DECEMBER 1995 VOLUME V NUMBER 4
by Jaime Tseng
The LT1160 and LT1162:
Medium Voltage,
N-Channel Bridge
Design Made Easy
IN THIS ISSUE . . .
COVER ARTICLE
The LT
®
1160 and LT1162:
Medium Voltage, N-Channel
Bridge Design Made Easy
........................................ 1
Jaime Tseng
Editor’s Page ................... 2
Richard Markell
LTC in the News .............. 2
DESIGN FEATURES
The LT1166: Power Output
Stage Automatic Bias
System Control IC............ 3
Dale Eagar
The LT1186: A Simple
Bits-to-Light Converter for
LCD Displays ................... 6
Anthony Bonte
Power Factor Correction,
Part Four: The Brains
Behind the Brawn ......... 10
Dale Eagar
How Do You Slew 200V/
µ
s
with a 250
µ
A Op Amp?
...................................... 12
George Feliz
DESIGN IDEAS
.................................17–26
(complete list on page 17)
New Device Cameos ....... 27
Design Tools .................. 28
Sales Offices ................. 28
any input condition. When both
inputs are high, both outputs are
actively held low.
Bipolar versus CMOS Drivers
Bipolar technology was chosen for
operation to 60V (absolute maximum)
due to its inherent greater-than-60V
junction-breakdown voltage. This is
particularly important in an N-chan-
nel topside driver, which must operate
above the supply rail. In addition,
bipolar technology provides consis-
tent drive capability, resulting in
transition times that increase moder-
ately with large increases in capacitive
load. The LT1160 switches 1,000pF
loads in 75ns, but for 10,000pF loads,
this only increases to 180ns, making
operation to 100kHz possible with
the largest MOSFETs. Figures 2a
shows a simplified schematic of the
LT1160 output stage; Figure 2b shows
how it delivers progressively more
current as the capacitive load in-
creases. On the other hand, CMOS
drivers’ switching times tend to be
fast at light loads but increase rapidly
at higher capacitive loads, as shown
in Figures 3a and 3b.
Introduction
The new LT1160 is a second-gen-
eration, half-bridge, N-Channel power
MOSFET driver. The LT1160, which
is derived from the LT1158 half-bridge
driver, has an increased voltage han-
dling capability and a higher output
drive current. The LT1162 is a full-
bridge version of the LT1160
Like its predecessor, the LT1160
effectively deals with the many prob-
lems and pitfalls encountered in the
design of high efficiency motor-con-
trol and switching regulator circuits.
This article will discuss these prob-
lems, along with the solutions the
LT1160 offers the bridge designer.
LT1160 Overview
Figure 1 is a block diagram of the
LT1160 (the LT1162 has two circuits
identical to that in Figure 1). New in
the LT1160 are two independent un-
dervoltage-detect circuit blocks that
disable the drivers when either the
supply voltage or the voltage across
the bootstrap capacitor drops below
its undervoltage trip point. Two input
pins control the switching of both
N-channel MOSFETs; both channels
are noninverting drivers. The inter-
nal logic prevents both outputs from
turning “ON” simultaneously under
continued on page 14
, LTC and LT are registered trademarks of Linear Technology Corporation.
Burst Mode, Bits-to-Nits and C-Load are trademarks of Linear Technology Corporation.
LINEAR TECHNOLOGY
LINEAR TECHNOLOGY
LINEAR TECHNOLOGY
2
Linear Technology Magazine • December 1995
EDITOR’S PAGE
by Richard Markell
On Being Analog in Silicon Valley
Authors can be contacted
at (408) 432-1900
LTC in the News...
Working in Silicon Valley is inter-
esting. The hottest technologies blow
through, virtually overnight. Some-
times the technologies take off, and
sometimes they self-destruct. But we
at LTC need to know what’s coming so
that we can design analog circuitry
that allows the fast-paced digital world
to talk to our linear world. The micro-
processor gurus might dispute it, but
the expertise required to optimize the
transient response of a DC/DC con-
verter is nontrivial. The IC designer
must know a multiplicity of disci-
plines, from transistor-level integrated
circuit simulation to the system-level
ESR requirements of capacitors. We
at LTC pride ourselves on being the
best in the industry at what we do,
whether it’s building applications cir-
cuits for customers, designing IC’s or
taking an order in customer service.
The design of ASICs, multimedia and
networks may be hot today, cold to-
morrow, but analog will always be
with us. We press on, growing at a
good clip, supplying the analog core
to the world of digital designers. We
are always looking for people who are
in love with analog circuit design.
Our feature article highlights the
LT1160 and LT1162 N-channel
MOSFET half- and full-bridge drivers,
with increased maximum operating
voltage (60V abs. max.) and higher
output drive current. Also new to
these topologies is undervoltage lock-
out, which disables the drivers when
either the supply voltage or the volt-
age across the bootstrap capacitor
drops below its trip point. Cross-con-
duction or “shoot-through” has been
completely eliminated. The LT1160
and LT1162 will see use as motor-
control circuits, high current,
high-efficiency switching regulators
and other half- and full-bridge driver
circuitry.
Another “control” device is the new
LT1186 Bits-to-Light converter for
LCD displays. The LT1186 is a com-
panion part to LTC’s LT1182/
LT1183/LT1184 CCFL/LCD contrast
switching regulator family of devices.
The LT1186 adds simple digital con-
trol for CCFL to a fixed frequency,
current mode switching regulator. The
device incorporates a micropower 8-
bit, 50µA full-scale, current output
DAC that provides the “bits-to-
lamp-current” control. The device
communicates in two modes: stan-
dard SPI and pulse mode.
The LT1352 is a new micropower
operational amplifier that can slew
200V/µs and settle to 0.1% in 700ns
for a 10V step. The LT1352 is stable
with any capacitive load and can drive
heavy loads (40mA) with low distor-
tion. These specifications are achieved
without compromising other specifi-
cations over the part’s ±2.5V to ±15V
supply range. Another Design Fea-
ture spotlights the LT1166, a power
output stage automatic bias system
control IC. The part is designed to
optimize amplifier output stage bias-
ing (and quiescent current) and to
eliminate the need for temperature
tracking, matching transistors or trim
pots. The device allows automatic bi-
asing of an amplifier’s output stage in
the Class AB operational region.
