1、ANS I/ I ES NA R P-2 7 2-00 Recommended Practice for Photobiological, Safety forLamps Spectroradiometry Consulting, P.O. Box 2747 La Plata, MD 20646-2747 USA (1997) Sliney, D. H. and Wolbarsht, M. L., Safety With Lasers and Other Optical Sources, Plenum, New York (1 980) 4.0 MEASUREMENT CONDITIONS 4
2、.1 Lamp Seasoning To maintain stable output during the measurement process and provide reproducible results, lamps shall be seasoned for an appropriate period of time. During the initial period of operation a lamp will change as its components come to equilibrium. If measurements are taken of an uns
3、easoned lamp, the variations with- in the measurement period and between measure- ments will be significant. As the output of a lamp gen- erally decreases over life, the seasoning time should be short to result in conservative hazard evaluations. Seasoning of lamps shall be done as stated in IESNA L
4、M-54. For the purposes of these standards, the lamp output at the end of the seasoning period is the initial output. For lamps not covered by the LM-54 standard, a study may be required to find the minimum time required to stabilize the operation of a source. 4.2 Test Environment Measurements shall
5、be made in a controlled environ- ment. The operation of sources and measurement equipment is impacted by environmental factors. Additionally, the formation of ozone in the measure- ment path may compromise accuracy and may pre- sent a safety hazard. 4.3 Temperature The ambient temperature will signi
6、ficantly influence the output of certain light sources; e.g., fluorescent lamps. The ambient temperature in which measure- ments are taken shall be maintained in accordance with the appropriate IESNA LMs noted in Section 3.1. 4.4 Drafts The characteristics of some light sources are signifi- cantly a
7、ffected by drafts. For the applicable light source, refer to the appropriate IESNA LM guides noted in Section 3.1. Other than normal convection air, air movement over the surface of test lamps should be reduced as much as possible consistent with safety considerations (ozone production). When the sy
8、stem under test provides interlocks that maintain circulation, measurements shall be performed with circulation. 4.5 Extraneous Radiation Careful checks should be made to ensure that extra- neous sources of radiation and reflections do not add significantly to the measurement results. Visually black
9、 surfaces can be highly reflective to UV and IR radiation. 2 ANSI / IESNA RP-27.2-00 Radiation from hot baffles must be considered in infrared measurements due to the large input angle subtended by baffles. Water cooled baffles and double baffles are two methods for addressing baffle heating. 4.6 La
10、mp and Lamp System Operation The lamp or lamp system shall be operated under conditions that are standardized. 4.6.1 Lamp Operation The input power to the test lamp shall be provided in accordance with the appropriate LM standard noted in Section 3.1. If no standard for the lamp type exists, the lam
11、p manufacturers recommendation for opera- tion should be used. 4.6.2 Lamp System Operation The power source for a complete lamp system shall be at the manufacturers principal specified operating voltage for the device. If the specified condition is a voltage range, the highest nominal value shall be
12、 used. The supply voltage should not vary more than plus or minus one half percent, (10.5 Yo), during the test. For AC operation, the supply frequency should not vary more than plus or minus one half percent, (k0.5 %), and the root mean square (RMS) value sum- mation of the harmonics components shou
13、ld not exceed three percent, (3 Yo) of the total. 4.6.3 Safety The systems tested are unknown in their optical haz- ard and shall be treated as hazardous unless proven otherwise. The possible formation of ozone presents a health hazard and shall be addressed for all sources with the exception of tho
14、se sources with an envelope known to block ozone forming UV. The operation of some light sources in a laboratory environment may require compromises in operation. No compromises shall be made that impact electrical safety. The sources under test may pose additional hazards (e.g., burn, explosion) wh
15、ich shall be addressed with appropriate measures. 5.0 MEASUREMENT INSTRUMENTATION 5.1 Recommended Measurement Instrumen- tation - Double Monochromator The measurement of a source for the purpose of haz- ard classification requires accuracy during calibration and testing. The detectors broad spectral
16、 response and high spectral resolution required to provide accu- rate weighting leads to stringent requirements for out- of-band stray light rejection. Calibration sources pro- vide wide spectral output, which needs to be rejected out of the pass-band. The ratio of out-of-band energy to pass-band en
17、ergy at 270 nm for tungsten halogen incandescent calibration lamps (e.g., the FEL) is 1. The double monochromator is the only instrument that provides the needed selectivity, and it is recommend- ed for hazard measurements involving UV and visible radiation. It is recognized that monochromator sys-
18、tems introduce limitations in convenience and speed. 5.1.1 Instrument Spectral Response The shape of the spectral response and the ratio of the measurement interval to the bandwidth will deter- mine whether the system is able to accurately mea- sure signals with narrow spectral extent, for example a
19、tomic emission lines. (See Kostkowski, Chapter 5.) A monochromator with perfect triangular spectral response used in a system that has a reporting inter- val that divides into the bandwidth integrally will accu- rately measure all signals regardless of their spectral shape. (See CIE 63, Section 1.8.
