Upload
maxx1nsane
View
214
Download
0
Embed Size (px)
Citation preview
8/7/2019 15731-41504di
1/7www.edn.com April 15, 2004 | ed n 69
ideasdesign
Edited by Bill Travis
Sensitive systems, such as thosein aircraft, must withstand faultconditions, thereby avoiding
component and system damage, be-cause a sensor failure could cause acatastrophic event to occur.A channelprotector, comprising two n-channel
MOSFETs connected in series with ap-channel MOSFET, can protect sen-sitive components from voltage tran-sients in the signal path, whether ornot the power supplies are present(Figure 1). The channel protectoracts as series resistor during normaloperation. If the input exceeds the pow-er-supply voltages, one of the MOSFETsturns off, clamping the output within thesupply rails, thus protecting the circuitryin the event of overvoltage or supply-lossconditions. Because channel protectors
work regardless of the presence of thesupplies, they are also ideal for applica-
tions in which correct power sequencingcannot be guaranteed and for hot-inser-tion rack systems. Figure 2 shows anADG465 channel protector with an inputsignal that exceeds the power-supply volt-age. The protector clamps the output sig-
nal, protecting the sensitive componentsthat follow the channel protector.
When a fault condition occurs, thevoltage on the input of the channelprotector exceeds a voltage set by thesupply-rail voltage minus the MOS-FETs threshold voltage.For a positiveovervoltage, this voltage is V
DD
VTN
, where VTN
is the threshold volt-
age of the NMOS transistor (typical-ly,1.5V).In the case of a negative over-voltage, the voltage is V
SSV
TP, where
VTP
is the threshold voltage of thePMOS device (typically,2V).Whenthe input of the channel protector ex-ceeds either of these voltages, the pro-
tector clamps the output within them.These devices offer bidirectional fault andovervoltage protection,so you can use theinputs or outputs interchangeably.Figure3 shows the voltages and MOSFET statesfor a positive-overvoltage event.
The output load limits the currentduring the fault condition to V
CLAMP/R
L
(Figure 4). If the supplies are off, theprotector limits the fault current tonanoamps. Figure 5 shows how you canuse the ADG466 channel protector toprotect the sensitive inputs of an instru-mentation amp from a sensor fault. Inapplications that require a multiplexer in
Data-acquisition system uses fault protectionCatherine Redmond, Analog Devices, Limerick, Ireland
NMOS
VDD
NMOS
PMOS
PMOS
VDDVSS
VSS
VIN VOUT
VOUT
VD1
VDD
VDD
VS1
VSS
ADG465VIN
VDD
OUTPUT CLAMPED
AT VDD
1.5V
F igure 2
The channel protector clamps overvoltage transients to a safe level.
NMOS NMOSPMOSPOSITIVEOVERVOLTAGE
(20V)
SATURATED NONSATURATED NONSATURATED
VDDVTN13.5V
VDD15V
VDD15V
VSS15V
NOTE: VTN = NMOS-THRESHOLD VOLTAGE (1.5V).
F igure 3
The voltages and MOSFET states appear like this during a positive-overvoltage event.
A channel protector can protect sensi-
tive circuitry from voltage transients.
Data-acquisition system
uses fault protection......................................69
Take steps to reduce
antiresonance in decoupling ......................70
Precision level shifter has
excellent CMRR ..............................................72
Celsius-to-digital thermometer works
with remote sensor........................................74
Quasiresonant converter uses
a simple CMOS IC..........................................74
Simple circuit serves
as milliohmmeter ..........................................78
Publish your Design Idea in EDN. See the
Whats Up section at www.edn.com.
F igure 1
The best ofdesign ideas! Check it out at:
www.edn.com
8/7/2019 15731-41504di
2/770 edn | April 15 , 2004 www.edn.com
ideasdesign
addition to channel protection,you canuse the ADG439F fault-protected, four-channel analog multiplexer
(Figure 6). These multiplex-ers use a series n-channel, p-channel,n-channel MOSFET connection. Duringfault conditions, the inputs or outputsappear as open circuits, protecting thesensor or signal source as well as theoutput circuitry.
