V03N3 - OCTOBER

LINEAR TECHNOLOGY
OCTOBER 1993
IN THIS ISSUE . . .
COVER ARTICLE
New LT1300 and LT1301
Micropower DC-to-DC
Converters ....................... 1
Steve Pietkiewicz
Editor's Page ................... 2
VOLUME III NUMBER 3
New LT1300 and LT1301
Micropower DC-to-DC
Converters
by Steve Pietkiewicz
Richard Markell
DESIGN FEATURES
The LTC1257 Provides a
Complete, Single-Supply,
12-Bit D/A in an
SO-8 Package .................. 3
Robert Reay
LT1204: High-Speed Video
Multiplexer with Cable
Driver has −90dB
Crosstalk......................... 5
John Wright
The World’s Lowest-Noise
Dual-JFET Op Amp, the
LT1113, Debuts ............... 8
Alexander Strong
A New Family of HighSpeed, Low-Power
Operational Amplifiers .. 10
George Feliz
DESIGN IDEAS .......... 17–28
(complete list on page 17)
DESIGN INFORMATION
Book Review:
Power Electronics—
Circuits, Devices and
Applications .................. 29
New Device Cameos ....... 30
Introduction
The new LT1300 and LT1301
micropower DC-to-DC converters
provide improvements in both electrical and physical efficiency, two key
areas of battery-based power-supply
design. Housed in 8-lead DIP or SOIC
packages, the devices feature a 1A
on-chip switch with a VCESAT of just
170mV. The internal oscillator frequency is set at 155kHz, allowing the
use of tiny, 5mm diameter surface
mount inductors along with standard
D-case size tantalum capacitors. A
complete 2-cell to 12V, 5V, or 3.3V
converter can fit in less than 0.4
square inches of PC board area.
The devices use Burst ModeTM operation to maintain high efficiency at
light load. The quiescent current is
only 120µA. It can be further reduced
to 10µA by taking the SHUTDOWN
pin high, which also disables the
device. The output voltage of the
LT1300 can be set at either 5V or 3.3V
via the logic controlled SELECT pin,
and the LT1301 output can be set at
either 5V or 12V using the same pin.
The I LIM pin allows the reduction of
peak switch current, normally 1A, to
approximately 400mA, increasing efficiency and allowing the use of even
smaller components in lighter load
applications.
Operation
Figure 1 is a block diagram of the
LT1300/1301. Refer also to Figure 2
for associated component hook-up.
When A1’s negative input, related
to the SENSE pin voltage by the appropriate resistor -divider ratio,
is higher than the 1.25V reference
voltage, A1’s output is low. A2, A3, and
continued on page 12
A2
CURRENT
COMPARATOR
VIN
SHUTDOWN SENSE
500k
1.25V
REFERENCE
+
OSCILLATOR
5.3µs ON
1.2µs OFF
–
+
R2
700Ω
+–
Q2
3×
18mV
A3
LTC in the News ............ 31
144k
Design Tools .................. 32
161k
–
ENABLE
A1
SLOW
COMPARATOR
DRIVER
BIAS
R1
3Ω
SW
Q1
500×
Q3
8500Ω
PGND
Sales Offices ................. 32
GND
SELECT
ILIM
Figure 1. LT1300/LT1301 block diagram
Burst Mode
TM
is a trademark of Linear Technology Corporation.
S1300_1.eps
EDITOR'S PAGE
Earwigs in the Drip Watering System?
Parsnips for Dinner!
by Richard Markell
I’ve been doing a lot of thinking
about earwigs lately. Earwigs are those
bugs that look like little lobsters, with
the little pincers on their tails (order
Dermaptera). They completely blocked
a filter in my drip watering system so
that I had to clean it out every week.
Not good. I spent many a sleepless
night trying to reason out how they got
into the pipes to be swept down to the
point where they were trapped by the
filter. Was there a hole underground?
Were earwigs spontaneously generated inside the pipes? Were they
delivered as protein-enriched water
from my small water company? A sort
of “big bug” theory of evolution?1
We all solve engineering problems,
both at work and at home. Jim Williams fixes Tektronix 547s; I engineer
systems in the garden that water my
parsnips. Is it the same thing? I got to
thinking about the creative process.
I’ve been tracing through the drip system in my mind, trying to figure out
how the earwigs got into a closed
system. Is this different from tracing
through a buck regulator circuit to
find the source of too much output
ripple? How about trying to find out
why the last two bits on an A/D converter system toggle? The creative
aspect of the thinking process is the
same in each endeavor.
This issue is packed with articles
on new products from LTC. Steve
Pietkiewicz and Dale Eagar provide
circuits and architectural insights into
the new LT1300 and LT1301 micropower DC-to-DC converters. Steve
provides the designer’s perspective and
Dale provides some novel circuits for
the LT1300 series.
George Feliz describes his new
series of high-speed operational
amplifiers, which combine the high
slew rates of current-feedback amplifiers with the benefits of traditional
operational amplifiers to create the
LT1354 through LT1365 series of amplifiers. In the related area of video
products, John Wright presents his
70MHz, four-input video multiplexer,
the LT1204. This product provides an
incredible 90dB of isolation at 10MHz.
Bob Reay offers a thorough introduction to LTC’s first digital-to-analog
converter, the LTC1257. This 12-bit
serial D/A converter typically achieves
1/4LSB DNL and 2LSBs INL error
without trimming. The part complements our line of serial A/Ds quite
nicely.
Alexander Strong introduces the
JFET -input LT1113 operational
amplifier. The LT1113 is the world’s
lowest-noise dual-JFET op amp; the
part was designed to amplify highimpedance capacitive transducers. Tim
Skovmand discusses 2-cell power management techniques and provides a
complete 2-cell to 3.3V, 5V and 12V
power-management system.
Also, there are many Design Ideas:
an LCD bias generator, a linear-phase
bandpass filter, a high-efficiency battery charger, a +5V to +3.3V converter
for desktop computers, and a LT1087
GTL terminator.
positioned in either direction for loading and unloading the blood samples.
The 20-inch-diameter rotor then had
to be accelerated to 5000 RPM at a
defined rate in 5.0 seconds. In addition, the rotational speed had to be
held to 720 RPM ±0.1 RPM during the
data-acquisition period to define the
sample rate in an A/D conversion system. This Analog-to-digital control
system comprised F/V and D/A converters, precision ampli-fiers and filters,
high-voltage H-bridge drivers, and a
microcontroller with digital filtering.”
Dave recently talked to an engineer who had a very interesting
application. He was building a
hydrophone amplifier to track the
movements of schools of fish. He had
gone so far as to have a taxidermist
make a plaster mold of a rainbow
trout. He and his co-workers spent
hours painting the prototype trout
to look very realistic. (This may be
taking precision to the extreme.)
He mounted the electronics (which,
of course, use LTC op amps for signal
conditioning) inside the plaster fish.
David and his wife, Donna, have
been married for seven years. They
have a 16-month-old baby girl, Dana,
who, like her father, is already taking
things apart to see how they work.
He enjoys golf, yard work (really?),
and playing with his daughter. You
can reach David through the LTC
Southeast Sales office listed on the
back of this magazine.
1
See the bottom of page 3 for the answer, “How they
got into the pipes.”
FAE Cameo: David Dinsmore
LTC now has twenty Field Application Engineers worldwide to assist our
customers in the design and selection
of circuits available from LTC. All of
our FAEs are available by phone and,
in certain situations, in person to help
you design your circuitry. This space
will profile one FAE per issue.
David Dinsmore works out of LTC’s
Southeast Sales office. He covers the
states of Texas, Oklahoma, Alabama,
Louisiana, Arkansas, Mississippi and
the western half of Tennessee. David’s
expertise is in the areas of data-acquisition systems and servo-motor control.
Dave relates, “One of my most interesting projects was a 3/4HP
motor-control system for a blood-analyzer centrifuge. It had to be precisely
2
Linear Technology Magazine • October 1993
DESIGN FEATURES
The LTC1257 Provides a Complete, SingleSupply, 12-Bit D/A
in an SO-8 Package
by Robert Reay
Introduction
Circuit Topology
Digital Section
The block diagram and pin diagram of the LTC1257 are shown in
Figure 2. The digital section consists
of a 12-bit shift register, control logic,
DAC register, and 5V logic regulator.
Signal lines include inputs CLK, DIN,
and LOAD, and output DOUT. The
data on the DIN pin is clocked into the
shift register and the last bit in the
shift register is shifted to DOUT through
a buffer on the rising edge of the clock.
The MSB is loaded first and the LSB
last. The DAC register loads the data
from the shift register when LOAD is
pulled low, and remains transparent
until LOAD is pulled high and
the data is latched. The logic regulator
limits the internal logic’s voltage
swing to 5V, reducing clock noise in
the analog section when VCC is at
15V. However, the DOUT pin will swing
from GND to VCC.
0.5
DNL ERROR (LSB's)
The new 8-pin SO LTC1257 requires no external components and
provides one of the smallest, easiestto-use 12-bit D/A (DAC) systems
available. The part includes an
output-buffer amplifier, a 2.048V
voltage reference, and an easy-to-use,
cascadable three-wire serial interface. The single supply voltage can
range from 4.75V to 15.75V and an
external reference can be used to override the internal reference to extend the output voltage range to
0V–12V. The power supply current is a
low 350µA max when operating from
a 5V supply, and the differential nonlinearity (DNL) is less than 1/2LSB
(see Figure 1), with no missing codes.
0.0
–0.5
0
512
1024
1536
2048
CODE
2580
3072
3584
Figure 1. LTC1257 differential nonlinearity (DNL) plot
Multiple LTC1257s can be daisychained by connecting the DOUT pin of
one chip to the DIN pin of the next, with
the CLK and LOAD signals remaining
common to all chips. The
serial data is clocked to all of the
DACs and the LOAD signal is pulled
low to update all of them simultaneously. The serial interface can be
clocked at any speed up to 1.4MHz.
4098
1257_1.eps
Analog Section
The 12-bit DAC ladder, the 2.048V
reference, and an op amp connected
as a voltage follower make up the
analog section. The DAC has been
optimized for excellent DNL performance in a small area. Figure 3 shows
the DAC ladder topology. A string of
equal-valued resistors is connected
between the reference pin and ground.
INTERNAL
LOGIC SUPPLY
5V REGULATOR
VCC
CLK
DIN
12-BIT SHIFT REGISTER
DOUT
12
LOAD
DAC REGISTER
12
–
VOUT
GND
DAC
+
CLK
1
8
VCC
DIN
2
7
VOUT
LOAD
3
6
VREF
DOUT
4
5
GND
1257_2b.eps
REF
2.048V REFERENCE
1257_2a.eps
Figure 2a. LTC1257 block diagram
Figure 2b. LTC1257 pin diagram
How they got into the pipes
The earwigs got in through the opening in the
anti-siphon valve. I solved the problem by putting a fine mesh (my wife’s old stocking) over the
valve.
Linear Technology Magazine • October 1993
3
DESIGN FEATURES
5V
22µF
VCC
CH0
CS
LTC1296
DOUT
8 ANALOG
INPUT CHANNELS
VREF
CH7
+
50k
–
µP
CLK
COM
DIN
REF+
– SSO
REF
50k
5V
74HC04
2N3906
0.1µF
VCC
VREF
DIN
100Ω
VOUT
CLK
LTC1257
LOAD
0.1µF
DOUT
GND
VCC
VREF
DIN
100Ω
VOUT
1257_3.eps
LTC1257
CLK
LOAD
0.1µF
DOUT
GND
1257_4.eps
Figure 3. Equivalent DAC topology
A switch connects each resistor tap to
the input of the op amp. For any given
code, the corresponding switch is
turned on, and the resistor ratio determines the fraction of the reference
voltage appearing at the output of the
buffer op amp. This topology guarantees no missing codes and the integral
nonlinearity is determined by the
matching of the resistors within the
string.
Because a 12-bit DAC would require 4096 resistors and a large layout,
we have developed a patented scheme
in which only the MSBs are determined by the resistor string and the
LSBs are decoded by a modified input
stage in the op amp. This design dramatically reduces the number of
resistors and the size of the DAC and
maintains excellent DNL performance.
The typical DNL error is
1/4LSB while the typical INL error is
2LSBs, without any costly trimming.
The internal voltage reference provides a constant 2.048V, making 1LSB
equal to 500µV. The reference is
connected to the DAC resistor ladder
4
Figure 4. Auto-ranging, 8-channel ADC with shutdown
and to the REF pin. The reference is a
bootstrapped bandgap circuit with a
typical temperature coefficient of
20ppm/°C. The voltage reference output is turned off when the pin is forced
above 2.5V, allowing an external, highprecision reference to be connected to
the REF pin and DAC resistor ladder.
By using the external reference, the
full scale voltage of the DAC can be
extended up to 12V. The external reference must be greater than 2.5V, less
than VCC − 2.7V, and must be capable
of driving the 10kΩ minimum DAC
resistor ladder.
The op amp buffer is connected as a
voltage follower and has a commonmode range that extends from ground
to within 2.7V of VCC while sourcing
2mA. An internal NMOS transistor
with a 200Ω equivalent impedance
pulls the output to ground. The output is protected against short circuits
and is able to drive a capacitive load of
up to 500pF without oscillation. Offset is 4mV max over temperature and
the settling time is 6µs maximum.
Applications
The LTC1257 is intended for applications where small size, low external
parts count, low supply current, singlesupply operation, and excellent DNL
performance are needed. Applications
include portable instrumentation, digitally controlled calibration, servo
controls, process-control equipment,
and automatic test equipment.
Two LTC1257s and a LTC1296 A/D
can be used to build an auto-ranging,
8-channel ADC system with shutdown, as shown in Figure 4. Two
LTC1257s are cascaded together and
provide the zero and full-scale reference voltages for the LTC1296 A/D
converter. The DAC outputs are filtered to remove the small amount of
digital noise from the outputs. VCC for
the DACs is supplied via a PNP transistor. The microprocessor writes the
two 12-bit full-scale and zero words to
the DACs via the DIN and CLK inputs,
then pulls LOAD low to update the
DAC outputs. With zero and full-scale
set, an A/D conversion can be made.
continued on page 15
Linear Technology Magazine • October 1993
DESIGN FEATURES
LT1204: High-Speed Video Multiplexer
with Cable Driver has −90dB Crosstalk
by John Wright
1 VIN 0
VIN 0
V+ 16
+1
The Challenge of
Video Multiplexers
+15V
+
75Ω
2
CFA
GND
75Ω
V0 15
VOUT
–
3 VIN 1
VIN 1
V – 14
+1
75Ω
4
5 VIN 2
VIN 2
FB 13
GND
–15V
RF
1k
RG
1k
SHUTDOWN 12
+1
75Ω
6
7 VIN 3
VIN 3
ENABLE 11
GND
LOGIC
A1 10
+1
75Ω
8
A0 9
REF
–15V
6.8k
6.2k
1204_1a.eps
Figure 1. LT1204 block diagram
Introduction
The LT1204 is the first in a series of
products developed by Linear Technology to solve difficult video switching
and distribution problems in the new
multimedia products. This new 4input multiplexer includes a 75MHz
current-feedback amplifier that will
directly drive 75 ohm cables. We have
developed a novel circuit technique to
expand the number of multiplexers for
large routing systems without degrading the signal integrity. This technique
simplifies cable terminations as the
number of inputs increases. Video
specifications, such as differential gain
and phase, gain flatness, and switching transients are at professional video
levels. Unlike many multiplexers,
which have 1V–2V input ranges, the
LT1204 supports large input and output signal levels, which make it ideal
for general-purpose analog signal se-
Linear Technology Magazine • October 1993
lection and multiplexing. The LT1204
is available in 16-lead PDIP and 16lead SOL packages.
