NJM4151 V-F / F-V CONVERTOR GENERAL DESCRIPTION PACKAGE OUTLINE The NJM4151 provide a simple low-cost method of A/D conversion. They have all the inherent advantages of the voltage-to-frequency conversion technique. The Output of NJM4151 is a series of pulses of constant duration. The frequency of the pulses is proportional to the applied input voltage. These converters are designed for use in a wide range of data conversion and remote sensing applications. NJM4151D NJM4151M FEATURES * Operating Voltage (8V to 22V) * Frequency Operation from (1.0Hz to 100kHz) * Package Outline DIP8, DMP8 * Bipolar Technology PIN CONFIGURATION PIN FUNCTION 1.Io (Current Source) 2.RS (Scale Factor) 3.VLo (Logic Output) 4.GND 5.Ro (One-Shot R,C) 6.Th (Threshold) 7.VI (Input Voltage) 8.V+ NJM4151D NJM4151M EQUIVALENT CIRCUIT V+ Q36 IO Q37 Q41 Q28 M Q35 Q34 10k R1 5K R2 Z N Y R Q32 VLO P X U T W V Q40 Q31 Q30 Q33 R6 RS R13 20k Q27 R14 20k Q26 R7 2K R8 2K VI Q1 Q3 Q2 R15 20K Q24 2K Q25 Th S Q21 Q4 R16 12K Q29 Q23 V+ R9 R3 Q14 Q15 Q12 Q5 Q6 2K R12 Q8 2K R10 2K Q9 Q11 Q10 R4 Q20 Q17 Q18 Q16 Ver.2012-10-24 Q43 Q22 Q13 R11 Q39 2K Q7 Q38 Q42 RO Q19 R5 GND -1- NJM4151 ABSOLUTE MAXIMUM RATINGS PARAMETER (Ta = 25C) SYMBOL RATINGS UNIT V+ 22 V Output Sink Current ISINK 20 mA Power Dissipation PD (DIP8) 500 (DMP8) 300 Input Voltage VI -0.2 to V+ V Operating Temperature Range Topr -40 to + 85 C Storage Temperature Range Tstg -40 to +125 C Supply Voltage mW mW ELECTRICAL CHARACTERISTICS (V+ = +15V, Ta = 25C) PARAMETER TEST CONDITION + TYP. MAX. UNIT 2.0 3.5 6.0 mA 15V < V < 22V 2.0 4.5 7.5 mA Conversion Accuracy Scale Factor VIN = 10V RS = 14.0k 0.90 1.00 1.10 kHz/V Drift with Temperature VIN = 10V - 100 - ppM/C Drift with V+ VIN = 1.0V + 8V < V <18V - 0.2 1.0 %/V Operating Current 8V < V < 15V MIN. + Input Comparator Offset Voltage Offset Current Input Bias Current Common Mode Range (Note 1) One-Shot Threshold Voltage, Pin 5 Input Bias Current, Pin5 Reset VSAT Current Source Output Current (Rs = 14.0k) - 5 10 mV - 50 100 nA - -100 -300 nA 0 to V+-3 0 to V+-2 - V 0.63 0.66 0.70 xV+ - -100 -500 nA Pin 5, I = 2.2mA - 0.15 0.50 V Pin 1, V+ = 0V - 138.7 - A Change with Voltage Pin 1, V+ = 0V to V+ = 10V - 1.0 2.5 A Off Leakage Pin 1, V+ = 0V - 1 50.0 nA Reference Voltage Logic Output VSAT VSAT 1.70 1.90 2.08 V Pin 3, I = 3.0mA Pin 2 - 0.15 0.50 V Pin 3, I = 2.0mA - 0.10 0.30 V - 0.1 1.0 A Off Leakage Note 1 : Input Common Mode Range includes ground. -2- Ver.2012-10-24 NJM4151 PRINCIPLE OF OPERATION Single Supply Mode Voltage-to-Frequency Conversion In this application the NJM4151 functions as a stand-alone voltage to frequency converter operating on a single positive power supply. Refer to Figure 1, the simplified block diagram. The NJM4151 contains a voltage comparator, a one-shot, and a precision switched current source. The voltage comparator compares a positive input voltage applied at pin 7 to the voltage at pin 6. If the input voltage is higher, the comparator will fire the one-shot. The output of the one-shot is connected to both the logic output and the precision switched current source. During the one-shot period, T, the logic output will go low and the current source will turn on with current I. At the end of the one-shot period the logic output will go high and the current source will shut off. At this time the current source has injected an amount of charge Q = I0T into the network RB-CB. If this charge has not increased the voltage VB such that VB > VI, the comparator again fires the one-shot and the current source injects another lump of charge, Q, into the RB-CB network. This process continues until VB > VI. When this condition is achieved the current source remains off and the voltage VB decays until VB is again equal to VI. This completes one cycle. The VFC will now run in a steady state mode. The current source dumps lumps of charge into the capacitor CB at rate fast enough to keep VB VI. Since the discharge rate of capacitor CB is proportional to VB/RB, the frequency at which the system runs will be proportional to the input voltage. Figure 1. Simplified Block Diagram, Single Supply Mode The NJM4151 VFC is easy to use and apply if you understand the operation of it through the block diagram, Figure 1. Many users, though, have expressed the desire to understand the workings of the internal circuitry. The