Introduction To Comparators, Their Parameters And Basic .

2y ago
16 Views
2 Downloads
482.11 KB
27 Pages
Last View : 1m ago
Last Download : 3m ago
Upload by : Rosa Marty
Transcription

AN4071Application noteIntroduction to comparators, their parameters and basicapplicationsBy Radim SmatIntroductionAfter operational amplifiers (op amps), comparators are the most generally used analog,simple integrated circuits.Operational amplifiers are well described in many publications and a lot of information canbe found regarding the design and proper use of these devices. On the other hand,information concerning comparators is much harder to find as they are often considered assimple devices. This note explains the main parameters of comparators and their limitationsfrom an application point of view.Figure 1.October 2012TS332 and TS334 micropower low-voltage rail-to-rail comparatorsDoc ID 022939 Rev 11/27www.st.com

ContentsAN4071Contents1Comparator substitution by an op amp . . . . . . . . . . . . . . . . . . . . . . . . . 42Comparator parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5Comparator classification by major parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . 534Datasheet parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.1Input common mode voltage range - VICM . . . . . . . . . . . . . . . . . . . . . . . . . 83.2Input offset voltage - VIO (VTRIP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.3CMRR and SVR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103.4Voltage gain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113.5Propagation delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11Hysteresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164.1Built-in hysteresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164.2External hysteresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174.3Dynamical hysteresis, example of oscillation issue and the solution . . . . 195Relaxation oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226Window comparator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247Revision history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262/27Doc ID 022939 Rev 1

AN4071List of figuresList of figuresFigure 1.Figure 2.Figure 3.Figure 4.Figure 5.Figure 6.Figure 7.Figure 8.Figure 9.Figure 10.Figure 11.Figure 12.Figure 13.Figure 14.Figure 15.Figure 16.Figure 17.Figure 18.Figure 19.Figure 20.Figure 21.Figure 22.Figure 23.Figure 24.Figure 25.Figure 26.Figure 27.Figure 28.TS332 and TS334 micropower low-voltage rail-to-rail comparators . . . . . . . . . . . . . . . . . . . 1Comparator pinout example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5Signal comparison by TS3011 with built-in hysteresis(blue IN , green IN-, red OUT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6Open drain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7Push-pull. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7Wired-OR configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7Voltage marking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Input-trip point VTRIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8Input offset voltage VIO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8VIO (VTRIP) measurement circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9TS3011 trip point voltage vs. common mode voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10TPLH diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12TPHL diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12TS3011 TPD vs. overdrive voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13TS3011 TPD vs. input common mode voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13Circuit for TPD measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14Trip point voltage definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Input hysteresis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16Non-inverting comparator with hysteresis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Inverting comparator with hysteresis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17Hysteresis circuit example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Dynamical hysteresis circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19Device oscillation without CFB capacitor (green OUT, blue IN ), time scale 500 ns/div. . . 20Case with CFB (green OUT, blue IN ), time scale 500 ns/div. . . . . . . . . . . . . . . . . . . . . . . 20Zoom of the signal (green OUT, blue IN ), time scale 20 ns/div. . . . . . . . . . . . . . . . . . . . . 21Relaxation oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22Window comparator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24Temperature control circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24Doc ID 022939 Rev 13/27

Comparator substitution by an op amp1AN4071Comparator substitution by an op ampTo use operational amplifiers in open loop as comparators is quite common. This especiallyapplies when an op amp is already used in the application, giving the user the opportunity touse a dual channel (or quad channel) op amp which can save space in the application. Thisis possible even if a better alternative is to use comparators that are optimized for thispurpose.The op amp is a device which is designed to be used with negative feedback.A major concern is to ensure the stability of such a configuration. Other parameters like slewrate and maximum bandwidth are trade-offs with current consumption and the architectureof an op amp.Comparators, on the other hand, are designed to operate in open loop configuration withoutany negative feedback. In most cases, they are not internally compensated. The speed(propagation delay) and slew rate (rise and fall time) are maximized. The overall gain is alsousually higher. The use of an op amp as a comparator leads to an unoptimized situation,where current consumption versus speed ratio is low. The opposite is even worse. Normally,a comparator cannot be used instead of an op amp. Most probably, the comparator showsinstability under negative feedback. Generally speaking, comparators and operationalamplifiers cannot substitute each other except for low performance designs.4/27Doc ID 022939 Rev 1