Figure 1 shows a block diagram of
the LT1204. Its truth table appears as
Table 1. The input buffers are actually
low-insertion-loss tee switches designed to give the excellent crosstalk
performance required for professional
video. The switches are internally connected to the noninverting input of the
CFA to reduce capacitance and improve the AC characteristics of the
signal path; the inverting input is external for easy gain adjustment. The
logic interface decodes channel select,
shutdown, and enable/disable, the
last of which puts the CFA into a true
high output impedance state. A reference (Pin 8) is available for optimizing
the internal logic circuitry for different
input signal ranges.
Before HDTV, CD Interactive, and
the proliferation of video products,
source selection was made during the
blanking period, and the effects of
switching transients were not visible.
As multimedia, new special effects,
and picture processing become popular, it has become necessary to switch
video “in picture”; in such cases the
nature of the switching transient is
critical. Switching techniques that
worked in the past now cause problems. Older bipolar ICs that switch
lateral PNP transistors in the signal
path take several microseconds to
settle, blurring the transition between
pictures. CMOS multiplexers, which
are bidirectional, suffer from poor output-to-input isolation and cause
transients to feed to the inputs. CMOS
MUXs have been built with breakbefore-make switches to eliminate the
talking between channels, but these
suffer from output glitches large
enough to interfere with the sync circuitry. By contrast, the LT1204 is
fabricated on LTC’s complementarybipolar process to attain good switching
characteristics, buffering, crosstalk,
and speed. Let’s look at these areas
one-by-one to see how they stack up.
Table 1. LT1204 truth table
A1
0
0
1
1
X
X
A0
0
1
0
1
X
X
Channel
ENABLE SHUTDOWN Selected
1
1
VIN0
1
1
VIN1
1
1
VIN2
1
1
VIN3
0
1
High-Z Output
X
0
OFF
5
DESIGN FEATURES
Switching Characteristics
Switching between channels is a
make-before-break condition where
both inputs are on momentarily. The
input with the largest positive voltage
determines the output level. If both
inputs are equal, there is only 40mV
of error at the input of the CFA during
the transition. The reference adjust
(Pin 8) allows the user to trade positive
input voltage range for switching time.
For example, on ±15V supplies, setting the voltage on pin 8 to −6.8V
reduces the switching transient duration to 50ns and the positive input
range from +6V to +2.35V; the negative input range is independent of the
setting on pin 8 and remains at −6V.
When switching composite video “in
picture,” this short (50ns) transient is
imperceptible, even on high quality
monitors. The reference pin has no
effect when the LT1204 is operating
on ±5V, and should be grounded in
this situation. Figure 2 is a scope
photograph of the output switching
transient with a 2MHz sine wave connected to VIN0 and VIN1.
puts. Disable crosstalk is measured
with all four inputs driven and the
part disabled. Crosstalk is critical in
many applications where video multiplexers are used. In professional video
systems, a crosstalk figure of −72dB is
a desirable specification.
The key to the outstanding
crosstalk performance of the LT1204
is the tee switch shown in Figure 3.
When the tee switch is on (Q2 off) Q1
and Q3 are a pair of emitter-followers
with excellent AC response for driving
the CFA. When the decoder turns off
the tee switch (Q2 on), the emitterbase junctions of Q1 and Q3 become
reverse biased while the Q2 emitter
absorbs current from I1. Not only do
the reverse-biased emitter -base
junctions provide good isolation, but
any signal at VIN0 coupling to the Q1
emitter is further attenuated by the
shunt impedance of the Q2 emitter.
Current source I2 routes current to
any ON switch.
Crosstalk performance is strongly
affected by the IC package and the PC
board layout, as well as by the circuit
design. The die layout uses grounds
between the inputs to isolate adjacent
channels, and the output and feedback pins are on opposite sides of the
die from the inputs. Laying out a PC
board that provides 90dB of isolation
from all crosstalk at 10MHz is not a
trivial task. That crosstalk level corresponds to a 30µV output below a 1V
input at 10MHz. We have fabricated a
V+
I1
LOGIC
PIN 9
Q3
VIN
+
Q1
TO
LOGIC
VOUT
CFA
Q2
–
V–
VOUT
PIN 15
RF
FB
Input Buffers
Crosstalk
The crosstalk (more accurately, all
hostile crosstalk) is measured by driving signal into any three of the four
inputs and selecting the fourth input
with logic control. This fourth input is
either shorted to ground or terminated in an impedance. All hostile
crosstalk is defined as the ratio in dB
of the signal, at the output of the CFA,
to the signal on the three driven in6
RG
I2
1204_3.eps
1204_2.eps
Figure 2. VIN0 and VIN1 connected to a 2MHz
sinewave, pin 8 voltage = – 6.8V
Figure 3. Tee switch
–20
–40
ALL HOSTILE CROSSTALK (dB)
The buffers isolate the inputs when
the make-before-break switching
occurs. The design of these input buffers included special attention to their
DC matching and dynamic characteristics. The DC input-offset match
between channels is more important
to the video engineer than is the
actual value of the input offset. A DC
mismatch as small as 3mV between
channels is just visible on a quality
video monitor. The typical VOS mismatch between channels on the
LT1204 is about 300µV.
VS = ±15V
RL = 100Ω
RS = 10Ω
–60
PDIP
–80
SOIC
–100
–120
1
10
FREQUENCY (MHz)
100
1204_4.eps
Figure 4. All hostile crosstalk of LT1204 in PDIP and SOIC
Linear Technology Magazine • October 1993
DESIGN FEATURES
demonstration board to show the component and ground placement
required to attain these crosstalk numbers. Figure 4 is a graph of all hostile
crosstalk for both the PDIP and SO
packages. It has been found empirically from these PC boards that
capacitive coupling across the package of greater than 3fF (yes, that is
0.003 picofarads) will diminish the
rejection; we recommend that you use
this proven layout in your designs.
V+
V–
V+
Multiplexer Expansion
TEE SWITCH
V–
+
TEE SWITCH
VIN3
75Ω
CFA
“OFF”
TEE SWITCH
CABLE
VOUT
–
75Ω
TEE SWITCH
RF
FB
RG
V–
75Ω
LT1204
“ON”
1204_5.eps
Figure 5. Active buffer drives FB node during disable
+2
DISABLE
0
SHUTDOWN
GAIN (dB)
LT1204s can be paralleled by shorting their outputs together to expand
the number of MUX inputs. This new
multiplexer uses a novel circuit (patent
pending) to ensure that the unselected
amplifiers do not load or alter the
cable termination and that there is no
shoot-through current when the outputs of two or more amplifiers are
shorted together. (Shoot-through current is a spike of power-supply current
caused by both amplifiers being on at
the same time.) When the LT1204 is
disabled (pin 11 low), the output stage
is turned off and an active buffer
senses the output voltage and drives
the feedback pin of the CFA (Figure 5).
This bootstraps the feedback resistors and raises the true output
impedance of the circuit. For the condition where RF = RG = 1kΩ, the output
impedance is typically raised to 25kΩ
during disable. If the part is shut
down, however, the bootstrapping is
inoperative, and the feedback resistors will load the output (ROUT = 2kΩ).
If the CFA is operated at a gain of +1,
the feedback resistor will not load the
output even in shutdown mode, because there is no resistive path to
ground; however, there will be a 6dB
loss through the cable.
Figure 6 is a frequency-response
plot showing the effect of using the
disable feature versus using the shutdown. In this example, four LT1204s
were connected together at their outputs to form a 16-to-1 MUX. The plot
shows the effect of the bootstrapping
circuit, which eliminates the improper
cable termination due to the feedback
resistors loading the cable.
VIN0
–2
VS = ±15V
RL = 150Ω
–4
–6
1
10
FREQUENCY (MHz)
100
1204_6.eps
Figure 6. 16-to-1 multiplexer response using disable feature versus shutdown feature
Continued on page 16
Linear Technology Magazine • October 1993
7
DESIGN FEATURES
The World’s Lowest-Noise Dual-JFET Op
Amp, the LT1113, Debuts
by Alexander Strong
The LT1113 joins the LTC family of
low-noise op amps as the lowest-noise
dual-JFET op amp available. This op
amp combines the low current noise
(less than 10fA/(Hz)1/2) of a FET op
amp with a maximum voltage noise of
6.0nV/(Hz)1/2. In addition, the LT1113
is stable for a gain of +1 and will
handle 1000pF load capacitances. All
of this is available in an 8-pin smalloutline surface mount package with
the standard pinout. Most op amps
require a relaxation of specifications
in the surface mount package due to
assembly shifts. For the LT1113CS8,
spec relaxation is not necessary. The
LT1113CS8 specs are identical to the
LT1113CN8. Table 1 highlights some
of the guaranteed specifications for
the low-cost grade.
Table 1. LTC1113CS8 guaranteed specifications: 100% tested, VS = ±15V, TA = 25°C
Voltage Noise @ 1kHz
VOS
AVOL (RL = 10kΩ)
GBWP @ 100kHz
Slew Rate
ISUPPLY per amp
6.0nV/√Hz Max
1.8mV Max
1000V/mV Min
4.5MHz Min
2.5V/µS Min
6.5mA Max
Current noise is derived from the
DC value of the FET input bias current,
or
In(fA/(Hz)1/2 ) = (2qIB)1/2.
where q = 1.6E–19
For a 300pA IB, a 9.8fA/(Hz)1/2
current noise is multiplied by the
transducer impedance, which adds to
the total output voltage noise of the
amplifier. The best low-noise bipolar
op amps cannot match the performance of the LT1113 where highimpedance transducers are used.
Figure 1 shows a comparison between
the LT1113 and the LT1124 bipolar
op amp. The LT1124 has 40% less
voltage noise at 1kHz than the LT1113
and is a better choice for transducer
impedances less than 1kΩ, but for
transducer impedances over 100kΩ,
the LT1113 is clearly the champ. The
dashed lines on Figure 1 show the
total noise with a parallel 1000pF
capacitance (since most hydrophone
and accelerometer transducers are
capacitive by nature). The graph shows
that when the LT1113 is used, the
total noise is dominated by the transducer impedance and not the amplifier,
as in the case of the LT1124.
Hydrophones Require
High Input Impedance
The combination of low voltage noise
and low current noise makes the
LT1113 suitable in applications where
low level signals need to be amplified
from high impedance capacitive transducers. Photo diodes, hydrophones,
and accelerometer transducers exhibit high impedances, which make
the op amp current noise dominate
the total output noise:
TOTAL 1kHz VOLTAGE NOISE DENSITY (nV/√Hz)
Introduction
1k
LT1124*
LT1113*
CS
100
–
RS
+
LT1124†
VO
CS
RS
LT1113†
10
LT1113*
LT1124*
BOTH SOURCE RESISTORS ONLY
1
100
10K 100K
1M
10M 100M
1K
SOURCE RESISTANCE (RS IN OHMS)
* PLUS BOTH SOURCE RESISTORS
†
SOURCE RESISTORS  1000pF SOURCE CAPACITANCE
Vn = √Vn2(OP AMP) + 4kT • 2RSOURCE + 2q IB • 2(RSOURCE)2
1113_1.eps
Figure 1. Comparison of LT1113 and LT1124
total output 1kHz voltage noise versus source
resistance
The two basic gain configurations
are shown in Figure 2. The noninverting gain configuration is used for
voltage-mode transducers such as hydrophones, whereas the inverting gain
configuration is used for charge transducers such as accelerometers. In
each example, a source-balancing impedance is added to improve the overall
performance of the amplifier. RB is
selected to cancel the voltage offset
due to the input bias current flowing
RF
R2
CB
CF
RB
–
–
R1
OUTPUT
+
CS
RS
CS
RS
+
TRANSDUCER
CB ≅ CS
RB = RS
RS > R1 OR R2
CB = CFCS
RB = RFRS
CB
RB
TRANSDUCER
Vn =
AV √Vn2(OP AMP) + 4kT • 2RTRANS + 2q IB • 2(RTRANS)2
8
OUTPUT
1113_2b.eps
1113_2a.eps
Figure 2. Noninverting and inverting gain configurations
Linear Technology Magazine • October 1993
DESIGN FEATURES
into the source impedance. This is
especially important as the operating
temperature increases, since FET
input bias currents are usually
uncancelled, and double for every 10
degrees C. A parallel capacitor CB
cancels the pole that is caused by the
amplifier input capacitance and RB.
Applications
Figure 3 shows a low-noise hydrophone amplifier with a DC servo. Here
one half of the LT1113 is configured in
the noninverting mode to amplify a
voltage signal from the hydrophone,
and the other half of the LT1113 nulls
errors due to voltage and current offsets of the amplifier and to impedance
mismatches. The value of C1 depends
on the capacitance of the hydrophone,
which can range from 200pF to
8000pF. The time constant of the servo
should be larger than the time constant of the hydrophone capacitance
and the 100MΩ source resistance.
This will prevent the servo from canceling the low-frequency signals from
the hydrophone.
Another popular charge-output
transducer is the accelerometer. Since
precision accelerometers are charge
output devices, the inverting mode is
used to convert the transducer charge
to an output voltage. Figure 4 is an
example of an accelerometer with a DC
servo. The charge from the transducer
is converted to a voltage by C1, which
should equal the transducer capaci-
tance plus the input capacitance of the
op amp. The noise gain will be 1 + C1/
CT. The low frequency bandwidth will
depend on the value of R1 (or
R1 × (1+R2/R3) for a Tee network). As
with the hydrophone example, the time
constant of the servo should be larger
than the time constant of the amplifier.
Conclusion
The LT1113 is not just another op
amp. As the lowest-noise dual-JFET
op amp in the market place, the
LT1113 should be the first choice
where high-impedance transducers
are used. The 8-pin surface mount
package will allow the LT1113 to
satisfy the most demanding board
space requirements.