circuit can be divided into five sections: the internal biasing network, input comparator, one-shot, voltage reference, and the output current source. The internal biasing network is composed of Q39-Q43. The N-channel FET Q43 supplies the initial current for zener diode Q39. The NPN transistor Q38 senses the zener voltage to derive the current reference for the multiple collector current source Q41. This special PNP transistor provides active pull-up for all of the other sections of the NJM4151. The input comparator section is composed of Q1-Q7. Lateral PNP transistors Q1-Q4 form the special ground-sensing input which is necessary for VFC operation at low input voltages, NPN transistors Q5 and Q6 convert the differential signal to drive the second gain stage Q7. If the voltage on input pin 7 is less than that on threshold pin 6, the comparator will be off and the collector of Q7 will be in the high state. As soon as the voltage on pin 7 exceeds the voltage on pin 6, the collector of Q7 will go low and trigger the one-shot. Ver.2012-10-24 -3- NJM4151 The one-shot is made from a voltage comparator and an R-S latch, Transistors Q12-Q15 and Q18-Q20 form the comparator, while Q8-Q11 and Q16-Q17 make up the R-S latch. One latch output, open-collector reset transistor Q16, is connected to a comparator input and to the terminal, pin 5. Timing resistor RO is tied externally from pin 5 to V+ and timing capacitor CO is tied from pin 5 to ground. The other comparator input is tied to a voltage divider R3-R5 which sets the comparator threshold voltage at 0.667V+. One-shot operation is initiated when the collector of Q7 goes low and sets the latch. This causes Q16 to turn off, releasing the voltage at pin 5 to charge exponentially towards V+ through RO. As soon as this voltage reaches 0.667 V+, comparator output Q20 will go high causing Q10 to reset the latch. When the latch is reset, Q16 will discharge CO to ground. The one-shot has now completed its function of creating a pulse of period T=1.1 ROCO at the latch output, Q21. This pulse is buffered through Q23 to drive the open-collector logic circuit transistor Q32. During the one-shot period the logic output will be in the low state. The one-shot output is also used to switch the reference voltage by Q22 and Q24. The low T. C. reference voltage is derived from the combination of a 5.5V zener diode with resistor and diode level shift networks. A stable 1.89 volts is developed at pin 2, the emitter of Q33. Connecting the external current-setting resistor RS = 14.0 from pin 2 to ground gives 135A from the collectors of Q33 and Q34. This current is reflected in the precision current mirror Q35-Q37 and produces the output current IO at pin 1. When the R-S latch is reset, Q22 and Q24 will hold the reference voltage off, pin 2 will be at 0V, and the current will be off. During the one-shot period T, the latch will be set, the voltage of pin 2 will go to 1.89V, and the output current will be switched on. TYPICAL APPLICATION 1. Single supply Voltage-to-Frequency Converter Figure 2 shows the simplest type of VFC that can be made with the NJM4151. Input voltage range is from 0 to +10V, and output frequency is from 0 to 10kHz. Full scale frequency can be tuned by adjusting Rs, the output current set resistor. This circuit has the advantage of being simple and low in cost, but it suffers from inaccuracy due to a number of error sources. Linearity error is typically 1%. A frequency offset will also be introduced by the input comparator offset voltage. Also, response time for this circuit is limited by the passive integration network RBCB. For the component values shown in Figure 2, response time for a step change input from 0 to +10V will be 135msec. For applications which require fast response time and high accuracy, use the circuits of Figure 3 and 4. Figure 2. Single Supply Voltage-to-Frequency Converter 2. Precision VFC with Single Supply Voltage For applications which require a VFC which will operate from a single positive supply with positive input voltage, the circuit of Figure 3 will give greatly improved linearity, frequency offset, and response time. Here, an active integrator using one section of the NJM3403A quad ground-sensing op-amp has replaced the RB-CB network in Figure 2. Linearity error for this circuit is due only to the NJM4151 