Instrumentation Applications for a Monolithic Oscillator

Application Note 93
February 2003
Instrumentation Applications for a Monolithic Oscillator
A Clock for All Reasons
Jim Williams
INTRODUCTION
Clock Types
Oscillators are fundamental circuit building blocks. A
substantial percentage of electronic apparatus utilize oscillators, either as timekeeping references, clock sources,
for excitation or other tasks. The most obvious oscillator
application is a clock source in digital systems.1 A second
area is instrumentation. Transducer circuitry, carrier based
amplifiers, sine wave formation, filters, interval generators and data converters all utilize different forms of
oscillators. Although various techniques are common, a
simply applied, broadly tunable oscillator with good accuracy has not been available.
Commonly employed oscillators are resonant element
based or RC types.2 Figure 1 shows two of each. Quartz
crystals and ceramic resonators offer high initial accuracy
and low drift (particularly quartz) but are essentially
untunable over any significant range. Typical RC types
have lower initial accuracy and increased drift but are
easily tuned over broad ranges. A problem with conventional RC oscillators is that considerable design effort is
required to achieve good specifications. A new device, the
LTC1799, is also an RC type but fills the need for a simply
applied, broadly tunable, accurate oscillator. Its accuracy
and drift specifications fit between resonator based types
and typical RC oscillators. Additionally, its board footprint,
a 5-pin SOT-23 package and a single resistor, is notably
small. Note that no external timing capacitor is required.
CLOCK TYPE
Quartz
TYPICAL
TYPICAL
FREQUENCY FREQUENCY
ACCURACY
RANGE
TUNABILITY
TEMPERATURE
COEFFICIENT
POWER SUPPLY
REJECTION RATIO
COMMENTS
0.005%
10kHz to
200MHz
Poor
0.5ppm/°C
Easily Achieved.
See Comments
1ppm/V
High Stability and Initial Accuracy at Expense of
Tunability. Essentially No Tunability. 1 • 10 – 9
Stability Achievable with Compensation Techniques
Ceramic
Resonator
0.5%
250kHz to
60MHz
Poor
30ppm/°C
20ppm/V
Lower Performance and cost than Quartz.
Essentially Untunable
LTC1799
1.5%
1kHz to
33MHz
Good
40ppm/°C
Plus Resistor
Temperature
Coefficient
500ppm/V
Add 10 to 50ppm/°C Temperature Coefficient,
Depending on Resistor Type. Extremely Small
Footprint— SOT-23 and 1 Resistor
Typical RC
Based Clock
10%
1Hz to
25MHz
Good
200ppm/°C
2500ppm/V
Requires Careful Design and Component Selection
for Best Results
Figure 1. LTC1799 Compared to Other Oscillators. Quartz and Ceramic Based Types Offer Higher Frequency Accuracy
and Lower Drift but Lack Tunability. RC Designs are Tunable but Accuracy, Temperature Coefficient and PSRR are Poor
, LTC and LT are registered trademarks of Linear Technology Corporation.
Note 1: Strictly speaking, an oscillator (from the Latin verb, “oscillo,”
to swing) produces sinusoids; a clock has rectangular or square wave
output. The terms have come to be used interchangably and this
publication bends to that convention.
Note 2: This forum excludes such exotica as rubidium and cesium
based atomic resonance devices, nor does it admit mundane but dated
approaches such as tuning forks.
AN93-1
Application Note 93
Figure 2 shows how simple to use the LTC1799 is. A single
resistor (RSET) programs the device’s internal clock and
pin-settable decade dividers scale output frequency. Various combinations of resistor value and divider choice
permit outputs from 1kHz to 33MHz.3 Figure 3 shows RSET
vs output frequency for the three divider pin states and the
governing equation. The inverse relationship between
resistance and frequency means that LTC1799 period vs
resistance is linear.
Figure 4 reveals that the LTC1799 has speciated into a
family. At present, there are two additional devices. The
LTC6900, quite similar, cuts supply current to 500µA but
gives up some frequency range. The LTC6902, designed
for noise smoothed, multiphase power applications,
has multiphase outputs and spread spectrum capability.
Spread spectrum clocking distributes power switching
over a settable frequency range, preventing significant
noise peaking at any given point. This greatly reduces EMI
concerns.
cuitry. The following text utilizes the device’s attributes in
a variety of such applications.
Platinum RTD Digitizer
A platinum RTD, used for RSET in Figure 5, results in a
highly predictable O1 output period vs temperature. O1’s
output, scaled via counters, is presented to a clocked,
period determining logic network which delivers digital
output data. Over a 0°C to 100°C sensed temperature,
1000 counts are delivered, with accuracy inside 1°C.
Extended range (sensor limits are – 50°C to 400°C) is possible by using a monitoring processor to implement linearity correction in accordance with sensor characteristics.4
1000
The LTC1799’s combination of simplicity, broad tunability
and good accuracy invites use in instrumentation cir-
3k ≤ RSET ≤ 1M
GND
1k
SET
DIV
AN93 F02
÷10
RSET = 10k •
OPEN
÷1
FREQUENCY
RANGE
FREQUENCY
ACCURACY
TEMPERATURE
COEFFICIENT
( )
10MHz
N • fOSC
100M
100
, N = 10
1
AN93 F03
Figure 3. RSET vs Output Frequency for the Three
Divider Pin States and Governing Equation.
Relationship between RSET and Frequency Is Inverse;
RSET vs Period has Linear Characteristic
Figure 2. LTC1799 Oscillator Frequency Is
Determined by RSET and Divider Pin (DIV).
Tunable Range Spans 1kHz to 33MHz
DEVICE
TYPE
MOST
ACCURATE
OPERATION
100k
1M
10M
10k
DESIRED OUTPUT FREQUENCY (Hz)
5V
÷100
÷1
1
V+
OUT
LTC1799
÷10
10
1kHz ≤ fOSC ≤ 33MHz
5V
÷100
100
RSET (kΩ)
A (Very) Simple, High Performance Oscillator
PSRR
COMMENTS
LTC1799 1kHz to 33MHz
1.5%
40ppm/°C + Resistor Drift 0.05%/V ISUPPLY = 1mA
LTC6900 1kHz to 20MHz
1.5%
40ppm/°C + Resistor Drift 0.04%/V Low Power (ISUPPLY = 500µA) Version of LTC1799
LTC6902 5kHz to 20MHz
1.5%
40ppm/°C + Resistor Drift 0.04%/V 2-, 3- or 4-Phase Outputs. Programmable Width Spread Spectrum
Frequency Modulation. Intended for Multiphase Power Supply
Applications
Figure 4. Oscillator Family Details. LTC6900 Is Low Power Version of LTC1799. LTC6902, Intended for Noise
Sensitive, High Power Switching Regulator Applications, Has Multiphase, Spread Spectrum Outputs. All Types
Have Excellent Tunability, Good Frequency Accuracy, Low Temperature Coefficient and High PSRR
Note 3: This deceptively simple operation derives from noteworthy
internal cleverness. See Appendix A, “LTC1799 Internal Operation” for
a description.
AN93-2
Note 4: Linearity deviation over –50°C to 400°C is several degrees.
See Reference 1.