C1
1250pF
R1
100MΩ
3.9k
R2
18k
R3
2k
C1
C4
2µF
V+
2
8
–
A
1/2 LT1113
100MΩ
200Ω
–
3
+
1
OUTPUT
7
6
R4
20M
5
R5
20M
1/2 LT1113
+
4
C5
2µF
V–
1µF
CT
HYDROPHONE
8
100MΩ
1MΩ
100k
2
–
7
6
B
1/2 LT1113
C1 = CT = 200pF TO 8000pF
DC OUTPUT ≤ 4mV AT TA < 70°C
OUTPUT VOLTAGE NOISE = 130nV/√Hz AT 1kHz
POWER SUPPLY RANGE = ±5V TO ±15V
+
5
1MΩ
ACCELEROMETER
3
CT
1MΩ
–
1/2 LT1113
1
+
4
1113_3.eps
OUTPUT = 0.8µV/pC*
= 8.0mV/G**
DC OUTPUT ≤ 2.7mV
OUTPUT NOISE = 6µV/√Hz AT 1kHz
*PICOCOULOMBS
**EARTH’S GRAVITATIONAL CONSTANT
Figure 3. Low-noise hydrophone amplifier with DC servo
Linear Technology Magazine • October 1993
1113_4.eps
Figure 4. Accelerometer circuit with DC servo
9
DESIGN FEATURES
A New Family of High-Speed,
Low-Power Operational Amplifiers
by George Feliz
Introduction
A new family of high-speed operational amplifiers from LTC utilizes a
novel circuit topology that blends
the high slew rate of current-feedback amplifiers with the benefits of
true voltage-feedback amplifiers. Compared to other devices with similar
bandwidths, these amplifiers offer
lower supply current, higher slew
rate, better DC specifications, faster
settling times, and lower input noise.
The family is built on LTC’s complementary bipolar process and
encompasses bandwidths from 12MHz
to 70MHz. The fastest parts are the
LT1363, LT1364, and LT1365, which
have 70MHz gain bandwidth and
1000V/µs slew rate, and consume
only 6mA of supply current per amplifier. The lowest power devices are
the LT1354, LT1355, and LT1356,
which draw only 1mA of supply current and provide 12MHz of bandwidth
and 400V/µs slew rate. In between
are the LT1357, LT1358, and LT1359,
which have 25MHz gain bandwidth
and 600V/µs slew rate and consume
2mA, and the LT1360, LT1361, and
LT1362, which have 50MHz bandwidth and 800V/µs slew rate, and
consume 4mA.
Table 1 summarizes the important specifications of each device. All
Table 1. Important specifications; LT1354, LT1357, LT1360, and LT1363 op amp
families
Parameter
VOS (max)
IB (max)
IOS (max)
AVOL (min)
Conditions
LT1354/5/6 LT1357/8/9 LT1360/1/2 LT1363/4/5 Units
VS = ±5V, ±15V
1
0.6
1
1.5
mV
VS = ±5V, ±15V
0.3
0.5
1
2
µA
VS = ±5V, ±15V
70
120
250
350
nA
VS = ±5V, 500Ω
10
10
3
3
V/mV
VS = ±15V, 1kΩ
20
20
4.5
4.5
V/mV
VS = ±15V
12
25
50
70
MHz
400
600
800
1000
V/µs
VS = ±15V
10V step, 0.1%
240
170
60
50
ns
VS = ±15V
86
86
86
86
dB
V S = ±5V
80
80
79
78
dB
VS = ±2.5V to ±15V 92
92
97
90
dB
10kHz
10
8
10
10
nV/√Hz
10kHz
0.6
0.8
1
1.2
pA/√Hz
12.5
12.5
13
13
V
VS = ±15V, 500Ω
VS = ±5V, 150Ω
3
3
3.2
3.4
V
VS = ±15V
1
2
4
6
mA
GBW
SR
tSETTLE
CMRR (min)
PSRR (min)
Noise voltage
Noise current
Output swing
ISUPPLY
Background
devices operate over a wide supply
voltage range of ±2.5V to ±18V and
are available in dual, single, and
quad versions. The single and dual
versions are available in 8-pin DIPs
and surface mount packages. The
quad versions are available in 14-pin
DIPs and narrow, S16 surface mount
packages.
V+
IO
–IN
IO
Q1
Q2
R
+IN
Q3
R
Q4
BIAS
+1
Q5
IO
Q6
Q7
CT
V–
Figure 1. Traditional high-speed amplifier—simplified schematic
10
OUTPUT
OUTPUT
STAGE
1363_1.eps
The new topology changes the relationship between slew rate and supply
current found in traditional voltagefeedback amplifiers. In order to
understand the circuit, let’s review a
traditional high-speed, voltage-feedback design, as shown in Figure 1.
The input pair Q1 and Q2 is resistively degenerated and feeds PNP
cascodes Q3 and Q4. Differential-tosingle-ended conversion is performed
by the mirror Q5-Q7. The collectors of
Q4 and Q5 form the high-impedance
gain node with capacitance CT, establishing the dominant pole. A
unity-gain buffer follows the gain
node. The slew rate is determined by
the current available to charge CT,
which, in this case, is IO. Since the
slew rate is IO/CT, it is directly proportional to supply current. In order to
obtain high slew rates, the circuit’s
quiescent currents must be large. For
instance, a 1mA current is required to
obtain 250V/µs with 4pF of capacitance. The bandwidth of the amplifier
Linear Technology Magazine • October 1993
DESIGN FEATURES
is determined by R and CT and is given
by 1/(2πRCT). The useable bandwidth is typically limited by other
poles in the frequency response, so R
is selected to obtain adequate stability. For R = 1kΩ and CT = 4pF, the
bandwidth is 40MHz. Once the slew
rate and bandwidth are chosen, the
DC performance is set by the degeneration of the input pair. With IO =
1mA and R = 1kΩ, a resistor mismatch of only 0.2% gives 1mV of
input offset voltage. The open-loop
gain is also reduced by the input
degeneration, as it is inversely proportional to R. For high-slew-rate designs, the input bias current is
high because it is also directly proportional to IO. In our example, if the
transistor beta is 100, the input bias
current would be 5µA and the offset
current would be about 500nA. Input
noise is also degraded due to the
resistor noise and reduced gain.
The New Topology
Figure 2 is a simplified schematic
of the new circuit. The circuit looks
similar to a current-feedback amplifier, but both inputs are high
impedance as in a traditional voltage-feedback amplifier. A complementary cascade of emitter followers
Q1–Q4 buffers the noninverting input and drives one side of a resistor.
The other side of the resistor is driven
by Q5–Q8, which form a buffer for
the inverting input. The input voltage
appears across the resistor, generating a current that is mirrored by
Q9–Q14 into the high impedance
node. Q15–Q18 form an output stage.
In a current-feedback amplifier, there
would be an external resistor connected between the output and the
emitters of Q3 and Q4, which would
be the inverting input.
The bandwidth, as before, is
1/(2πRCT), but the current available
to slew CT is the differential input
voltage divided by 2R, so the slew rate
is independent of the operating currents of the input stage (which can
then be reduced). For a 4V input
step, a 2kΩ resistor, and a 4pF capacitor, the slew rate is ideally 500V/
µs. The high slew rate and balanced
input stage significantly reduce settling time. As with current-feedback
am-plifiers, the RC time constant sets
both the small and large signal responses. Figure 3 shows the large
signal response of the family of
amplifiers.
The new topology provides significantly better DC performance. The
offset does not rely on tight resistor
matching as it does in the conventional circuit, but rather on transistor
matching and errors in the current
mirrors. A mismatch of 4% between
the input devices gives 1mV of input
V+
Q10
Q11
Q17
Q9
Q7
–IN
Q5
Q3
2R
Q6
Q2
Q1
Q16
+IN
RC
Q8
Q4
OUTPUT
CC
Q15
Q18
Q12
CT
Q13
Q14
V–
1363_2.eps
Figure 2. New high-speed amplifier—simplified schematic
Linear Technology Magazine • October 1993
LT1354
LT1357
LT1360
LT1363
Figure 3. Large signal step response: LT1354,
LT1357, LT1360, LT1363
offset voltage; this degree of matching is more readily achieved than the
0.2% resistor matching required for
the conventional circuit. A 1µA mirror error will show up as 2mV of VOS
for R = 2kΩ. It is practical to reduce
the value of R in this circuit, increasing the open-loop gain, decreasing
offset contributions from the current
mirrors, and reducing noise. Input
bias currents are reduced in two ways:
the input devices can run at lower
currents because the slew rate is
independent of their operating current, and the NPN and PNP base
currents tend to cancel. For example,
if the input devices run at 200µA, the
input bias current will be less than
1µA, even if the betas of the NPN and
PNP mismatch by a factor of two. The
balanced inputs of the new topology
also provide excellent rejection of
power-supply and common-mode
variations.
The circuit has one other noteworthy feature—it is able to drive
capacitive loads and remain stable.
The RC–CC network across the output
stage is bootstrapped when the amplifier is driving a light or moderate
load and has no effect under normal
operation. When driving a capacitive
load, the network is incompletely
bootstrapped and adds to the compensation network. The added
capacitance provided by CC slows
down the amplifier and the zero created by R C adds phase margin,
continued on page 16
11
DESIGN FEATURES
L1*
10µH
2×
AA
CELL
SHUTDOWN
+
100µF
SELECT
SHDN
TRACE A
1A/DIV
VIN
SW
LT1300
NC
ILIM
GND
SENSE
PGND
+
D1
1N5817
5V OUTPUT
200mA
C1
100µF
TRACE B
1A/DIV
*SUMIDA CDS 4-100LC (708) 956-0666
COILCRAFT 3316-223 (800) 322-2645
5µs/DIV
S1300_3.eps
S1300_2.eps
Figure 3. Switch pin current with ILIM floating or
grounded
Figure 2. 2-cell to 5 volt DC/DC converter delivers >200mA with a 2.0 volt input
LT1300, continued from page 1
high, and the entire cycle repeats. If
switch current reaches 1A, causing
A2 to trip, switch on-time is reduced
and off-time increases slightly. This
allows continuous mode operation
during bursts. A2 monitors the
voltage across 3Ω resistor R1. Q2’s
collector current is set by the emitterarea ratio to 0.6% of Q1’s collector
current. When R1’s voltage drop exceeds 18mV, A2’s output goes high,
truncating the on-time portion of the
oscillator cycle and increasing off-time
to about 2µs, as shown in Figure 3,
trace A. Eighteen millivolts across R1
corresponds to a switch current of 1A.
This peak current can be reduced by
tying the ILIM pin to ground, causing
15µA to flow through R2 into Q3’s
collector. Q3’s current causes a
10.4mV drop in R2, so that only
7.6mV is required across R1 to turn
off the switch. This corresponds to a
400mA switch current, as shown in
the oscillator are turned off, drawing
no current. Only the reference and A1
consume current, typically 120µA.
When the voltage at the SENSE pin
decreases enough to overcome A1’s
6mV hysteresis, A1’s output goes
high, enabling the oscillator, A2, and
A3. Quiescent current increases to
2mA as the device prepares for highcurrent switching. Q1 then turns on
in a controlled saturation for
(nominally) 5.3µs or until current comparator A2 trips, whichever comes
first. After a fixed off-time of (nominally) 1.2µs, Q1 turns on again. Q1’s
switching causes current to alternately
build up in L1 and dump into output
capacitor C1 via D1, increasing the
output voltage. When the output is
high enough to cause A1’s output to go
low, switching action ceases. C1 is left
to supply current to the load until VOUT
decreases enough to force A1’s output
Figure 3, trace B. The reduced peak
switch current reduces I2R losses in
Q1, L1, C1, and D1. You can increase
efficiency by doing this, provided the
output-current reduction is acceptable. Lower peak currents also extend
alkaline battery life, due to the alkaline cells’ high internal impedance.
Five Volts from Two Cells
Figure 2’s circuit provides 5V from
a 2-cell input. Shutdown is effected
by taking the SHUTDOWN pin high.
VIN current drops to 10µA in this
condition. This simple boost topology
does not provide output isolation,
and in shutdown the load is still
connected to the battery via L1 and
D1. Figure 4 shows the efficiency of
the circuit with a range of input
voltages, including a fresh battery
(3V) and an “almost-dead” battery
(2V). At load currents below a few
milliamperes, the 120µA quiescent cur-
90
88
VIN = 4.0V
EFFICIENCY (%)
86
84
VIN = 3.0V
82
VIN = 2.5V
80
VIN = 2.0V
78
A = 20mV/DIV
AC COUPLED
B = 5V/DIV
C = 1A/DIV
76
74
1
10
100
LOAD CURRENT (mA)
500
20µs/DIV
S1300_5.eps
S1300_4.eps
Figure 5. Burst ModeTM operation in action
Figure 4. Efficiency of Figure 2’s circuit
12
Linear Technology Magazine • October 1993
DESIGN FEATURES
5.0
OUTPUT
2× E91
ALKALINE
4.0
2× L91
LITHIUM
3.5
3.0
2.5
BATTERY
2.0
1.5
1.0
0.5
0
0
1
2
3
4 5 6 7
TIME (HOURS)
8
9 10 11
S1300_6.eps
Figure 6. Two Eveready L91 lithium AA cells
provide approximately twice the life of E91
alkaline cells at a 100mA load current
rent of the device becomes significant,
causing the fall-off in efficiency detailed in the figure. At load currents in
the 20mA to 200mA range, efficiency
flattens out in the 80% to 88% range,
depending on the input. Figure 5 details circuit operation. VOUT is shown in
trace A. The burst-repetition pattern is
clearly shown as VOUT decays, then
steps back up due to switching action.
Trace B shows the voltage at the switch
node. The damped, high-frequency
waveform at the end of each burst is
due to the inductor “ringing off,” forming an LC tank with the switch and
diode capacitance. It is not harmful
and contains far less energy than the
high-speed edge which occurs when
the switch turns off. Switch current is
shown in trace C. The current comparator inside the LT1300 controls
peak switch current, turning off the
switch when the current reaches approximately 1A.
Although efficiency curves present
useful information, a more important
measure of battery-powered DC/DC
converter performance is operating life.
Figures 6 and 7 detail battery life tests
with Figure 2’s circuit at load currents
of 100mA and 200mA, respectively.
Operating-life curves are shown using
both Eveready E91 alkaline cells and
new L91 “Hi-Energy” lithium cells.
These lithium cells, new to the market,
are specifically designed for high-drain
applications. The performance advantage of lithium is about 2:1 at 100mA
load current (Figure 6), increasing to
2.5:1 at 200mA load (Figure 7). Alkaline cells perform poorly at high drain
rates because their internal impedance ranges from 200mΩ to 500mΩ,
causing a large voltage drop within the
L1*
22µH
2×
AA
CELL
+
SHUTDOWN
SELECT
SHDN
47µF
VIN
SW
SENSE
PGND
4.0
*COILCRAFT 1608-223 (800) 322-2645
S1300_8.eps
Figure 8. Lower-power applications can use smaller components. L1 is tallest
component at 3.1mm
2× E91
ALKALINE
EFFICIENCY (%)
2.5
BATTERY
1.5
VIN = 3V
84
VIN = 2.5V
82
VIN = 2V
80
78
1.0
76
0.5
0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
TIME (HOURS)
S1300_7.eps
Figure 7. Doubling load current to 200mA
causes E91 Alkaline battery life to drop by
2/3; L91 lithium battery shows 2.5:1
difference in operating life
Linear Technology Magazine • October 1993
4.0
2× E91
ALKALINE
3.5
2× L91
LITHIUM
3.0
2.5
2.0
1.5
1.0
BATTERY
0.5
0
74
0
OUTPUT
4.5
86
3.0
2.0
5.0
88
2× L91
LITHIUM
3.5
+
33µF
90
OUTPUT
D1
MBRS140T3
5V OUTPUT
50mA
LT1300
ILIM
GND
5.0
4.5
OUTPUT/BATTERY VOLTAGE (V)
cell. The alkaline cells feel quite warm
at 200mA load current, the result of
I2R losses inside the cells.