In this issue we present the fourth
and concluding part of the “Power
Factor Correction” series. Also pro-
vided with this issue is an index of
circuits presented in the first five
years of the Linear Technology Magazine.
This issue also has the usual selec-
tion of Design Ideas, carefully chosen
and tested in our Applications Lab.
Here we have circuits ranging from GTL
terminators to rail-to-rail filters.
Once again Linear Technology
reports another quarter of record
sales and profits. Net sales
for LTC’s first quarter ending
October 1, 1995 were a record
$87,005,000, an increase of 50%
over net sales of $58,082,000 for
the previous year. The company
also reported record net income
for the quarter of $30,520,000 or
$0.39 per share, an increase of
71% over $17,830,000 or $0.24
per share reported for the first
quarter of the last year.
More good news was reported in
Investor’s Business Daily
September 29th issue, which
featured a corporate profile under
their New America section. Bob
Swanson talked about the con-
tinuing need for analog circuitry.
“We cheer the growth of digital, but
when the rubber hits the road, we
live in an analog world and we
don’t see this ever changing.
In Financial World October 10,
1995 issue Linear Technology
Corporation was included on its
list of “America’s Best 100 Growth
Companies.” LTC was ranked
#22 out of all the companies
considered: #3 for semi-
conductors and # l for analog!
In the “What’s Hot for Product”
category, Electronic Design
October 13, 1995 issue high-
lighted the LTC1410 in a feature
article titled “12-Bit IC ADCs
Relieve Error Budgets.” The
LTC1410 is the fastest 12-bit
Successive Approximation (SAR)
based ADC to date. The part uses
a novel design technique to
achieve excellent power and cost
performance. The LTC1410 joins
LTC’s fast growing line of A to D
and D to A products for high
performance applications.
Linear Technology Magazine • December 1995
3
The LT1166: Power Output Stage
Automatic Bias System Control IC
1166_1.eps
>4mA
>4mA
GATE
GATE
IN OUTINPUT OUTPUT
R
2
R
1
V
1
V
2
I
SENSE
I
SENSE
LT1166
V+
V
Q
2
Q
1
1166_2.eps
V
1
V
2
V
3
V
+
V
V
3
IS ADJUSTED
SO THAT
V
1
*
V
2
= 0.0004
WHEN THE OUTPUT
CURRENT IS ZERO
V
1
= V
2
= 20mV
Figure 1. Basic LT1166 circuit configuration Figure 3. LT1166 Voltage-control loop Figure 5. LT1166’s can be easily paralleled
1166_4.eps
MAGICAL
OPERATING
POINT
CROSSOVER
DISTORTION QUIESCENT
POWER
BIAS
00
Introduction
Class AB amplifiers are popular
because of their “near Class A” per-
formance and their ability to operate
on considerably less quiescent power
than Class A. Class AB amplifiers are
easy to construct, rugged and reli-
able. However there is an aspect of
these amplifiers that can cause per-
plexity, consternation and finally hair
loss—their bias scheme. The problem
is that the very parameter that makes
Class AB so good, namely low quies-
cent current, is poorly controlled. The
LT1166 offers control over the quies-
cent current directly, removing the
necessity of temperature tracking,
matching transistors or trim pots.
Functional Description
The LT1166 (Figure 1) controls the
Class AB output stage by incor-
porating two control loops, the
current-control loop and the voltage-
control loop. The current-control loop
(Figure 2) operates independently of
the voltage loop while keeping the
set bias and the nerves to take the
catastrophic results of misalignment.
The LT1166 removes all excess cross-
over distortion caused by improperly
set quiescent current, while signifi-
cantly reducing the distortion caused
by the effects of nonlinear transcon-
ductance in the output transistors.
Figure 4 shows how the magic point
for a Class AB amplifier is visualized
(keep in mind that the magic point is
temperature sensitive in many de-
signs, leading the designer to
temperature tracking tricks, which
are not needed with the LT1166.)
Makes Random
Transistor Pairs into
Parallelable Modules
The LT1166, combined with
external transistors, implements a
unity-gain buffer with an offset volt-
age of ±50mV. This buffer in turn
becomes a component that can easily
be paralleled to increase the output
power (see Figure 5).
Terra Firma!
California, a place where the
ground is anything but stable, is now
the birthplace of “The Rock,” a virtual
ground with an attitude. The Rock
(detailed in Figure 6) is a 5V, ±0.4%
1166_3.eps
V
1
V
2
+
V
3
2
V
3
2
OUT
by Dale Eagar
product of V1 and V2 constant. The
voltage loop (Figure 3) maintains the
output voltage at the input voltage
level by driving both gates up or down.
The two loops, although mutually in-
dependent, act in harmony to provide
a component insensitive, tempera-
ture insensitive, simple Class AB bias
network.
Crossover Distortion
Adjusting the bias point of a Class
AB amplifier is quite like walking a
tight rope. Quiescent current needs
to be set precisely; if it is adjusted too
low, the amplifier exhibits crossover
distortion, whereas if it is adjusted
too high, the amplifier becomes a very
effective heater. Setting the bias of a
Class AB amplifier is often reserved
for the “Medicine Man,” an individual
with both the smarts to know how to
DESIGN FEATURES
Figure 2. LT1166 Current-control loop
1166_5.eps
IN OUTPUT
LT1166
LT1166
Figure 4. “Magical” operating point of
Class AB amplifier
4
Linear Technology Magazine • December 1995
DESIGN FEATURES
low noise virtual ground. The Rock
will source or sink 1A of current with
ease. In this circuit, the current limits
are set to 1.5A for both sourcing and
sinking currents. Changing the val-
ues of R5 and R6 to 0.33 increases
the output current by a factor of
three. This circuit, properly built, will
easily withstand an earthquake of
magnitude 8, with unmeasurable
output deviation.
Shaker-Table Driver/
Brute Audio Amplifier
In the event that you don’t live in
California, you might need the circuit
shown in Figures 7a–7c, the 1.8kW
shaker-table driver/audio amplifier.