Application Note 93
5V
19.1k* (TRIMMING OPTIONAL—SEE TEXT)
5V
RSET
PLATINUM
RTD
0°C = 10k
100°C = 13.85k
5.208MHz (IDEAL)
LTC1799 OUT
DIV
–
O2
A1
LT1219L
DATA OUTPUT
0°C = 2604 COUNTS
100°C = 3604 COUNTS
0.1µF
+
0°C = 100kHz (10µs PERIOD)
100°C = 72.27kHz (13.837µs PERIOD)
100k
* = 0.1% METAL FILM RESISTOR
RTD = MINCO S19827PS12
RSET
NC
DIV LTC1799 OUT
SQUARE WAVE
÷ 10
74HC90
74HC90 ÷ 10
÷5
= 1/4 74HC00
RESET
O1
AN93 F05
Figure 5. Platinum RTD Digitizer Accurate within 1° Over 0°C to 100°C. Platinum RTD Value Is Linearly
Converted to Period by LTC1799. Logic and Second LTC1799 Clock Digitize Period into Output Data Bursts.
A1 Drives RTD Shield at RSET Potential, Bootstrapping Pin Capacitance to Permit Remotely Located Sensor
If the RTD is at the end of a cable, the cable shield should
be driven by A1 as shown. This bootstraps the cable shield
to the same potential as RSET, eliminating jitter inducing
capacitive loading effects at the RSET node.5
Figure 6 shows operating waveforms. The RTD determines O1’s output (Trace A), which is divided by 100 and
assumes square wave form (Trace B). The logic network
combines with O2’s fixed frequency to digitize period
measurement, which appears as output data bursts
(Trace␣ C). The logic also produces a reset output (Trace␣ D),
facilitating synchronization of monitoring logic.
A = 5V/DIV
As shown, accuracy is about 1.5°C, primarily due to
LTC1799 initial error. Obtaining accuracy inside 1°C involves simulating a 100°C temperature (13,850Ω) at the
sensor terminals and trimming RSET for appropriate output. A precision resistor decade box (e.g., ESI DB62)
allows convenient calibration.
Thermistor-to-Frequency Converter
Figure 7’s circuit also directly converts temperature to
digital data. In this case, a thermistor sensor biases the
RSET pin. The LTC1799 frequency output is predictable,
although nonlinear. The inverse RSET vs frequency relationship combines with the thermistor’s nonlinear characteristic to give Figure 8’s data. The curve is nonlinear,
although tightly controlled.
B = 5V/DIV
5V
C = 5V/DIV
V+
OUT
LTC1799
D = 5V/DIV
RT
GND
SET
fOSC = 10MHz • 10k
RT
10
RT = YSI #44011
DIV
NC
AN93 F07
HORIZ = 100µs/DIV
AN93 F06
Figure 6. Platinum RTD Biased LTC1799 Produces Output
(Trace␣ A) which Is Divided by 100 (Trace B) and Gated with
5.2MHz Clock. Resultant Data Bursts (Trace C) Correspond to
Temperature. Reset Pulse (Trace D), Preceding Each Data Burst,
Permits Synchronization of Monitoring Logic
Figure 7. Simple Temperature-to-Frequency
Converter Biases RSET with Thermistor. Frequency
Output Is Predictable, Although Nonlinear
Note 5: The RSET node, while not unduly sensitive, requires management of stray capacitance. See Appendix B, “RSET Node Considerations” for detail.
AN93-3
Application Note 93
1400
1200
FREQUENCY (kHz)
Isolated, 3500V Breakdown, Thermistor-to-Frequency
Converter
MAX
TYP
MIN
This circuit, building on the previous approach, galvanically isolates the thermistor from the circuit’s power and
data output ports. The 3500V breakdown barrier between
the thermistor and power/data output ports permits operation at high common mode voltages. Such conditions
are often encountered in industrial measurement situations.
1000
800
600
400
200
0
Figure 9’s pulse generator, C1, running around 10kHz,
produces a 2.5µs wide output (Trace A, Figure 10). Q1-Q2
provide power gain, driving T1 (Trace B is Q2’s collector).
T1’s secondary responds, charging the 100µF capacitor to
a DC level via the 1N5817 rectifier. The capacitor powers
O1, which oscillates at the sensor determined frequency.
O1’s output, differentiated to conserve power, switches
Q4. Q4, in turn, drives T1’s secondary, T1’s primary
receives Q4’s signal and Q3 amplifies it, producing the
circuit’s data output (Trace C). Q3’s collector also lightly
–20 –10 0 10 20 30 40 50 60 70 80 90
TEMPERATURE (°C)
AN93 F08
Figure 8. LTC1799 Inverse Resistance vs Frequency Relationship
and Nonlinear Thermistor Characteristic Result in above Data.
Curve Is Nonlinear, Although Tightly Controlled
5V
POWER
DRIVER
PULSE
GENERATOR
1N4148
1k
100Ω
226k*
–
1000pF
5%
Q4
2N3904
Q2
ZTX-749
1k
1k
C1
LT1671
2Ω
1N5817
Q1
2N2369
3
1
4
6
+
750k
5V
50pF
5V
750k
T1
ISOLATION/
POWER
TRANSFORMER
4.3k
C2
Q3
2N3904
1k
+
200Ω
–
* = 1% METAL FILM RESISTOR
RT = YSI 44006
T1 = BI TECHNOLOGIES HM-41-11510
CKT GROUND
FLOATING COMMON
< 4.5V LOCKOUT
6
BAT85
OUTPUT DATA
DEMODULATOR
3
LT1635
50pF
RSET
DIV
O1
NC
ISOLATED
TEMPERATURE
SENSOR AND DIGITIZER
5V
820pF
1k
LTC1799 OUT
GND
BAT85
470Ω
V+
100µF
750k
DATA
OUT
1N4690
5.6V
1N4148
+
RT
2
1, 8
0.2V
THERMISTOR
SENSOR
OUTPUT
VALUE (Ω) TEMPERATURE (°F) FREQUENCY (Hz)
5k
109
2.01M
10k
77
1.01k
20k
47
505k
30k
31
337k
40k
20
253k
50k
12
203k
60k
6
168k
70k
–1.3
145k
80k
–4.7
127k
90k
–8.5
113k
100k
–12
101k
AN93 F09
Figure 9. A Galvanically Isolated Thermistor Digitizer. C1 Sources Pulsed Power to Thermistor Biased LTC1799 via Q1, Q2 and T1.
LTC1799 Output Modulates T1 through Q4. Q3 Extracts Data, Presents Ouput. T1’s 3500V Breakdown Sets Isolation Limit
AN93-4
Application Note 93
modulates C1’s negative input (Trace D), synchronizing
T1’s primary drive to the data output. C2 prevents erratic
circuit operation below 4.5V by removing Q1’s drive.
Relative Humidity Sensor Digitizer-Hetrodyne Based
Figure 11 converts the varying capacitance of a linearly
responding relative humidity sensor to a frequency output. The 0Hz to 1kHz output corresponds to 0% to 100%
sensed relative humidity (RH). Circuit accuracy is 2%,
plus an additional tolerance dictated by the selected sensor grade. Circuit temperature coefficient is ≈400ppm/°C
and power supply rejection ratio is <1% over 4.5V to 5.5V.
Additionally, one sensor terminal is grounded, often beneficial for noise rejection.
C1’s continuous clocking, while maintaining O1’s isolated
DC power supply, generates periodic cessations in the
frequency coded output. These interruptions can be used
as markers to control operation of monitoring logic.