The reduced-power circuit shown
in Figure 8 can generate five volts at
currents up to 50mA. Here the ILIM pin
is grounded, reducing peak switch
current to 400mA. Lower profile components can be used in this circuit.
The capacitors are C-case size solid
tantalum and inductor L1 is the tallest
component at 3.2mm. The reduced
peak current also extends battery life,
since the I2R loss due to internal battery impedance is reduced. Figure 9
details efficiency versus load current
for several input voltages and Figure
10 shows battery life at a 50mA load.
Note that the L91 lithium battery lasts
only about 40% longer than the alkaline. The higher cost of the lithium
cells makes the alkaline cells more
OUTPUT/BATTERY VOLTAGE (V)
OUTPUT/BATTERY VOLTAGE (V)
4.5
1
10
LOAD CURRENT (mA)
100
S1300_9.eps
Figure 9. Efficiency of Figure 8’s circuit
0
2
4
6
8 10 12 14 16 18 20 22 24
TIME (HOURS)
S1300_10.eps
Figure 10. 50mA load and reduced
switch current are kind to E91 AA
alkaline battery; the advantages of
L91 lithium are not as evident
13
DESIGN FEATURES
A 4-Cell Application
A 4-cell pack is a convenient,
popular battery size. Alkaline cells
are sold in 4-packs at retail stores
and four cells usually provide
sufficient energy to keep battery replacement frequency reasonable.
Generating 5V from four cells, however, is a bit tricky. A fresh 4-cell
pack has a terminal voltage of 6.4V,
but at the end of its life, the pack’s
terminal voltage is around 3.2V;
hence, the DC/DC converter must
step the voltage either up or down,
depending on the state of the batter-
ies. A flyback topology with a costly,
custom-designed transformer could
be employed, but Figure 11’s circuit
gets around these problems by using
a flying-capacitor scheme along with
a second inductor. The circuit also
isolates the input from the output,
allowing the output to go to zero volts
during shutdown. The circuit can be
divided conceptually into boost and
buck sections. L1 and the LT1300
switch comprise the boost or step-up
section, and L2, D1, and C3 comprise
the buck or step-down section. C2 is
charged to VIN and acts as a level shift
between the two sections. The switch
node toggles between ground and VIN
+ VOUT, and the L2–C2–diode node
toggles between −VIN and VOUT + VD.
Figure 12 shows efficiency versus
load current for the circuit. All four
L1*
27µH
energy-storage elements must handle
power, which accounts for the lower
efficiency of this circuit compared to
the simpler boost circuit in Figure 2.
Efficiency is directly related to the
ESR and DCR of the capacitors and
inductors used. Better capacitors cost
more money. Better inductors do not
necessarily cost more, but they do
take up more space. Worst-case RMS
current through C2 occurs at minimum input voltage and measures
0.4A at full load with a 3V input. C2’s
specified maximum RMS current
must be greater than this worst-case
current. The Sanyo capacitors noted
specify a maximum ESR of 45mΩ
with a maximum ripple current rating of 2.1A. The Gowanda inductors
specify a maximum DCR of 58mΩ.
C2**
100µF
+
cost-effective in this application. A pair
of Eveready AAA alkaline cells (type
E92) lasts 96.6 hours with 5mA load,
very close to the rated capacity of the
battery.
NC
4×
AA
CELLS
+
5V/3.3V
C1**
100µF
SHUTDOWN
ILIM
SELECT
L2*
27µH
VIN
SW
1N5817
LT1300
SHDN
GND
SENSE
PGND
C3**
100µF
5V OR
3.3V
220mA
+
*L1, L2 = GOWANDA GA20-272K (716) 532-2234
**C1, C2, C3 = SANYO OS-CON 16SA100M (619) 661-6835
S1300_11.eps
Figure 11. 4-Cell to 3.3V or 5V converter output goes to zero when in shutdown.
Inductors may have, but do not require coupling; a transformer or two separate
units can be used
84
80
EFFICIENCY (%)
L1*
22µH
3.3V
OR 5V
INPUT
82
78
76
VIN = 3V
74
72
+
VIN = 4V
SHUTDOWN
100µF
70
VIN = 5V
68
VIN
SW
1N5817
LT1301
ILIM
GND
VIN = 6V
66
SELECT
SHDN
SENSE
PGND
12V OUTPUT
+
47µF
64
1
10
100
LOAD CURRENT (mA)
S1300_12.eps
Figure 12. Efficiency of up-down
converter in Figure 11
14
*L1 = SUMIDA
CD75-220K
(708) 956-0666
S1300_13.eps
Figure 13. LT1301 Delivers 12V from 3.3V or 5V input
Linear Technology Magazine • October 1993
DESIGN FEATURES
LT1301 Outputs
Five or Twelve Volts
Conclusion
90
The new LT1300 and LT1301 are
full-function, micropower step-up
DC/DC converters optimized for
battery-powered operation. The converters have been optimized for a
complete surface mount solution with
high efficiency, low quiescent current, and a low parts count.
88
VIN = 5V
86
EFFICIENCY (%)
The LT1301 is identical to the
LT1300 in every way except output
voltage. The LT1301 can be set to a
5V or 12V output via its SELECT pin.
Figure 13 shows a simple 3.3V or 5V
to 12V step-up converter. It can generate 120mA at 12V from either 3.3V
or 5V inputs, enabling the circuit to
provide V PP on a PCMCIA card
socket. Figure 14 shows the circuit’s
efficiency. Switch voltage drop is a
smaller percentage of input voltage
at 5V than at 3.3V, resulting in the
higher efficiency at 5V input.
84
VIN = 3.3V
82
80
78
76
74
10
LOAD CURRENT (mA)
1
100
S1300_14.eps
Figure 14. Efficiency of Figure 13’s circuit
LTC1257, continued from page 4
When the LTC1296 receives a shutdown command, the SSO pin of the
ADC goes high, the PNP turns off, and
the system shuts off.
Figure 5 shows how an LT1021-10
can be used to override the LTC1257’s
internal reference. The supply voltage
for both the reference and the DAC is
set to 15V, and the full-scale voltage
for the DAC becomes 10V.
The circuit in Figure 6 is a 12-bit,
single 5V-supply temperature-control
system with shutdown. An external
temperature is monitored by a J-type
thermocouple. The LT1025A provides
the cold junction compensation for the
thermocouple and the LTC1050 chopper op amp provides signal gain. The
47kΩ, 1µF capacitor filters the chopping noise before the signal is sent to
the A/D converter. The LTC1297 A/D
converter uses the reference of the
LTC1257 after it has been filtered to
set full scale. After the A/D measurement is taken, CS is pulled high and
everything except the LTC1257 is powered down, reducing the system supply
current to about 350µA. A word can
then be written to the LTC1257 and its
output can be used to as a temperature-control signal for the system being
monitored.
Conclusion
The LTC1257 is one of the smallest,
easiest-to-use 12-bit DACs available
today. With its cascadable 3-wire serial
interface, built-in reference and voltage
buffer, single supply operation, low supply current, and small package size, the
chip is a natural choice to reduce system complexity and cost.
5V
100k
10k
10µF
2N3906
VCC
–IN
CS
DOUT
0.1µF
15V
–
+
IN
LT1021-10
GND
+IN
VIN
J
GND
CLK
COMM
VREF
47k
0.1µF
DATA
µP
LTC1297
ADC
LT1025A
CS/
POWER DOWN
CLK
DAC LOAD
GND
OUT
VCC
CONTROL
OUTPUT
VOUT
VREF
DIN
LTC1257
CLK
+
1µF
1µF
1µF
100k
LTC1050
µP
–
LOAD
74k
VREF
DOUT
VCC
DIN
GND
1k
1257_5.eps
CONTROL
OUTPUT
VOUT
CLK
LTC1257
DAC
LOAD
DOUT
GND
1257_6.eps
Figure 5. DAC with external reference
Linear Technology Magazine • October 1993
Figure 6. 12-bit single 5V control system with shutdown
15
DESIGN FEATURES
LT1204, continued from page 7
Table 2. LT1204 Performance
Switchable Gain Amplifier
An example of the flexibility of the
LT1204 can be seen in a switchable
gain amplifier (Figure 7), which can
use either the shutdown or disable
feature. This circuit maintains a relatively constant output voltage of 1VPP
±330mV over a 128-to-1 change in
input level. When LT1204 #1 is selected, an input attenuator alters the
VIN
1
3
5
7
13
499Ω
249Ω
+
+
+
+
–
LT1204
#1
Parameter
Conditions
Bandwidth
0.1dB gain flatness
Slew rate
Differential gain
Differential phase
Channel select time
Enable time
Disable time
Input voltage range
Input offset voltage
Output swing
Supply current
Supply current in shutdown
Output limit current
AV = +2, RL = 150Ω
AV = +2, RL = 150Ω, no peaking
AV = +10
AV = +2, RL = 150Ω, NTSC
AV = +2, RL = 150Ω, NTSC
AV = +10, VIN = 0.5V
AV = +10, VIN = 0.5V
AV = +10, VIN = 0.5V
Pin 8 = 0V, VS ≥ ±10V
RL = 400Ω
Pin 12 = 0V
VOUT = 1VPP
1
3
5
7
13
124Ω
+
+
+
+
–
70MHz
>30MHz
1000V/µs
0.04%
0.06°
120ns
100ns
50ns
± 6V
5mV
13.5V
19mA
1.5mA
55mA
1.5k
100
124Ω
Typical Value
LT1204
#2
1.5k
1204_7.eps
Figure 7. Switchable Gain amplifier
AV = 1, 2, 4, 8
Conclusions
input signal by 1, 0.5, 0.25, or 0.125
to form an amplifier with a gain of 16,
8, 4, or 2. LT1204 #2 is connected to
the same attenuator, and when it is
enabled (LT1204 #1 disabled), it has a
gain of +1 instead of +16. The second
LT1204 is used to extend the gain
range to 1, 0.5, 0.25, and 0.125.
Performance
Table 2 summarizes the major performance specifications of the LT1204.
The LT1204 combines a
fast-switching multiplexer with a highspeed current-feedback amplifier for
gain adjustment and cable driving.
Switching transients have been greatly
reduced in amplitude and duration, all
hostile and disable crosstalk have been
reduced below –90dB at 10MHz, and
the unique disable feature eases system expansion. The high performance
of this multiplexer makes it ideal for
the newest multimedia products.
LT1363, continued from page 11
LT1354
LT1354
LT1357
LT1357
LT1360
LT1360
LT1363
LT1363
1363_5.eps
1363_4.eps
Figure 4. Small signal step response, CLoad = 200pF:
LT1354, LT1357, LT1360, LT1363
ensuring stability. Figure 4 shows the
family of amplifiers driving a 200pF load
and Figure 5 shows the large signal response with a 10,000pF load. Note that
the slew rate when driving a large capacitive load is limited by the short-circuit
current limit of the amplifier.
16
Figure 5. Large signal step response, CLoad = 10,000pF:
LT1354, LT1357, LT1360, LT1363
Conclusion
The new topology offers significant
improvement in circuit performance.
It achieves higher slew rates with lower
supply currents and improves DC
specifications and noise with higher
input-stage transconductance, lower
operating currents, and balanced input stages. The family of parts is
also stable with capacitive loads and
can be used in any voltage feedback
application.
Linear Technology Magazine • October 1993
DESIGN IDEAS
The LT1300: Two-Cellsto-Real-World Interface
by Dale Eagar
Introduction
The LT1300 micropower, high-speed
step-up DC/DC converter opens
up many new applications to the
user, such as those requiring high
efficiency in battery-operated equipment. The LT1300 can be used to
produce high voltages for many spe-
DESIGN IDEAS...
The LT1300: Two-Cellsto-Real-World Interface . 17
Dale Eagar
Using a Fast Analog
Multiplexer to Switch Video
Signals for NTSC
“Picture-in-Picture”
Displays ........................ 21
Frank Cox
High-Efficiency (>90%)
NiCad Battery-Charger
Circuit Programmable for
1.3A Fast Charge or 100mA
Trickle Charge .............. 23
Brian Huffman
cialized tasks with high efficiency. Here
are three such applications. In the first
application, the LT1300 is used to
produce 325VDC while drawing a mere
200 microamps from two C-size cells.
D1
MUR1100
LTC1163: 2-Cell
Power Management ....... 25
INTR
VN2222
PULSE
1300a_2.eps
Figure 2. FET inverter/microprocessor
interface
T1
R2
1k
D3
1N758A
3V
+ C3
Q1
ZTX788
3V
100µF
6.3V
3V
Randy G. Flatness
Philip Karantzalis
10k
D5
1N5718
D2
1N4148
A High-Efficiency, 5V to
3.3V/5A Converter ......... 26
+VLOGIC
C1
47pF
V1
R2868
Tim Skovmand
A Linear-Phase
Bandpass Filter for Digital
Communications ........... 28
The LT1300 and transformer T1
form a flyback converter to step up the
voltage from 3V to 325V. The secondary winding of T1 connects through
D1 (an MUR1100) to C1, a holding
capacitor for the 325VDC, which, in
turn, is applied to the anode of the
photoelectric sensor tube V1. The
LT1300 SW pin senses the voltage on C1, as scaled by the turns
ratio, through T1. The voltage on the
primary winding is programmed to be
10.6V, translating to 325V on C1.
When C1 has charged to 325V, the
An interesting characteristic of flame
is that it emits short-wavelength ultraviolet light (<260nm). This
short-wavelength light falls into a window of the light spectrum that is
relatively empty. Tungsten light, fluorescent light, and sunlight below the
atmosphere are almost totally devoid
of spectral energy in this window. The
circuit shown in Figure 1 uses
a photoelectric sensor with a sufficiently high cathode work function to
make it blind to anything with a wavelength longer than 260nm (such as
normal UV, visible light, or infrared.)
Cathode work function is a measure
of how hard it is to free an electron
from an atom; when related to light
illuminating a cathode, it specifies the
minimum energy of a photon that can
Mitchell Lee
Jon A. Dutra
Theory of Operation
Flame Sensor
An LTC1087-Based
1.2V GTL Terminator ..... 24
A Dual-Output LCD-Bias
Voltage Generator ......... 27
liberate an electron. UV photons
have higher energy than visible light.
