RIGHTS OF THE GOVERNMENT

The invention described herein may be manufactured and used by or for the Government of the United States for all governmental purposes without the payment of any royalty.

BRIEF SUMMARY

The present invention is directed to a device and method for checking the photometric linearity of optical instruments, and a method for reconstructing the input-output relationship for calibration purposes when excessive nonlinear performance is detected.

In one form, the device comprises a variable attenuator and a step attenuator inserted in series optically. The variable attenuator is uncalibrated and manually adjustable over the full range of zero to 100 percent attenuation. The step attenuator, on the other hand, is both accurately calibrated and spatially homogeneous as to attenuation irrespective of the size of the light beam passing therethrough. When the step attenuator is inserted into the path of the light beam used by the intensity measuring optical instrument, at diverse settings of the variable attenuator, the output response to the insertion indicates the relative nonlinearity between the input and output of the instrument at the point prescribed by the variable attenuator setting.

The method by which a check is made for nonlinearity and an accurate input-output response curve is reconstructed consists of an iterative sequence commencing with a tabulation of output responses over the full output response range of the instrument. The itensity measuring optical instrument is first adjusted to respond correctly at the opposite ends of the scale, i.e. zero and full scale, with the intermediate nonlinearities being the subject of the inquiry. At each intermediate setting of the variable attenuator the output is recorded both with and without the step attenuator inserted. A calculation is performed to ascertain the apparent step, a ratio between the output response with and without the step attenuator. If the apparent step is substantially identical to the calibrated value of the step attenuator, the instrument is linear. Otherwise, the correct input-output response of the instrument must be reconstructed from the measured data.

The actual input-output response is reconstructed recursively using interpolated or smoothed values of the apparent step from the foregoing measurement sequence and the accurately known value of the step attenuator.

DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic representation of a spectrophotometer type optical instrument.

FIG. 2 illustrates one embodiment of the complete attenuation device inserted within a cell holder of a spectrophotometer.

FIG. 3 is a cross-sectional schematic of the embodiment in FIG. 2, taken at section line 3--3.

FIGS. 4, 5 and 6 are various embodiments of the step attenuator.

FIG. 7 is a plot of apparent step verses indicated output.

FIG. 8 contains a reconstructed input-output plot of the spectrophotometer undergoing calibration.

DETAILED DESCRIPTION

Optical instruments which are used to measure light intensity are normally capable of being adjusted to set the zero end output response and the full scale output response. When the instrument is operating properly the intermediate output responses are expected to be linear once the zero and full scale ends are adjusted. For a variety of reasons the instrument may actually be nonlinear, introducing errors into the output data which are not readily detectable.

It is undoubtedly possible to calibrate light intensity measuring instruments, representative of which are the photometer and spectrophotometer. Unfortunately, conventional calibration processes are generally complicated, intensive, demand specialized test equipment or require disassembly of the instrument.

The invention here disclosed provides a comparatively simple device and various methods for not only checking the calibration but also correcting the input-output response curve for nonlinear characteristics detected during the calibration check.

Turning now to FIG. 1, there appears a schematic of a conventional spectrophotometer. The particular one depicted provides a response based upon the transmittance of material within a cuvette over the full wavelength spectrum being investigated. More particularly, tungsten lamp 1 provides broad spectrum luminous energy which is reflected and focused by substantially identical mirrors 2 and 3. The two focused beams, 4 and 6, are generally homogeneous in luminous energy density, though the invention is not limited by that constraint. Beams 4 and 6 are reduced in size by identical apertures 7 and 8 so that the remaining beam passes through a small segment of reference cuvette 9 and sample cuvette 11, respectively. Reference cuvette 9 is generally used to represent complete transmittance or full scale output of the response. Sample cuvette 11 contains the material whose spectral absorption characteristics are sought.

Motor 12 rotates semicircular shaped mirror segment 13 causing the luminous energy reaching conventional monochromator 14 to alternate between reference beam 16 and sample beam 17. The output of monochromator 14 is detected by photomultiplier 18 while its spectral scan is under the control of wavelength scanner control 19. Amplifier and processor 21 synchronizes the detected signals from multiplier 18 with the position of rotating mirror 13 to coordinate the signal sent to output means 22, typically a strip chart recorder, X-Y recorder or visual display. Reference signals from wavelength scanner control 19 are also sent to output means 22, allowing output responses to be presented in terms of the luminous energy wavelength. Fundamentally, the device depicted in, and described with reference to, FIG. 1 is well known by those practicing in the art.