Output frequency vs thermistor characteristics are included in Figure 9.
This is basically a hetrodyne circuit. Two oscillators, one
variable, one fixed, are mixed, producing sum and difference frequencies. The variable oscillator is controlled by
the capacitive humidity sensor. The demodulated difference frequency is the output.6 The hetrodyne frequency
subtraction approach permits a sensed 0% RH to give a
0Hz output, even though sensor capacitance is not zero at
RH = 0%.
A = 5V/DIV
B = 10V/DIV
C = 10V/DIV
D = 1V/DIV
HORIZ = 10µs/DIV
AN93 F10
C1, the sensor controlled variable oscillator, runs between
the indicated output frequencies for the RH sensor excursion noted. The RH sensor is AC coupled, in accordance
with its manufacturer’s data sheet.7 Reference oscillator
O1 is tuned to C1’s nominal 25% RH dictated frequency.
Figure 10. Isolated Thermistor Digitizer’s Waveforms Include
C1’s Output (Trace A), Q2’s Collector Drive to T1 (Trace B),
Data Output (Trace C) and C1’s Negative Input (Trace D).
C1’s Negative Input (Trace D) Is Lightly Modulated by Q3,
Synchronizing Transformer Power Drive to Data Output
17.8k*
5V
1µF
–
2pF
SENSOR
680Ω
0.1µF
C1
LT1394
10k
FILTER-DEMODULATOR
+
49.9k**
5V
*
**
SENSOR
0% RH
100% RH
= 0.1% METAL FILM RESISTOR
= 1% METAL FILM RESISTOR
= PANAMETRICS MC-2
= 196.7pF
= 227.8pF
0.31pF/RH
= 1N4148
O1
= 1N5712
100% RH = 174.4kHz
NOMINAL
49.9k**
25% RH = 193.5kHz
NOMINAL
SENSOR
OSCILLATOR
49.9k**
25% RH TRIM
5V
2k
100k
Q1
2N2369
48.7k*
OUT
5V
1k
REFERENCE
OSCILLATOR
RSET
LTC1799
5V
0.02µF
100pF
0.1µF
51pF
10M
+
51k
–
Q
OUTPUT
SCALING
74C90 ÷ 5
MIXER
199.7kHz 0.1µF
10k
NOMINAL
Q
C2
LT1671
74C90 ÷ 5
OUT
0% TO 100% RH =
0Hz TO 1kHz
AN93 F11
Figure 11. Hetrodyne Based Humidity Transducer Digitizer Has Grounded Sensor, 2% Accuracy. Capacitively Sensed
Hygrometer Beats Humidity Dependent Oscillator (C1) Against Stable Oscillator O1. Difference Frequency Is Demodulated
by Q1, Converted to Pulse Form at C2. Counters Scale Output for 0kHz to 1kHz = 0% to 100% Relative Humidity
Note 6: Hetrodyne techniques, usually associated with communications circuitry, have previously been applied to instrumentation. This
circuit’s operation was adapted from approaches described in
References 2, 3 and 4.
Note 7: DC coupling introduces destructive electromigration effects.
See Reference 6.
AN93-5
Application Note 93
The two oscillators are mixed at Q1’s base (Figure 12,
Trace A). Q1 amplifies the mixed frequency components,
although collector filtering attenuates the sum frequency.
The RH determined difference frequency, appearing as a
sine wave at Q1’s collector (Trace B), remains. This
waveform is filtered and AC coupled to zero crossing
detector C2. AC hysteretic feedback at C2’s input (Trace C)
produces clean C2 output (Trace D). Counter based scaling at C2’s output combines with slight sensor padding
(note 2pF value across the sensor) to provide numeric
output frequency correspondence to RH. Calibration involves simulating the RH sensor’s 25% value and trimming O1 for a 250Hz output. The simulated value may be
built up from known discrete capacitors or simply dialed
out on a precision variable air capacitor (General Radio
1422D).
A = 2V/DIV
B = 2V/DIV
C = 1V/DIV
D = 5V/DIV
HORIZ = 50µs/DIV
Figure 12. Sensor and Stable Oscillators are Mixed at Q1’s Base
(Trace A); Difference Frequency Appears at Q1’s Collector
(Trace␣ B). Filtering and AC Hysteresis at C2’s + Input (Trace C)
Produce Clean Response at C2’s Output (Trace D)
Relative Humidity Sensor Digitizer-Charge Pump
Based
Figure 13 also digitizes the capacitive humidity sensor’s
output but has better specifications than the previous
circuit. Circuit accuracy is 0.3%, plus the selected sensor
grade’s tolerance. Temperature coefficient is about
300ppm/°C and power supply rejection ratio is 0.25% for
5V ±0.5V. Compromises include a floating sensor and
somewhat more complex circuitry.
When evaluating circuit operation, it is useful to consider
that C1’s frequency changes inversely with sensor capacitance; its period is linear vs sensor capacitance. This
would normally corrupt the desired linear output relationship between frequency and RH. Practically, because the
sensor’s excursion range is small compared to its 0% RH
value, the error is similarly small. This term almost entirely
accounts for the circuit’s stated 2% accuracy.
80.6k*
OUT
13.3k*
7
4V
S
4V
12
CHARGE
PUMP
1µF
L
A1
LTC1050
+
100k
RH = 25% TRIM
(204.5pF)
LT1634
4V
RESET
COMPARATOR
–
562k*
TO ALL 4V
POINTS
Q1
VN2222L
5V
14
5V
200Ω
4V
1000pF †
1µF
13
D
4
SENSOR
1k*
CLOCK
10k
BAT85
5V
8
11
0.1µF
10k
17
5V
O1
VIN
GND
4V
16
RSET
LTC1799
1N4148
125kHz
LTC1043
AN93 F12
INTEGRATOR
300k*
4V
C1
+V
LT1671
–
4V
1N5712
110k*
* = 1% METAL FILM RESISTOR
†
= WIMA, TYPE MKP-2
SENSOR = PANAMETRICS MC-2
0% RH = 196.7pF
100% RH = 227.8pF
0.31pF/RH
5V
+
300pF
Q
Q
OUT
0% TO 100% RH =
0Hz TO 1kHz
20k
RH = 100% TRIM
(227.8pF)
AN93 F13
Figure 13. Hygrometer Digitizer Has 0.3% Accuracy, Although Sensor Must Float Off-Ground. Humidity
Sensor Determines Charge Delivered to A1 Integrator During Each Charge Pump Cycle. Resultant A1 Output
Ramp Is Reset by Level Triggered C1 via Q1. Output Frequency, Taken at C1, Varies with Humidity
AN93-6
Application Note 93
O1 (Trace A, Figure 14) clocks an LTC1043 switch array
based charge pump. This configuration alternately connects the AC coupled RH sensor to a 4V reference derived
potential and then discharges it into A1’s summing point.