In Figure 7a, we develop a slice of
power by connecting the LT1166 to
two power MOSFETs, two current
sense resistors and two current
1166_7a.eps
15VA*
12.5VA*
100V
1k
100
1k
7.5k
–12.5VA*
–15VA*
–100V
100
100
CURRENT LIMIT ±6.8A
*FROM FIG. 7c
IRF9240
0.22
IRF230
0.22
OUT
300W
RMS 
INTO
16
1µF
1µF
IN 2
1
8
9
3
6
5
4
2N3904
2N3906
GATE
GATE
LT1166
SENSE
I
LIM
OUT
I
LIM
SENSE
100
1166_7B.eps
OUTPUT
0.22
0.22
0.22
0.22
TO
FIG. 7c
0.22
0.22
+
1k
1k10k
9.1k
3
2
6
INPUT
±10Vp-p
1800W RMS
INTO 2.6
1250W RMS
INTO 4
625W RMS
INTO 8
EXTENDED POWER
SUPPLY A
V
= 10
VOLTAGE AMP
POWER SLICES OF FIG. 7A
U1
LT1360
Figure 6. “The Rock,” a 5V,
±
0.4% tolerance,
±
1A low noise voltage source
Figure 7b. Paralleling power slices to increase output drive
sources. The power slice shown in
Figure 7a will deliver 300W of sine
wave power into 16 when powered
from ±100VDC. The MOSFETs and/
or the sense resistors can be increased
or decreased to get output power in
almost any flavor.
Figure 7b details how six power
slices can be paralleled to put out a
cool, clean 1800W of Richter or 1250W
of rock and roll. The LT1360 amplifier
operates in the extended supply mode
(see “Extending Op Amp Supplies to
Get More Output Voltage,” Linear
Technology, IV:2, pp. 20–22).
Figure 7a.
±
100V, 6A “power slice”
1166_6.eps
U1
LT1431
12V
12V
12V
R2
2k
R4
1k
R7
1k
100
100
R1
1k
Q1
IRFZ34
Q2
IRF9Z30
C3
270µF
5V
OUT
OUT
R5
1
2W
R6
1Ω
2W
R3
100
4
56
C1
1µF
C2
1µF
R8
100
2.5V
5k
5k
78 3 1
2
IN
U2
LT1166
TOP
DRIVE
TOP
SENSE
TOP
I
LIM
BOTTOM
I
LIM
BOTTOM
SENSE
BOTTOM
DRIVE
1
8
7
3
6
5
4
+
Linear Technology Magazine • December 1995
5
1166_9.eps
TRIM R.
POTENTIOMETER
1166_7c.eps
OUTPUT
OF FIG. 7B
LT1004
2.5
LT1004
2.5
– 15VA*
*TO FIG. 7A
15VA*
15V
15V
1k
R1
390
R2
390Ω
LT1360
4
7
C2
3300pF
C1
0.1µF
12.5VA*
12.5VA*
U1 OF FIG. 7B
R3
1k
DESIGN FEATURES
Figure 7c. Connection diagram for the
±
15V floating supply
Figure 8a. Ring-tone generator schematic diagram
1166_8A.eps
0.33µF
11k 10k 10k
7
0.22
0.22 RING-TONE
OUTPUT
300W/SLICE
POWER SLICES
OF FIGURE 7a
2.5VA*
51k 51k
–2.5VA*
V
SS
RV
CC
CMOS 555
0.01µF
6
2
3
*TO FIG. 7c
10.15µF
0.047
1
20Hz
84
0.01µF
360k
6
5
OUT 3
2
360k
+
LT1078
+
LT1078
TO FIG. 8b
1166_8B.eps
LT1004
– 2.5
LT1004
– 2.5
15VA
–15VA
15V
FROM
FIGURE 8a
15V
1k
1k
4
12.5VA
–12.5VA
LT1078
0.1µF
8
LT1004
– 2.5
LT1004
– 2.5
7.5k
7.5k
3.3k
0.1 
100V
– 2.5VA
2.5VA
Figure 9. Trim pots are no longer required
Conclusion
With the advent of the LT1166,
biasing Class AB amplifiers becomes
trivial. Dependencies on temperature
tracking, component matching and
alignment pots have succumbed to
cool, clean, closed-loop control of
operating points. Suddenly, MOSFETs
look a lot friendlier to the power am-
plifier designer. Suddenly, pots have
bitten the dust.
Finally, Figure 7c shows how to
connect the ±15V floating supply and
also details the PSRR snubber for the
LT1360 (R1, R2, R3 and C2).
Ring-Tone Generation
Building upon an earlier Linear
Technology Magazine article (see
above) we show a power ring-tone
generator. This generator is detailed
in Figures 7a, 8a and 8b. This circuit
starts at 300W (or less if you use
smaller FETs and bigger resistors in
Figure 7a) and can be stacked up to
any desired power level. With this
circuit, you could easily build a ringer
capable of ringing the phones of every
U.S. politician simultaneously (not
that this would do any good).
Figure 8b. Power supply circuitry for
Figure 8a
6
Linear Technology Magazine • December 1995
The LT1186: A Simple Bits-to-Light
Converter for LCD Displays by Anthony Bonte
Introduction
Many current-generation portable
and handheld instruments use back-
lit liquid crystal displays (LCDs).
Cold-cathode fluorescent lamps
(CCFLs) provide the highest available
efficiency for backlighting the dis-
play. These lamps require high voltage
AC to operate, mandating an efficient
high voltage DC/AC converter. In ad-
dition to good efficiency, the converter
should deliver the lamp drive in the
form of a sine wave. The sine wave
excitation helps to minimize RF emis-
sions and also provides optimal
current-to-light conversion in the
lamp. The circuit should also permit
lamp intensity control from zero to
full brightness with no hysteresis or
“pop-on.”
Manufacturers offer an array of
monochrome and color displays.