SENSOR 1
REFERENCE
+
IN AMP
ANALOG OUT
TO ACTUATOR
ADG466
ADC
DAC
DSP
F igure 5
In this circuit, the ADG466 channel protector guards the sensi-
tive inputs of an instrumentation amplifier from a sensor fault.
SENSOR 1
SENSOR 4
REFERENCE
+
IN AMP
ANALOG OUT
TO ACTUATOR
ADG439F
ADC
DAC
DSP
F igure 6
A multiplexer in a data-acquisition system protects the signal source as well
as the output circuitry.
OVERVOLTAGEOPERATION
(SATURATED)
NONSATURATEDOPERATION
PMOS
V
NMOS
N+ N+ N+EFFECTIVE
SPACE-CHARGEREGION
N-CHANNEL
VD VDD15V
VT = 1.5V
VG VT =13.5V
13.5V
P
VG VS
IOUT
RL VCLAMP
20V
Figure 4
The output load limits the current to VCLAMP
/RLduring a fault condition.
To maintain power integrity on pcboards, you need multiple capaci-tors to decouple the power-distri-
bution system. A typical configurationmight comprise five capacitors connect-ed in parallel between the power and theground traces or planes. To providebroadband decoupling per-formance, assume the indi-vidual values of the capacitors are 470,1, 10, 100, and 220 nF (Figure 1). Thisparallel network provides 801-nF total
capacitance to the power-distributionsystem. If you measure each capacitor
with a vector-network analyzer, you canidentify each capacitors SRF (self-res-onant frequency). Figure 2 is a plot ofeach capacitors SRF, as well as the SRFof the overall parallel connection. EachSRF can cause antiresonance in the par-allel decoupling configuration.Theantiresonance occurs when one ca-
pacitor is still capacitive, while anotherhas become inductive.
470 nF
0.801 F
TOTAL
1 nF 10 nF 100 nF 220 nF
POWER
GROUND
F igure 1
Take steps to reduce antiresonance in decouplingDale Sanders, X2Y Attenuators, LLC, Farmington Hills, MI
A typical decoupling configuration uses several
multilayer-ceramic capacitors connected in
parallel.
400 nF(801 nF TOTAL)
POWER
RETURN
G1 G2
B
A
F igure 3
A 400-nF X2Y capacitor yields a total decou-pling capacitance of 800 nF.
10
0
10
20
30
40
50
60
70
80
VOLTAGE
(dBV)
10 kHz 10 kHz 1 MHz 10 MHz 100 MHz 1 GHz 10 GHz
FIVE STANDARD CAPACITORS: 470, 1, 10, 100,220-nF (801-nF TOTAL CAPACITANCE)
ONE STANDARD 1206 10-nF CAPACITOR
ONE STANDARD 1206 220-nF CAPACITORBOARD S21
ONE STANDARD 1206 1-nF CAPACITOR
ONE STANDARD 1206 1000-nF CAPACITOR
ONE STANDARD 1206 470-nF CAPACITOR
Figure 2
Measurements with a vector-network analyzer reveal undesirable antiresonance effects.
8/7/2019 15731-41504di
3/772 edn | April 15 , 2004 www.edn.com
ideasdesign
A way to considerably reduce the an-tiresonance effects is to use a single 400-nF X2Y capacitor for decoupling. (Ca-
pacitors using X2Y technology areavailable, for example, from JohansonDielectrics (www.johansondielectrics.com). You measure the capacitance rat-ing for an X2Y componentfrom line to ground; in oth-er words, from an A or a B terminal toeither of the G1 or G2 terminals in Fig-ure 3. So, the total capacitance a 400-nFX2Y component supplies, connected asin Figure 3 would be double the capac-itance rating, or 800 nF. Figure 4 showsthat a single X2Y capacitor with the
same total capacitance as in Figure 1provides the same broadband decou-
pling as the standard decoupling con-figuration but without the antireso-
nance effects. In addition, because X2Ycomponents come in the same package
sizes as standard capacitors (1812, 1210,1206, 0805, and 0603), the use of X2Y
components saves pc-board space andreduces layout complexity.