20、4.2.1 or Koctkowski 1997, Section 5.9.) Deviations from this may lead to errors in measured energy. The spectral response of the system is determined by a spectral scan of a nar- row isolated spectral line, e.g., filtered laser or atomic emission, using scan steps much smaller than the instruments b
21、andwidth. The resulting spectrum is a mirror image of the systems spectral response. The spectral response is what would be found by holding the instrument at a single wavelength and noting the response to a monochromatic source whose wave- length is varied around the that wavelength. (See Kostkowsk
22、i 1997, Section 4.9.) The systems ability to accurately measure the energy in a narrow band sig- nal is the sum of the spectral responses at each reported wavelength. The variation across the summed spectrum is the potential error in total mea- sured signal and shall be included in the uncertainty a
23、nalysis. The result of hazard evaluations will be influenced by the instruments characteristics. The bandwidth of the monochromator will change the weighted results of any spectrum with varying levels. All finite bandwidth instruments report signal at the wrong wavelength, leading to errors in a wei
24、ghted sum. Table 1 lists the recommended bandwidth for 2 per- cent upper bound of uncertainty in weighted sums. 3 ANSI / IESNA RP-27.2-00 Range (nm) ! Bandwidth (FWHM) 1 5 4nm Table 1. Recommended Bandwidths A more complete analysis that takes the source spec- trum into account may be used to relax
25、the suggested bandwidth accuracy. The results of analysis shall be included in the stated uncertainty of the measurement. Note: Systems that constantly integrate the signal during the spectral scan will not experience errors in total measured power from spectral response shape or from the ratio of b
26、andwidth to reporting interval. (See CIE 63, Section 1.8.4.2.2.) Large bandwidths will still lead to errors in weighted results with this type of instrument. gation of uncertainty the percent wavelength uncer- tainty must be multiplied by a sensitivity factor of 78. The wavelength accuracy of the mo
27、nochromator used for hazard testing should be sufficient to provide weighted results with an error arising from wavelength inaccuracy less than two percent (2 %). The needed accuracy therefore depends on the region of the spectrum and the weighting function used. Table 3 summarizes the suggested acc
28、uracy which will bound the error to approximately 2 percent. Only unweighted power is used above 1050 nm. Any bandwidth is permissible providing that it does not intro- duce an unacceptable error in the integrated power. A more complete analysis that takes the source spec- trum into account may be u
29、sed to relax the suggest- ed wavelength accuracy. (See Example 4.) The results of analysis shall be included in the stated uncertainty of the measurement. 5.1.3 Stray Radiant Power 5.1.2 Wavelength Accuracy The wavelength accuracy of the instrument used to determine the spectral ouput of a source ha
30、s a great impact on the weighted values. The photobiological weighting functions change at an extreme rate, e.g., 250 percent over 3 nm at 300 nm for S(h), the UV Hazard Weighting Function. If a reasonable limit of error is desired, then the measured energy must be assigned to its proper wavelength
31、so that it is appro- priately weighted. Table 2 is an example showing the change in weight- ed results from a line when the line is moved by 0.1 nm. The measured values are calculated by assuming a spectroradiometer with a triangular response, 2 nm bandwidth, and 1 nm reporting interval. The sums of
32、 the measured values are equal as the line moves because of the principles described in Section 5.4. The weighted measurement changes by 2.6 percent for a wavelength change of 0.1 nm. The wavelength deviation in this example is 0.033 percent causing an error in weighted result of 2.6 percent. In the
33、 propa- The absolute calibration of spectroradiometers requires the use of sources with broad spectral output and high energy. If the spectral rejection is insufficient, additional energies from other parts of the spectrum will be included in the calibration. The result of this type of error is to u
34、nder-calibrate the spectoradiome- ter and leads to lower readings of the potential haz- ard. Typical ratios between the total energy and the signal passed by the monochromator are on the order of lo4. To obtain one percent accuracy, rejection of out of band radiation needs to be on the order of IO6.