6
D4
1N5718
VIN
3
PULSE
R1
100k
SHUTDOWN
U1
LT1300
C2
0.01
NC
SW
7
5 I
LIM
SEL
V1 = HAMAMATSU R2868
FLAME SENSOR
2
HAMAMATSU (408) 261-2022
T1 = COILTRONICS CTX02-12186
COILTRONICS (407) 241-7876
Q1 = ZETEX ZTX788
ZETEX (516) 543-7100
C1 = 47pF > 500V
SENSE
4
GND PGND
1
8
C4
0.47µF
1300a_1.eps
Figure 1. Flame detector
Linear Technology Magazine • October 1993
17
DESIGN IDEAS
3V
C5
0.01
R7
3M
R21
5.1Ω C8
0.1
3V
R8
100k
POT 1
SENSITIVITY
ADJ
1M
4
R9
30k
U1
11
3V
C4
0.1
R1
300k
R2
300k
R4
10k
U1B
LT1179
+
C2
1µF
U1A
LT1179
C3
0.15
R10
1M
D1
1N4148
+
R3
300k
PULSE
C1
0.68
FLAME
+ C6
–
1µF
–
R6
1M
SMOOTHED
R5
10M
3V
ALARM
–
R11
560k
+
R15
750k
R14
1.2M
R12
390k
3V
R16
510k
C7
4.7µF
3V
R22
180Ω
A1
LED1
D2
1N4148
R20
10k
U1D
LT1179
R17
390k
3V
ALARM
DEVICE
R13
560k
+
–
3V
U1C
LT1179
Q1
2N2222
3V
R18
470k
U2
LT1004-1.2 R19
100k
D3
1N4148
REPLACE
U1 = LT1179
A1 = ARCHER 273-065A
1300a_3.eps
Figure 3. Discriminator circuitry
feedback loop comprised of D3, R2,
and Q1 kicks in and charges C4
through D4. When the voltage at C4
exceeds 3.3V, the LT1300 goes into
its wait mode. In wait mode the
LT1300 consumes only 100µA of current. The LT1300 stays in wait mode
until the voltage on C4 falls below
3.3V, at which time the LT1300 turns
on to burst recharge both C1 and C4.
Burst ModeTM operation ensures 30Hz
oscillation in this system. This rate is
determined by the value of C4, the
internal sense resistance to ground
in the LT1300 (approximately 1 MΩ),
and the amount of overcharge C4
gets when charging.) D5 is a Schottky
catch diode to keep reverse current
out of U1.
When illuminated with a photon of
sufficient energy, the photoelectric
tube’s cathode liberates an electron.
18
The tube V1 has 325V across its terminals to get sufficient energy into
a liberated photo electron to ionize
the gas that fills the tube. Once the
gas in the tube ionizes, there are more
electrons available; they cause a chain
reaction in the tube that causes the
tube to avalanche. When the tube
avalanches, most of the charge on C1
is transferred to C2 and the voltage
across C1 drops to a fraction of its
original 325V. When C2 has charged
to 3.6V, all the excess charge residing
in C1 gets dumped through D2 into
the battery. The voltage across C2 is
the output signal called PULSE. PULSE
asserts the shutdown pin of the
LT1300, allowing the plasma in the
photoelectric tube to quench. Figure 2
shows an interface circuit that enables the PULSE signal to interrupt a
microprocessor.
For you analog purists, Figure 3
shows a discriminator circuit with
low battery detect for a complete 3V
flame alarm. The discriminator is
needed because the photo detector
occasionally detects a cosmic ray
or some rare room-light photon.
The discriminator consists of four
sections.
Filter Section
(U1a, R1–R6, C1–C4)
In this section, the PULSE signal is
filtered with a third-order Gaussian
lowpass filter. Each PULSE is converted into a Gaussian-shaped pulse
about 300ms wide. These Gaussianshaped pulses are superimposed and
can be seen on the SMOOTHED test
point. The net effect of this filtering is
to accumulate PULSEs that are close
together (<300ms) into a single DC
Linear Technology Magazine • October 1993
DESIGN IDEAS
voltage. When the input PULSE happens frequently, the SMOOTHED test
point goes positive, indicating many
recently detected photons.
Threshold-Setting Section
(U1b, POT1, R1–R9, C5)
POT1 sets the threshold of the
alarm; U1b compares the threshold
with the SMOOTHED signal from the
previous section and pulls its output
low when it detects more voltage on
SMOOTHED than on the threshold.
This output is the test point labelled
FLAME.
Pulse-Stretcher Section:
(D1, R10–R13, C6, U1c)
When FLAME goes low, C6 discharges through D1 and is slowly
recharged through R10. U1c is a
comparator with hysteresis that outputs a high signal (test point ALARM)
when FLAME goes low. Because of the
time constant of R10 and C6, ALARM
stays high for about one second after
FLAME goes high. The ALARM output
also drives the base of Q1 through D2
to sound the sounding device A1.
Infinite-Input-Impedance
Voltage Buffer
In the flame detector circuit (Figure 1), it is difficult to measure the
voltage across C1 because almost any
load invalidates the meter reading.
This next application for the LT1300
is a voltage buffer that overcomes
this measurement problem. This is a
four-terminal, unity-gain buffer, as
shown functionally in Figure 4. The
input impedance is essentially infinite, the input bias current is negligible,
Linear Technology Magazine • October 1993
Theory of Operation
U1 monitors the voltage difference
between the circuit’s noninverting input and output and attempts
to make it zero. If the voltage on the
noninverting input is less than the
voltage on the noninverting output,
U1’s output goes positive, turning Q1
on slightly. Q1 acts as a current
source, discharging C3. When the
+OUT
+
–
E
ISOLATION
+IN
+
–OUT
–IN
1300a_4.eps
Figure 4. Voltage buffer block diagram
C7
1000pF
R10
100M
D3
MUR1110
–IN
Low-Battery-Detect Circuit
(R14–R19, C7, D3, U1d)
This sub-circuit is a divider/comparator/oscillator that is activated
when the battery voltage drops to 2V.
The output is a positive-going 1/4second signal called REPLACE.
REPLACE initially occurs every three
seconds at a battery voltage of 2V, but
the frequency of repetition increases
as the battery voltage drops. At a
battery voltage of 1.5V, the REPLACE
signal frequency is approximately 2Hz.
REPLACE also sounds the sounding
device A1 through R20 and Q1. R21
and C8 for m a trash compactor to decouple U1.
The complete circuit, including the
detector (Figure 1) and the discriminator (Figure 3) consume a mere
300µA of supply current from 3V. The
circuit lends itself very well to both
battery operation and two-wire remote operation. Battery life is more
than two years when powered by two
C-size alkaline cells. In full sun, the
and the input offset voltage is less
than 0.05 volts. The output voltage
tracks the input voltage from 0V to
520V. For safety (and to isolate the
input capacitance) a 100MΩ resistor
is placed in series with the input,
but with the ±570pA of input bias
current (over temperature) for the
LT1097, this translates into only
±57mV of additional offset. The input
impedance of this buffer measures
four trillion ohms when measured
with a 100-to-400 volt input. The
detailed circuit is shown in Figure 5.
detector can easily “see” a cigarette
lighter 30 feet away and still discriminate it from the sun.
T1
SW1
3V
+
3V
1.5V
C6
100µF
6.3V
6
VIN
NC
5
ILIM
SW
7
7
Z1
15V
U2
LT1300
U1
LT1097
4
SHUTDOWN
SENSE
4
2
C2
0.01
R3
1M
1
3
R4
100k
C4
220pF
D1
1N4148
D2
1N4148
8
R8
1M
+ OUT
3V
C1
1000pF
R1
100M
+ IN
+
C5
100µF
6.3V
R11
C8
20Ω
0.1
1.5V
3V
R9
5.1Ω
–OUT
2
R2
1M
3
–
U1
LT1097
6
R7
1k
R5
10k
C3
0.01
Q1
2N3904
+
R6
10k
T1 = COILTRONICS CTX02-12179
COILTRONICS (407) 241-7876
R1, R10 = VICTOREEN SLIM-MOX100
VICTOREEN (216) 248-9300
1300a_5.eps
Figure 5. Voltage buffer schematic
19
DESIGN IDEAS
C2
15pF
3V
T1
C1
100µF
6.3V
+
CCFL
6
VIN
PWM
IN
3
SW
7
SHUTDOWN
D1
1N5718
U1
LT1300
NC
5 I
LIM
SENSE
SEL
GND PGND
2
1
4
8
CCFL LAMP = JKL BF650-20B
JKL (800) 897-3056
T1 = COILTRONICS CTX02-12189
COILTRONICS (407) 241-7876
C2 = 15pF, 500V
1300a_6.eps
Figure 6. CCFL driver
voltage on C3 falls below approximately 0.6V, U2 is enabled. When it is
enabled, U2 turns its switch on (U2
pin 7 pulls low, to near 0V). This
causes approximately 3V to be imposed across the primary winding of
T1. The magnetizing inductance of
the primary winding of T1, across
which a voltage is applied, requires a
steadily increasing current. At the
same time, C4 is charging through
D2. When the current flowing through
the switch of the LT1300 reaches 1
amp, the LT1300 switches off. The
magnetizing inductance of the primary winding of T1, seeing that the
LT1300 is attempting to discontinue
current flow, takes over by swinging
positive in voltage until it finds something that will take the 1 amp of
magnetizing current. While the primary winding is finding somewhere
to put the magnetizing current, the
secondary winding takes it upon itself to do the same, but due to its
turns ratio with the primary winding,
it moves 100 times faster and 100
times as far as the primary winding.
T1’s secondary dumps a significant
portion of the magnetizing energy
into C7 via D3, thus forming a flyback
inverter.
Z1 dissipates the energy stored in
T1’s inductance. During the flyback
time, C4 charges C3 through D1.
The voltage across C3 exceeds
0.6V, shutting down U2. U2 stays
shut down until Q1 discharges C3 to
restart the sequence.
When the +output voltage is more
positive than the +input voltage, the
output of U1 goes low, Q1 stays off, R8
keeps C3 charged to more than 0.6V,
and U2 stays shut down. The parallel
combination of R10 and the load resistance (e.g., 10MΩ in a handheld
voltmeter) discharges C7 and the
+output and the +input voltages are
again equal. The current output of
this circuit is limited to a safe value
(1mA at 50V, 0.1mA at 500V), even
when the +input is attached to +500V.
We do not recommend increasing
the value of C7, because at higher
voltages it may become a shock
hazard. Battery life 40 hours for a
pair of AA alkaline batteries driving
10MΩ at 500V.
Cold-Cathode
Florescent Lamp Driver
CCFLs seem to be the latest craze;
they offer high brightness, long life,
small size, and produce white light.
Figure 6 shows a CCFL driver circuit.
Theory of Operation
This is a forward/flyback inverter
optimized for minimum parts count.
When enabled, U1 charges the primary winding of T1 to 1 amp, and lets
go. T1 then flies back, exciting many
hundreds of volts across its secondary winding, which, in turn, ionizes
the CCFL. Because the initial current
through the CCFL is only in one direction, C2 takes on a DC potential. As
the circuit runs, the voltage across C2
stabilizes at about 100VDC.
Additionally, C2 removes the DC component from the tube current,
extending tube life. The nonlinear V/I
characteristic of the CCFL, in conjunction with C2, forces the converter
to run in both forward and flyback
modes simultaneously. The light intensity can be pulse-width modulated
LIGHT LEVEL
PROGRAM
3V
R7
100M
D1
1N4148
D1
2N3904
R5
1k
ON
SUNSET
OFF
C2
470µF
6
+
5
–
U1B
LT1178
R6
1M
7
+
PULSE
WIDTH
OUT
C3
0.01
3V
R7
10Ω
C4
0.1
SAWTOOTH
C1
0.022
8
R1
300k
2
LT1178
3
4
–
U1A
LT1178
1
+
R3
270k
R2
270k
3V
R4
200k
1300a_7.eps
Figure 7. Electronic light stick controller
by modulating the shutdown pin.
When the shutdown pin is pulled
high, the LT1300 goes into its shutdown mode, where it draws only 10µA
of input current.
Electronic Light Stick
Camping in November with my
kids has its own unique problems,
even if we aren’t camping in six feet of
snow. Although we had the usual
light sources, something was missing, namely a light that simulates the
natural sunset at bedtime to wind the
kids down for the night. The circuit in
Figures 6 and 7 (see explanation
below) details a high efficiency fluorescent lantern with a built in sunset
feature.
The function of the circuit is as
follows:
● To turn on: switch SW1 into the
ON position.
● To turn off fast: switch SW1 into
the OFF position.
● To simulate sunset: 1. turn light
ON. 2. switch SW1 into the
SUNSET position.
This application uses the circuitry
of both Figure 6 and Figure 7. The
pulse-width output of Figure 7 drives
the pulse-width input of Figure 6.
continued on page 24
20
Linear Technology Magazine • October 1993
DESIGN IDEAS
Using a Fast Analog Multiplexer
to Switch Video Signals for NTSC
“Picture-in-Picture” Displays
by Frank Cox
Introduction
The majority of production1 video
switching consists of selecting one
video source out of many for signal
routing or scene editing. For these
purposes, the video signal is switched
during the vertical interval in order to
reduce visual switching transients.
The image is blanked during this time,
so if the horizontal and vertical synchronization and subcarrier lock are
maintained, there will be no visible
artifacts. Although vertical-interval
switching is adequate for most routing functions, there are times when it
is desirable to switch two synchronous video signals during the active
(visible) portion of the line to obtain
picture-in-picture, key, or overlay effects. Picture-in-picture or active video
switching requires signal-to-signal
transitions that are both clean and
fast. A clean transition should have a
minimum of pre-shoot, over-shoot,
ringing, or other aberrations commonly lumped under the term
“glitching.”
Using the LT1204
A quality, high-speed multiplexer
amplifier can be used with good
results for active video switching. The
important specifications for this application are small, controlled
switching glitch, good switching speed,
low distortion, good dynamic range,
wide bandwidth, low path loss, low
channel-to-channel crosstalk, and
good channel-to-channel offset matching. The LT1204 specifications match
these requirements quite well, especially in the areas of bandwidth,
distortion, and channel-to-channel
crosstalk (which is an outstanding
90dB at 10MHz). The LT1204 was
evaluated for use in active video
switching with the test setup shown
in Figure 1. Figure 2 shows the video
waveform of a switch between a 50%
white level and a 0% white level about
30% into the active interval and back
again at about 60% of the active
interval. The switch artifact is brief
and well controlled. Figure 3 is an
expanded view of the same waveform.
When viewed on a monitor, the switch
artifact is just visible as a very fine
line. The lower trace is a switch between two black level (0V) video signals
showing a very slight channel-tochannel offset, which is not visible on
Figure 2. Video waveform switched from 50%
white level to 0% white level and back
Linear Technology Magazine • October 1993
OFF AIR VIDEO
SOURCE OR VIDEO
PATTERN
GENERATOR
SYNC STRIPPER,
SAMPLE PULSE
GENERATOR
SAMPLE PULSE
SCOPE
50%
75Ω
75
75
LT1204
75
OUTPUT
MONITOR
LOOP
THROUGH
75
INPUTS
COX_1.eps
Figure 1. “Picture-in-picture” test setup
the monitor. Switching between two
DC levels is a worst-case test, as
almost any active video will have
enough variation to totally obscure
this small switch artifact.