These displays vary in size, operating
lamp current, lamp-voltage range and
power consumption. The small size
and battery-powered operation asso-
ciated with LCD-equipped apparatus
dictate low component count and high
efficiency. Size constraints place limi-
tations on circuit architecture, and
long battery life is a priority. All com-
ponents, including pc board and
hardware, usually must fit within the
LCD enclosure, with a height restric-
tion of 5mm–10mm.
Handheld instruments are gener-
ally characterized by dedicated
functionality and reduced processing
power in comparison with notebook
or laptop computers. These portable
systems often use general-purpose,
low cost microcontrollers such as the
Intel 80C51 or Motorola 68HC11.
These microcontrollers are popular
because of their versatile I/O capabil-
ity. A simple interface that allows the
microcontroller to easily control the
operating lamp current and display
brightness is required. To address
the requirements of these portable
instruments, Linear Technology
introduces the LT1186 DAC program-
mable CCFL switching regulator.
BAT
8V TO 28V
16
15
14
13
12
11
10
9
I
CCFL
DIO
CCFL V
C
AGND
SHDN
CLK
CCFL V
SW
BULB
BAT
ROYER
V
CC
I
OUT
D
OUT
CCFL
PGND
CS
LT1186
LAMP
UP TO 6mA
10 6
L1
R1
750
L2
100µH
321 5
4
+
+
C7, 1µF
C1*
0.068µF
C5
1000pF
R2
220k
R3
100k
C3A
2.2µF
35V
C4
2.2µF
TO MPU
FROM MPU
V
IN
3.3V
OR 5V
1186_1.eps
C3B
2.2µF
35V
C2
27pF
3kV
+
Q2* Q1*
SHUTDOWN
ALUMINUM ELECTROLYTIC IS RECOMMENDED FOR C3A AND C3B. 
MAKE C3B ESR 0.5 TO PREVENT DAMAGE TO THE LT1186 HIGH-SIDE 
SENSE RESISTOR DUE TO SURGE CURRENTS AT TURN-ON
C1 MUST BE A LOW LOSS CAPACITOR, C1 = WIMA MKP-20
Q1, Q2 = ZETEX ZTX849 OR ROHM 2SC5001
0µA TO 50µA I
CCFL
CURRENT GIVES 
0mA TO 6mA LAMP CURRENT 
FOR A TYPICAL DISPLAY.
FOR ADDITIONAL CCFL/LCD CONTRAST APPLICATION CIRCUITS, 
REFER TO THE LT1182/83/84/84F DATA SHEET
D1
BAT85
D
IN
FROM MPU
FROM MPU
1
2
3
4
5
6
7
8D1
1N5818
L1 = COILTRONICS CTX210605
L2 = COILTRONICS CTX100-4
*DO NOT SUBSTITUTE COMPONENTS
COILTRONICS (407) 241-7876
New LTC CCFL Part
The LT1186 is a fixed frequency,
current mode switching regulator that
provides a simple digital control func-
tion for cold-cathode fluorescent
lighting. The regulator is a compan-
ion part to Linear Technology’s
LT1182/LT1183/LT1184 CCFL/LCD
contrast switching regulator family.
The LT1186 supports grounded-lamp
and floating-lamp configurations. The
IC includes an efficient high current
switch, an oscillator, output-
drive logic, control circuitry and a
micropower 8-bit 50µA full-scale cur-
rent output DAC.
The DAC provides simple
“bits-to-lamp-current control” and
communicates in two interface modes:
standard SPI mode and pulse mode.
On power-up, the DAC counter resets
to half-scale and the DAC configures
itself to SPI or pulse mode, depending
on the CS (chip select) signal level. In
SPI mode, the system microprocessor
serially transfers the present 8-bit
data and reads back the previous
Figure 1. Up to 90% efficient floating CCFL with SPI control of lamp current
DESIGN FEATURES
Linear Technology Magazine • December 1995
7
returned to the user to verify the last
programmed current level. An input
of one corresponds to 1 LSB of DAC
output current (195.3nA) and an in-
put of 255 corresponds to the full-scale
DAC output current of 50µA.
Listings 1 and 2 on pages 8–9 are
the complete C and assembly language
code that controls the evaluation setup
of Figure 2. My thanks and apprecia-
tion go to Tommy Wu at Linear
Technology for developing this soft-
ware. Any questions or comments
regarding the software should be ad-
dressed to Tommy, in care of this
publication. The software is fairly ge-
neric and is easily adaptable to other
microcontrollers or microprocessors.
Conclusion
The LT1186 is a novel Bits-
to-Nits
TM
* converter for driving
cold-cathode fluorescent lamps in
handheld or portable instruments
with LCD displays. Lamp current is
digitally programmed in either of two
interface modes: standard SPI mode
and pulse mode, which provides
simple push-button up/down con-
trol. The LT1186 represents another
step in Linear Technology’s commit-
ment to providing efficient, compact
and economical backlight solutions.
connected between the BAT and Royer
pins, provides feedback, and a single
capacitor (C7) on the V
C
pin provides
stable loop compensation.
Lamp current is programmed by
driving the I
CCFL
pin from the I
OUT
pin.
The transfer function between lamp
current and input programming cur-
rent must be empirically determined
and is dependent on a myriad of fac-
tors, including lamp characteristics,
display construction, transformer
turns ratio and the tuning of the
Royer oscillator. The LT1186 control
circuitry is powered by either a 3.3V
or 5V logic supply, and the input
battery voltage range for this example
is 8V to 28V. Lower input battery
voltage operation is accomplished by
increasing the turns ratio of trans-
former L1 and retuning the Royer by
tweaking the values of C1 and C2.
Electrical efficiency with typical CCFL
lamps ranges between 80%–90% at
full load with lamp current being vari-
able from 200µA to 6mA for a typical
display. The component values cho-
sen are an interactive compromise
in maximizing photometric output
versus input power.
Figure 2 is a block diagram for an
Intel 80C31 microcontroller-based
evaluation setup for the LT1186. The
lamp current is controlled by a gen-
eral PC keyboard interface. The user
inputs a number between 0 and 255
to set a new input programming cur-
rent, and therefore, a new lamp
current level. The previous code is
8-bit data from the DAC. In pulse
mode, the upper six bits of the DAC
are configured for increment-only
(1-wire interface) or increment/dec-
rement (2-wire interface) operation,
depending on the D
IN
signal level.
