Solar Irradiance Meter Calibration: Use the Supplied Sensitivity

Quick Answer: A thermopile solar irradiance meter converts its low-level voltage to W/m2 using the sensitivity assigned to that instrument. Use the supplied calibration value, with matching voltage units. Do not apply this raw-signal calculation to a 4-20 mA, scaled-voltage or RS485 version whose output has already been converted.

A plausible irradiance reading can still use the wrong multiplier

A photovoltaic monitoring station can show a smooth daytime curve even when its solar irradiance meter is scaled incorrectly. The same error then affects comparisons between irradiance and plant output. A meteorological data logger may also retain a believable trend while every observation carries a proportional error.

Before changing a calibration setting, identify the ordered output version. The NiuBoL thermopile solar irradiance sensor is offered with several output options. A raw millivolt input, a conditioned current loop and a digital register require different receiver settings; selecting the same sensor name in software does not make those settings equivalent.

Use the instrument sensitivity, not the family range

For the raw thermopile signal, sensitivity expresses how much voltage the detector produces per unit of irradiance. If sensitivity is S in uV/(W/m2), irradiance G in W/m2 equals the measured voltage in uV divided by S. Convert millivolts to microvolts before dividing.

These are calculation examples, not a calibration certificate for a delivered sensor. The family sensitivity range of 7-14 uV/(W/m2) is not a menu from which the installer chooses a convenient value. Obtain the value associated with the instrument serial number and enter it with the correct units.

Some loggers request sensitivity; others request its reciprocal as a multiplier. Read the input definition. Entering a sensitivity where the software expects a conversion factor can produce an error even when the number was copied correctly from the calibration record.

Different output versions need different receiver settings

If a project requires Modbus RTU, confirm that protocol for the ordered RS485 version and obtain its register documentation. RS485 describes the electrical interface, not the complete application protocol. A raw millivolt sensor has no Modbus registers unless a separate converter provides them.

The listed 0-20 mV output option does not, by itself, define an irradiance conversion factor or the maximum electrical signal of every configuration. Check the ordered assembly's signal limits against the logger's input range. Also distinguish an input's ability to survive a voltage from its resolution and uncertainty when measuring a small thermopile signal.

What the published solar irradiance meter specifications establish

Obtain power consumption, cable material, connector details and the protection rating for the ordered assembly. Do not infer an IP rating or an ISO classification from a photograph or from a single stability figure. Overall uncertainty also involves calibration, direction of incoming radiation, temperature and the acquisition equipment.

Check the input chain before adjusting the sensor

Record the serial number, sensitivity or configured output mapping, logger input type and engineering units. For a raw channel, compare the logger's voltage reading with a suitable reference instrument under stable conditions. Verify polarity and terminal assignment using the delivered wiring diagram; a changing wire color between assemblies is not a reliable basis for reconnection.

A voltage offset matters more at low irradiance because it represents a larger share of the signal. Investigate terminal connections, input configuration and cable routing before forcing a software zero. An offset correction should not conceal intermittent wiring or a wrong input range.

During a side-by-side field comparison, align the sensor planes, avoid transient shadows and keep the domes clean. Allow the instruments to settle. A difference during rapidly changing cloud conditions is not sufficient evidence of a calibration error, particularly when instruments have different response times. The solar radiation installation and abnormal-data guide covers the installation checks that should precede a calibration decision.

Project Decision FAQ

Q1. Can I use 10 uV/(W/m2) for every thermopile sensor?

No. Ten is only the illustrative value used here. Use the sensitivity associated with the delivered instrument. A family range describes variation between instruments, not a universal calibration constant.

Q2. Should a digital W/m2 reading be divided by sensitivity again?

Not when the delivered register already reports irradiance in W/m2. Apply only the documented digital scaling. Reapplying a raw thermopile conversion would convert the value twice.

Q3. Does annual stability of +/-2% mean the station is accurate to +/-2%?

No. Annual stability describes one performance characteristic. Installed uncertainty also depends on calibration, environmental effects, mounting and the logger. It cannot be established from stability alone.

Q4. Will one-second polling resolve fast cloud changes?

Polling can capture available updates, but it does not shorten the detector response. This family specifies 99% response within 35 s. Define acquisition and averaging with that physical response in mind.

Q5. What should be included in an order or replacement request?

Specify output type, supply, required cable and mounting plane, and identify the receiving logger. Request the applicable calibration information and signal mapping. For an existing station, include the current input settings so the replacement does not inherit an incompatible multiplier.

Summary

The useful calibration question is not whether a solar irradiance meter produces a daytime curve, but whether its signal is converted correctly. NiuBoL can review the required output and receiver interface before ordering. Keep the delivered calibration information with the instrument so a later logger change does not silently alter the irradiance record.