35、 Methods to check for out-of-band stray light and remedies are detailed in Annex B. 5.1.4 Input Optics Recommendation: Integrating Sphere A properly designed and suitably coated integrating sphere will address the three issues detailed below. The integrating sphere will be the most convenient S(h) i
36、s the UV Hazard Weighting Function Table 2. Example of Error in Weighted Value for Wavelength Error 4 - I Range (nrn) I Wavelength accuracy I ANSI / IESNA RP-27.2-00 5.1.5 Calibration Sources Table 3. Recommended Wavelength Accuracy method for irradiance measurements. The random reflectance of the c
37、oating depolarizes the incoming light and proper design can closely match a cosine response. The multiple reflections within the inte- grating sphere consistently fill the radiometers input. 5.1.4.1 Polarization The calibration source and the device tested may not have the same polarization. The int
38、ernal optics of a monochromator will generally be selective to polariza- tion. Selectively rejecting some of the calibration or test signal can lead to an erroneous determination. Input optics for the double monochromator shall be selected to depolarize the input light to maintain a valid calibratio
39、n. The random reflections in an integrating sphere will depolarize the incoming light. 5.1.4.2 Angular Response The calibration of the system for irradiance is typically accomplished at a moderate distance with a small source, leading to a small input angle. Measurements may then be performed on sou
40、rces that subtend a greater angle. Irradiance is the measure of flux on a sur- face element. For both calibration and test source to be comparable, the angular response of the input optics should conform to a cosine function, the response vary- ing as the cosine of the angle to the normal. Proper de
41、sign of an integrating sphere can closely match a cosine response. 5.1.4.3 Monochromator Input The throughput of a monochromator will vary with position and input angle. The response of the detector will be affected by the position of the flux on the detec- tor. For a calibration to be valid the cal
42、ibration and the test signal must have the same angular extent. The multiple reflections within the integrating sphere consistently fill the spectroradiometers input. I The current ANSIIIESNA RP27.1-96 requires no spectrally varying weighting of power above 1050 nrn. Intended change in the next revi
43、sion will introduce a spec- trally varying weighting up to 1200 nm, and an appropriate bandwidth limit will be required up to 1200 nm. The suggested sources for calibration are the deuteri- um discharge lamp and the calibration grade tungsten halogen lamp (e.g., the FEL) with NIST type base.The deut
44、erium lamp may vary in output level while main- taining its spectral shape. Therefore the calibration of the system in the 200 nm to 250 nm region using the deuterium lamp shall be adjusted by comparison to the calibration level from the FEL between 250 nm and 350 nm. The wavelength below which the
45、deuterium shape is used shall be at as short a wavelength as practical considering the noise in the FEL calibration. 5.2 Broadband Detectors When faced with short duration pulsed or low intensi- ty sources, it is convenient to use broadband detec- tors. Broadband hazard sensors attempt to match the
46、weighting spectra by the use of filterS.The matching is never exact and leads to some amount of error. Spectral error is uncertainty only to the extent that the sources spectrum or the detectors spectral response is unknown. If the source spectrum is unknown, then the point of largest percentage dev
47、iation between the detector and the action spectrum must be assumed as the uncertainty. When both the detectors response and sources spectrum are known, straightfotward calculations can generate a correction factor. Used with an appropriate correction factor, the broadband detector provides a valid
48、method for measurements under the RP-27 series of standards. It is incumbent upon the radiometrist to show that the correction fac- tor is valid in each specific case. Variations that lead, or may lead, to changes in spectra require re-deter- mination of the correction factor. 5.2.1 An Example of Co
49、mbined InstrumentTesthg The combination of weighted broadband radiometers and a double monochromator can improve the mea- surement process in many instances. A survey of spa- tial, temporal, or item to item variations of a source type can be rapidly accomplished using the filtered detector. Representative samples, mean, high and low, can then be measured using the double monochromator system. The results are analyzed to determine the correction factor and the factors variation. Further tests can then be accomplished using the filtered detector with the improved knowledge of th