A1, an integrator, responds with a ramping output, Trace␣ B
of Figure 14. When A1’s output exceeds C1’s negative
input voltage, C1’s Q output (Trace C, Figure 14) goes
high, triggering Q1 and resetting the ramp. AC feedback to
C1’s negative input (Trace D) ensures long enough Q1 ontime for complete ramp resent. This action’s repetition
rate depends on RH sensor value. The A1-C1 loop is
synchronized to the charge pump’s clocking by O1’s
output path to C1’s latch input. In theory, if the charge
pump, offset term (25% trim current) and ramp amplitude
are tied to the same potential, this circuit does not require
a voltage reference. In practice, the sensor’s extremely
small capacitance shifts magnify the effect of charge
pump errors vs supply, necessitating powering the
LTC1043 from the 4V reference. Once this is done, the
mentioned points are tied to the 4V reference. Note that the
5V powered O1’s output must be level shifted to drive the
LTC1043.
A trimmed DC offset current (100k potentiometer) into
A1’s summing junction compensates the RH sensor’s
offset term (e.g., 0% RH ≠ 0pF). Output frequency is
scaled by the 20kΩ trim at C1 so 0% to 100% RH = 0Hz
to 1kHz. Trimming involves substituting capacitance for
the sensor’s known 100% and 25% values and trimming
the appropriate adjustments. The adjustments are somewhat interactive, necessitating repetition until convergence occurs. A precision variable capacitor (General
A = 5V/DIV
B = 1V/DIV
C = 5V/DIV
D = 5V/DIV
HORIZ = 100µs/DIV
AN93 F14
Figure 14. LTC1799 Clock (Trace A) Drives Humidity
Sensor Based Charge Pump, Producing A1 Output
Ramp (Trace B). C1 Q Output, Trace C, Biases Q1,
Resetting Ramp. AC Feedback at C1 (Trace D) Permits
Complete Ramp Reset, Sets Output Pulse Width
Radio type 1422D) is invaluable in this regard, although
acceptable results are possible with built-up calibrated
discrete capacitors.
Relative Humidity Sensor Digitizer-Time Domain
Bridge Based
Figure 15, also a relative humidity (RH) digitizer, features
1% accuracy, PSRR of 1% over 4.5V to 5.5V, temperature
coefficient of 350ppm/°C and a ground referred sensor.
Additionally, the circuit’s trim scheme accommodates
wide tolerance grade RH sensors. The circuit is basically
a time domain bridge; it subtracts time intervals representing sensor and sensor offset values to determine
sensor value extrapolated to RH = 0%. This measurement
is digitized and scaled so zero to 100 counts equals 0% to
100% RH at the output.
O1’s nominal 12.77MHz output, conditioned by a counter
chain and an inverter configured gate, presents a 12.4kHz,
2.5µs pulse (Trace A, Figure 16) to Q1A and Q1B. The
transistors’ collectors fall (Trace B = Q1A collector – Trace
C = Q1B collector) to zero volts. When the base drive
ceases, both collectors ramp towards 5V. Trace B’s ramp
slope varies with the RH sensor’s capacitance; Trace C’s
ramp slope represents the sensor’s offset value (0% RH ≠
0pF). C1 and C2 switch when their associated ramp inputs
cross the comparators’ common DC input potential. The
comparator outputs (Trace D = C2, Trace E = C1) define a
“both high” time region proportional to the ramp slopes’
difference and, hence, an offset corrected version of
sensor value. This time interval is gated with O1’s output,
providing Trace F’s data output.
Circuit operation is fairly straightforward, although some
details bear mention. Q1, a dual transistor, promotes
cancellation of the individual transistors’ VCE vs temperature terms, minimizing their error contribution. The unit
specified, a 2-die type, minimizes crosstalk; monolithic
types should not be substituted. Similarly, a dual comparator should not be substituted for the single types
specified for C1 and C2. Also, the comparators operate at
high source impedance relative to their input characteristics but symmetry provides adequate error cancellation.
Finally, the 5.6k resistor combines with the output gates’
input capacitance, forming a ≈20ns lag. This delay prevents false output data transients when the ramps are
resetting.
AN93-7
Application Note 93
Trimming procedure is similar to the previous RH circuit.
It involves substituting capacitance for the sensor’s known
100% and 25% values and trimming the indicated adjustments. The adjustments are somewhat interactive, neces12.4kHz
5V
2.5µs
1/2 74HC10
12
10
9
13
5V
÷ 1024
14
74HC592
12.77MHz
NOMINAL
74HC4040
B C D E F G H
11
÷ 32
15
1k
RH = 100% TRIM
IN
CLOCK
5V
10k*
5V
1.4M*
SENSOR RAMP
OUT
+
C1
LT1671
0.1µF
Q1A
sitating repetition until convergence occurs. A precision
variable capacitor (General Radio type 1422D) is invaluable for this work, although acceptable results are possible with calibrated discrete capacitor assemblies.
Q
RSET
LTC1799
PULSE SUBTRACTION
AND GATING
O1
DIV
–
SENSOR
5.6k
1/2 74HC10
3.24M*
825k*
10k
TIME
DOMAIN
BRIDGE
5V
DATA OUT
0% TO 100% RH =
0 TO 100 COUNTS
BRIDGE
COMPARATORS
0.1µF
5V
100k
25% TRIM
10k
1.8M*
OFFSET RAMP
Q1B
150pF †
+
C2
LT1671
–
Q
* = 1% METAL FILM RESISTOR
†
= WIMA FKP-02
Q1A, Q1B = CENTRAL SEMICONDUCTOR CMXT3904
(DO NOT SUBSTITUTE MONOLITHIC DUAL TYPE)
SENSOR = PANAMETRICS MC-2
0% RH = 196.7pF
100% RH = 227.8pF
AN93 F15
0.31pF/RH
Figure 15. Humidity Transducer Digitizer Has Grounded Sensor, 1% Accuracy; Trim Scheme Allows Low
Tolerance Sensors. Clocked Q1A-Q1B Configurations Produce Ramp Outputs. Q1A Ramp Slope Varies with
Humidity Sensor Value, Q1B Ramp Represents Sensor’s Offset (0% RH ≠ 0pF). C1, C2 Digitize Ramp
Times. Gate Extracts Time Difference, Presents 0 to 100 Counts Out for 0% to 100% Relative Humidity
A = 10V/DIV
B = 2V/DIV
C = 2V/DIV
D = 10V/DIV
E = 10V/DIV
F = 10V/DIV
HORIZ = 10µs/DIV
AN93 F16
Figure 16. Humidity Sensor Time Domain Bridge Waveforms. Gate
(Figure 15, Upper Left) Clocks (Trace A) Q1A and Q1B. Sensor and Offset
Ramps Are Traces B and C. C1 and C2 Outputs are Traces D and E. Gate
Extracts C1-C2 Time Difference, Presents Trace F’s Digitized Output
AN93-8
Application Note 93
40nV Noise, 0.05µV/°C Drift, Chopped Bipolar
Amplifier
O1’s 37kHz output is divided down to form a 2-phase
925Hz square wave clock. This frequency, harmonically
unrelated to 60Hz, provides excellent immunity to harmonic beating or mixing effects which could cause instabilities. S1 and S2 receive complementary drive, causing
A1 to see a chopped version of the input voltage. A1
amplifies this AC signal. A1’s square wave output is
synchronously demodulated by S3 and S4. Because these
switches are synchronously driven with the input chopper, proper amplitude and polarity information is presented to A2, the DC output amplifier. This stage integrates
the square wave into a DC voltage, providing the output.
The output is divided down (R2 and R1) and fed back to the
input chopper where it serves as a zero signal reference.
