IN SITU CALIBRATION OF PHOTODIODES
The present invention relates to the in situ calibration of light detectors such as photodiodes. The invention is particularly but not exclusively concerned with the in situ calibration of photodiodes integrated on an optics device.
A wide variety of optical chips are currently being manufactured which include an optical device such as a dispersive optical component for example in the form of an array waveguide grating (AWG). Such optical devices include photodiodes for monitoring light output from the optical components. Such photodiodes are generally manufactured in arrays of a predetermined number, such as 64, 128, 256 or even 512. In order to keep the costs of the optical device down, the photodiodes are manufactured to certain tolerances which mean that they do not have a perfectly defined opto-electrical response. That is, each photodiode in the array can have a slightly differing opto-electrical response to others in the array. Moreover, the opto-electrical response can change with time and also in differing environmental conditions in which the chip might be used.
There is therefore a need to calibrate such arrays so that the signals which are derived therefrom can be corrected for such variations and errors before being used in subsequent processing steps which are intended to define the optical operation of the device.
According to one aspect of the present invention there is provided a system for calibrating an array of light detectors, each detector having an opto-electrical response according to which incident light is converted to an electrical value, the system comprising: a light source arranged to illuminate the array of detectors; a reference detector having a known opto-electric response and being located such that it is illuminated by the light source simultaneously with at least one detector of the array; a measurement device arranged to measure the electrical value output from the detector under calibration and the reference value output from the reference detector whilst those detectors are being illuminated by the light source
and operable to compare said values to derive a sensitivity correction factor (bj)for said detector under calibration.
The measurement device can also be arranged to measure a set of electrical values output from a detector under calibration at differing illumination levels to determine a saturation correction factor (aj) for that detector.
The measurement device can also be arranged to measure the electrical value output from each detector in conditions of no illumination to derive an offset correction factor (q) for each detector.
In the preferred embodiment, these correction factors are utilised in the following equation to provide a corrected output y for a given signal from a photodiode x:
y = a; x2 + bj x + Cj (Equation 1 )
Another aspect of the invention provides a method of calibrating an array of light detectors, each detector having an opto-electrical response according to which incident light is converted to an electrical value, the method comprising: selectively illuminating at least one detector under calibration simultaneously with a reference detector having a known opto-electric response; comparing the electrical value output from the detector under calibration and the electrical value output from the reference detector to derive therefrom a sensitivity correction factor for said detector under calibration.
A further aspect of the invention provides a method of operating an optics device including a dispersive optics element arranged to direct a plurality of optical channels onto an array of light detectors, the method comprising: measuring electrical values output from the array of detectors; applying correction factors to said electrical values, said correction factors having been obtained by calibration of said array, in situ, in the optics device; and using the corrected electrical values to represent incident light on the detectors in subsequent processing steps.
Still further, the invention provides in another aspect an optics device comprising: an array of light detectors, each detector being operable to convert incident light to an electrical value according to an opto-electrical response for that detector; a dispersive optics element for directing a plurality of optical channels onto the array of detectors; a light source selectively controllable to illuminate the array of detectors in a calibration mode; and at least one reference detector having a known opto-electrical response and being located such that it is illuminated by the controllable light source simultaneously with at least one detector of the array, whereby the opto-electrical response of said at least one detector can be ascertained.
For a better understanding of the present invention and to show how the same may be carried into effect, reference will now be made by way of example to the accompanying drawings in which:
Figure 1 is a schematic diagram of a multi-chip module; Figure 2 is a schematic diagram of an optical chip forming part of the multi- chip module;
Figure 3 is a schematic diagram of a processor chip; and
Figure 4 is a graph illustrating the opto-electric response of a photodiode.
Figure 1 illustrates in highly schematic form a multi-chip module 2 which is an integrated package containing at least two chips, an optical chip 4 and a processing chip 6. The optical chip 4 comprises an optical processing component 8 which can for example be an array waveguide grating (AWG). It receives an input light signal 10 in the form of one or more optical channels and generates a light output signal 12 in the form of a plurality of separated optical channels 12a ... 12n. It will be appreciated that Figure 1 is highly schematic and is not an attempt to represent accurately the optical operation of the optical component 8. Light output from the optical component 8 is incident on a photodiode array 14 which comprises a plurality of n photodiodes together with associated amplification
circuitry (not shown) for generating electrical signals relating to the light incident on the photodiodes. These electrical signals are denoted 16 in Figure 1 and are supplied to the processor chip 6.
The substance of this invention concerns the photodiodes in the photodiode array 14 and in particular their calibration, it will be appreciated that photodiodes can be used in a variety of different optical applications and to the extent that any particular application is described herein is by way of example only.
Each photodiode has an opto-electrical response according to which incident light is converted to an electrical value. In order to accurately represent the optical behaviour of the optical component 8 it is important that the electrical values which are output from the photodiodes do accurately represent the incident light. A number of factors mitigate against this.
Photodiodes are subject to leakage currents. That is, even in a situation where there is no light incident on a photodiode, an electrical current may be detected due to leakage within the photodiode itself.
Photodiodes are sold with a nominal sensitivity, that is a nominal electrical value per unit of incident light power. However due to manufacturing variations the precise sensitivity can vary from photodiode to photodiode, and can vary with time and in different environmental conditions.