AN40712Comparator parametersComparator parametersComparator classification by major parameters Propagation delay Current consumption Output stage type (open collector/drain or push-pull) Input offset voltage, hysteresis Output current capability Rise and fall time Input common mode voltage range.Besides major parameters, comparators are classified by other parameters such as inputbias current, common mode and power supply rejection ratio, sample/hold function, andstartup time.Figure 2.Comparator pinout example /546##). 6## ).!- A single device has, ordinarily, five pins: two for power supply VCC ,VCC-, two as inputs IN ,IN- and one for the output OUT. It is possible to have an extra pin for standby function.When VIN( ) VIN(-), the output is in high state, if VIN( ) VIN(-), the output is in low state.Doc ID 022939 Rev 15/27

Comparator parametersAN4071Figure 3 shows the comparison made by a TS3011 comparator between a 20 m VPP inputsignal (blue) applied on IN and a 50 mV DC reference voltage (green) applied at IN-. Theoutput signal (red) amplitude is 5 V. As can be seen, sometimes, the input signal exceedsthe reference voltage though the output remains low. This is caused by the internalhysteresis voltage of the TS3011. Hysteresis is discussed in Section 4.Figure 3.Signal comparison by TS3011 with built-in hysteresis(blue IN , green IN-, red OUT)Concerning the output configuration, there are two main types of comparators: withpush-pull and open collector (or open drain). Push-pull stage output levels are typically VCCand 0 V (voltage drop from power supply pins can be neglected). In the case of opencollector configuration, an external pull-up resistor is used. Such a configuration allows anexternal voltage to be applied, different from VCC, to drive the high level state. Pull-upconfiguration can be used as a simple voltage level translator. The second advantage ofa device with open collector is that more outputs can be connected together. This is usefulfor wired-OR configuration systems. It is also possible to find comparators with a differentialoutput stage. These devices are less common and mainly used in telecommunicationsystems as transmission line drivers. For example, the TS3021, TS3011 and TS861 areSTMicroelectronics comparators with push-pull output stage while the LMV331, TS7221and TS331 are open drain configuration.6/27Doc ID 022939 Rev 1

AN4071Comparator parametersFigure 4.Open drainFigure 5.Push-pull6%846 !- Figure 6.!- Wired-OR configuration6%84!- Doc ID 022939 Rev 17/27

Datasheet parametersAN40713Datasheet parameters3.1Input common mode voltage range - VICMThe VICM is defined as the voltage range where both inputs must remain to guarantee thefunctionality of the device. When the input common mode range covers both power supplyrails 0 V and VCC, it is a rail-to-rail input stage.Regarding the use of a rail-to-rail (R2R) comparator, from a technical point of view, if theR2R input capability is not needed, it is always better to choose a comparator without R2Rinput stage. The power consumption is, in that case, obviously smaller and the device cost islower.3.2Input offset voltage - VIO (VTRIP)The input offset voltage (VIO) can be defined as the differential input voltage to apply inorder to be at the toggling level. Input offset voltage limits the resolution of comparators.Therefore, for very small signals (in the same order as the VIO), the comparator toggles atan undesired value or does not toggle at all.Figure 7.Voltage marking6##6).6/546).!- Figure 8.6).Input-trip point VTRIP642)0Figure 9.642)06).6 ).T6)/6).T6/546/54TT!- 8/27Input offset voltage VIODoc ID 022939 Rev 1!-