Figure 3. Expaneded view of rising edge of
LT1204 switching from 0% to 50% (50ns
horizontal division)
21
DESIGN IDEAS
50ns/DIV
COX_p3.eps
Figure 4. Expanded view of “brand-x” switch
0%–50% transition
Video-Switching Caveats
In a video processing system that
has a large bandwidth compared to
the bandwidth of the video signal, a
fast transition from one video level to
another (with a low-amplitude glitch)
will cause minimal visual disturbance.
This situation is analogous to the
proper use of an analog oscilloscope.
In order to make accurate measurement of pulse waveforms, the
instrument must have much more
bandwidth than the signal in question (usually five times the highest
frequency). Not only should the glitch
be small, it should be otherwise well
controlled. A switching glitch that
has a long settling “tail” can be more
troublesome (that is, more visible)
than one that has more amplitude
but decays quickly. The LT1204 has
a switching glitch that is not only low
in amplitude but well controlled and
quickly damped. Refer to Figure 4,
which shows a video multiplexer that
has a long, slow-settling tail. This sort
of distortion is highly visible on a
video monitor.
Composite video systems, such as
NTSC, are inherently band-limited
and thus edge-rate limited. In a
sharply band-limited system, the introduction of signals that contain
significant energy higher in frequency
than the filter cutoff will cause distortion of transient waveforms (see Figure
5). Filters used to control the bandwidth of these video systems should
be group-delay equalized to minimize
this pulse distortion. Additionally, in
a band-limited system, the edge rates
of switching glitches or level-to-level
transitions should be controlled to
prevent ringing and other pulse aberrations that could be visible. In
practice, this is usually accomplished
with pulse-shaping networks (Bessel
filters are one example). Pulse-shaping networks and delay-equalized
filters add cost and complexity to
video systems and are usually found
continued on page 29
1
fC
Some Definitions—
“Picture in picture” refers to the
production effect in which one video
image is inserted within the boundaries of another. The process may
be as simple as splitting the screen
down the middle or it may involve
switching the two images along a
complicated geometric boundary.
In order to make the composite
picture stable and viewable, both video
signals must be in horizontal and
vertical sync. For composite
color signals, the signals must also be
in subcarrier lock.
“Keying” is the process of switching
among of two or more video signals,
triggering on some characteristic of
one of the signals. For instance, a
chroma keyer will switch on the
presence of a particular color.
Chroma keyers are used to insert a
portion of one scene into another.
In a commonly used effect, the TV
weather person (the “talent”) appears to be standing in front of a
computer generated weather map.
Actually, the talent is standing in
front of a specially colored background; the weather map is a
separate video signal, which has
been carefully prepared to contain
none of that particular color. When
the chroma keyer senses the keying
color, it switches to the weather
map background. Where there is
no keying color, the keyer switches
to the talent’s image.
RISE TIME ≈
DELAY ≈
1
2fC
N
(WHERE N IS ORDER OF FILTER)
2fC
COX_5.eps
Figure 5. Pulse response of an ideal sharp-cutoff filter at frequency fC
22
Linear Technology Magazine • October 1993
DESIGN IDEAS
High-Efficiency (>90%) NiCad BatteryCharger Circuit Programmable for 1.3A
Fast Charge or 100mA Trickle Charge
by Brian Huffman
Battery-charger circuits are of
universal interest to laptop,
notebook, and palmtop computer
manufacturers. High efficiency is desirable in these applications to
minimize the power dissipated in the
surface mount components. The circuit shown in Figure 1 is designed to
charge four NiCad cells at a 1.3A fast
charge or a 100mA trickle charge,
with efficiency exceeding 90%. This
circuit can be modified easily to handle
up to eight NiCad cells.
The circuit uses an LTC1148 in a
step-down configuration to control
the charge rate. The LTC1148 is a
synchronous switching-regulator
controller that drives external,
complementary power MOSFETs using a constant off-time current mode
architecture. When the LTC1148’s
P-drive output pulls the gate of Q1
low, the P-channel MOSFET turns on
and connects one side of the inductor
to the input voltage. During this period, current flows from the input
through Q1, through the inductor,
and into the battery. When the
LTC1148 P-drive pin goes high, Q1 is
turned off, and the voltage on the
drain of Q1 drops until the clamp
diode is for-ward biased. The diode
conducts for a very short period of
time, until the LTC1148 internal circuitry senses that the P-channel is
fully off, pre-venting the simultaneous
conduction of Q1 and Q2. Then the Ndrive out-put goes high, turning on
Q2, which shorts out D1. Now the
inductor current flows through the Nchannel MOSFET instead of through
the diode, increasing efficiency. This
type of switching architecture is known
as synchronous rectification.
During the fast-charge interval,
the resistor divider network (R4 and
R5) forces the LTC1148’s feedback
pin (VFB) below 1.25V, causing the
LTC1148 to operate at the maximum
output current. R3, a 100mΩ resistor, senses the current and sets it at
approximately 1.3A, according to
equation 1. When the batteries are
disconnected, the error amplifier
forces the feedback pin to 1.25V,
limiting the output voltage to 8.1V.
Diode D2 prevents the batteries from
discharging through the divider network when the charger is shut down.
In shutdown mode, the circuit draws
less than 20µA from the input supply.
The dual-rate charging is controlled
by Q3, which can be toggled between
fast charge and trickle charge. The
trickle-charge rate is set by resistor
R1. Figure 2 is a graph showing the
continued on page 24
+
VIN
8V - 15V
C1
1µF
C2
0.1µF
3
VIN
PDRIVE
0V = NORMAL
>1.5V = SHUTDOWN
10
6
R1
51
“1”
TRICKLE
CHARGE
Q3
VN2222LL
R2
1k
C4
3300pF
X7R
4
C5
200pF
NPO
SHUTDOWN
ITH
SENSE +
1
NDRIVE
SGND
11
PGND
12
1
4
D2
MBRS340T3
2
R3
3 100mΩ
VOUT
8
7
SENSE –
LTC1148
9
VFB
CT
Q1
Si9430DY
C3
22µF x2
25V
L1
50µH
C6
0.01µF
R4
274k
1%
+
14
C7
100pF
Q2
Si9410DY
C1 = (TA)
C3 = AVX (TA) TPSD226K025R0200 ESR = 0.200 IRMS = 0.775A
C8 = AVX (TA) TPSE227M010R0100 ESR = 0.100 IRMS = 1.149A
Q1 = SILICONIX PMOS BVDSS = 20V RDSON = 0.125 CRSS = 400pF QG = 25nC θJA = 50°C/W
Q2 = SILICONIX NMOS BVDSS = 30V RDSON = 0.050 CRSS = 160pF QG = 50nC θJA = 50°C/W
D1, D2 = MOTOROLA SCHOTTKY VBR =40V
R3 = KRL SP-1/2-A1-0R100J Pd = 0.75V
L1 = COILTRONICS CTX50-4 DCR = 0.175 IDC = 1.350A KOOL Mµ CORE
D1
MBRS140T3
R5
49.9k
1%
VBATT
4 CELLS
C8
220µF
10V
Brian_1.eps
VOUT = 1.25V • (1 + R4/R5) = 8.1V
FAST CHARGE = 130mV/R3 = 1.3A (EQ. 1)
TRICKLE CHARGE = 100mA (SEE FIGURE 2)
ALL OTHER CAPACITORS ARE CERAMIC
COILTRONICS (407) 241-7876
KRL (809) 668-3210
Figure 1. Schematic diagram: 4 cell, 1.3 amp battery charger implemented in surface mount technology
Linear Technology Magazine • October 1993
23
DESIGN IDEAS
An LT1087-Based 1.2V GTL Terminator
by Mitchell Lee
A recent development in high-speed
digital design has resulted in a new
family of logic chips called Gunning
Transition Logic (GTL). Because of the
speeds involved, careful attention must
be paid to the transmission-line
characteristics of the interconnections between these chips; active
termination is required.
The termination voltage is 1.20V,
and currents of several amperes are
common in a complete system. One
method of generating 1.2V is to use a
linear regulator operating from 3.3V
or 5V. Unfortunately, this method
suffers from two major drawbacks.
First, the minimum adjust voltage,
without the aid of a negative supply,
is 1.25V for most adjustable linear
regulators. Second, most low-voltage
linear regulators do not feature low
dropout characteristics, rendering
them unusable on a 3.3V input. The
LT1087 solves both of these problems
with an output that can be adjusted
to less than the reference voltage and
a low-dropout architecture.
Figure 1 shows the complete circuit. The LT1087 features feedback
sense, which, in its original application, was used for remote Kelvin
sensing. In the GTL terminator circuit,
the sense pins are used to adjust the
internal 1.25V reference downward.
The result is a 1.20V, 5A regulator,
with 2% output tolerance over all conditions of line, load, and temperature.
To minimize power dissipation, a 3.3V
input source is recommended.
1.20V/5A
2%
+
VOUT
–SENSE
VIN = 3.3V
VIN
LT1087CT
+SENSE
GND
+
10µF
22µF
Ta
4.42Ω
1%
1.21kΩ
1%
GTL_1.eps
Figure 1. Schematic diagram: 1.2 volt GTL
termination voltage generator
NiCad, continued from page 23
1400
1200
1200
1000
1000
IOUT (mA)
1400
IOUT (mA)
value of R1 for a given trickle-charge
output current. The trickle-charge current can also be varied by using an op
amp to force the threshold pin voltage
within its 0V–2V range. Figure 3 shows
the output current as a function of
threshold pin voltage.
800
600
800
600
400
400
200
200
0
0
0
1
2
R1 (kΩ)
3
4
0
1.0
0.5
1.5
THRESHOLD PIN VOLTAGE (V)
2.0
Figure 2. LTC1148 output currentBrian_2.eps
voice
trickle charge set resistance (R1)
Figure 3. LTC1148 output current Brian_3.eps
versus
forced threshold pin voltage
is at or below 1 volt, the light is off.
D1 and R5 charge and hold C2 when
SW1 is in the ON position. R5 and
SW1 discharge and hold C2 when
SW1 is in the OFF position. The combination of D2, R6, and U1b discharge
C2 when SW1 is in the SUNSET position. The discharging of C2 when in
the SUNSET mode is doubly exponential, causing the tail end of the
simulated sunset to go very slowly (a
good idea, because kids have a logarithmic response to light). The first
exponential aspect of the SUNSET decay is implemented by R6 and C2,
which form an exponential RC time
constant. The second exponential aspect of the SUNSET decay is
implemented because R6 is driven by
U1b pin 7, whose duty factor is
changing,causing the off time to decrease exponentially as the light level
fades. The output of U1b is a pulsewidth-modulated level gating the light
driver on and off. The lamp is illuminated when U1b’s output is low. C3 is
a trash compactor and R7 and C4 form
a trash compactor to decouple U1
from the high-frequency ripple generated by the switcher.
LT1300, continued from page 20
Theory of Operation
U1a, R1–R4, and C1 form a
sawtooth oscillator for pulse-width
modulating the light (implementing
light levels less than 100%). U1b acts
as a comparator, comparing the
sawtooth output of the oscillator with
the programmed light level (as seen
on the + terminal of C2). C2 is the
holding cap that programs the light
level; when it is charged to 2.5V, the
light is on 100% of the time. As
the voltage on C2 drops below 2.5V,
the overall light level decreases,
because the light is being pulse-width
modulated. When the voltage on C2
24
Linear Technology Magazine • October 1993
DESIGN IDEAS
LTC1163: 2-Cell
Power Management
The LTC1163 1.8V to 6V high-side
MOSFET driver allows inexpensive
N-channel switches to be used to
efficiently manage power in 2-cell systems such as palmtop computers,
portable medical equipment, cellular
telephones and personal organizers.
Any supply voltage above 3V, such
as: 3.3V, 5V or 12V, can be generated
by step-up converters powered from a
2-cell supply. Step-up regulators are
typically configured as shown in
Figure 1. An inductor is connected
directly to the 2-cell battery pack
and switched by a large (1A) switch.
The inductor current is then passed
through a low-drop Schottky rectifier
to charge the output capacitor to a
voltage higher than the input voltage.
Unfortunately, when the regulator is
shut down, the inductor and diode
remain connected and the load may
leak significant current in standby.
One possible solution to this problem is to add a low RDS(ON) MOSFET
switch between the battery pack and
D1
L1
+
2-CELL
BATTERY
PACK
by Tim Skovmand
+
CIN
STEP-UP
SWITCHING
REGULATOR
SHUTDOWN
LOAD
+
COUT
1163_1.eps
the input of the regulator to completely disconnect it and the load
from the battery pack. MOSFET
switches, however, cannot operate
directly from 2-cell battery supplies because the gate voltage is
limited to 3V with fresh cells and 1.8V
when the cells are fully discharged.
The LTC1163 solves this problem
by generating gate drive voltages
which fully enhance high-side
N-channel switches when powered
from a 2-cell battery pack as shown
in Figure 2. The standby current
with all three drivers switched off
is typically 0.01µA. The quiescent
current rises to 85µA per channel
with the input turned on and the
charge pump producing 10V (above
ground) from a 3V supply. The
surface mount MOSFET switches
shown are guaranteed to be less
than 0.1Ω with VGS = 5V and less than
0.12Ω with VGS = 4V and therefore
have extremely low voltage drops.
Figure 1. Typical step-up converter topology
+
+ 2-CELL
BATTERY
PACK
IN1
CONTROL
LOGIC
OR µP
100µF
6.3V
RFD14N05LSM
VS
OUT1
LTC1163
IN2
RFD14N05LSM
OUT2
IN3
RFD14N05LSM
OUT3
GND
100µH
MBRS120T3
3.3V
1
22µH
22µH
MBRS120T3
MBRS120T3
12V
3
1
47
1
2
3
5V
3
LT1173CS8
7
LT1109CS8-12
4
7
8
+
10µF
20V
LT1109CS8-5
4
8
8
+
22µF
16V
39k
+
24k
4
5
220µF
6.3V
1163_2.eps
Figure 2 Complete 2-cell to 3.3V, 5V, and 12V power management system
Linear Technology Magazine • October 1993
25
DESIGN IDEAS
A High-Efficiency,
5V to 3.3V/5A Converter
The next generation of notebook
and desktop computers is incorporating more 3.3V ICs alongside 5V
devices. As the number of devices
increases, the current requirements
also increase. Typically, a highcurrent 5V supply is already available.
Thus, the problem is reduced to deriving 3.3V from 5V efficiently in a
small amount of board space.
High efficiency is mandatory in
these applications, since converting
5V to 3.3V at 5A using a linear regulator would require dissipating
over 8W. This wastes power and
board space for heat sinking.
The LTC1148 synchronous switching-regulator controller accomplishes
the 5V to 3.3V conversion with high
efficiencies over a wide load-current
range. The circuit shown in Figure 1
provides 3.3V at efficiencies greater
than 90% from 5mA to 5A (over
three decades of load current). The
efficiency of the circuit in Figure 1 is
plotted in Figure 2.