Gain, in this case 1000, is set by the R1-R2 ratio. Because
A1 is AC coupled, its DC offset and drift do not affect
overall circuit offset, resulting in the extremely low offset
and drift noted. A1’s input damper minimizes offset voltage contribution due to nonideal switch behavior.
Figure 17’s circuit, adapted from Reference 7, combines
the low noise of an LT1028 with a chopper based carrier
modulation scheme to achieve an extraordinarily low
noise, low drift DC amplifier. DC drift and noise performance exceed any currently available monolithic amplifier. Offset is inside 1µV, with drift less than 0.05µV/°C.
Noise in a 10Hz bandwidth is less than 40nV, far below
monolithic chopper stabilized amplifiers. Bias current, set
by the bipolar LT1028 input, is about 25nA. The circuit is
powered by a single 5V supply, although its output will
swing ±2.5V. Additionally, a carefully selected chopping
frequency prevents deleterious interaction with 60Hz related components at the amplifier’s input. These specifications suit demanding transducer signal conditioning
situations such as high resolution scales and magnetic
search coils.
74C90 ÷ 10
74C74 ÷ 4
Q
Q
925Hz
5k
925Hz
TRIM
37kHz
TO
Ø1
POINTS
10Ω
TO
Ø2
POINTS
22µF
+
24.9k**
5V
5V
22µF
+
LTC1799 OUT
DIV RSET
47µF*
10Ω
8
INPUT
74HC04-X6
+
–
9
3
2
470pF
Ø2
10k
100k
S3
S4
16
V–
0.01µF
** = 1% METAL FILM RESISTOR
* = SANYO 6TPB47M-6.3V
10Ω
240k
–
A2
LT1097
14
15
100Ω
= BAT85
47µF*
≈ –3.5V TO
“V – ” POINTS
1µF
A1
LT1028
11
10
1µF
1
S1
S2
100µF
Ø2
6
7
+
10Ω
Ø1
+
5V
+
O1
4
+
V–
V–
Ø1
NOISE = 40nVP-P 0.1Hz TO 10Hz
OFFSET = 0.5µV
DRIFT = 0.05µV/°C
R2 +1
GAIN =
R1
OPEN-LOOP GAIN = >10 8
I BIAS = 25nA
SWITCHES = LTC201A QUAD
0.47µF
OUTPUT
R2
10k
R1
10Ω
AN93 F17
Figure 17. 5V Powered, Chopped Bipolar Amplifier. Noise Is ≈40nV with 0.05µV/°C Drift. DC Input Is Carrier
Modulated, Amplified by A1, Demodulated to DC and Fed Back from A2. 925Hz Carrier Clock Prevents Interaction
with 60Hz Line Originated Components. Negative Supply, Derived via Charge Pump, Allows Zero Volt Output Swing
AN93-9
Application Note 93
Normally, this single supply amplifier’s output would be
unable to swing to ground. This restriction is eliminated by
powering the circuit’s negative rail from a charge pump.
O1’s 37kHz output excites the charge pump, comprised of
paralleled logic inverters and discrete components. Deliberate 10Ω loss terms combine with the specified 47µF
capacitors to form a very low noise power source. These
precautions eliminate charge pump noise which might
otherwise degrade amplifier noise performance.
Figure 18, a noise plot of the amplifier in a 0.1Hz to 10Hz
bandwidth, shows about 40nV of peak-to-peak noise. A1
and the 60Ω resistance of S1-S2 contribute about equally
to form this noise. When using this amplifier, it is important to realize that A1’s bias current flowing through the
input source impedance causes additional noise. In general, to maintain low noise performance, source resistance should be kept below 500Ω. Fortunately, transducers
such as strain gauge bridges, RTDs and magnetic detectors are well below this figure.
45nV Noise, 0.05µV/°C Drift, Chopped FET Amplifier
Figure 19 replaces the previous circuit’s input stage with
a pair of extremely low noise J-FETs. In most other
respects, circuit operation is similar. Noise increases very
slightly, to ≈45nV, but bias current decreases to only
500pA—50 times lower than the previous circuit. The
noise performance is especially noteworthy—it is almost
17 times better than currently available monolithic chopper stabilized amplifiers and nearly equals the best bipolar
designs. Other performance specifications, appearing in
the figure, are similar to Figure 17.
The 925Hz clock is retained, although this ±15V powered
design uses zeners to derive internal ±5V points. The clock
and logic run from 5V and the LTC201 switches use ±5V.
The switches low voltage rails reduce charge injection,
minimizing its effect on offset voltage. RC damper networks further attenuate parasitic switch behavior effects,
resulting in the 1µV offset specification.
Noise measured over Figure 20’s 50 second interval is about
45nV in a 0.1Hz to 10Hz bandwidth. This is spectacularly
low noise for a J-FET based design and is directly attributable to the input pairs’ die size and current density.8
Clock Tunable, Filter Based Sine Wave Generator
A feedback loop enclosed resonator can be made to
oscillate. Figure 21’s sine wave generator takes advantage
of this and eliminates the need for an amplitude control
loop. This circuit, a mildly modified form of the Regan
resonant bandpass loop, is clock tunable and produces
sine and cosine outputs.9
The LTC1060 switched capacitor filter is set up as a clock
tunable bandpass filter with a Q of 10. O1 clocks the filter
at 100kHz, resulting in a 1kHz bandpass. C1, switched by
the sine output, supplies square wave drive to the filter
input in regenerative fashion. The loop is self-sustaining,
resulting in continuous sine wave outputs at the indicated
points. Zener bridge clamping of C1’s output stabilizes
square wave amplitude applied to the filter and, hence, the
sine wave outputs. This form of amplitude control eliminates AGC loop settling times and potential instabilities.
Changes in O1’s clock frequency permit bandpass tuning,
with no amplitude shifts during or after tuning.
40nV
AN93 F18
10 SECONDS
Figure 18. Noise in a 0.1Hz-10Hz Bandwidth Is about 40nV with 0.05µV/°C Drift
Note 8: See References 8 and 9.
AN93-10
Note 9: This circuit draws heavily on a scheme originated by Tim
Regan. See Reference 10.
Application Note 93
15V
–15V
5V
2k
–5V
TO LTC201 V + PIN
2k
TO LTC201 V – PIN
1µF
+
+
5V
5V
18.5kHz
V+
74C90 ÷ 10
DIV LTC1799 OUT
RSET
5V
1N751
1µF
1N751
74C74 ÷ 2
Q
Q
925Hz
54.2k*
TO
Ø1
POINTS
15V
Ø1
TO
Ø2
POINTS
8
900Ω**
3k
INPUT
900Ω**
6
7
0.002µF
Ø2
S1
S2
11
10
9
Ø2
1µF
1
–
2.2k
900Ω**
0.01µF
1µF
A1
LT1056
+
10M
–15V
3k
3
2
S3
S4
100k
240k
14
15
16
10k
= TOSHIBA 2SK147
A2
LT1097
OUTPUT
+
Ø1
0.47µF
1µF
* = 0.1% METAL FILM RESISTOR
** = 1% METAL FILM RESISTOR
= LTC201 QUAD
–
NOISE = 45nVP-P 0.1Hz TO 10Hz
OFFSET = 1µV
DRIFT = 0.05µV/°C
R2 +1
GAIN =
R1
OPEN-LOOP GAIN = 10 9
I BIAS = 500pA
R2
10k
R1
10Ω
AN93 F19
Figure 19. FET Input Version of Figure 17 Has 500pA Bias Current. 925Hz Clock Is Retained, Noise Increases Slightly to ≈45nV
A = 20nV/DIV
HORIZ = 5s/DIV
AN93 F20
Figure 20. Chopped FET Input Amplifier Noise Is ≈45nV in 0.1Hz to 10Hz Bandwidth
AN93-11
Application Note 93
Figure 22 shows operating waveforms. The bandpass
filter, responding to C1’s clamped output (Trace A), produces sine (Trace C) and cosine (Trace B) outputs. Distortion, Trace D, dominated by filter clock residue, is 2%.