AN4071Datasheet parametersIn other words, input offset voltage can be represented by a voltage source applied in serieswith one input of an ideal comparator. Consequently, the output doesn't toggle whenVIN VIN-, as in the case of an ideal comparator, but the threshold level is shifted by theinput offset value VIO. Input offset voltage rises in the input stage as a consequence oftransistor imbalance. For comparators with built-in hysteresis, VIO is defined as the averagevalue of VTRIP and VTRIP-, and the hysteresis of VHYST VTRIP - VTRIP- where VTRIP (respectively VTRIP-) is the input differential voltage for which the output switches from low tohigh state (respectively high to low state).MeasurementInput offset voltage and trip points can be measured using the circuit shown in Figure 10.The first DC source sets the power supply VCC and the second sets VICM, the commonmode voltage. A 100 m VPP triangle signal is applied on the voltage divider (1/101). Thevoltage divider is necessary to obtain a good accuracy on the VIO reading by the scope, andto use the function generator with an amplitude that it can handle. The triangle signal shouldbe low frequency (20 Hz); higher frequency can lead to error in the VIO measurementcaused by propagation delay of the device. When the output changes its state, the actualinput voltage value VIN is read from the scope, from VIN we can simply deduce VIO VIN /101.Pay attention to the fact that the oscilloscope probes and waveform generator ground are onthe inverting input pin of the comparator. Therefore, VICM and VCC power supplies must befloating from earth-ground, or an isolation transformer must be used. The advantage is thatthe VICM voltage does not need to be subtracted from the VIO reading. VICM can be easilychanged without having to adjust the offset of the input signal, making the measurementmore comfortable.Figure 10. VIO (VTRIP) measurement circuit02/"% N&6).6)#- 6## N&02/"%!- Doc ID 022939 Rev 19/27

Datasheet parametersAN4071Figure 11. TS3011 trip point voltage vs. common mode voltage!- Use passive oscilloscope probes with 10:1 dividing ratio. For increased accuracy on VIOreading and decreased noise background, the probe applied at VIN can be replaced bya coaxial cable. Unlike propagation delay (TPD) measurement, the cable capacity doesn'thave an impact here because the frequency of VIN is low.Figure 11 is an example of the measured VIO vs. VICM for the TS3011 which is a high speedrail-to-rail comparator from STMicroelectronics. The effect of the two input stages can beclearly seen. Each stage has a different VIO with its own dependency on VICM. One stage isoperating up to VICM VCC - 0.7 V while the second one covers the range from 0.7 V to VCC.The overall curve is the merging of both stages. It can be understood from the presence oftwo VTRIP curves that the TS3011 has a built-in hysteresis.3.3CMRR and SVRThe “common mode voltage rejection ratio” (CMRR) describes the relationship betweeninput offset voltage VIO and the input common voltage VICM. It is defined as VICM over VIOvariation ratio, and is usually represented in logarithmic scale.Equation 1CMRR [dB] 20 log ( ΔVICM /ΔVIO )CMRR is calculated with two values of input offset voltages measured for two different inputcommon mode voltages (usually 0 V and VCC).The “supply voltage rejection” (SVR) is a parameter describing the relation between theinput offset voltage VIO and the power supply voltage. The power supply voltagemodification may affect, more or less, the bias of the input differential transistor pairs.It means that the input offset voltage is also slightly modified.10/27Doc ID 022939 Rev 1