current source output conductance. Frequency offset is due only to the op-amp input offset and can be nulled to zero by adjusting RB. This technique uses the op-amp bias current to develop the null voltage, so an op-amp with stable bias current, like the NJM3403A, is required. -4- Ver.2012-10-24 NJM4151 Figure 3. Precision Voltage-to-Frequency Converter Single Supply 3. Precision Voltage-to-Frequency Converter In this application (Figure 4) the NJM4151 VFC is used with an operational amplifier integrator to provide typical linearity of 0.05% over the range of 0 to -10V. Offset is adjustable to zero. Unlike many VFC designs which lose linearity below 10mV, this circuit retains linearity over the full range of input voltage, all the way to 0V. Trim the full scale adjust pot at VI = -10V for an output frequency of 10kHz. The offset adjust pot should be set for 10Hz with an input voltage of -10mV. The operational amplifier integrator improves linearity of this circuit over that of Figure 2 by holding the output of the source, Pin 1, at a constant 0V. Therefore linearity error due to the current source output conductance is eliminated. The diode connected around the op-amp prevents the voltage at NJM4151 pin 7 from going below 0. Use a low-leakage diode here, since any leakage will degrade the accuracy. This circuit can be operated from a single positive supply if an NJM3403A ground-sensing op-amp is used for the integrator. In this case, the diode can be left out. Note that even though the circuit itself will operate from a single supply, the input voltage is necessarily negative. For operation above 10kHz, bypass NJM4151 pin 6 with 0.01F. Figure 4. Precision Voltage-to-Frequency Converter Ver.2012-10-24 -5- NJM4151 4. Comparison of Voltage-to-Frequency Application Circuits Table 1 compares the VFC applications circuits for typical linearity, frequency offset, response time for a step input from 0 to 10 volts, sign of input voltage, and whether the circuit will operate from a single positive supply or split supplies. Table 1 Linearity Frequency Offset Response Time Input Voltage Single supply Split Supply Figure 2 Figure 3 Figure 4 1% +10Hz 135msec + yes - 0.2% 0 10msec + yes - 0.05% 0 10msec yes yes 5. Frequency-to-Voltage Conversion The NJM4151 can be used as a frequency-to-voltage converter. Figure 5 shows the single-supply FVC configuration. With no signal applied, the resistor bias networks tied to pins 6 and 7 hold the input comparator in the off state. A negative going pulse applied to pin 6 (or positive pulse to pin 7) will cause the comparator to fire the one-shot. For proper operation, pulse width must be less than the period of the one-shot, T=1.1 ROCO. For a 5Vp-p square-wave input the differentiator network formed by the input coupling capacitor and the resistor bias network will provide pulses which correctly trigger the one-shot. An external voltage comparator such as the NJM311 of NJM2901 can be used to "square-up" sinusoidal input signals before they are applied to the NJM4151. Also, the component values for the input signal differentiator and bias network can be altered to accommodate square waves with different amplitudes and frequencies. The passive integrator network RBCB filters the current pulses from the pin 1 output. For less output ripple, increase the value of CB. Figure 5. Single Supply Frequency-to-Voltage Converter -6- Figure 6. Precision Frequency-to-Voltage Converter Ver.2012-10-24 NJM4151 6. Precision Frequency-to-Voltage Converter For increased accuracy and linearity, use an operational amplifier integrator as shown in Figure 6, the precision FVC configuration. Trim the offset to give -10mV out with 10Hz in and trim the full scale adjust for -10V out with 10kHz in. Input signal conditioning for this circuit is necessary just as for the single supply mode, and scale factor can be programmed by the choice of component values. A tradeoff exists between output ripple and response time, through the choice of integration capacitor C1. If C1 = -0.1F the ripple will be about 100mV. Response time constant tR=RB*C1. For RB = 100k and C1 = 0.1F, tR = 10ms. PRECAUTIONS 1. The voltage applied to comparator input pins 6 and 7 should not be allowed to go below ground by more than 0.3 volt. 2. Pins 3 and 5 are open-collector outputs. Shorts between these pins and V+ can cause overheating and eventual destruction. 