10
0
10
20
3040
50
60
70
80
VOLTAGE
(dBV)
10 kHz 10 kHz 1 MHz 10 MHz 100 MHz 1 GHz 10 GHz
FIVE STANDARD CAPACITORS: 470, 1, 10, 100,220nF (801-nF TOTAL CAPACITANCE)
BOARD S211 X2Y 1206 400 nF
F igure 4
The single X2Y decoupling capacitor displays no antiresonance effects.
Most designers makelevel shifters withop amps and 1%-
tolerance discrete resis-tors. Discrete-resistormismatching limits theop amps CMMR (com-mon-mode rejection ra-tio) to 40 dB, so you can-not use op amps incircuits that require highCMRR. Differential am-plifiers contain precisionmatched internal resis-tors, so ICs such as theINA133 can readily
achieve CMRRs of ap-proximately 90 dB. Theycan offer such highCMRR by trim-ming internal matchedresistors. Assume thateach input in the circuit ofFigure 1 has an associated noise voltage(V
N1, V
N2, and V
NREF). The transfer func-
tion of the amplifier circuit isV
OUT(V
REFV
NREF)(V
IN 2V
N2)
(VIN1V
N1). Note that the reference volt-
age shifts the output signal, either single
or differential.Once this level shifting oc-curs, you can turn your attention to the
noise cancellation. Careful cabling anddifferentially coupling the signal into thedifferential amplifiers inputs force thenoise on the signal inputs to be equal(V
N1V
N2). The input noise is a com-
mon-mode signal,so the differential am-
plifier rejects it to the best of its ability(nominally, 90 dB). Now, V
OUT
VIN2V
IN1V
REFV
NREF.
Now, you need to elim-inate the reference noise
to obtain a clean level-shifted signal. You couldconnect the X end of C
1to
ground to shunt the ref-erence noise to ground,but this solution may beineffective because thesource impedance of thereference is low. When,however,you connect theX end of C
1to the V
IN1sig-
nal source,the differentialamplifier acts as a lowpass
filter and rejects the refer-ence noise. This circuitkeeps the input imped-ance of the differentialamplifier low (approxi-mately 25 k for theINA133) to facilitate
matching.Thus,you must keep the signalsource impedance low to prevent gain er-rors. The source impedance should beless than 1/1000 the input impedance tominimize gain error. If this situationdoesnt occur naturally, then it is best to
buffer the inputs.
Precision level shifter has excellent CMRRRonald Mancini, Texas Instruments, Bushnell, FL
VIN1
VIN2
VOUT
VREF
C1
25k 25k SENSE
25k
25k
X
+V
+IN
V
IN
INA133
_
+
Figure 1
C1
allows the level shifter to act as a lowpass filter that rejects the reference noise.
8/7/2019 15731-41504di
4/774 edn | April 15 , 2004 www.edn.com
ideasdesign
You can use a single-supply system to pre-cisely measure the
temperature at a remotelocation with less than1C error over a 0 to100C range (Figure 1).The circuit includes T
1, a
low-cost AD590 tempera-ture sensor; IC
1, an
AD8541 rail-to-rail am-plifier; four resistors; atrimming potentiometer;and an ADC. You canomit the ADC if you needan analog output. Youcould replace the trim-ming potentiometer withan AD8400 or AD5273digital potentiometer for easier calibra-tion. The feedback resistor, R
F, should be
a precision resistor to minimize the scale-factor error, but the accuracy of the re-
maining resistors is not critical. You canchoose the grade of the AD590 sensor toachieve the required accuracy.
The AD590 provides an output currentproportional to absolute temperature (1A/K).In this application, the circuit off-sets and scales the output to provide afull-scale range of 0 to 5V with a scale fac-tor of 50 mV/C over the chosen tem-perature range of 0Cthe freezingpoint of waterto 100C, the boilingpoint of water. The AD8541 is a low-cost,
low-power, rail-to-rail operational am-plifier.It has a high common-mode volt-age range and extremely low bias cur-rents. You can calibrate out its 1-mV
typical offset, the resistor,and AD590 er-rors. The output swing of the amplifieris 25 mV to 4.965V with a single 5V pow-er supply, limiting the output by about0.5C on either end.