AN4071Datasheet parametersEquation 2SVR [dB] 20 log ( ΔVCC /ΔVIO )Note that, the higher CMMR and SVR are the better.MeasurementThe circuit is the same as the one used for VIO (Figure 10).3.41.CMRRVIO at low (L) VICM rail and VIO at high (H) VICM rail are measuredthen CMRR 20 log [(VICM(H) - VICM(L)) / ( VIO(H)-VIO(L) )]2.SVRVIO at low VCC and VIO at high VCC are measuredthen SVR 20 log [(VCC(H)- VCC(L)) / ( VIO(H)-VIO(L) )]Voltage gainVoltage gain AVD indicates the overall device gain. Higher gain means better small inputsignal resolving capability which can be an advantage in certain applications. Commoncomparators have an AVD in the range of 200 V/mV (106 dB). 1 mV input signal amplified by106 dB leads to theoretical amplitude of 200 V. In reality, the output signal swing is limited byVCC. Note that the AVD doesn’t affect external hysteresis as the output is always in high orlow state and never between (unlike an operational amplifier, a comparator is not used in itslinear region).3.5Propagation delayPropagation delay TPD is one of the key parameters for many applications because it limitsthe maximal input frequency which can be processed. Voltage comparison of analog signalsrequires a minimum amount of time.TPD is defined as the time difference between the moment the input signal crossing thereference voltage and the moment the output state changes (usually when the output signalcrosses 50% of VCC, if nothing is specified).A graphical interpretation is shown in Figure 12 and 13. For TPLH an input square signalfrom -100 mV to VOV, called overdrive voltage and referenced to VICM VTRIP , is appliedon the non-inverting input. The inverting input is connected at VICM voltage. Initial signalcondition 100 mV below VICM VTRIP ensures a sufficient saturation for the input stage. Forthe falling edge propagation delay measurement (TPHL), the input signal goes from 100 mV to -VOV referenced to VICM VTRIP-.Doc ID 022939 Rev 111/27

Datasheet parametersAN4071Figure 12. TPLH diagram6).6 )#- 642)0TIME6/6 M66/54 40,( 6##TIME!- Figure 13. TPHL diagram6). M66/66)#- 642)06/54 TIME40(, 6##TIME!- 12/27Doc ID 022939 Rev 1

AN4071Datasheet parametersThe input signal is referenced to VICM VTRIP and not only to VICM because, in the case ofsmall signal overdrive, an error in TPD measurement can occur due to non-zero VIO (VTRIP and/or VTRIP- 0). In reality, we can neglect VIO without significant impact on TPD for VOVvalues of 20 mV and higher, considering that the offset value reaches around 1 mV andconsidering the exponential dependency of TPD on overdrive voltage, as shown inFigure 14.Figure 14. TS3011 TPD vs. overdrive voltage!- Figure 15. TS3011 TPD vs. input common mode voltage!- Doc ID 022939 Rev 113/27

Datasheet parametersAN4071MeasurementA basic circuit for TPD measurement is shown in Figure 16. One DC power supply is usedfor VCC biasing, and a second one for VICM voltage. If the measurement is performed atVICM 0 V, the source can be removed and the inverting pin directly connected to ground.To keep low source impedance and prevent from parasitic oscillation during switching, a 100nF bypass capacitor is connected close to the positive supply pin VCC of the comparator.A second bypass capacitor should be connected to the comparator input pin IIN- when VICMsource is used. A 50 Ω resistor minimizes the effect of input pin capacitance and avoidssignal reflection on the line matching the impedance at inputs of the comparator with theimpedance of generator (VIN).It is necessary to define exactly the measurement conditions, especially the loadcapacitance CLOAD which has a big impact on output signal edge speed, consequently, alsoon the TPD value (measured at 50% of VOUT). The C LOAD represents the overall capacitiveloading at the comparator output including loading capacitor, oscilloscope probe capacityand parasitic capacity of the PCB track.Figure 16. Circuit for TPD measurement6## N&02/"% P&6).02/"% N&6)#-#,/! P&!- Many pulse generators, despite great performance in the time-base, are not able to provideaccurate signal amplitude, especially for low output voltage. Generating 5 mV overdrive maybe a problem because such a value is often below the generator accuracy. Therefore, forsmall overdrive measurement, it is more suitable to place a 50 Ω attenuator (divider bridge),instead of a single 50 Ω resistor, and increase the generator amplitude. In this way, a goodamplitude accuracy can be obtained with low overdrive values.High speed signal processingFor high-speed signal applications, more attention (at PCB level) must be paid to: proper lowresistive grounding, short tracks and quality SMD capacitors having low ESR. Bypasscapacitor stores charge and provides supplementary source when spikes occur on the VCCline. Each real capacitor has resonant frequency where its impedance reaches the lowestvalue. If the input signal frequency is far from the resonant frequency, impedance stronglyincreases and the capacitor loses bypassing capability. Placing different capacitors withdifferent resonant frequencies therefore allows a wide frequency bandwidth to be covered.14/27Doc ID 022939 Rev 1