At an output current of 5A the
efficiency is 90%; this means only
1.8W are lost. This lost power is
distributed among RSENSE, L1, and
the power MOSFETs; thus heat sinking is not required.
The LTC1148 series of controllers
use constant off-time current-mode
architecture to provide clean startup, accurate current limit, and
excellent line and load regulation. To
maximize the operating efficiency at
low output currents, Burst ModeTM
operation is used to reduce switching
losses. Synchronous switching, combined with Burst ModeTM operation,
yields very efficient energy conversion over a wide range of load currents.
The top P-channel MOSFETs in
Figure 1 will be on 2/3 of the time
with an input of 5V. Hence, these
devices should be carefully examined
by Randy G. Flatness
to obtain the best performance. Two
MOSFETs are needed to handle the
peak currents safely and enhance
high-current efficiency. The LTC1148
can drive both MOSFETs adequately
without a problem. A single N-channel
MOSFET is used as the bottom
synchronous switch, which shunts
the Schottky diode. Finally, adaptive anti-shoot-though circuitry
automatically prevents cross conduction between the complementary
MOSFETs, which can kill efficiency.
The circuit in Figure 1 has a noload current of only 160µA. In
shutdown mode, with pin 10 held
high (above 2V), the quiescent current reduces to less than 20µA with
all MOSFETs held off DC. Although
the circuit in Figure 1 is specified
at a +5V input voltage, the circuit
will function from 4V to 15V
without requiring any component
substitutions.
VIN
5V
C1
1µF
+
C2
0.1µF
3
0V = NORMAL
>2V = SHUTDOWN
10
4
R1
470Ω
C4
3300pF
C5
680pF
NPO
Q1
Si9430DY
SHUTDOWN
LTC1148-3.3
8
SENSE +
6
Q2
Si9430DY
VIN
1
PDRIVE
SENSE –
ITH
CT
SGND
11
7
L1
27µH
C3
100µF
20V
X2
R2
20mΩ
VOUT
3.3V/5A
C7
0.01µF
NDRIVE
PGND
12
C1 = TANTALUM
C3 = SANYO (OS-CON) 20SA100M ESR = 0.037Ω IRMS = 2.25A
C6 = AVX (TA) TPSE227K01R0080 ESR = 0.080Ω IRMS = 1.285A
Q1, Q2 = SILICONIX PMOS BVDSS = 20V DCRON = 0.100Ω Qg = 50nC
Q3 = SILICONIX NMOS BVDSS = 30V DCRON = 0.050Ω Qg = 30nC
D1 = MOTOROLA SCHOTTKY VBR = 30V
R2 = KRL NP-2A-C1-0R020J Pd = 3W
L1 = KOOL Mµ CORE, 16 GAUGE
100
Q3
Si9410
D1
MBRS140T3
C6
220µF
10V
X2
+
EFFICIENCY (%)
+
80
1148_1.eps
COILTRONICS (408)241-7876
KRL BANTRY (603) 668-3210
SILICONIX (800) 554-5565
Figure 1. LTC1148-3.3 high-efficiency 5V to 3.3V/5A step-down converter
26
90
70
1
10
100
IOUT (mA)
1000
10000
1148_2.eps
Figure 2. Efficiency for 5V to 3.3V synchronous
switcher
Linear Technology Magazine • October 1993
DESIGN IDEAS
A Dual-Output LCD-Bias
Voltage Generator
With the many different kinds of
LCD displays available, systems
manufacturers often want the
option of deciding the polarity of the
LCD-bias voltage at the time of
manufacturing.
The circuit in Figure 1 uses the
LT1107 micropower DC-to-DC converter with a single inductor. The
LT1107 features an ILIM pin that enables direct control of maximum
inductor current. This allows the use
of a smaller inductor without the risk
of saturation. The LT1111 could also
be used, with a resulting reduction in
output power.
Circuit Operation
The circuit is basically an ACcoupled boost topology. The feedback
signal is derived separately from the
outputs, so loading of the outputs
does not affect loop compensation.
Since there is no direct feedback from
the outputs, load regulation performance is reduced. With 28 volts out,
from 10% to 100% load (4mA to
40mA), the output voltage sags by
about 0.65 volts. From 1mA to 40mA
load, the output voltage sags by about
1.4 volts. This is acceptable for most
displays.
Output noise is reduced by using
the auxiliary gain block (AGB) in the
feedback path. This added gain effectively reduces the hysteresis of the
comparator and tends to randomize
output noise. With low-ESR caps for
C2 and C4, output noise is below
30mV over the output load range.
Output power increases with VBATTERY,
from about 1.4 watts out with 5 volts
in to about 2 watts out with 8 volts or
more. Efficiency is 80% over a broad
output-power range.
If only a positive or negative output
voltage is required, the two diodes
and two capacitors associated with
the unused output can be eliminated.
The 100kΩ load is required on each
output to load a parasitic voltage
doubler created by the capacitance of
diodes D2 and D4. Without this minimum load, the output voltage can go
up to almost 50% above the nominal
value.
VBATTERY
4V to 16V
(OPTION
SEE TEXT)
+
100k
30Ω
VIN
AO
10µF
16V
C3
LT1107CS8
FB
SW2
5V CONTRAST ADJ.
1M POT.
C2
D1
ILIM
SW1
+VO
24V TO 32V
(0mA TO 40mA)
+
+
+
D2
100k
D4
–VO
–24V TO –32V
(0mA TO 40mA)
+
10µF
16V
C1
L1
D3
1N4148
+
VIN
3V to 12V
C4
100k
SET
GND
1.25V
1.43M
0.01
2.32M
1N4148
10k
SHUTDOWN IN
“1” = OFF
100k
L1 = COILTRONICS CTX 50-4
C1, C2, C3 and C4 = 22µF, 35V LOW ESR
by Jon A. Dutra
Component Selection
The voltage at the switch pin SW1
swings from zero volts to VOUT plus
two diode drops. This voltage is AC
coupled to the positive output
through C1 and D1 and to the negative output through C3 and D3. The
full output current flows through C1
and C3. Most tantalum capacitors are
not rated for current flow, and their
use can result in field failures. Use a
rated tantalum or a rated electrolytic
for longer system life. At lower output
currents or higher frequencies, using
monolithic ceramics is also feasible.
One could replace the 1N5819
Schottky diodes with 1N4148 types
for lower cost, with a reduction
in efficiency and load-regulation
characteristics.
Shutdown
The circuit can be shut down in
several ways. The easiest is to pull the
set pin above 1.25 volts; however, this
consumes 300µA in shutdown conditions. A lower power method is to
turn off VIN to the LT1107 by means
of a high-side switch or by simply
disabling a logic supply. This drops
quiescent current from the VBATTERY
input below 10µA. In both cases VOUT
drops to zero volts. In the that event
+VOUT does not need to drop to zero,
C1 and D1 can be eliminated.
Output Voltage Adjustment
The output voltage can be adjusted
from any voltage above VBATTERY up to
46 volts with proper passive components. Output voltage can be
controlled by the user with DAC,
PWM, or potentiometer control. By
summing currents into the feedback
node, the output voltage can be adjusted downward.
= 1N5819 or MBR140
LCD_1.eps
This design idea originally appeared in EDN
magazine.
Figure 1. Schematic diagram LT1107 dual-output LCD-bias generator
Linear Technology Magazine • October 1993
27
DESIGN IDEAS
A Linear-Phase Bandpass Filter for
Digital Communications
by Philip Karantzalis
Bandpass filters with linear passband phase are useful for a variety of
data communications tasks, the most
noteworthy of which may be in
modulation-demodulation (modem)
circuitry. Modems generate signals
that must be processed without phase
distortion to allow error-free transmission and reception of information
(or the closest approach to that ideal
we can achieve).
Figure 1 shows a linear-phase
bandpass filter using the LTC1264
high-frequency, universal switched-
capacitor filter building block. This
filter is an eighth-order narrow
bandpass filter, centered at 50kHz for
a 1MHz clock input, with flat group
delay in its passband. The fCLK-tofCENTER frequency ratio is 20:1. Figure
2 shows the filter’s narrow-band gain
response and Figure 3 shows the
passband group delay.
An interesting feature of linearphase bandpass filters is that their
response to a step input produces a
short transient sine wave burst with
a symmetrical envelope. Figure 4
RlB, 30.1k
RhB, 107k
R1B, 60.4k
1
VIN
R2B, 10.7k
2
R3B, 29.4k
3
R4B, 13.3k
4
5
6
7
7.5V
0.1µF
8
R4A, 11.5k
9
R3A, 12.4k
10
R2A, 10.7k
11
12
INV B
INV C
HP B
HP C
BP B
BP C
LP B
LP C
SB
LTC1264
SC
V–
GND
V+
FCLK
SA
SD
LP A
LP D
BP A
BP D
HP A
HP A
INV A
INV D
24
23
22
21
R2C, 12.4k
R3C, 37.4k
R4C, 10.7k
CC
20
0.1µF
19
18
–7.5V
1MHz
17
16
15
14
R4D, 10.7k
CD
R3D, 29.4k
RhD, 27.4k
Table 1. Capacitor selection guide
VS
±7.5V
±7.5V
±5V
±5V
R2D, 10k
VOUT
13
RlD, 100k
R1A, 16.2k
shows a comparison of the transient
responses to a step input for the
linear phase bandpass filter of Figure
1 and a bandpass filter with a similar
passband and nonlinear phase response. The response of a bandpass
filter to a step input is a simple qualitative test for determining the linearity of
its phase response, although in
data transmission systems the measurements are usually made with
eye-diagrams and constellation
displays.
The maximum clock frequency for
the filter is 2MHz with ±7.5 volt supplies. This allows bandpass filters
with center frequencies up to 100kHz
to be realized without significant
phase distortion in the passband.
Capacitor C, across R4 in sections
C and D, minimizes gain and phase
variations when the filter is used with
clock frequencies greater than
1.4MHz. For ±5 volt supplies, the
maximum clock frequency is 1.6MHz.
Use the Table 1 as a guide for the
selection of capacitor C.
fCLK
1.8MHz
2.0MHz
1.6MHz
1.4MHz
CC = C D
3pF
5pF
5pF
3pF
1264_1a.eps
10
100
0
95
–10
90
–20
85
DELAY (µs)
GAIN (dB)
Figure 1. LTC1264 linear-phase, 8th order bandpass filter
–30
–40
–50
75
70
–60
65
–70
60
–80
55
–90
10
50
46
100
200
FREQUENCY (kHz)
1264_2.eps
Figure 2. Filter gain versus frequency
28
80
47
48
49 50 51 52
FREQUENCY (kHz)
53
54
1264_3.eps
Figure 3. Filter group delay versus frequency
Figure 4. Step response—
Top: nonlinear-phase filter
Bottom: linear-phase filter
Linear Technology Magazine • October 1993
DESIGN INFORMATION
Book Review:
Power Electronics—
Circuits, Devices and Applications
by Muhammad H. Rashid, Englewood Cliffs, New Jersey,
Prentice Hall Publishing. Second Edition, 1993.
This is the first book review to
appear in the pages of Linear Technology. As is our custom, we shall present
the reader with tools that help him to
do his job. We hope that by reviewing
and recommending a few really useful books each year, we will help you,
our customers, become even better
design engineers.
While such subjects as power and
DC/DC converters are certainly worthy of life-long study, one has to start
somewhere. I am a novice in this area,
but I chose to review this book in
order to bring an unbiased viewpoint
to the subject.
The book is quite a good fundamental text on power electronics.
Power electronics is the study and
application of solid-state electronics
for the control and conversion of electric power. Rashid walks the reader
through many of the basic to not-sobasic concepts inherent in the design
of all types of power circuits, from
large “electric company” type circuitry
to DC/DC converters. For example,
Rashid carefully covers the important parameters of power transistors
in his chapter on these devices.
I noted, in particular, his discussions
on the storage-time constant and the
saturating charge in bipolar transistors. His treatment of the gate-drive
requirements for MOSFETs will help
the non-expert designer gain the expertise required for complex switcher
designs.
The text also includes chapters on
DC/DC converters, Advanced Modulation Techniques, and Resonant
Converters. There is a chapter on
Protection of Devices and Circuits
that should be required reading for all
designers, whether of power supplies
or A-to-D converters.
Simulation of circuits, a topic often
given the “black magic” treatment
until graduate school or one’s first
job, receives some ink in Rashid’s
text. Rashid examines device models
in PSPICE for diodes, thyristors,
power transistors, and other active
devices as well as for circuits
ranging from buck regulators
to pulse-width-modulated inverters.
These simulations are included to
verify design examples and to
introduce the reader to SPICE.
This is an excellent way to introduce
not only the concept of modeling but
the idea that each SPICE model represents a device or circuit composed
of devices in the real world. Designers
must build circuits from these realworld devices, so the model must
match the device; though such models are simplifications and may omit
some real-world complexities.
I would advise the reader to breadboard any circuits gleaned from this
text. Although it is quite a good introduction to power circuitry, no text
can bring home the problems encountered when real circuits are
switching amperes in the vicinity of
nodes that must sense millivolts.
Rashid’s book is a good, concise
text that includes all the right
topics at a technical level that will
be useful for many power-supply
designers. The book also is a good
introduction to power electronics
for would-be power supply experts.
It deserves a look.
——RM
Conclusion
important. When the LT1204 is used
for active video switching between
two flat-field video signals (a very
critical test) the switching artifacts
are nearly invisible. When the LT1204
is used to switch between two live
video signals, the switching artifacts
are invisible.
“Picture-in-Picture” continued from page 22
only on expensive equipment. Where
cost is a determining factor in system
design, the exceptionally low amplitude and brief duration of the
LT1204’s switching artifact make it
an excellent choice for active video
switching.
Active video switching can be accomplished inexpensively and with
excellent results when care is taken
with both the selection and application of the high-speed multiplexer.
Both fast switching and small, wellcontrolled switching glitches are
1
Linear Technology Magazine • October 1993
Video production, in the most general sense,
means any purposeful manipulation of the video
signal, whether in a television studio or on a
desktop PC.
29
NEW DEVICE CAMEOS
New Device Cameos
LT1161: Quad, Protected,
High-Side MOSFET Driver
High-side switching in hostile environments, such as industrial
control, avionics, and automotive
applications, requires a ruggedized
N-channel MOSFET driver. The
LT1161 provides a full 100% operating voltage margin in 24V systems
and can survive supply voltages from
−15V to +60V (+75V on the gate)
without damage.
Each of the four LT1161 switch
channels has a completely selfcontained charge pump to fully
enhance an N-channel MOSFET
switch with no external components.
Also included in each switch channel
is a drain-sense comparator which is
used to sense switch current.
The LT1161 independently protects and restarts each MOSFET.
Latch-off current limit can also be
implemented with the LT1161.
The LT1161 carries an industrial
temperature grade and is available in
either 20-lead plastic DIP or 20-lead
SO surface mount packages.
