Acta Polytechnica Vol. 43 No. 212003 Measurement of Light Intensity Based on a Flip-flop Sensor M. Kolli4r Thts paper deals uith a new type of system for measuring light intensitl uith the use of a flip-ftop sensor controlled by a so-called slow-rue aoltage control puke. A photod,iod,e was usedfor quantification of the measured light intensity in the structure of the ftip-Jtop. The theoretical cons'iderations are cornpared uith experimental results, and good agreement is reported. Keywords: measurement, tlip-flop sensor, light intensity, equiualent aoltage. I Introduction The circuit in Frg. I was introduced in [l] as a flip-flop sensor. The flip-flop sensor is part of a class of silicon sensors with a digital oulput. A standard flip-flop consisting from rwo transistors and two resistors (see tig. l) is characterized by two stable states. I and 0. One of the authors of the patent flip-flop sensor was Lian [1], who showed that a flip-flop sensor can be used for measuring non-electrical quantities and derived a formula for calculating the equivalent voltage of the flip-flop sensor controlled by a slow-rise control pulse. The principle of mea- surement is based on the measured non-electrical quantity breaking the value symmetry of the inverters relative to the morphological symmetry axis passing through points K and Z (see Fig. l). Fig. l: Flip-flop sensor However measured quantity can be compensated by a voltage U N =U Nn in such a way that by repeated connection to source U (t) the 50 Vo state I I ] is restored, so that the magni- tude of the measured non-electrical quantiry will be reflected in the voltage U"u, which we will call the equivalent voltage. Howeve4 it is not necessary to stick to the custom of using sensorial elements, as shown in Fig. l It should be noted that in voltage control we also distin- guish between pulses with a fast or slow-rise segment of the control pulse (see FiS. 2). Control with a slow-rise segment of the control pulse is characterized by the ratio U * I \ being such that the currents Fig. 2: Voltage control pulse passing through the capacitors are negligible compared to the transistor currents of the flip-flop sensors. The notion negligible should be understood in its relative sense. In prac- tice, this condition is satisfied if 6r,62 > R,C, and 6,, 62> R2C2 at the same time. The goal of this paper is to show that light intensity can be measured with the use of a flip-flop sensor. 2 Equivalent voltage As described above, the asymmetry of the flip-flop sensor U(t) can be compensated by the equivalent voltage U"r []. If we assume mismatches in the load resistors ,Rr,Rr, saturation currents irrrir, or mismatches in the orrent coefficient B,', B, of the transistors of the flip-flop sensot then the formula for the equivalent voltage has the form [, 2]: A u(t) | A flip-flop circuit with a photodiode for measuring light intensity is shown in Frg. 3. In this case, the value ofthe equivalent voltage can be cal- culated using formula (2): where rRi = ftr + ft, V7 is a thermal voltage and R< R1. In equation (2), l represents the photo-current caused by a mea- 25 (l) [t,^, -^') IUttr=tR."'"|ffin] (2) lp2 )u"1u I IR,t- ilc,II lUr -'l- | - IV lz lr ?z 81 (3) Acta Polytechnica Vol. 43 No. 2/2003 t Vet) Fig. 3: Flip-f1op sensor with a photodiode D sured light intensity. Let us assume that the bipolar transis­ tors of the flip-flop were matched, as were the load resistors. Then the part of the equivalent voltage caused by any distur­ bances and calculated by equation (1) is negligible compared to the expression IR in (2). However, this must be understood in its relative sense, because the assumption is not valid if the value of the equivalent voltage into which the disturbances and the value corresponding to the light intensity are re­ flected are from same range. As shown in section 3 "Proposed solution and experimental results", the contribution of this inaccuracy to the relative error in the result as a function of the light intensity may be particularly relevant in the lower part of the range. 3 Proposed solution and experimental results Fig.4 shows a system for measuring light intensity. Here the feedback is fed from the outputs ofthe flip-flop to the col­ lector resistor R2• As can be seen, in Fig. 4, sampling circuit S/H, the reverse counter and DAC are connected in feedback. It is necessary to distinguish between the controlling pulses, with the symbol <1> in Fig. 4, and the controlling pulses with the shape shown in Fig. 2. The pulses <1> control the digital circuits as sampling circuits, reverse counter and DAC, while the pulses with the shape in Fig. 2 are needed for correct functioning of the flip-flop. The principle of operation is based on the measured Iight intensity breaking the value symmetry of the flip-flop, but as shown above this can be compensated by the equivalent voltage. In this case, the asymmetry will be reflected in the number of pulses read by the reverse counter. The light inten­ sity as a function of time will be equivalent to the number of the pulses read as function of time. As known, the dependence between the voltage, given by expression IR, and the light intensity is in general non-Iinear. This also predicts a non-linear dependence between the light intensity and number of pulses read. One way is to use a poly­ nomial function, the coefficient of which can be found using the well-known least squares method. In our experiment, a polynomial function was used, as follows 2 3 E = aO + aPNE + ~UNE + a3UNE + 4 5 U6+a4UNE + aSUNE + a6 NE, where E represents Iight intensity, UNE = IR and coeffi­ cients: a, =0.1100 [xIV, a2= 5.1936 IxlV2, a3= -0.6859 lxIV3 , a4 =0.0467 IxIV4, as=-0.0012 IxlVs, a6= 1.0082·1O-s lxf0. l.S~---------------, 5[%] 05 oL-~~~_---'_---'-_~_~_-L-_-' o 200 400 600 800 1000 1200 1400 1600 E(Ix) Fig. 5: Relative error as a function of E D Fig. 4: The complete sensor system 26