AN4071Datasheet parametersSuch a bypass combination, made from 100 nF, 10 nF and 1 nF in parallel, eliminatesunwanted spikes on the VCC line much better than only one 100 nF capacitor.Each mm of the track plays a role. Bypass capacitors must be placed as close as possible tothe comparator supply pin. Place the smallest capacitor closer to the supply pin than thebigger one. Removing GND copper under signal track minimizes overall capacitive loading.Input source impedance shouldn't exceed 1 kΩ, otherwise undesirable oscillations canappear. In addition, too high input impedance in series with parasitic PCB capacity and inputcomparator capacity produces additional RC constant which means additional propagationdelay.Concerning time measurements on high-speed comparators, remember that for a highspeed signal the oscilloscope and probe can cause significant error in measurementaccuracy when their bandwidth is too low. The measured TRISE (TFALL) value is affected bythe rise time of the scope and the probe by:Equation 3TRISE (TRISE SIGNAL2 TRISE SCOPE2 TRISE PROBE2)1/2Doc ID 022939 Rev 115/27

Hysteresis4AN4071HysteresisFor slow time changing input signal, an output oscillation can appear while the input signalremains close to the reference voltage. Also low amplitude signal on high impedance cancause oscillations due to noise background. Such unwelcome behavior can be solved byhysteresis. The principle of hysteresis consists of two different input threshold voltagesdepending on actual output state.4.1Built-in hysteresisMany comparators have built-in hysteresis. Typical hysteresis value is a few mV. This isenough to suppress output undesired toggling in most cases but it doesn't impactsignificantly the resolution of the comparator. For comparators with built-in hysteresis, theaverage lower and upper threshold voltage is computed and referred as input offset voltageVIO, the VTRIP and VTRIP- difference is referred as hysteresis voltage VHYST and is shown inFigure 17.Figure 17. Trip point voltage definition6(9346).642)06)/642)06/54!- Figure 18. Input hysteresis6/546) 6 42)06)/642)0!- 16/27Doc ID 022939 Rev 1

AN40714.2HysteresisExternal hysteresisIf the device doesn't include built-in hysteresis, or if a large hysteresis is required, a positivefeedback network can be implemented. Figure 19 shows a non-inverting and Figure 20 aninverting comparator with hysteresis.Figure 19. Non-inverting comparator with hysteresis6 /542 &"2)6 /(6 ##6 ). 3)'.!,6 2%&6/546 ).6 /,6 4( 6 4(!- Figure 20. Inverting comparator with hysteresis6/542 &"6 2%&2)6 /(6##6 /546 ). 3)'.!,6 ).6 /,64(6 4(!- For a non-inverting comparator circuit, neglecting input offset voltage and the effect of inputbiasing current, the input threshold voltages are:Equation 4V TH RI ---------- ( V R EF – V OL ) V R EFR FBV TH –RI ---------- ( V REF – V OH ) V REFR FBVOL is the saturation voltage in low state and VOH is the saturation voltage in high state atthe comparator output. The VTH - VTH- difference determines the hysteresis voltage (VHYST)while their average determines the middle of hysteresis (VTRIP).Equation 5V HYST V TH – V TH –RI ---------- ( V OH – V OL )R FBRIV TH V T H –V OL V OH V TRIP ----------------------------------------- V REF ---------- V REF – --------------------------- R FB22In Equation 5 the influence of reference voltage on the trigger voltage level can be seen. Thetrip point voltage VTRIP (middle of hysteresis) is equal to the reference voltage VREF onlywhen the second part of the equation equals zero, it means when VREF is set just to theDoc ID 022939 Rev 117/27