Clock Tunable, Memory Based Sine Wave Generator
This circuit generates a variable frequency sine wave by
continuously clocking a sine coded lookup table memory.
The memory’s state is converted to an analog output by a
51k
110k
DAC. A strength of this technique is its rapid, high fidelity
response to frequency and amplitude change commands.
O1, set to one of three output frequencies dictated by its
digital control inputs, clocks the 74HC191 counters. These
counters parallel load a 2716 EPROM programmed to
produce an 8-bit (256 states) digitally coded sine wave.
The program, developed by Sean Gold and Guy M. Hoover,
appears in Figure 24.10 The 2716’s parallel output is fed to
a DAC, producing the analog output.
SINE
OUT
5.1k
5V
BANDPASS
FILTER
100k
100kHz
HPA
INVA
S1A
BPA
OUT
CLKA
O1
LTC1060
–
+
LPA
+
COS
OUT
RSET
LTC1799 DIV
NC
CLOCK
+
AGND
+V
–V
5V
–5V
12
COMPARATOR
47µF
AMPLITUDE
CLAMP
100k
510Ω
–
C1
L LT1713
LT1004 100k
2.5V
4.7k
3.9k
3.9k
5V
–5V
9
+
1k
1k
–V
–V
= 1N4148
AN93 F21
Figure 21. The Regan Resonant Bandpass Loop. A Bandpass Filter, Driven by C1’s Oscillation Loop, Continuously
Rings at Resonance. Clock Controls Output Frequency. Zener Bridge Clamp Sets Sine and Cosine Output Amplitude
A = 10V/DIV
B = 2V/DIV
C = 2V/DIV
D = 2%
DISTORTION
HORIZ = 200µs/DIV
AN93 F22
Figure 22. Bandpass Filter, Responding to C1’s Loop Enforced Excitation (Zener Clamp Output, Trace A), Produces Sine
(Trace␣ C) and Cosine (Trace B) Outputs. Distortion (Trace␣ D), Dominated by Switched Capacitor Filter Clock Residue, Is 2%
Note 10: See Reference 11.
AN93-12
Application Note 93
LOGIC INPUTS
LO = ENABLED
HI = OFF
50Hz
60Hz
5V
400Hz
100k
* = 0.1% METAL FILM RESISTOR
5V
= 1N4148
SWITCHES = LTC201
(
)
77.7k*
64.9k*
PROGRAMMABLE
CLOCK
97.6k*
RSET
COUNTERS/SINE MEMORY
DIV
LTC1799
O1
OUT
5V
VCC
CLK
QA
A0
QB
A1
ENABLE
QC
A2
CLR
QD
A3
P/U
74HC191
VCC
Q0
Q1
Q2
Q3
Q4
Q5
Q6
Q7
OUTPUT DAC
5V
RIPPLE CLOCK
2716
MSB
VCC
P/U
QA
A4
ENABLE
QB
A5
CLR
QC
A6
QD
A7
74HC191
LSB
WR
5V
CLK
5V
CLR
REFOUT
SINE
OUTPUT
4VP–P
LTC1450
REFIN
X1 X2 REFLO
CSMSB
CSLSB
CE
A8 A9 A10
OE
AN93 F23
Figure 23. Counter Driven, Sine Encoded Memory Produces 0.75% Distortion Sinewave via D/A
Converter. LTC1799 Oscillator Frequency, Controlled by Digital Inputs, Sets Output Frequency
Trace A in Figure 25 is the sine wave output, in this case
tuned to 60Hz. Distortion, appearing as Trace B, is mostly
composed of clock residue and measures about 0.75%. In
Figure 26, the digital inputs abruptly change output frequency to 400Hz and then promptly return it to 60Hz.
These frequency shifts occur crisply, with no alien components or untoward behavior. Amplitude shifts, accomplished by driving the DAC’s reference input (see LTC1450
data sheet), are similarly well behaved. Figure 27 shows
Trace B’s amplitude faithfully responding to Trace A’s DAC
reference input step. As before, the lack of control loop
time constants promotes uncorrupted response.
Figure 24. Sinewave Generation Code for the Memory
AN93-13
Application Note 93
CLOCK
NOTCH CENTER
R FREQUENCY FREQUENCY
210k
4.75kHz
60Hz
249k
3.96kHz
50Hz
31.5k 31.72kHz
400Hz
A = 2V/DIV
R1
R2
B = 0.75%
DISTORTION
79.3
f
= 1.234, CLOCK =
f NOTCH
1
10k*
10k*
R1
1.24k*
HORIZ = 2ms/DIV
AN93 F25
VIN
Figure 25. Sinewave Output (Trace A) and Its Distortion
(Trace␣ B). Clock Related Products Are Evident in Distortion
Presentation
R2
1k* 1
8
2
7
–5V
3
LTC1062
4
6
5
5V
10k*
5V
fCLK
–
A1
LT1006
VOUT
+
R
5V
A = 1V/DIV
(UNCAL)
DIV LTC1799 OUT
Q1
TP0610L
3.3k
–5V
O1
* = 1% METAL FILM RESISTOR
HORIZ = 2.5ms/DIV
AN93 F26
Figure 26. Fast Oscillator Frequency Shifting Permits Crisp
Sinewave Output Frequency Change
A = 2V/DIV
AN93 F28
Figure 28. A Clock Tuned, Highly Selective Notch Filter.
LTC1799 Oscillator Sets Notch Center Frequency According
to Table. R Value Could be Switched Under Digital Control
Figure 29 shows notch performance at a 60Hz center
frequency. Response is down over 45dB at 60Hz, with
steep slopes on either side of the notch. This characteristic
is maintained as center frequency is clock-tuned over
broad ranges.
B = 2V/DIV
HORIZ = 5ms/DIV
AN93 F27
Figure 27. Trace B’s Sinewave Amplitude Instantaneously and
Faithfully Responds to DAC Reference Input Step, Trace A
NOTCH DEPTH (dB)
0
–10
–20
–30
–40
–50
–60
Clock Tunable Notch Filter
Figure 28 shows a quick, clean way to tune a notch filter’s
center frequency by varying a single resistor, which could
be switched. The LTC1062 switched capacitor filter and A1
form a clock tunable notch (see LTC1062 data sheet). O1,
running from the 5V supply, furnishes the clock, which is
level shifted by Q1 to drive the ±5V powered LTC1062. In
this case, three common notch frequencies are listed;
others are selectable by tuning O1 in accordance with the
equivalency listed.
