structures) that shade or sometimes reflect sunlight onto the detector
DNI high
Suspect electrical problem first (see the section on signal greater than physical limit)
Not many physical reasons for NIP to have prolonged increased output— can have temporary increase when sun is next to the edge of a cloud
Contaminated NIP window, especially due to frost or dew, typically reduces reading, but may cause increased signal with specific solar geometry
GHI low
Dirty/iced dome
Possibly bad electrical connections
Nearby obstructions, trees, poles, or structures shade the pyranometer
Alignment problem—pyranometer tilted to north
GHI high
If persistent, suspect electrical problem first
Could be leveling problem, tilted toward sun
Water droplets (dew) on dome can focus sun’s rays
Ice or frost on dome opposite the sun’s location can reflect sun’s rays onto detector
Nearby reflective surfaces, such as buildings and poles
Artificial lights (at night)
DHI low
Dirty/iced dome
Possibly bad electrical connections
Correct PSP for thermal-offset problem—Model 8–48, B/W, will not exhibit this behavior
Pyranometer not level, tilted to north
DHI high
Suspect misaligned solar tracker (dome must be in full shade of ball)
PSP and PIR not coplanar
Pyranometer not level, tilted to south
Possible ground loop or other electrical problems (see the section on no signal or unstable signal)
Nearby reflective surfaces, such as buildings and poles
Artificial lights (at night)
Upwelling GHI low
Dirty/iced dome
Possibly bad electrical connections
Field of view of local terrain biased by dark surfaces—water pooling?
Unlevel mounting platform
Upwelling GHI high
Fresh snow cover can reflect up to 98% of the GHI
Frost on dome can reflect more radiation onto pyranometer detector
Possible ground loop or other electrical problems
Field of view of local terrain biased by bright surfaces—water pooling?
Unlevel mounting platform
Tracker failure (if pyrheliometer and shaded pyranometers are mounted on the same tracker, then DNI = 0 and DHI = GHI, where GHI is indicative of clear sky irradiance levels)
Tracker alignment problem?
Uniform soiling of optics
Suspect communications failure
Modem failure
Loss of electrical power to data logging system
Card not initialized at last site visit (card was full)
Card battery failure
Failed connection between logger and card storage module
Loss of electrical power to data logger system for more than 10 days
Incorrect data logger time (clock drift? improper switch to daylight time?)
Incorrect time zone
Incorrect longitude
Assignment of hours (0–23 vs 1–24)
Instrument alignment (pyranometer not level?)
Data buffer overrun due to corrupt memory?
Based on past experience and an understanding of the irradiance measurement processes, the following actions are suggested for the previous failure modes. The following information should complement the existing instructions for routine maintenance
Note: Excessive dew, frost, snow, or dust on the PSP or PIR can indicate ventilator failure.
Contamination of optical surface
Wash dome or window with distilled water, wipe dry
Wash dome or window with alcohol, wipe dry
Warm iced dome or window with palm of hand if necessary—do not scrape ice from dome or window
Moisture inside the dome or window?
Check PSP desiccant canister seal is tight
Check PSP desiccant canister window is not cracked
Remove PSP sunshade and check dome collar screws are tight
Check NIP window screws are tight
Change desiccant if granule colors are changed showing moisture saturation.
Align Solar Tracker (north/south and base is level)
Adjust shading balls to shade PSP and PIR domes
PSP/PIR is level using circular spirit level
Signal cables not caught on tracker or mounting fixture
For Campbell Scientific Data Loggers that use Loggernet program in monitor mode to view all data channels
Compare with expected readings
Open channel value = “6999”
Note lightning can burn out delicate thermopile windings
Confirm electrical power available
Sun shade must be level with black detector or base of PIR dome
The sun shade can shadow or reflect radiation onto detector if not properly positioned at or slightly below the base of the outer dome
Confirm cable is not cut, stretched, or kinked
Check electrical connection at data logger wire panel and at radiometer connector
Hot and cold cycling loosens some wire connections
Check electrical connection at data logger, enclosure, and ground rod for corrosion or loose fittings
Visually inspect radiometer serial numbers at each mounting location and compare with data logger program locations (Monitor mode) or current inventory listing
Compare calibration sticker information with data logger program locations (Monitor mode) or current inventory listing
The red LED indicator on power supply should be illuminated
Check logger battery voltage is at least 9.2 Vdc using Monitor mode
Check the “on–off” switch is in the “on” position
Note confusing label on CR10X–1M wiring panel—switch is pointing away from “on” label when in operating position
Confirm CR10X–1M logger clock is within 3 sec of GMT
3.137.171.121