Photodiodes can be subject to saturation above certain incident light levels. That is, although a linear opto-electrical response is expected from the photodiodes, this linearity can vanish above certain optical levels.
There is described in the following a system for the in situ calibration of photodiodes which are integrated on a silicon chip or otherwise highly embedded and difficult to remove or access directly.
Figure 2 shows the optical chip 4 modified in accordance with the described embodiment of the invention. As in Figure 1 , reference numeral 8 denotes an optical component. Reference numeral 14 denotes a photodiode array to be calibrated. Three light emitting diodes denoted 20 are also arranged on the optical chip 4. In addition, two reference photodiodes 22 are arranged in the vicinity of the photodiode array 14 and the light emitting diodes 20. The light emitting diodes 20 are controlled by signals from the processor chip 6 although these are not shown in Figure 2 for the sake of clarity. Outputs from the photodiodes in the photodiode array 14 and from the reference photodiodes 22 are likewise supplied to the processor chip 6, again the signal lines being omitted from Figure 2 for the sake of clarity. The reference photodiodes are high quality such that their opto-electrical response does not fail or change with time. However the photodiodes in the photodiode array 14 do not need to be of such high quality because they can be calibrated in situ in accordance with the techniques now to be described.
The light emitting diodes 20 are used to shine uniform diffused light onto the photodiode array and simultaneously onto the reference photodiodes 22. Outputs from the photodiode array can then be compared with outputs from the reference photodiodes, thus enabling the extraction of calibration coefficients for photodiodes in the photodiode array 14. In particular, the aim is to extract three calibration coefficients for each photodiode i, aι, bj, q.
The calibration coefficient aj corrects for photodiode saturation. The calibration coefficient bj corrects for sensitivity variations. The calibration coefficient q corrects for leakage currents.
Figure 3 is a schematic diagram illustrating the control signals and responses exchanged between the optical chip 4 and the processor chip 6. The processor chip 6 comprises a controller 6a and a memory 6b. The controller 6a emits LED control signals 30 which turn on the light emitting diodes and which can control the light power output from the light emitting diodes by adjusting the value of the
electrical current supplied thereto. The controller 6a receives a response in the form of an electrical value from the reference photodiodes on signal line 32 and also responses from the photodiodes 0 ... n in the photodiode array 14 along signal line 34. It will be appreciated that distinct signal lines 30, 32 and 34 are shown by way of explanation only and that in practice a common data bus could be utilised.
The memory 6b holds the calibration coefficients which are obtained in a manner to be described in more detail in the following.
Obtaining the calibration coefficient q (for leakage current) will first be described. To obtain this coefficient, the light emitting diodes 20 are turned off and it is ensured that there is no optical input to the optical component 8. Thus in principle the photodiodes of the photodiode array 14 should be in darkness and there should be no output. A measurement of the electrical current from each photodiode i in the photodiode array 14 is taken along signal line 34 to generate a value for the calibration coefficient q for that particular photodiode.
To generate the sensitivity calibration coefficient bj, the light emitting diodes 20 are turned on by the LED control signal 30 to shine uniform diffused light simultaneously onto the photodiodes of the photodiode array 14 and onto the reference photodiodes 22. The electrical currents output from the reference photodiodes are read and are compared with the electrical current output from each photodiode of the photodiode array 14. The coefficient bf is a ratio of the electrical value from the reference photodiodes and each photodiode i in the photodiode array.
The effect of saturation and generation of a saturation calibration coefficient
will now be described with reference to Figure 4. Figure 4 is a graph illustrating the effects of saturation in a photodiode. The vertical axis labelled I represents electrical current output from the photodiode. The horizontal axis represents light incident on the photodiode. The dotted line in Figure 4 represents the expected
linear response of a photodiode having a fixed sensitivity for the full range of expected light input. In fact, a real photodiode often behaves differently at increased light levels, as shown by the full line in Figure 4. For the purposes of the present calibration technique, this is modelled as a quadratic curve with the square coefficient being obtained by taking three or more current values , l
2, l
3 at respectively increasing light values and modelling a curve of best fit on the three values. For the in situ calibration of the photodiode array 14, the required differing light values are obtained by supplying different current levels to the LEDs 20 thereby adjusting their light output. The precise level of incident light is obtained by reading the response from the reference photodiode 22, which has a known opto-electrical response, for each light output level from the LED. This allows the parameters for the curve of Figure 4 to be generated.
The calibration coefficients thus obtained are held in matrix form in the memory 6b. Matrix A denotes factors where no correction of the photodiodes is to be applied. Matrix B denotes a first correction matrix and matrix C denotes a second correction matrix. The calibration coefficients are utilised in accordance with the following equation:
yi = a, Xj2 + bj Xj + q (Equation 1 )
where Xj is the signal read from a particular photodiode in use and yj is the corrected signal which will be used in subsequent processing. Thus in use it will be appreciated that the signal Xj is supplied along signal line 34 and the output yi is supplied as the output signal from the multi-chip module 2.
By providing for two correction matrices, matrix B and matrix C, it is possible to perform in situ calibration while the product is in use. That is, the product can be utilised using a first correction matrix, matrix B while a second correction matrix is being updated to take into account ageing or differing environmental conditions. Thus, operation of the device is not interrupted. In particular, the device does not
have to be removed from the field or interfered with in order to allow calibration to take place.