3. Reference voltage terminal pin 2 is connected to the emitter of an NPN transistor and is held at approximately 1.9 volts. This terminal should be protected from accidental shorts to ground or supply voltages. Permanent damage may occur if current in pin 2 exceeds 5mA. 4. Avoid stray coupling between NJM4151 pins 5 and 7, which could cause false triggering. For the circuit of Figure 2, bypass pin 7 to ground with at least 0.01F. If false triggering is experienced with the precision mode circuits, bypass pin 6 to ground with at least 0.01F. This is necessary for operation above 10kHz. PROGRAMMING THE NJM4151 The NJM4151 can be programmed to operate with a full scale frequency anywhere from 1.0Hz to 100kHz. In the case of the VFC configuration, nearly any full scale input voltage from 1.0V and up can be tolerated if proper scaling is employed. Here is how to determine component values for any desired full scale frequency. 1. Set Rs = 14k or use a 12k resistor and 5k pot as shown in the figures. (The only exception to this is Figure 4.) 2. Set T=1.1R0C0 = 0.75 [1 / f0] where f0 is the desired full scale frequency. For optimum performance make 6.8k < R0 < 680k and 0.001F < C0 < 1.0F 3. a) For the circuit of Figure 2 make CB = 10-2 [1 / f0] Farads. Smaller values of CB will give faster response time, but will also increase frequency offset and nonlinearity. b) For the active integrator circuits make C1 = 5 x 10-5 [1 / f0] Farads. The op-amp integrator must have a slew rate of at least 135 x 10-6 [1/C1] volts per second where the value of C1 is again give in Frads. 4. a) For the circuits of Figure 2 and 3 keep the values of RB and RB as shown and use an input attenuator to give the desired full scale input voltage. b) For the precision mode circuit of Figure 4, set RB = V10/100A where V10 is the full scale input voltage. Alternately the op-amp inverting input (summing node) can be used as a current input with full scale input current I10 = -100A. 5. For the FVCS, pick the value of CB or C1 to give the optimum tradeoff between response time and output ripple for the particular application. DESIGN EXAMPLE I. Design a precision VFC (from Figure 4) with f0 = 100kHz and V10 = -10V. 1. Set RS = 14.0k. 2. T = 0.75 (1/105)=7.5sec Let R0 = 6.8k and C0 = 0.001F 3. C1 = 5 x 1-5 (1/105) = 500pF Op-amp slew rate must be at lease SR=135 x 10-6 (1/500pF)=0.27V/sec 4. RB = 10V/100A = 100k II. Design a precision VFC with f0=1Hz and V10 = -10V. 1. Let RS = 14.0k 2. T = 0.75(1/1) = 0.75sec Let R0 = 680k and C0 = 1.0F 3. C1 = 5 x 10-5 (1/1) F = 50F 4. RB = 100k Ver.2012-10-24 -7- NJM4151 III. Design a single supply FVC to operate with a supply voltage of 8V and full scale input frequency f0 = 83.3Hz. The output voltage must reach at least 0.63 of its final value in 200msec. Determine the output ripple. 1. Set RS = 14.0k 2. T = 0.75 (1/83.3) = 9msec Let R0 = 82k and C0 = 0.1F 3. Since this FVC must operate from 8.0V, we shall make the full scale output voltage at pin 6 equal to 5.0V. 4. RB = 5V/100A = 50k 5. Output response time constant is tR < 20msec Therefore CB < tR / RB = 200 x 10-3 / 50 x 103 = 4F Worst case ripple voltage is: VR = 9mSx135A / 4F = 304mV IV. Design an opto-isolated VFC with high linearity which accepts a full scale input voltage of +10V. See Figure 7 for the final design. This circuit uses the precision mode VFC configuration for maximum linearity. The NJM3403A quad op-amp provides the functions of inverter, integrator, regulator, and LED driver. Figure 7. Opto-Isolated VFC -8- Ver.2012-10-24 NJM4151 APPLICATION EXAMPLE 1. FSK Demodulator FSK(Frequency Shift Keying) data demodulator shows as an example. This is one of continuous input of two frequencies. Transmission of this signal is often used a telephone line. Therefore, the fluctuation amplitude, noise may occur. Therefore, it may experience noise and amplitude fluctuations. FSK demodulator must sense the frequency of which was entered. And it is "0" "1" must be shown at the level of logic. Figure 8 is a circuit diagram FSK. It uses the 2-channel operational amplifier and NJM4151. The FSK signal through a filter to remove noise and high frequency. NJM4151 is the frequency-voltage converter configuration. DC output voltage through a low-frequency filter is converted to a logic output