AN93-14
–70
30
60
90
120
FREQUENCY (Hz)
150
AN93 F29
Figure 29. Notch Characteristic at 60Hz Center
Frequency. Response Is Essentially Identical as
Center Frequency Is Tuned over Broad Range
Application Note 93
Clock Tunable Interval Generator with 20 × 106:1
Dynamic Range
An accurate interval generator with large dynamic range
appears in Figure 30.11 The circuit is made up of a clock,
a counter and a dual flip-flop. Clock frequency and counter
modulo are programmable. A trigger input is passed to
flip-flop 1’s Q output (Trace A, Figure 31) synchronously
with O1’s clock (Trace C). This output going low sets flipflop 2’s Q output, a circuit output, high (Trace B). Simultaneously, flip-flop 2’s Q output resets the 4060 counter,
allowing it to accumulate clock pulses (again, Trace C).
When enough clock pulses occur to set the selected 4060
output high, flip-flop 2’s clear input (Trace D) is pulled low,
ending the circuit’s output width. The output width is
settable by O1’s frequency and the counter’s modulo, both
variable over many decades. As shown, the interval is
÷100
5V
÷10
programmable over 800 nanoseconds to 16 seconds,
although other counters can extend this range. Interval
accuracy and stability is almost entirely dependent on O1’s
programming resistor.
8-Bit, 80µs, Passive Input, A/D Converter
In general, monolithic A/D converters have replaced discrete types. Occasionally, specific desirable circuit characteristics dictate a discrete design. Examples of such
special cases include the need for a passive analog input,
output data format, control protocol or economic constraints. Figure 32’s 8-bit design has 90ppm/°C drift
(<1LSB 0°C to 70°C) and converts in 80µs. The circuit
consists of a current source, an integrating capacitor, a
comparator, logic and a clock.12
14-BIT COUNTER
÷1
CLOCK
1 2 3 4 5 6 7
13 14 15
MSB
LSB
74HC4060
IN
RESET
DIV
LTC1799 OUT
RSET
O1
5V
3.3k
20k
INTERVAL
WIDTH
TUNE
5V
TRIGGER
INPUT
CLR1 PRE1 CLK1
1D
74HC74
CLR2 2D CLK2 PRE2 2Q
1Q
2Q
Q OUTPUT
Q OUTPUT
AN93 F30
= FAIRCHILD NC7504
TRIGGER INPUT SYNCHRONIZATION
AND SET-RESET LOGIC
Figure 30. 1% Accurate Interval Generator with 20 × 106:1 Dynamic Range. Flip-Flop Output, Set
by Trigger Input, Resets when Counter Times Out. LTC1799 Oscillator Controls Timing Sequence
A = 5V/DIV
B = 5V/DIV
C = 5V/DIV
D = 5V/DIV
HORIZ = 2µs/DIV
AN93 F31
Figure 31. Trace A’s Trigger Pulse Sets Circuit Output (Trace B) High. LTC1799 Oscillator (Trace C) Clocks Counter until Selected
Counter Output Biases Inverter Low (Trace D), Resetting Circuit Output (Trace B). Reset Sequence Intensified for Photographic Clarity
Note 11: Pedestrian laboratory argot for interval generator is “one
shot.”
Note 12: This circuit is a modern incarnation of the earliest electronic
A/D known to the author. See Reference 13.
AN93-15
Application Note 93
5V
5V
LT1389
2.5V
10k*
COMPARATOR
EIN
0V TO 2V
–
1k
Q1
+
10k
Q
C1
LT1671
Q2
5k
CALIBRATE
MINIMIZE TRACE
CAPACITANCE
STATUS
1 = CONVERSION
COMPLETE
0 = BUSY
29.4k*
DIV
OUT
2k
Q
O1
LTC1799
0.01µF †
CLOCK
RAMP RESET
CURRENT
SOURCE
RSET
≈3.2MHz
Q3
VN2222L
OUTPUT
GATING
CLK SYNCH
CONV COMMAND
74HC74
DUAL
FLIP-FLOP
PRE1
CLR1
5V
D1
* = 1% METAL FILM RESISTOR, 10ppm/°C
†
= WIMA MKM-2
PNP = THERMALLY MATE 2N5087
CLOCK
SYNCHRONIZER
= 1N4148
MINIMIZE TRACE
CAPACITANCE
Q1 Q1 CLK1 PRE2 Q2
CLK2
CLR2
D2
Q2
DATA OUT
5V
CONVERT COMMAND
20µs MINIMUM
AN93 F32
Figure 32. Simple 8-Bit A/D Converter Has Passive, High Impedance Input.
Additional Features Include 80µs Conversion Time, Accurate 0°C to 70°C Operation
Applying a pulse to the convert command input causes
flip-flop Q1 output to go high (Trace A, Figure 33) when the
CLK1 input is clocked by O1. This turns on Q3, resetting
the 0.01µF capacitor (Trace B). Simultaneously, Q1 goes
low, pulling the CLK2-CLR2 input down. C1’s Q output, the
circuit’s status output (Trace C), also goes low and C1’s Q
output rises high. This logic state prevents any of O1’s
clock pulses from being transmitted to the circuit’s data
output (Trace D). When the convert command falls, Q1
goes low, Q3 turns off and the 0.01µF capacitor begins to
ramp. Concurrently, Q1 goes high, allowing clock pulses
to appear at the data output. When the ramp crosses EIN’s
voltage, C1’s outputs exchange state, pulling the CLK2CLR2 line low and data output pulses cease. Thus, the O1
originated clock burst appearing at the data output is
directly and solely proportional to EIN. For the arrangement shown, 256 pulses appear for a 2V full-scale input.
Conversion time decreases with the time required for the
ramp to cross EIN. A full-scale conversion requires 80µs,
linearly descending to 8µs at 0.1 scale.
A = 5V/DIV
B = 2V/DIV
C = 5V/DIV
D = 5V/DIV
HORIZ = 20µs/DIV
AN93 F33
Figure 33. 8-Bit A/D Converter Waveforms (EIN = 1V) Include
Synchronized Convert Command (Trace A), Reference Ramp
(Trace B), Status Output (Trace C) and Data Output (Trace␣ D).
Conversion, Initiable when Status Output Is High, Begins when
Command Line Goes Low
Flip-flop 2, connected as a logic buffer, duplicates the high
impedance diode-2kΩ node’s logic state. As such, this
node’s trace capacitance should be minimized. This is
facilitated by locating the diodes and 2k resistor adjacent
to the CLK2-CLR2 inputs. Circuit trimming is accomplished by applying a 2V input and adjusting O1’s frequency output (“calibrate”) for 256 data output pulses per
conversion.
Note: This Application Note was derived from a manuscript originally
prepared for publication in EDN magazine.
AN93-16
Application Note 93
REFERENCES
1. Minco Products, Inc., Bulletin TS-102(N), Minco Products, Inc., 2002.
8. Toshiba, “2SK147 Datasheet,” Toshiba Corporation,
Tokyo, Japan.
2. Benjaminson, Albert, “The Linear Quartz Thermometer—A New Tool for Measuring Absolute and Differential Temperatures,” Hewlett-Packard Journal, March
1965.
9. Williams, J., “Practical Circuitry for Measurement and
Control Problems,” Linear Technology Corporation,
Application Note 61, August 1994.
