I. Comparison of Key Data

Instrument A:

The main peak center wavelength is 313.2nm, the half-width is 5.6nm (standard narrowband UVB), the ultraviolet concentration is 305~320nm , and the visible light energy is extremely low;

Instrument B:

The main peak extends to the 360-380nm UVA range , with multiple sets of mercury visible light characteristic peaks appearing (405/436/546/579nm).

II. Analysis of 5 Key Potential Causes

1. Inconsistent spectrometer range and test band settings

Instrument A: Set to 200~400nm UV dedicated range (high-sensitivity UV channel) , only amplify UV band signals, visible light is filtered/attenuated by the instrument's optical path, so only a single peak of 313nm is displayed;

Instrument B: Full-band visible light + ultraviolet with 200~1650nm wide range , no ultraviolet cutoff filter, all characteristic spectral lines of natural mercury from low-pressure mercury lamp (254, 313, 365, 405, 546, 579nm) are displayed, and the visible light sensitivity is much higher than that of Instrument A, and the visible light peaks visible to the human eye are enhanced .

In fact, this LIGHTBEST PL9W is a 313nm narrow-spectrum UVB fluorescent lamp with a low-pressure mercury substrate. It has trace amounts of 254UVC, 365UVA, and mercury visible light clutter peaks. Instrument A blocks visible light, while instrument B collects clutter peaks across the entire wavelength range .

2. The integration time and gain level are different (sensitivity differences).

Instrument A: High gain in ultraviolet light, low channel compression in visible light, only collecting UVB signals ;

Instrument B: Screenshot shows that the integration time is 8128ms, with a high sensitivity setting . The ultra-long integration time amplifies and displays all trace stray light and mercury impurity spectral lines. Instrument A has a short integration time and is calibrated with ultraviolet light, so stray peaks are directly swallowed by the baseline.

3. Different configurations of the front optical filter

Instrument A has a built-in visible light cutoff filter (<400nm transmits, >400nm cuts off) , so visible light above 400nm cannot enter the detector, and the spectrum has no visible light peak;

Instrument B has no cutoff filter and is a full-spectrum direct pass-through instrument , which collects all the visible light characteristic peaks of mercury, resulting in a completely different spectrum.

4. Minor differences in the working state of the lamps during testing

Voltage fluctuations: A. The test voltage is stable at 220V; B. Instantaneous voltage fluctuations cause changes in mercury discharge intensity and alter the proportion of impurity peaks.

Lamp preheating: UV fluorescent lamps need 3-5 minutes to stabilize. A: Sufficient preheating results in stable phosphor output at 313nm; B: Upon initial power-on testing, mercury line noise is relatively high.

5. Inconsistency between wavelength calibration and instrument factory calibration.

The two spectrometers have different factory calibration curves for their spectrometer gratings and CCD detectors .

A has undergone UV-specific calibration (200–400nm) before leaving the factory.

B is a universal full-spectrum calibration (200–1650nm). The UV/Vis response coefficients are different, and the peak heights of the same energy appear very different on the spectrum.

III. Summary

The lamp tubes remained the same; the difference lay in the testing bands , filters , and gain parameter settings of the two instruments. One instrument only measured ultraviolet light, while the other measured the entire spectrum, resulting in visually distinct spectra of two different lamps. Based on the parameters of instrument A: main peak 313.2nm, half-width 5.6nm, it was confirmed that the LIGHTBEST lamp tube and the Philips PL-S9W/01 narrowband UVB (medical 311~313nm) spectrum were identical .

IV. Unified Retesting Scheme

To eliminate discrepancies, the two instruments were standardized: wavelength range of 200~400nm, 400nm cutoff filter added, integration time unified, lamp preheating for 5 minutes, and voltage stabilization at 220V. After retesting, the main peak of both spectra fell at 313nm, and the curves basically overlapped.