levels. Output op amp will work with hysteresis. The value of parts such as resistors, please determine the following as an example. Two input frequencies, the magnitude of the noise, the response time. FSK INPUT (1070Hz or 1270Hz) Figure 8. FSK demodulator 2. Motor Control Changes in the axis of rotation is changed to pulse train by Trans Deuce. Pulse frequency is proportional to the rotation speed. Pulse train can be changed to a DC voltage proportional to the frequency by the NJM4151. Please refer to Figure 9. Trans Deuce Motor Speed Control frequency NJM4151 Pulce output Figure 9. Motor Control Ver.2012-10-24 -9- NJM4151 3. A/D Converter One example is to connect the microprocessor to analog converter output. A / D converer is required are as follows. Good noise rejection, high stability, low cost. Many industrial systems are sufficient for an 8-bit1/256. A typical microprocessor works with 8-bit input. Data from the AD converter should be parallel with TTL logic levels. Digital data should be controlled by logic input. Typically, the input is called STATUS. Request conversion rate depends on applied field. Digital voltmeter is converted into digital signals per second to 2.3 times. AD converter for the radar pulse must be converted within 100 nanoseconds. AD converter can be configured with a parallel 8-bit output by using the NJM4151. If the slow conversion of the AD converter by using the NJM4151, noise removal is good, better linearity and better isolation can be configured at low cost. Also, you can easily increase the degree of decomposition. If you use the NJM4151, the conversion time is 1s to 10ms. Please refer to Figure 10. Single-digit or further of NJM4151 will be coupled to the counter through a photo coupler. Nch V/F Converter And Multiplication Timing and Counter 8bit parallel output micro processor Positive /Negative Voltage N Trance Deuce Optical coupled Isolator Figure 10. Connect to Micro Processor In many applications, the counter circuit / timing make time division for each channel. Photo coupler is not absolutely necessary. Help to separated the counter / timing circuit and VF-encoder. Thus, VF encoder does not matter even if the contact point away. Therefore, can be used in distance away from the microprocessor. Encoder output pulse train must be parallel to the digital signal. The 8-bit counter is set to 0. Then, only a fixed time to count the pulses from the encoder. Condition that the gate is ON, the counter is only for the specified duration. The higher the input voltage to the encoder, the encoder output frequency is higher. The digital readout of the counters at the end of a certain duration will be directly proportional to the input frequency. To make the best operating, the duration and the encoder scale factor recommend choose to become the maximum count of 255(as an 8-bit binary) at maximum input voltage. Figure 11 is a data conversion system of N channels. A number of transducers is coded each by VFC. N-channel multiplexer is coupling to 8-bit counter. At that time, any one of the encoder output go through with photo-coupler. Channel selection method is as follows. Microprocessor, or manually enter the address of the digital. Timing will be performed by a number of different things. The most accurate way is as follows. It is to be used as a low-frequency by dividing high frequency with the accurate high frequency crystal oscillator. A similar method, it attached divider circuit to a tuning fork oscillator . The most inexpensive way to make the timing is as follows. It is way to drive a one-shot use the optocouplers from the line voltage for AC. Maximum error is less than 0.5% so AC voltage is about 0.2% in most areas.Please note that this method is affected because of the AC line transient response. In addition, NJM555 (Timer IC) can also be used in the astable mode. - 10 - Ver.2012-10-24 NJM4151 from computer Optical coupled Isolator Reset 8bit counter Timing watch Commercial power / Timer 8bit latch Digital output Single stable multi and Timing circuit separated pulse Positive /Negative Voltage common MUX address Optical coupled Isolator frequency Trance Deuce 1 N channel multiplexer V/F N channel encode system Figure 11. Multi channel ADC system Figure 12 shows 1-channel ADC. Commercial power supply frequency has been used to generate the timing. NJM556 (Dual timer IC) is used to strobe and reset pulse generator. Latch circuit is the same as the output data of a counter circuit when the strobe pulse is high level. The counter data is latched when the strobe pulse is a low level. Counter Latch Reset Commercial power Digital output Single stable multi Trance Deuce V/F convetor frequency output Commercial power Encode common for encode common for computer common and counter circuit common Figure 12. 