One embodiment of the calibration checking device appears in FIG. 2, where body 23 is tapered at shoulders 25 for fixed insertion into conventional cuvette holder 24. FIG. 3 is a cross-sectional schematic of the complete device taken at section lines 3--3 of FIG. 2. Body 23 of the embodiment is cylindrical above shoulder 25, and of square cross section below shoulder 25 to permit insertion and fixed orientation in cuvette holder 24. Rectangular openings 26 on both sides of cuvette holder 24 allow unobstructed passage of light beams directly through the cuvette, replaced here by the lower end of the calibration checking device. Vertically disposed in body 23 of the device are two rectangular channels 28 and 29 within which are slidably located two attenuator elements, variable attenuator element 31 and step attenuator element 32. Both attenuator elements are capable of being slid vertically in their respective channels and positioned transverse to passage 27 so as to obstruct portions thereof. For purposes of this embodiment the bottom edge of variable attenuator element 31 has a diagonal edge, 30.

The embodying structure appearing in FIGS. 2 and 3 was reduced to practice with readily available construction materials, wood for body 23 and variable attenuator element 31. Step attenuator 32 was fabricated from conventional slide glass material.

To undertake a calibration check the device is inserted into the cuvette holder, as shown in FIG. 2, and substituted in place of sample 11. Once aligned, sample beam 17 freely passes through passage 27 when attenuator elements 31 and 32 are withdrawn from the passage. Variable attenuator 31 is then inserted to a depth sufficient to obstruct part of sample beam 17, as exemplified by its location in FIGS. 2 and 3. Since the variable attenuator does not contribute directly to the calibration accuracy, no significant precision is actually associated with locating the variable attenuator. Step attenuator 32 has only two positions in the calibration process. A first position where it is completely withdrawn from passage 27, and a second, fully inserted position, where it completely covers passage 27.

The essential operating principle of the device is one of perturbation with a step attenuator of high accuracy at points of various output magnitude to compare the known input change with the indicated output response. Consider, for purposes of elaboration, that the step attenuator is designed to attenuate 10%, i.e. transmit 90%. If the variable attenuator is randomly set to produce an output response at 71% of full scale (FS) with the step attenuator absent, the output should drop to 63.9% (71-7.1) when the 10% step attenuator element is inserted into the device. Any output different from 63.9% indicates a nonlinearity in the region of the 71% output response. Alternately, the apparent step, 63.9/71=0.9, can be calculated from the test data and compared to the precisely known transmittance of the step attenuator. Note that the actual relationship between the input, as set by the variable attenuator, and the output response, showing 71%, is not sought directly.

Body 23 of the device is made from an opaque material. Likewise, variable attenuator element 31 is fabricated from a material which is opaque at all wavelengths over which the spectrophotometer in FIG. 1 will operate. Step attenuator 32 is partially transmissive, the particular embodiment described hereinafter being in the range of 90%, and as a general rule is constant in magnitude and of known accuracy for all the sample beam wavelengths being evaluated. In FIG. 4 step attenuator element 32 is shown as slide glass, but could, as depicted in FIGS. 5 and 6, be made in the form of a grid, 33, or a mesh, 34. In addition to correcting for the effects of reflection and diffraction, the grid and mesh type attenuators would of necessity have to attenuate homogeneously over the whole of any area exposed to sample beam 17. This point is apparent when one reflects on the sample beam appearing schematically in FIG. 3, since by definition attenuated sample beam 36 must be an accurate and unvarying fraction of beam 37, irrespective of where variable attenuator element 31 is positioned.

The method by which the device described hereinbefore is utilized to check the calibration of a spectrophotometer of the type depicted in FIG. 1 is preferably exemplified by analyzing the test results and calculations attributable to an actual test of one embodiment. In this particular case, inquiry was directed to a single wavelength of 5500 angstroms with step attenuator element 32 fabricated from slide glass having a transmittance of 91.6%.

              TABLE I______________________________________Oi      Of with just the variableIndicated output signal         signal with bothIndicated output                        ##STR1##attenuator inserted        attenuators inserted                       Apparent Step(% of full scale)        (% of full scale)                       (%)______________________________________100          91.4           91.495           86.6           91.289           81.2           91.284           76.4           90.978           70.9           90.973           66.3           90.867           60.9           90.962           56.4           91.057           51.8           90.851           46.3           90.745           41.0           91.040           36.3           90.734           30.8           90.729           26.2           90.422           19.9           90.518           16.2           90.212           10.8           89.6 9           8.0            89.0______________________________________

Table I contains the raw measurement data. As in normal use the spectrophotometer is initially preset so that the inputs and outputs correspond at zero and full scale. O_i is the initial output set by means of variable attenuator element 31 in a generally descending sequence. O_f is the final output, after effects of step attenuator 32 have been introduced. The apparent step, S_a, is defined hereinbefore as the ratio of O_f /O_i. For the slide glass attenuator which transmits 91.6%, any departure in S_a from that amount indicates local nonlinearity in the input-output relationship, notwithstanding the fact that the zero and full scale points remain unaltered from the original adjustment. The S_a data from Table I is plotted on the graph in FIG. 7.