This circuit can derive its power froma single 5V power supply. The output ofthe AD590 varies from 273.15 to 373.15A as the temperature varies from 0 to100C. The positive input of the AD8541has an offset of 4V to provide sufficientheadroom for the AD590. The series
combination of R1
and R2
develops a 1V drop, andyou adjust R
2to provide a
nominal current of 353.15A. Thus, the currentthrough the feedback re-sistor, R
F, varies from80
to20 A as the temper-ature varies from 0 to100C. The voltage across
this resistor varies from4 to1V. The 4V offsetcauses the output voltageof the amplifier to varyfrom 0 to 5V.
To guarantee the accu-racy of 1C throughoutthe range,you need to per-form a calibration proce-
dure. At a known temperature, such as25C, adjust trimming potentiometer R
2
to obtain the desired voltage at the out-put of the amplifier, 1.250V, or the de-
sired code at the output of the ADC,400H.Once you perform the calibration,
you can calculate the temperature in Cel-sius at any measured point inside therange by multiplying the output voltageby 20. Because the sensor has a currentoutput, it is immune to voltage-noisepickup and voltage drops in the signalleads; you can thus use it at a remote lo-cation. You should use a twisted-pair orshielded cable.
AD8541
AD590
IC1
R4
R3
R1RF
AD7476
50kR21k
2k50k
200k
OUT DIGITAL-DATA
OUTPUT
GND
VDD
VIN
LONGWIRES
+T1
_
+
Celsius-to-digital thermometer
works with remote sensorElana Lian and Chau Tran, Analog Devices, Wilmington, MA
This system precisely measures temperature at a remote
location, with less than 1C error over a 0 to 100C range.
Figure 1
Figure 1 shows a flyback power sup-ply that has low noise and uses a sim-ple CMOS 4093 IC for its control.
The electrical noise of a converter arisesmainly when current switches on. Dioderecovery and charging parasitic capaci-tances create high di/dt,which is the maincause of noise. The converter in Figure 1
(pg 76) has a low noise level, because itslowly switches current on at nearly zero
voltage. The converter works in theboundary between discontinuous andcontinuous mode and switches on whenthe drain voltage is at its lowest value. Toavoid working with low gate voltages,which would cause excessive MOSFETlosses, ZD
1conducts and enables the in-
put gate of the 4093 when the voltage is
high enough. When the supply starts, theauxiliary nonisolated winding through D
3
keeps the gate input high. When theMOSFET is on, current increases linearlyuntil the base of Q
5starts to conduct,and
this transistor turns the MOSFET off.Theflyback operation then starts,and the pri-mary energy charges the output capaci-tors. During this phase of operation, D
5
and R6keep Q
5conducting and the MOS-
FET off.When the energy has discharged,
Quasiresonant converter uses a simple CMOS ICFrancesc Casanellas, Aiguafreda, Spain
(continued on pg 78)
8/7/2019 15731-41504di
5/776 edn | April 15 , 2004 www.edn.com
ideasdesign
C2100 F
C11 F
C647 pF
R24.7k
R122k
R68.2k
R154.7k
R12470k
R10470
OC1
4N35
R9
1kC9
1 nF
R7
1k
R16
100
R510k
ZD315V
ZD14.7V
C533 F
Q1
Q3
Q4
D3
D4
D5
D1
D2
D6
Q2
114
7
23
2222
2907
Q5
2369
4093
2369
8
910
12
1311
5
614
R8
C3
C4
C7
R3
R4
VCC
VOUT
R11
ZD2
TL431
R13
R17
R14
C8
C10
+
+
Figure 1
Using a simple CMOS IC, this flyback power-supply circuit exhibits extremely low noise.