8-bit ADC system (for 1-channel) Ver.2012-10-24 - 11 - NJM4151 This pulse is used to the microprocessor status monitor. After the counter output is latched, the counter is reset by a reset pulse. Then becomes the first status again. This alternate get started the counter under the action of microprocessor instructions. The best logic sequence is determined by individual applications. To illustrate the calculation of the scale factor, try to assume the resolution of 8 bits at 1 / 60 s convert time. Maximum output will be 1+2+4+8+16+ 32+64+128=255. The maximum output frequency is 255 x 60Hz, and 15.3kHz. Accuracy and stability of the ADC is determined by NJM4151 (VFC) chip, an external R,C components, and the timing circuit. Please discuss the error at the following contents: offset, scale factor, linearity, noise. Please consider that there is uncertainty of 0.195% for to put into the prediction error. The approximate ADC has linearity error of 1 / 2 LSB at corresponding to 0.195%. NJM4151 can easily design by linearity of 0.05% or linearity of 0.1%. Most of linearity is not an issue. Offset drift is determined by input offset drift of the external OP-AMP. So in most cases is very small. The largest source of error are gain drift of NJM4151, and temperature coefficient of external R,C components. NJM4151 has a gain drift of 100ppm / C average approximately. Error goal of 100ppm / C can be improved by choosing the external components very well. At 200ppm / C stability of the timing, gain drift of the total will be 400ppm / C. This is the maximum gain variation of 2.8% at 70C from 0C. If the ADC gain drift is too high, you should try and see the power of the microprocessor. Coupled to the input of the VFC at the reference value of zener diodes of low drift at periodically. To calculate the gain error, impress the digital value of the supplied reference voltage to microprocessor and memory circuit. It is easy. The correct term of scale factor calculated is used to correct gain drift. The computer can be used to remove noise. It is time and again to convert the same signal voltage. Next, remove the ones far off the average. Then calculate the average. Using computer technology to all of these applications are not required. However, it is often used to improve the accuracy of the system without increasing the cost of the hardware set. NJM4151 is achieved ADC by a very low cost channel. The accuracy of the system conforms to the resolution of the 8-bit. If necessary, Computer receives the data conversion will reduce in many errors. It can be used to improve the overall accuracy of the system. 4. ADC for proportional meter The accuracy of the systems is better on used proportional method, for measurement equipment. Use the NJM4151, Figure 13 shows the proportion by weight measurement system. First, the strain gauge bridge circuit, and the cells of weight detecting detect the weight. Then generates a difference voltage. The Difference voltage (VD) for the converting bridge approximated VBR / RT in a small range of resistance changes of strain - gauge. Necessary information is R / RT and the voltage (VB) across the bridge must be stable. Instrumentation - amplifier amplify the low input differential voltage (VD). It is used to eliminate the common mode input voltage. The voltage across the bridge (VB) converted to frequency (fB). Then, the output voltage amplitude (V0) converted to frequency (f0). Hereby, ADC is designed for proportional meter. The voltage across the bridge converter is 1Hz / V. Amplifier voltage converter is 1kHz / V conversion factor. Proportional counter and frequency meter type of proportional will be used to convert frequency ratio f0/fB. The first advantage of conversion technology by the proportional method is that it is independent of the voltage across the bridge. Frequency (fB) is proportional to the