The LTC1165 Triple, Inverting
1.8V to 6V MOSFET Driver
The LTC1165 triple, inverting 1.8V
to 6V gate driver makes it possible
to switch either supply- or groundreferenced loads through low RDS(ON),
N-channel switches from as little as
1.8V (two discharged cells). The inverting inputs make it possible to
directly replace P-channel MOSFET
switches while maintaining systemdrive polarity. The LTC1165 contains
three on-chip charge pumps, so
that less expensive, lower R DS(ON)
N-channel MOSFETs can be used to
replace high-side P-channel switches.
The three charge pumps have been
designed to be very efficient and require no external components. The
standby current with the three inputs switched off is typically 0.01µA.
The quiescent current rises to 95µA
per channel with the input turned on
and the charge pump producing 11V
30
from a 3.3V supply. The LTC1165 is
available in both 8-pin DIP and 8-pin
SOIC packaging.
LTC1143: Dual, 3.3V/5V HighEfficiency Step-Down Regulator
The LTC1143 is a dual-output
switching-regulator controller in a
narrow, 16-lead SOIC package that
minimizes board area. The LTC1143
utilizes two independent currentmode regulator blocks to provide
simultaneous 3.3V and 5V outputs
with individual shutdown controls.
The LTC1143 extends battery life
by providing high efficiencies at load
currents ranging from milliamps to
amps. Both LTC1147-based regulator
blocks use current-mode architecture
with Burst ModeTM operation.
The LTC1143 is ideal for low-cost
applications requiring 3.3V and 5V
simultaneously, with high conversion efficiencies and the smallest
component count and board space.
LTC1142: Dual, 3.3V/5V UltraHigh-Efficiency Synchronous
Step-Down Regulator
The LTC1142 is a dual-output,
synchronous switching-regulator
controller in a compact, 28-lead SSOP.
Two independent regulator blocks
simultaneously provide 3.3V and 5V
outputs with individual shutdown
controls.
Both LTC1148-based regulator
blocks use current-mode architecture
with Burst Mode TM operation to
provide extremely high operating efficiency, typically greater than 90%,
over the entire load range. The
LTC1142 extends battery life by providing high efficiencies at load currents
ranging from milliamps to amps.
LT1203/LT1205: 170MHz Dual
and Quad Video Multiplexers
The LT1203 is a wideband, twoinput video multiplexer designed for
sub-pixel switching and broadcastquality routing. The LT1205 is a
four-input, two-output multiplexer
designed for multi-input expansion.
These multiplexers act as SPDT video
switches with 10ns transition times at
toggle rates up to 30MHz. The −3dB
bandwidth is 170MHz; gain flatness
is −0.1dB to greater than 60MHz.
Many parts can be tied together at
their outputs by using a shutdown
feature that reduces the power dissipation and raises the output
impedance to 5MΩ. Output capacitance in shutdown mode is only 1pF
and the LT1203 peaks less than 2dB
with a 50pF load. Channel crosstalk
and shutdown isolation are greater
than 80dB at 10MHz. An on-chip
logic buffer interfaces to fast TTL or
CMOS edges to reduce switching transients to 50mV with a 20ns duration.
The LT1203 and LT1205 outputs are
protected against shorts to ground.
The LT1203 is available in 8-lead
PDIP and SO packages and the LT1205
is available in the 16-lead SO package.
LTC1327, LTC1337, LTC1349,
and LTC1350 Ultra-Low-Power
RS232 Transceivers
Four new RS562/RS232 transceivers for notebook and palmtop
computers feature the industry’s lowest power consumption. Each
transceiver is a three-driver/fivereceiver device that offers a complete
PC serial port solution for 3V or 5V
systems and draws only 300µA quiescent supply current.
On-chip charge pumps allow operation from standard logic supplies.
The charge pumps require only four
space-saving 0.1µF capacitors and
can supply up to 12mA of extra current for external circuitry. The
transceivers operate up to 120k baud
with a 1000pF/3kΩ loads. The parts
are pin compatible with the LT1137.
All transceivers feature LTC’s proprietary output structure, which
withstands repeated ESD strikes of up
to ±10kV using the human body model.
The LTC1327 is an RS562 transceiver designed for 3V systems. In the
SHUTDOWN mode, the supply curLinear Technology Magazine • October 1993
NEW DEVICE CAMEOS
rent drops to less than 1µA. The
LTC1337 is the RS232 version of the
LTC1327—it operates from a 5V
supply.
The LTC1349 is a 5V-supply RS232
transceiver that features two lowpower receivers that remain active
while in the SHUTDOWN mode. The
two receivers together draw only 30µA
of supply current. The LTC1350 is
the RS562, 3V-supply version of the
LTC1349.
All four circuits are available in 28pin DIP, SSOP, and SOIC packages.
Four New Products Combine
Low-Power RS562 and RS485
Transceivers On One Chip
Four new RS562/RS485 transceivers each contain two interface
ports that can operate in either RS485
mode, RS562 mode, or in a combination of the two. A loopback mode
performs a diagnostic self test; a
shutdown mode reduces the supply
current to 15µA. The typical unloaded
supply current is only 600µA in
normal operation.
The RS562 transceivers operate to
120k baud and are fully compliant
with RS562 overvoltage specifications.
The RS485 transceivers operate to
10M baud and are fully compliant
with all RS485 specifications. All driver
outputs feature short-circuit protection and thermal shutdown. An enable
pin allows the RS485 driver outputs
to be forced into three-state operation,
which is maintained when the outputs are forced beyond the supply
rails or the power is off. Both receiver
inputs and driver outputs feature
±10kV ESD protection.
A +5V supply for VDD and VCC and
a −5V supply for VEE are required for
RS562 operation. However, if RS232
voltage levels are required, VDD can be
raised to +(6.5 to 10V) and VEE lowered to −(6.5 to 10V).
The LTC1321 can be configured as
two RS562 transceivers, as one RS562
transceiver and one RS485 transceiver, or as two RS485 transceivers.
The LTC1322 can configured as four
RS562 transceivers, as two RS562
transceivers and one RS485 trans-
Linear Technology Magazine • October 1993
ceiver, or as two RS485 tranceivers.
The LTC1334 and LTC1335 are pin
compatible with the LTC1321 and
LTC1322 respectively, except that VCC
is internally connected to VDD and the
unused pin becomes a receiveroutput enable.
All four circuits are available in
24-pin DIP and SOIC packages.
LT1413 Dual, Single-Supply
Precision Op Amp
The LT1413 is an improved, lowcost version of Linear Technology’s
industry-standard LT1013 dual,
single-supply op amp. The LT1413 is
optimized for single 5V applications:
the input goes below ground and the
output swings to ground while sinking current (no output pull-down
resistors are needed). Phase-reversal
protection circuitry keeps the proper
phase at the output, even when the
input is significantly below ground.
The specifications achieved at supply voltages of V+ = +5V, V− = 0V with
the low-cost plastic DIP and SO-8
grades are 280µV maximum offset
voltage (380µV maximum in the SO8), 0.8nA maximum offset current,
1.4 million voltage gain, 0.5µV/°C
drift, 950kHz gain-bandwidth product, 0.55µV peak-to-peak noise from
0.1Hz to 10Hz, and 140dB channel
separation. The output delivers in
excess of 10mA load current (sourcing or sinking), even though the supply current per amplifier is only 330µV
quiescent. A full set of specifications is
also provided on the LT1413 at
±15V supplies.
For further information on the
above or any other devices
mentioned in this issue of Linear
Technology, use the reader service
card or call the LTC literatureservice number: 1-800-4-LINEAR.
Ask for the pertinent data sheets
and application notes.
Information furnished by Linear Technology Corporation is believed to be accurate and reliable.
However, no responsibility is assumed for its use.
Linear Technology makes no representation that
the circuits described herein will not infringe on
existing patent rights.
LTC in the News . . .
Fiscal 1993 Again a Record
On July 20, 1993, Linear Technology Corporation announced
that net sales for its fiscal year
ended June 27, 1993 were a record
$150,867,000, an increase of 26%
over fiscal 1992. The company also
reported record net income for the
year of $36,435,000 or $0.99 per
share.
Net sales for the fourth quarter
of 1993 were a record $42,813,000,
a 31% increase over the fourth
quarter of the previous year. Net
income for the quarter was
$10,838,000 or $0.29 per share,
an increase of 49% over the fourth
quarter of the previous year. A
cash dividend of $0.05 per share
will be paid on August 18, 1993 to
LTC shareholders of record on
August 2, 1993.
According to Robert H.
Swanson, President and CEO,
“1993 was another excellent year
for us. Bookings, revenues, profits,
and cash generated from the business all grew handsomely. We
introduced a record number of
new products, brought up a
second fabrication line, and commenced construction on our first
offshore plant in Singapore. This
balance of financial performance,
product support and introduction,
and sales penetration continues to
be our major strategic focus.”
More Rankings and Ratings
Recognition of Linear Technology's performance continues.
California Business ranked LTC as
43rd in its Top 500 companies in
California, and also ranked LTC
among its Top 50 in Technology.
The Los Angeles Times included
Linear Technology in its Top 100
Companies in California. Locally, LTC
was included in the San Francisco
Chronicle’s and the San Francisco
Examiner’s Top 100 Bay Area Companies, and in the San Jose Mercury
News Silicon Valley Top 150.
Burst ModeTM is a trademark of Linear Technology
Corporation.
31
DESIGN TOOLS
World Headquarters
Applications on Disk
Linear Technology Corporation
1630 McCarthy Boulevard
Milpitas, CA 95035-7487
Phone: (408) 432-1900
FAX: (408) 434-0507
NOISE DISK
This IBM-PC (or compatible) progam allows the user to
calculate circuit noise using LTC op amps, determine the
best LTC op amp for a low noise application, display the
noise data for LTC op amps, calculate resistor noise, and
calculate noise using specs for any op amp.
Available at no charge.
SPICE MACROMODEL DISK
This IBM-PC (or compatible) high density diskette contains
the library of LTC op amp SPICE macromodels. The
models can be used with any version of SPICE for general
analog circuit simulations. The diskette also contains working circuit examples using the models, and a demonstration
copy of PSPICETM by MicroSim.
Available at no charge.
Technical Books
1990 Linear Databook — This 1440 page collection
of data sheets covers op amps, voltage regulators,
references, comparators, filters, PWMs, data conversion
and interface products (bipolar and CMOS), in both commercial and military grades. The catalog features well over
300 devices.
$10.00
1992 Linear Databook Supplement — This 1248 page
supplement to the 1990 Linear Databook is a collection of
all products introduced since then. The catalog contains full
data sheets for over 140 devices. The 1992 Linear Databook
Supplement is a companion to the 1990 Linear Databook ,
which should not be discarded.
$10.00
Linear Applications Handbook — 928 pages full of
application ideas covered in depth by 40 Application Notes
and 33 Design Notes. This catalog covers a broad range of
“real world” linear circuitry. In addition to detailed, systemsoriented circuits, this handbook contains broad tutorial
content together with liberal use of schematics and scope
photography. A special feature in this edition includes a 22
page section on SPICE macromodels.
$20.00
1993 Linear Applications Handbook Volume II —
Continues the stream of “real world” linear circuitry initiated
by the 1990 Handbook. Similar in scope to the 1990 edition,
the new book covers Application Notes 41 through 54 and
Design Notes 33 through 69. Additionally, references and
articles from non-LTC publications that we have found
useful are also included.
$20.00
U.S. Area
Sales Offices
CENTRAL REGION
Linear Technology Corporation
Chesapeake Square
229 Mitchell Court, Suite A-25
Addison, IL 60101
Phone: (708) 620-6910
FAX: (708) 620-6977
NORTHEAST REGION
Linear Technology Corporation
One Oxford Valley
2300 E. Lincoln Hwy., Suite 306
Langhorne, PA 19047
Phone: (215) 757-8578
FAX: (215) 757-5631
NORTHWEST REGION
Linear Technology Corporation
782 Sycamore Dr.
Milpitas, CA 95035
Phone: (408) 428-2050
FAX: (408) 432-6331
SOUTHEAST REGION
Linear Technology Corporation
17060 Dallas Parkway
Suite 208
Dallas, TX 75248
Phone: (214) 733-3071
FAX: (214) 380-5138
SOUTHWEST REGION
Linear Technology Corporation
22141 Ventura Boulevard
Suite 206
Woodland Hills, CA 91364
Phone: (818) 703-0835
FAX: (818) 703-0517
Interface Product Handbook — This 200 page handbook
features LTC’s complete line of line driver and receiver
products for RS232, RS485, RS423, RS422 and AppleTalk 
applications. Linear’s particular expertise in this area involves low power consumption, high numbers of drivers
and receivers in one package, 10kV ESD protection of
RS232 devices and surface mount packages.
Available at no charge.
Monolithic Filter Handbook — This 232 page book comes
with a disk which runs on PCs. Together, the book and disk
assist in the selection, design and implementation of the
right switched capacitor filter circuit. The disk contains
standard filter responses as well as a custom mode. The
handbook contains over 20 data sheets, Design Notes and
Application Notes.
$40.00
SwitcherCAD Handbook — This 144 page manual, including disk, guides the user through SwitcherCAD—a
powerful PC software tool which aids in the design and
optimization of switching regulators. The program can cut
days off the design cycle by selecting topologies, calculating operating points and specifying component values and
manufacturer's part numbers.
$20.00
International
Sales Offices
FRANCE
Linear Technology S.A.R.L.
Immeuble “Le Quartz”
58 Chemin de la Justice
92290 Chatenay Mallabry
France
Phone: 33-1-46316161
FAX: 33-1-46314613
GERMANY
Linear Technology GMBH
Untere Hauptstr. 9
D-8057 Eching
Germany
Phone: 49-89-3194710
FAX: 49-89-3194821
JAPAN
Linear Technology KK
5F YZ Building
4-4-12 Iidabashi Chiyoda-Ku
Tokyo, 102 Japan
Phone: 81-3-3237-7891
FAX: 81-3-3237-8010
KOREA
Linear Technology Korea Branch
Namsong Building, #505
Itaewon-Dong 260-199
Yongsan-Ku, Seoul
Korea
Phone: 82-2-792-1617
FAX: 82-2-792-1619
SINGAPORE
Linear Technology Pte. Ltd.
101 Boon Keng Road
#02-15 Kallang Ind. Estates
Singapore 1233
Phone: 65-293-5322
FAX: 65-292-0398
TAIWAN
Linear Technology Corporation
Rm. 801, No. 46, Sec. 2
Chung Shan N. Rd.
Taipei, Taiwan, R.O.C.
Phone: 886-2-521-7575
FAX: 886-2-562-2285
UNITED KINGDOM
Linear Technology (UK) Ltd.
The Coliseum, Riverside Way
Camberley, Surrey GU15 3YL
United Kingdom
Phone: 011-44-276-677676
FAX: 011-44-276-64851
LINEAR TECHNOLOGY CORPORATION
1630 McCarthy Boulevard
Milpitas, CA 95035-7487
(408) 432-1900
Literature Department 1-800-4-LINEAR
AppleTalk is a registered trademark of Apple Computer, Inc.
©32
1993 Linear Technology Corporation/ Printed in U.S.A./20K
Linear Technology Magazine • October 1993