HysteresisAN4071centre of output voltage swing. Otherwise VREF and VTRIP are different and the hysteresis isnot centred on VREF.For calculating RI, RFB and VREF, two formulas, obtained from Equation 4 and 5, can beused:Equation 6V OH – V OLR FB---------- ------------------------V H YSTRIR FBRIV REF ------------------------------------- ( V OH V OL ) ----------------------- V TRI P2 ( R I R FB )R I R FBExample of real design of comparator using external hysteresis network:Task: find resistor values and voltage reference in order to implement 400 mV hysteresis on1.2 V threshold voltage. Use push-pull comparator TS3021 powered by 5 V source.Solution: knowing the output voltage level swing and required hysteresis, calculate theresistor ratio and choose the appropriate resistor (in the range of hundreds of kΩ). Then,substituting resistor values and trip point voltage in the second formula, calculate thereference voltage.For 5 V output swing and hysteresis 400 mV, the ratio of feedback and input resistor is:Equation 7V OH – V OLR FB5V--------- --------------------------- ------------ 12.50.4VV HYSTRIAppropriate resistors are RFB 300 kΩ and RI 24 kΩ. The VREF can now be easilycalculated:Equation 8R FBR1V R EF ------------------------------------- ( V OH V OL ) ----------------------- V TRIP 2 ( R I R FB )R I R FB452.4 103 10- 5 ------------------------------------------------- 1.2 0.185 1.111V 1.296V ------------45452 ( 2.4 10 3 10 )2.4 10 3 10Figure 21 shows the final circuit. A reference voltage is generated by the voltage dividersupplied from the 5 V source of the comparator. With resistors 68 kΩ and 24 kΩ from theE24 series, the VREF is equal to 1.3 V.18/27Doc ID 022939 Rev 1

AN4071HysteresisFigure 21. Hysteresis circuit example 6)./54 643 N&!- 4.3Dynamical hysteresis, example of oscillation issue and thesolutionDynamical hysteresis is another way to eliminate parasitic oscillation during the transitionperiod. When the input signal changes slowly around the reference voltage, an outputoscillation may occur. A small capacitor C FB applied between the output and the noninverting pin boosts up the input signal to go over (or below) the reference voltage faster andin consequence to help reduce oscillations.The circuit example in Figure 22 shows the usage of the CFB. To induce some outputoscillations (on purpose only), consider the case of the application shown in Figure 22: theimpedance of the 0.9 V reference voltage is too high because of the 6.8 kΩ resistor. First,consider the situation without CFB when VIN exceeds VIN- and the output toggles to highstate. The fast output edge, together with the parasitic PCB capacity Cparasitic between theIN- and OUT pin, causes a positive voltage peak to IN-. VIN- now becomes higher than VIN .Consequently, the output is returning to low state. The negative peak now goes back toIN-, VIN- is lower than VIN and the output returns to high. This leads to repetitiveoscillations, as shown in Figure 23. The frequency of the oscillations is related to the TPDof the comparator, here it is 10 MHz (TPD 50 ns).Figure 22. Dynamical hysteresis circuit#&" 6 P& 6/54 6 6 K(Z43 # PARASITIC!- Doc ID 022939 Rev 119/27

HysteresisAN4071Figure 23. Device oscillation without CFB capacitor (green OUT, blue IN ), time scale500 ns/div.Applying a 22 pF feedback capacitor between the IN and OUT pin stops the oscillations, asshown in Figure 24 and Figure 25. This feedback capacitor creates a peak ( 150 mV for20 ns) on the IN pin which securely eliminates the effects of the peak on VIN- coming fromCparasitic. The feedback capacitor CFB implements a dynamic hysteresis.Figure 24. Case with CFB (green OUT, blue IN ), time scale 500 ns/div.20/27Doc ID 022939 Rev 1

AN4071HysteresisFigure 25. Zoom of the signal (green OUT, blue IN ), time scale 20 ns/div.Main general contributors leading to device oscillation Parasitic capacity/inductance onboard: long narrow wires, signal and output track closetogether, device plugged into socket Fast (sharp) output edges: faster edge means higher dV/dT and therefore biggerimpact of parasitic capacities and inductances on the board High impedance on input pins: sensitivity to noise, increased effect of parasiticstructures and signal crosstalk Poor grounding High power supply impedance: inappropriate or missing bypass capacitor No hysteresis (static or dynamic) implemented Use of high-speed comparator where it isn't necessary.Doc ID 022939 Rev 121/27