The LT1186 control circuitry oper-
ates from a logic supply voltage of
3.3V or 5V. The IC also has a battery
supply voltage pin that operates from
4.5V to 30V. The LT1186 draws 6mA
typical quiescent current. An active
low shutdown pin reduces total sup-
ply current to 35µA for standby
operation and the DAC retains its last
setting. A 200kHz switching frequency
minimizes magnetic component size,
and the use of current mode switch-
ing techniques gives high reliability
and simple loop frequency compen-
sation. The LT1186 is available in a
16-pin, narrow-body SO.
Typical Application
Figure 1 is a complete floating CCFL
circuit using the LT1186 in standard
SPI mode to control operating lamp
current from a microprocessor. A
floating-lamp circuit allows the
transformer to provide symmetric, dif-
ferential drive to the lamp. Balanced
drive eliminates the field imbalance
associated with parasitic lamp-to-
frame capacitance and reduces
“thermometering” (uneven lamp in-
tensity along the lamp length) at low
lamp currents.
Q1, Q2, L1 and C1 form the core of
a current-fed, Royer-class converter.
The LT1186 power switch and L2
create a switch-mode current source
to drive the Royer. The 200kHz switch-
ing frequency allows the use of only
100µH for L2. Base drive for Q1 and
Q2 is provided by L1’s tickler winding
and R1. R1 is selected to guarantee
sufficient base drive with minimum
betas for Q1 and Q2. R2, R3 and C5
comprise an external divider network
for the open-lamp protection circuit.
The divider monitors the voltage
across the Royer converter and com-
pares it with an internal 7V clamp
voltage between the BAT and BULB
pins. Monitoring the primary-side
converter current through a high-
side sense resistor and amplifier,
CS
CLK
D
IN
D
OUT
5A0-A15
D0-D7
4
RS232
FROM PC
1
2
8
7
9
10
31
P1.4
P1.3
P1.0
P1.1
TX
RX
LT1186
ROYER
CONVERTER
LAMP
ROM
80C31
1186_2.eps
Note:
* Bits-to-Nits is a trademark of Linear Technology
Corporation. 1 Nit = 1 Candela/meter
2
. Nit is
derived from the Latin word nitere meaning “to
shine.”
DESIGN FEATURES
Figure 2. LT1186 Floating CCFL system using an Intel 80C31 microcontroller
8
Linear Technology Magazine • December 1995
Listing 1. LT1186.C
#include <stdio.h>
#include <reg51.h>
#include <absacc.h>
extern char lt1186(char); /* external assembly function in lt1186a.asm */
sbit Clock = 0x93;
main()
{int number=0;
int LstCode;
Clock = 0;
TMOD = 0x20; /* Establish serial communication 1200 baud */
TH1 = 0xE8;
SCON = 0x52;
TCON = 0x69;
while(1) /* Endless loop */
{printf(“\nEnter any code from 0 - 255: “);
scanf(“%d”,&number);
if((0>number)|(number>255))
{number = 0;
printf(“The number exceeds its range. Try again!”);
}else
{LstCode = lt1186(number);
printf(“Previous # %u”, (LstCode&0xFF)); /* AND the previous
number with 0xFF to turn off sign extension */
}
number = 0; /* Reinitialize number */
}
}
C and Assembly Language Program Code for LT1186 Evaluation
The LT1186.C program demon-
strates the use of the LT1186
DAC programmable CCFL switch-
ing regulator.
The example circuit uses
the popular 80C51 family of
microcontrollers to interface between
the user and the LT1186.
The LT1186 DAC algorithm is
written in assembly language,
and is contained in a file named
LT1186A.ASM; it is called as a func-
tion from the MAIN function of
LT1186.C. The user inputs an integer
from 0 to 255 on a keyboard and the
LT1186 adjusts the I
OUT
programming
current to control the operating lamp
current and the brightness of the
LCD display. The assembly language
program, LT1186A.ASM, receives
the D
IN
word from the main C pro-
gram function, lt1186( ). Assembly
to C interface headers, declarations
and memory allocations are listed
before the actual assembly code.
DESIGN FEATURES
Linear Technology Magazine • December 1995
9
DESIGN FEATURES
Listing 2. LT1186A.ASM
; Port p1.4 = CS
; Port p1.3 = CLK
; Port p1.1 = Dout
; Port p1.0 = Din
;
NAME LT1186_CCFL
PUBLIC lt1186, ?lt1186?BYTE
?PR?ADC_INTERFACE?LT1186_CCFLSEGMENT CODE
?DT?ADC_INTERFACE?LT1186_CCFLSEGMENT DATA
RSEG ?DT?ADC_INTERFACE?LT1186_CCFL
?lt1186?BYTE: DS 2
RSEG ?PR?ADC_INTERFACE?LT1186_CCFL
CS EQU p1.4
CLK EQU p1.3
DOUT EQU p1.1
DIN EQU P1.0
lt1186: setb CS ;set CS high to initialize the LT1186
mov r7, ?lt1186?BYTE ;move input number(Din) from keyboard to R7
mov p1, #01h ;setup port p1.0 becomes input
clr CS ;CS goes low, enable the DAC
mov a, r7 ;move the Din to accumulator
mov r4, #08h ;load counter 8 counts
clr c ;clear carry before rotating
rlc a ;rotate left Din bit(MSB) into carry
loop: mov DIN, c ;move carry bit to Din port
setb CLK ;Clk goes high for LT1186 to latch Din bit
mov c, DOUT ;read Dout bit into carry
rlc a ;rotate left Dout bit into accumulator
clr CLK ;clear Clk to shift the next Dout bit
djnz r4, loop ;next data bit loop
mov r7, a ;move previous code to R7 as character return
setb CS ;bring CS high to disable DAC
ret
END
;NOTE: When CS goes low, the MSB of the previous code appears at Dout
10
Linear Technology Magazine • December 1995
Power Factor Correction, Part Four:
The Brains Behind the Brawn
The Control Room
In the three previously published
parts of our series on power factor
correction, we investigated the
workings of the “power transfer
mechanism” as used in the ever so
popular “boost-mode power factor
correction conditioner.” This power
transfer mechanism is shown in Fig-
ure 1 as the “engine room.” Also shown
is the “control room,” where the deci-
sions are made and actions taken to
make the PFC power conditioner be-
have like a resistor (the ultimate
disguise for a switching power sup-
ply). Having had a brief tour of the
engine room in an earlier article,
we can summarize its workings in a
few rules:
1. The amount of power intercepted
from the input line is related to
the duty factor of the boost
engine.