bridge voltage (VB). Output frequency (f0) is proportional to the VB. f0/fB is independent of VB, only proportional to R / RT. Read Span 000 to 999 -10V Typical value Converter Instruments Amplifier Converter 1kHz/V Scale Proportional Frequency Meter or Proportional Counter Parallel Digital Output Figure 13. Proportional meter system - 12 - Ver.2012-10-24 NJM4151 5. Temperature to Frequency converter Temperature is a physical variable that is most commonly measured. Therefore, the transmitter (transducer) is used. However, it each has advantages and disadvantages. Sensitivity, measuring range, linearity, reproducibility, cost, depends on the transducer. Monolithic transistor pair can be used as temperature sensors. Figure 14 shows the circuit. The advantage of used this circuit linearity and repeatability. AD converter is used so that VFC displayed digital to temperature. Technique of temperature measurement based on a constant characteristic of the transistor. One of dual transistor acts as a temperature-sensitive element. Another dual transistor supplies a constant current to temperature-sensitive portions. High precision amplifier stabilized the voltage difference between the base-emitter's voltage of temperature-sensitive portions. Also, it is amplified at good linearity. It is used to zero to the voltage drop of a fixed I1RE, when the output voltage of the amplifier is 0C. This extra voltage canceled out the absolute temperature. It is necessary to obtain Celsius temperature, when need to read of the absolute temperature can be ignored. VBE=KT/q * 1n * Ic/Is' Except Ic/Is >> 1 K= Boltzmann constant (1.38062 * 10-23 Joule/K) T= Absolute temperature (K) q= Charge of Electron (1.60219 * 10-19 coulomb) Is= Reverse saturation current ( 1.87 * 10-19A) Ic= Collector current When running in the range of collector current is 1A to 1mA, monolithic transistor pair is close to this formula. Transistor pair is high . It is matched pair has a low offset voltage. In addition, IS low, and is also consistent. VBE integrity is within 1mV at a typical value. In addition, the drift is 3V / C, the difference in is less than 10%. The differential input to a precision amplifier, V+IN-V-IN=KT/q * 1n * I1RE+KT/q * 1n * I2/IS2 =KT/q (1n * I2/I1 + 1n * IS1/IS2) - I1RE When Transistor pairs are consistent, IS2 IS1 is the same and the second term 1n1 is zero. The above equation, V+IN-V-IN=KT/q * 1n * I2/IS2 - I1RE KT/q is 86.171V / C. It is 198.4V / K when ratio of I1 versus I2 are 10 versus 1. It subtracted 273.15C when the absolute - temperature want back to Celsius. It is 54.193mV, when Constant current (I1) and RE multiplied. In that case, the amplifier's input voltage is changed to Celsius. 0.1V / C is convenient as the scale factor. It generates a voltage of 0 to 7V at 0 to 70 C ranges. High-precision amplifier input common mode rejection takes 0.5 to 0.6. Then take 504 as an amplifier. The amplifier input bias current should be low in relation to the emitter current of transistor temperature sensitive pair. This temperature measurement method can be applied to other applications. Sensitivity can be adjusted easily by changing the gain of the amplifier or changing the ratio of the collector current. Linearity of output voltage and temperature are within 0.01% linearity in the range of 0 to +70C at circuit of Figure 14. Application will be at large by introducing microprocessors. It can change the sensitivity of the transducer at the computer, and program can also be automated. Ver.2012-10-24 - 13 - NJM4151 Temperature pair detector Instruments Amplifier r V/F Converter Constant current circuit Vo=0.1V/C ( 0 C to +70 C ) Condition: I1=10A, Amplifier gain=170 Figure 14. Temperature to Frequency converter - 14 - Ver.2012-10-24 NJM4151 TYPICAL CHARACTERISTICS VIN-fo Characteristics (VFC) Scale Factor Characteristics Ambient Temperature Characteristics VIN-fo Characteristics (VFC) Input Frequency vs. Output Voltage (FVC) fo vs. Operating Voltage [CAUTION] The specifications on this databook are only given for information , without any guarantee as regards either mistakes or omissions. The application circuits in this databook are described only to show representative usages of the product and not intended for the guarantee or permission of any right including the industrial rights. Ver.2012-10-24 - 15 -