It is readily apparent to one skilled in the art that S_a values in range 38, at the lower end of the output scale, are more susceptible to error since the output change attributable to inserting the step attenuator is a fraction of an already low output value. Recording or display errors become significant unless appropriate linear amplification is added when operating in this region.

Having detected the existence of a nonlinearity, a further refinement of the method uses the apparent step (S_a) data plotted in FIG. 7, interpolating between points when necessary, to reconstruct the actual input-output relationship between the intensity of beam 17 after it is attenuated by sample 11 and the output response appearing in output means 22. Obviously, as the number of O_i and O_f data samples increase greater refinement is injected into the interpolation process, and improved accuracy results in the reconstructed input-output relationship.

              TABLE II______________________________________   Interpolat-Oi ed Sa             Of ' = Oi × Sa '                         Ii If (%(% of full   Sa '             (% of full  (% of full                                 of fullscale   (%)       scale)      scale)  scale)______________________________________100%    91.4      91.4%       100%    91.6%91.4    91.2      83.36       91.6    83.9183.36   90.85     75.73       83.91   76.8675.73   90.82     68.78       76.86   70.4068.78   90.85     62.49       70.40   64.4962.49   90.97     56.84       64.49   59.0756.84   90.78     51.60       59.07   54.1151.60   90.70     46.80       54.11   49.5646.80   90.92     47.55       49.56   45.4042.55   90.85     38.66       45.40   41.5938.66   90.70     35.07       41.59   38.0935.07   90.70     31.80       38.09   34.8931.80   90.58     78.81       34.89   31.9628.81   90.40     26.04       31.96   29.2826.04   90.44     23.55       29.28   26.8223.55   90.48     21.31       26.82   24.5721.31   90.42     19.27       24.57   22.5019.27   90.31     17.40       22.50   20.6117.40   90.22     15.70       20.61   18.8815.70   90.00     14.13       18.88   17.2914.13   89.83     12.69       17.29   15.8412.69   89.68     11.38       15.84   14.5111.38   89.60     10.20       14.51   13.2910.20   89.32     9.11        13.29   12.189.11    89.12     8.12        12.18   11.158.12    89.00     7.22        11.5    10.227.22                          10.22______________________________________

Reconstruction of the input-output relationship involves an interactive sequence best described with reference to Table II. I_i and I_f represent the initial and final values of light intensity input, with the variable attenuator alone and with both attenuators inserted, respectively. Since the effect of the step attenuator is known, I_f is always 91.6% of I_i. S_a ', interpolated S_a, is for present purposes a straight line average between data points in FIG. 7, though it could also be smoothed by any one of many mathematic or graphic techniques. O_f ' is equal to the multiple of O_i and the S_a ', with the value for S_a ' taken directly from FIG. 7 at the appropriate O_i value.

Beginning with I_i and O_i at 100%, the adjusted full scale relationship, the first O_f ' value is determined by taking S_a ' (91.4%) from FIG. 7 and multiplying it with the O_i value; the result being O_f ' at 91.4% as appears in the table. Since the input I_f is always 91.6% of I_i, the corresponding input for output O_f ' at 91.4% is I_f at 91.6%. These values for O_f ' and I_f then serve as the respective starting points for the ensuing calculation of O_f ' and I_f. For example in the second row of the table, O_i is now 91.4% (O_f ' from the previous row), S_a ' is read from FIG. 7 as 91.2%. O_f ' is then calculated to be 83.36%. The input tabulation is independent of the output tabulation with the second row now showing I_i at 91.6% (the previous I_f) and I_f calculated to be 83.91% (91.6%×91.6%). The process is repeated for the subsequent rows.

The reconstructed input-output relationship for the embodiment shown and tabulated appears in FIG. 8. The input values I_i are taken along the abscissa while the corresponding outputs O_i are along the ordinate. The actual values are obtained from Table II beginning at 100%, then proceeding row by row, matching the corresponding values of input I_i and output O_i. The point at the origin of the plot is based on the zero point adjustment of the spectrophotometer prior to undergoing calibration checking. Straight line 39 represents the ideal input-output line, while points 41 describe the actual path of the relationship.

The actual input-output relationship described by data points 41 in FIG. 8 provides a direct means by which the user of the spectrophotometer can readily ascertain the actual magnitude of the light energy passing through sample 11. The actual amounts are those values of I_i which correspond to displayed outputs O_i as defined by the relationship curve joining data points 41.

The input-output test was performed on a spectrophotometer at a wavelength of 5500 angstroms. One skilled in the art will recognize that comparable analysis must be performed at other wavelengths and scale settings over the full operating range of the instrument to completely verify its correct operation.

It is equally apparent that the process disclosed herein is highly repetitive, and therefore, readily amenable to computer processing. Furthermore, the burdensome reconstruction portion of the process is likely to be used less frequently than the calibration check sequence in which apparent step S_a alone is determined. If S_a indicates acceptable accuracy, reconstruction of the actual input-output relationship is of marginal benefit to the instrument user.