2. When the duty factor is at 0%, no
power is intercepted from the
input line.
3. Under all conditions, increasing
the duty factor will increase the
amount of power intercepted
from the input line.
4. The boost engine is essentially
lossless, so all power intercepted
from the input line ends up in
the load.
5. The load voltage is to be held
constant.
6. The load voltage shall be set
above the peak input voltage,
where the peak input voltage is
1.414 times the maximum RMS
input power-line voltage; this is
where the power factor is to be
corrected.
If the peak input voltage exceeds
the output voltage at any time, the
output diode (internal to the boost
engine) takes it upon itself to forward
conduct, transferring power from the
input to the output without any
by Dale Eagar
EMI
FILTER AC
BRIDGE BOOST
ENGINE LOAD
POWER PATH (ENGINE ROOM)
AC
POWER
LINE
0
0
10
2030 40 50 60 70 80
90
100
CONTROL ROOM
pfc4_1.eps
+
~
~
INPUT
VOLTS DUTY
FACTOR OUTPUT
VOLTS
INPUT
AMPS
00
Figure 1. Boost PFC block diagram
DESIGN FEATURES
controller to watch the input volts
and amps while adjusting the duty
factor. The duty factor is constantly
adjusted to make the input amps
equal to some constant (K) times the
input volts. Presto resistor! Looking
into the input terminals, one would
see a pure resistor. The input would
look resistive at a value of Z
IN
= 1/K.
This impedance would be fairly con-
stant for frequencies from DC to the
fastest frequency at which the con-
troller could accurately keep up with
things. In the real world there is this
guy named Nyquist who says, among
other things, that the controller would
make serious blunders if he were to
try to control the duty factor faster
than 1/2 of the switching frequency.
With the switching frequency set to
somewhere around 100kHz, the in-
put looks resistive from DC to many
kiloHertz, which is good enough for
the 60Hz line and many of its nasty
harmonics.
Careful inspection of the output
would reveal an output impedance of
V
OUT2
/K, a big number. The whole
system would look like a constant
power source, with the power level set
to K times V
IN2
. These observations
are entered in the first row of Table 1.
authorization from the control room.
Pure Insubordination! The only way
to prevent this disgraceful act by
the output diode is to keep the out-
put voltage out of reach of the peak
input voltage.
At the Console
In the control room is a control
console, a panel with meters and a
duty factor control. Sitting in the con-
trol seat is the controller, an intelligent
entity. The controller is given the task
of watching the three meters (Figure
1) and controlling the duty factor to
keep things in check.
The controller is given two tasks to
perform, namely:
1. Make the input look resistive.
2. Keep the output voltage constant
at 386V, regardless of the load
current or input line voltage.
Careful inspection of the above two
tasks may surprise you. Let’s look at
each task separately.
Task One: Making the
Input Look Resistive
It would not be too difficult to teach
the controller to make the input look
resistive. One would need to teach the
Linear Technology Magazine • December 1995
11
Task-2: Keeping the Output
Voltage Constant at 386V
Looking at task-2 in the absence of
any task-1 constraints, we teach our
controller to adjust the duty factor to
keep the output voltage constant, re-
gardless of the load current or input
line voltage. Having visualized such a
system, we can look at its character-
istics. Looking into the input port at
any constant load current, we see a
negative input impedance. (As the
input line voltage goes up, the input
current goes down.) Looking into the
output port we see an impedance of
zero. (Any finite increase in output
current causes no change in output
voltage.) Finally, looking at the power
intercepted from the input line we
have a constant at V
OUT
× I
OUT
regardless of input voltage. These ob-
servations occupy row two of Table 1.
Putting Them Together
It should be apparent that no de-
vice can exhibit positive impedance
and negative impedance at the same
time. However, we know how to
cheat—we take the reciprocal of time
to get frequency.
This trick doesn’t need smoke and
mirrors, unless, of course, you hook
it up wrong, in which case you get
seven years of bad luck. This is a trick
in the frequency domain.
The first step is to loosen up the
output specification a little. We can
see that the power delivered from a
sine wave into a resistive load is any-
thing but constant (see Figure 2). We
at low input currents where input
ripple is less critical.
The boost converter PFC provides
no isolation from the line, has high
inrush current, has a very high out-
put voltage, has 120Hz ripple on the
output and exhibits slow load-step
response. All of these difficulties of
the boost converter are small com-
pared to the problems in the design of
an input EMI filter if other technolo-
gies were used.
Isolation, hold-up time and regula-
tion are performed by the switching
power supply that follows the PFC
conditioner. Inrush protection is pro-
vided by a negative tempco thermistor
or other external circuitry.
It is advantageous to synchronize
the PFC conditioner and the following
switching power supply 180 degrees
out of phase to reduce the ripple
current on the output capacitor and
reduce overall system noise. The Lin-
ear Technology PFC family of parts all
have synchronizing ability, and the
LT1508 and LT1509 controllers have
both a PFC conditioner and a switch-
ing power converter on one chip, fully
synchronized at 180 degrees.
The input impedance of a function-
ing PFC will be negative from DC to
about 10Hz; from there the imped-
ance goes reactive, then to positive
resistance at frequencies above 10Hz.
The output impedance of the PFC
conditioner is near zero ohms from
DC to about 10Hz; above 10Hz the
output impedance looks like a vari-
able resistor in parallel with the output
capacitor’s reactance.