8/7/2019 15731-41504di
6/778 edn | April 15 , 2004 www.edn.com
ideasdesign
D5stops conducting, as do the secondary
diodes, so no recovery problems exist.
The time constant of R5 and C5 keepsthe MOSFET off for a while. The outputcapacitance of the MOSFET plus theparasitic capacitance of the primary res-onate with the primary inductance andthe voltage decreases.R
5and C
5allow the
MOSFET to turn on when the voltage
has reached the minimum value. Thevalues are valid only for this case. Thecircuit ofFigure 1 not only minimizes
turn-on losses, but also reduces electri-cal noise. Voltage regulation uses tradi-tional techniques, using a TL431. Theoptocoupler current adds to the shuntcurrent. Because the MOSFET turns onwhen current is zero, the gate resistormay be high, so parasitic capacitances
charge slowly, further reducing switch-ing noise. The circuit around Q
4is op-
tional; you can use it in most power sup-
plies. It kills the current glitch when Q3turns on. It is more effective than theusual RC circuit,and it allows a low dutycycle at low loads. Note that many of thecomponent values in Figure 1 are un-designated; you should determine thesevalues to fit the application.
When I was recently debugging adesign, I discovered that a shortcircuit existed from a ground
plane to a power plane.I did not have ac-cess to a milliohmmeter or an equivalenttester for locating this type of short cir-cuit.So, I logged onto the Internet to findan easily constructible milliohmmeter. I
found the answer in a manufacturersdata sheet,which outlined the basic four-wire method of making low-resistancemeasurements. The method uses a volt-age-reference IC as the input stage for acontrolled constant-current source. Aquick dig in the old component bucketrevealed a supply of LM317 variable-volt-
age regulators. These ICs provide 1.25Vbetween their VOUT
and VADJ
terminals, aconstant voltage to attack the constant-current problem. The other problem toattack was the output-voltage range ofthe constant-current source. The circuitI was working on used a 3.3V supply, soI had to limit the voltage to 3.3V. An
Simple circuit serves as milliohmmeterAM Hunt, Lancaster Hunt Systems Ltd, Shepperton, UK
(continued from pg 74)
8/7/2019 15731-41504di
7/780 edn | April 15, 2004 www.edn.com
ideasdesign
LM317, configured as aconstant-current source,delivers an output voltage
equal to the in-put if the outputresistance is too high. Be-cause I wanted to use abench supply or a 9V bat-tery, the voltage would fryany 3.3V logic on theboard. Ideally, I wantedvoltage to be limited to1.5V. So,I came up with theconfiguration in Figure 1.
IC1
controls the base ofthe npn Darlington transis-
tor,Q1. The IC regulates thevoltage across the selected resistor to formthe constant-current source. The currentsource delivers either 10 or 100 mA, de-pending on which emitter resistor is in thecircuit.The purpose of S
1is to give longer
battery life. You can calibrate the currentsource by strapping a resistive load be-
tween test points A and B and measuringthe voltage across the resistor using aDVM (digital voltmeter). I used 5 and10 and set one S
2position for 10 mA
and the other for 100 mA. To measure asmall resistance, you attach test points Aand B across the resistance. You set the
DVM on a millivolt range.The DVM reads a voltagethat is proportional to the
resistance under test.If youcalibrate the circuit as sug-gested, then the reading is10/V on the 100-mArange and 100/V on the10-mA range.
To track down pc-boardshort circuits, attach theunit with test points A andB across the suspectedshorted signals.Attach oneDVM probe to test point A
and use the other to
probe the circuit. Con-stant voltage along a trace indicates thatno current is flowing and that the trace isnot the source of the short circuit. Lookfor high readings on the trace with thelow reading and low readings on the tracewith the high reading, to locate thesource of the short circuit.
LM317
IC1
S1A
S1B
VIN
VOUT
P1
100
Q1
BD636
P2
10
R1
68
R2
6.8
VADJ
+
+
3V
BATTERY
9V
BATTERY
OR BENCH
SUPPLY
A
B
F igure 1
Make your own milliohmmeter, using a voltage-regulator IC and some resistors.