Relaxation oscillator5AN4071Relaxation oscillatorA relaxation oscillator belongs to the regenerative circuits group. One subgroup ismultivibrators which are furthermore classified as monostable, bistable and astable. Therelaxation oscillator is an astable multivibrator.Figure 26. Relaxation oscillator 62 2 /542 N& # !- The circuit in Figure 26 shows a representative circuit of relaxation oscillator based on theTS3021 comparator. It uses both positive and negative feedback. Positive feedbackproduces voltage hysteresis which has already been described in Section 4.2 Thresholdvoltages VLOW (from low to high) and VHIGH (from high to low) on the inverting input aregiven by R2, R3 and R4 resistors together with output voltage given by power supply voltage.Considering zero voltage drop at the output see Equation 9:Equation 9VLOW VCC R2 R4 / (R3 R2 R 4) and VHIGH VCC R2 / (R2 R3 R4)For R2 R3 R4: VLOW 1/3 VCC and VHIGH 2/3 VCCVoltage on the non-inverting input is generated by charging and discharging capacitor C1from the comparator output via resistor R1 in the feedback: V - (V - V) e-t/τ V - 2/3 V e-t/τ1. While C is charging: V1C1(t)CCCCLOWCapacitor C1 is charged at voltage VHIGH after time t1:Equation 10VHIGH VCC - 2/3 VCC e-t1/τ2/3 VCC

AN4071 Comparator parameters Doc ID 022939 Rev 1 5/27 2 Comparator parameters Comparator classification by major parameters Propagation delay Current consumption Output stage type (open collector/drain or push-pull) Input offset voltage, hysteresis Output current capability Rise and fall time Inp

Related Documents:

Low Power Low Offset Voltage Quad Comparators General Description The LM139 series consists of four independent precision voltage comparators with an offset voltage specification as low as 2 mV max for all four comparators. These were designed specifically to operate from a single power supply over a wide range of voltages. Operation from split .File Size: 400KB

Low Power Low Offset Voltage Quad Comparators General Description The LM139 series consists of four independent precision voltage comparators with an offset voltage specification as low as 2 mV max for all four comparators. These were de-signed specifically to operate from a single power

Low Power Low Offset Voltage Quad Comparators General Description The LM139 series consists of four independent precision voltage comparators with an offset voltage specification as low as 2 mV max for all four comparators. These were designed specifically to operate from a single power suppl

Conversion: Resolution: n ref outin V V Integer 2 n V ref V 2 1 14.7 Analogue to Digital Converters ADC's If the conversion rule for a DAC is n ref out in V . these use a number of comparators in parallel. o In general, to generate n bits, it takes 2n-1 comparators. So, for something modest like 12 bits, it takes 4,095 comparators! Not .

Linear Measurement Comparators Electronic comparators Advantages of Electrical and Electronic comparator 1. It has less number of moving parts. 2. Magnification obtained is very high. 3. Two or more magnifications are provided in the same instrument to use various ranges. 4.

LMx39-N, LM2901-N, LM3302-N Low-Power Low-Offset Voltage Quad Comparators 1 Features 3 Description The LMx39-N series consists of four independent 1 Wide Supply Voltage Range precision voltagecomparators with an offset LM139/139A Series 2 to 36 VDC or 1 to 18 VDC specifica

an offset voltage specification as low as 2 mV maximum for LM339A, LM239A, and LM139A. Each comparator has been designed specifically to operate from a single power supply over a wide range of voltages. Operation from split power supplies is also possible. These comparators also have a unique

2 API R. ECOMMENDED. P. RACTICE. 500. 1.2.4. Section 9 is applicable to locations in which flammable petroleum gases and vapors and volatile flammable liquids are processed, stored, loaded, unloaded, or otherwise handled in petroleum refineries. 1.2.5 . Section 10 is applicable to location s surrounding oil and gas drilling and workover rigs and production facilities on land and on marine .