Table 1. Task-1 and task-2 characteristics
Input Output Power
Task impedance impedance transferred
Make Input
Resistive 1/K V
2OUT
/K K V
IN
Constant V
OUT
Negative Zero V
OUT
× I
OUT
continued on page 25
DESIGN FEATURES
would have to have an infinitely large
output capacitor to realize zero out-
put ripple when there is input ripple
current. Infinite capacitors went out
of vogue back in the Cretaceous era.
First we loosen up the output voltage
specification to allow a few volts of
ripple, bringing the output capacitor
down to a realistic size. Second, we
further loosen the output specifica-
tion to allow the output voltage to
swing tens of volts during heavy load
steps.
We do this by teaching the controller
to keep the output voltage at 386V
when averaged over any 3-second
interval. This is in effect telling the
controller to make the input imped-
ance negative for frequencies from DC
up to several Hertz. We are instructing
the controller to adjust the K value in
the task-1 operation slowly to keep the
output voltage at an average of 386V.
See Figure 3 for the input impedance
versus frequency characteristics.
Summary/Hindsight
Power factor correction is incorpo-
rated into a product design for one of
several reasons:
A. Government standards, such as
IEC555, compliance with which
is required to sell product in
several worldwide markets.
B. The marketing boys upstairs feel
that they can sell it as a feature.
C. Management needs to keep the
power supply design group out of
trouble until the next major
design.
The boost converter is the best way
to perform PFC at any power greater
than 50W. The crowning advantage of
the boost topology is its continuous
input current when running in the
continuous mode. The loss of
continuity of input current in the
discontinuous mode is unimportant,
as discontinuous mode only happens
VOLTS
WATTS
AMPS
INPUT
POWER
INPUT
VOLTS
AND
AMPS
TIME
Figure 2. Input power waveform
POSITIVE
R
IN
NEGATIVE
NOTE: WHEN THE
INPUT RESISTANCE
IS AT ZERO OHMS
THE INPUT STILL
HAS IMPEDANCE
BECAUSE AT THAT
POINT THE INPUT
LOOKS REACTIVE.
FREQUENCY
0
0
10Hz
pfc4_3.eps
Figure 3. Input impedance (resistive
component) versus frequency
12
Linear Technology Magazine • December 1995
FREQUENCY (Hz)
0.0001
0.001
0.010
0.1
THD+N (%)
20K
1352_2.eps
100 1K 10K
A
V
= 1
A
V
= –1
How Do You Slew 200V/µs
with a 250µA Op Amp? by George Feliz
Introduction
The LT1354/LT1365 family of low
power, high-slew-rate op amps
changed the relationship between slew
rate and supply current found in tra-
ditional voltage-feedback operational
amplifiers. These parts slew from
400V/µs with 1mA supply current up
to 1,000V/µs with 6mA supply cur-
rent. The LT1352 dual op amp is a
precocious new sibling that extracts
200V/µs from a mere 250µA of supply
current (see Figure 1).
The new design shares the family
circuit topology, but adds a frugal
output stage that enthusiastically
drives heavy loads with up to 40mA of
current. A primary benefit, unheard
of with other low power op amps, is
the ability to drive large signals into
heavy loads with low distortion. A plot
of distortion for 20V peak-to-peak
signals with a 1k load is shown in
Figure 2. Table 1 shows that this
remarkable output performance is
achieved without compromising other
specifications over its 2.5V to 15V
supply voltage range. The settling time
for a 10V step is only 700ns for 0.1%
and 1250ns for 0.01%. In addition to
performance, the LT1352 is a user-
friendly C-Load op amp, which is
stable with any capacitive load.
output swing and/or create instabil-
ity with a capacitive load.
The solution used on the LT1352
was to create a pair of composite
transistors formed by transistors
Q19–Q21 and Q22–Q24. The current
mirrors attached to the collectors of
emitter followers Q19 and Q22 pro-
vide additional current gain. The ratio
of transistor geometries Q20 to Q21
and Q23 to Q24, and resistor ratios
R2 to R3 and R4 to R5 set the gain to
five at low currents and to a maxi-
mum of 23 at high currents. The
variation in gain is required in order
to have low quiescent current yet be
able to provide output drive on de-
mand. There is no output-swing loss
with this output stage, as the swing is
limited at the high impedance node.
The dynamics of a composite are not
as benign as an emitter follower, so
compensation is required and is pro-
vided by R6, C1 and R7, C2.
The C-Load stability is provided by
the R
C
, C
C
network between the out-
put stage and the high impedance
node. When the amplifier is driving a
light or moderate load the output can
follow the high impedance node and
the network is bootstrapped and has
no effect. When driving a heavy load
such as a big capacitor or small-value
How We Did It
A simplified schematic of the cir-
cuit is shown in Figure 3. 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 resistor R1. 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, generat-
ing currents in Q3 and Q4, which are
mirrored by Q9–Q11 and Q13–Q15
into the high impedance node. Tran-
sistors Q17–Q24 form the output
stage. Bandwidth is set by R1, the
g
m
’s of Q3, Q4, Q7 and Q8, and the
compensation capacitor C
T
.
The current available to slew C
T
is
proportional to the voltage that ap-
pears across R1. This method of “slew
boost” achieves low distortion due to
its inherent linearity with input step
size. Since large slew currents can be
generated without increasing quies-
cent current, C
T
can be increased and
R1 can be decreased. Lowering R1
reduces the input noise voltage to
14nV/Hz and helps reduce input
offset voltage and drift.