3. Hewlett-Packard Company, “Model 2801A Quartz Thermometer Operating and Service Manual,” HewlettPackard Company, 1969.
4. Tektronix, Inc., “Type 130 LC Meter Operating and
Service Manual,” Tektronix, Inc., 1959
5. Yellow Springs Instrument Company, “YSI Precision
Thermistors,” Yellow Springs Instrument Company.
6. Panametrics, Inc., “MiniCap 2—Relative Humidity
Sensor,” Panametrics, Inc., 2000.
7. Williams, J., “Measurement and Control Circuit Collection,” Linear Technology Corporation, Application
Note 45, June 1991.
10. Regan, Tim, “Introducing the MF-10: A Versatile Monolithic Active Filter Building Block,” National Semiconductor Corporation, Application Note 307, August
1982.
11. Williams, J., “Step-Down Switching Regulators,” Linear Technology Corporation, Application Note 35,
August 1989.
12. Linear Technology Corporation, LTC1799 Data Sheet,
Linear Technology Corporation, August 2001.
13. Wilkinson, D.H., “A Stable Ninety-Nine Channel Pulse
Amplitude Analyzer for Slow Counting,” Proceedings
of the Cambridge Philosophical Society, Cambridge,
England, 46,508, 1950.
APPENDIX A
LTC1799 INTERNAL OPERATION
As shown in Figure A1, the LTC1799’s master oscillator is
controlled by the ratio of the voltage between the V+ and
SET pins and the current entering the SET pin (IRES). The
voltage on the SET pin is forced to approximately 1.13V
below V+ by the PMOS transistor and its gate bias voltage.
This voltage is accurate to ±7% at a particular input
current and supply voltage (see Figure A2). The effective
input resistance is approximately 2k.
A resistor RSET, connected between the V+ and SET pins,
“locks together” the voltage (V + – VSET) and current, IRES,
variation. This provides the LTC1799’s high precision. The
master oscillation frequency reduces to:
 10kΩ 
ƒMO = 10MHz • 

 RSET 
The LTC1799 is optimized for use with resistors between
10k and 200k, corresponding to master oscillator frequencies between 0.5MHz and 10MHz. Accurate
frequencies up to 20MHz (RSET = 5k) are attainable if the
supply voltage is greater than 4V.
To extend the output frequency range, the master oscillator signal may be divided by 1, 10 or 100 before driving
OUT (Pin 5). The divide-by value is determined by the state
of the DIV input (Pin 4). Tie DIV to GND or drive it below
0.5V to select ÷1. This is the highest frequency range, with
the master output frequency passed directly to OUT. The
DIV pin may be floated or driven to midsupply to select
÷10, the intermediate frequency range. The lowest frequency range, ÷100, is selected by tying DIV to V+ or
driving it to within 0.4V of V+. Figure A3 shows the
AN93-17
Application Note 93
relationship between RSET, divider setting and output
frequency, including the overlapping frequency ranges
near 100kHz and 1MHz.
The CMOS output driver has an on resistance that is
typically less than 100Ω. In the ÷1 (high frequency) mode,
the rise and fall times are typically 7ns with a 5V supply and
11ns with a 3V supply. These times maintain a clean
1
square wave at 10MHz (20MHz at 5V supply). In the ÷10
and ÷100 modes, where the output frequency is much
lower, slew rate control circuitry in the output driver
increases the rise/fall times to typically 14ns for a 5V
supply and 19ns for a 3V supply. The reduced slew rate
lowers EMI (electromagnetic interference) and supply
bounce.
VRES = 1.13V ±25%
(V+ – VSET)
V+
PROGRAMMABLE
DIVIDER
(÷1, 10 OR 100)
+
RSET
3
SET
–
+
–
MASTER OSCILLATOR
ƒMO = 100MHz • kΩ •
VBIAS
2 GND
5
V+
GAIN = 1
IRES
OUT
DIVIDER
SELECT
IRES
(V + – VSET)
5µA
DIV
THREE-STATE
INPUT DETECT
IRES
4
5µA
GND
AN93 FA01
Figure A1. LTC1799 Master Oscillator Frequency Is Controlled by Ratio of Voltage Between V + and SET
and Current Entering SET. Pin-Programmable Frequency Divider Permits Output Frequency Ranging
1.4
TA = 25°C
1000
1.3
1.2
÷100
100
V + = 3V
RSET (kΩ)
VRES = V + – VSET
V + = 5V
1.1
1.0
÷10
÷1
MOST
ACCURATE
OPERATION
10
0.9
0.8
1
10
100
1000
IRES (µA)
AN93 FA02
Figure A2. V + – VSET Variation with IRES
AN93-18
1
1k
10k
100k
1M
10M
DESIRED OUTPUT FREQUENCY (Hz)
100M
AN93 FA03
Figure A3. RSET vs Desired Output Frequency
for Three Output Divider Settings
Application Note 93
APPENDIX B
RSET NODE CONSIDERATIONS
The RSET node is the LTC1799’s sole analog input.
Figure␣ B1, a partial LTC1799 block diagram (see Appendix␣ A
for more detail) shows that the node is a MOSFET source
and an amplifier input. Equivalent input resistance is about
2kΩ and the point sits approximately 1.13V below the
LTC1799 V + pin. Excessive stray capacitance or noise at
RSET will influence amplifier operation, causing master
oscillator jitter. Stray capacitance at RSET should be limited to <10pF and signal lines, particularly those operating
at high speed, should be routed away from RSET. A simple
guideline is to place the programming resistor directly at
RSET. In cases where RSET is a transducer (e.g., a temperature sensor), it may be desirable to locate the transducer
at the end of cable. Maintaining low effective capacitance
at the RSET node requires “bootstrap” driving the cable
shield at the RSET potential (Figure B2). This negates the
effect of shield capacitance, because charge cannot transfer between it and RSET. An amplifier capable of driving the
shield is required but this is accomodatable. Text Figure 5
is a practical incarnation of this technique.
V+
EXTERNAL
PROGRAMMABLE
RESISTOR
RIN ≈ 2k
CSTRAY MAXIMUM = 10pF
+
–
RSET PIN
MASTER
OSCILLATOR
TO OUTPUT
VIA PROGRAMMABLE
DIVIDER
AN93 FB01
Figure B1. RSET Pin Has Effective Input Resistance of ≈2k. Stray
Pin Capacitance Must be <10pF to Avoid Output Frequency Jitter
V+
+
TRANSDUCER
(PROGRAMMING
RESISTOR)
A1
LT1219L
–
0.1µF
100k
DIV
LTC1799
RSET
OUT
AN93 FB02
Figure B2. A1 Senses RSET Voltage, Bootstraps Cable Drive
Potential. Arrangement Prevents Cable Capacitance from
Influencing RSET Node Because Charge Cannot Transfer.
100k Resistor Isolates A1’s Input and Trace Capacitance
Information furnished by Linear Technology Corporation is believed to be accurate and reliable.
However, no responsibility is assumed for its use. Linear Technology Corporation makes no representation that the interconnection of its circuits as described herein will not infringe on existing patent rights.
AN93-19
Application Note 93
AN93-20
Linear Technology Corporation
an93f LT/TP 0203 2K • PRINTED IN USA
1630 McCarthy Blvd., Milpitas, CA 95035-7417
(408) 432-1900 ● FAX: (408) 434-0507
●
www.linear.com
 LINEAR TECHNOLOGY CORPORATION 2003