The output stage buffers the high
impedance node from the load by
providing current gain. The LT1352
and its relatives are single-gain-stage
amplifiers in order to remain stable
with capacitive loading and to achieve
their high slew rate with low quies-
cent supply current. The simplest
output stage would be two pairs of
complementary emitter-followers,
which would provide a current gain of
Beta
NPN
× Beta
PNP
. Unfortunately, this
gain is insufficient for driving even a
2k load when running at the low
current levels of the LT1352. Adding
another emitter-follower can reduce
DESIGN FEATURES
Figure 2. THD + noise, 20V
P–P
, R
L
= 1k
Figure 1. Large-signal response, A
v
=1
Linear Technology Magazine • December 1995
13
OUT
+IN
–IN
V
+
V
R1
1k C
C
R
C
C
T
C
1
C
2
R6
R7
R2
R4 R5
R3
Q14
Q23
Q1
Q2
Q3
Q4
Q7
Q8
Q6
Q5
Q10
Q9
Q22
Q20
Q12
Q21
Q17
Q13
Q18
Q19
Q16
Q24
Q15
Q11
1352_3.eps
Table 1. Important specifications for the LT1352 at 25
°
C
Parameter Conditions Minimum Typical Maximum Units
VOS VS = 5V, ±15V 300 600 µV
VS = ±2.5V 400 800 µV
IBVS = ±2.5V to ±15V 20 50 nA
IOS VS = ±2.5V to ±15V 5 15 nA
AVOL VS = ±15V, 1k20 60 V/mV
VS = ±5V, 1k20 60 V/mV
VS = ±2.5V, 5k20 60 V/mV
GBW VS = ±15V 2.2 3 MHz
VS = ±5V 2.0 2.7 MHz
VS = ±2.5V 2.5 MHz
SR VS = ±15V 125 200 V/µs
VS = ±5V 40 50 V/µs
tSETTLE 10V step, 0.1% 700 ns
10V step, 0.01% 1250 ns
CMRR Vs = ±15V 80 97 dB
VS = ±5V 78 84 dB
VS = ±2.5V 68 75 dB
PSRR VS = ±2.5V to ±15V 90 106 dB
Noise voltage VS = ±2.5V to ±15V,10kHz 14 nV/
Hz
Noise current VS = ±2.5V to ±15V,10kHz 0.5 pA/
Hz
Output Swing VS = ±15V, 1k12.0 13.0 V
VS = ±5V, 5003.1 3.5 V
VS = ±2.5V, 2k1.3 1.7 V
ISUPPLY VS = ±15V 250 325 µA
VS = ±5V 220 290 µA
Figure 3. LT1352 simplified schematic
resistor, the network is incompletely
bootstrapped and adds to the com-
pensation at the high impedance node.
The added capacitance provided by
C
C
slows down the amplifier and the
zero created by R
C
adds phase margin,
ensuring stability. The small-signal
DESIGN FEATURES
Figure 4. Small-signal step response, A
v
=1,
C
L
= 500 pF
Figure 5. Large-signal step response, A
v
= 1,
C
L
= 10,000 pF
response with a 500pF load is shown
in Figure 4 and the large-signal re-
sponse with 10,000pF is shown in
Figure 5. Note that the slew rate in
Figure 5 is limited to 4V/µs by the
short-circuit current limit of 40mA.
Conclusion
The LT1352 is a dual op amp with
an unmatched combination of low
supply current, large-signal perfor-
mance, stability and DC precision. If
a figure of merit equal to the slew rate
divided by supply current is calcu-
lated, the LT1352 stands alone among
voltage-feedback op amps with
800V/µs-mA.
14
Linear Technology Magazine • December 1995
1160_2.eps
+
V
OUT
I
OUT
C
LOAD
=
1000 - 9000pF
1160_1.eps
V+
V+BOOST
IN TOP
DC-100kHz
IN BOTTOM
DC-100kHz
UV OUT
BOTTOM
UVLOCK
TOP
GATE
DR
BOTTOM
GATE
DR
CBOOTSTRAP
DRIVE
FEEDBACK
TO 
LOAD
SOURCE
DRIVE
FEEDBACK
HV
TOP
UVLOCK
+
+
CONTROL
LOGIC
Figure 2a. Bipolar driver topology
The best solution to this problem is
to use a floating topside N-channel
gate driver operating from a boot-
strap capacitor. Although widely used,
the technique of bootstrapping must
be carefully implemented, since
the gate voltage applied to the top
MOSFET derives directly from the
voltage on the bootstrap capacitor. If
this voltage becomes too low, it can
cause problems for the MOSFET. For
example, if at high duty cycles the
output is not swinging low enough to
fully recharge the capacitor, the
topside gate drive will be starved,
leading to overdissipation.
Life Above the Supply Rail
The first problem designers face in
N-channel half-bridge circuits is de-
veloping the topside gate drive, which
must swing at least 10V above the
supply rail for standard MOSFETs.
This challenge alone is so demanding
that designers often resort to more
expensive and less efficient P-chan-
nel MOSFETs on the topside in low
voltage bridge circuits. However, even
this solution is no longer straight-
forward when the supply voltage
exceeds the maximum gate-to-source
voltage (V
GS
) rating for the MOSFET
(typically 20V).
Figure 1. LT1160 block diagram with pins shown in actual package sequence
LT1160, continued from page 1
SWITCHING TIME (ns)
0
0V
V
OUT
15V
600 800 1000 1200 1400 1600400200
1000pF 9000pF
1000pF 9000pF
SWITCHING TIME (ns)
0
–1.2A
I
OUT
0A
1.2A
1160_2b.eps
600 800 1000 1200 1400 1600400200
1000pF
1000pF
9000pF
9000pF
Figure 2b. Bipolar driver switching time increases only 150ns for an increase of 8,000pF in
load capacitance
DESIGN FEATURES
Linear Technology Magazine • December 1995
15
In the LT1160, a floating under-
voltage detection circuit monitors the
voltage across the bootstrap capaci-
tor and a separate undervoltage
detection circuit monitors the supply
voltage. If, as a result of operation at
high duty cycles, the voltage across
the bootstrap capacitor drops below
approximately 8.7V, the top driver is
pulled low to prevent overdissipating
the top MOSFET, and also to allow
the capacitor to be recharged. Simi-
larly, if the supply voltage drops below
about 8.3V, both drivers are pulled
low to prevent overdissipating either
MOSFET. The LT1162 has separate
and independent undervoltage detec-
tors for the two half-bridge drivers.
Shoot-Through Unwelcome
Probably the most frustrating ele-
ment to address in a synchronous
The most common techniques rely
on fixed delay times to create a dead-
time between conduction of the top
and bottom MOSFETs. Although this
can work, the delay time must take
into account temperature changes,
supply variations, and production tol-
erances of the MOSFETs and other
MOSFET drive circuit is the timing
between the top and bottom drives to
prevent cross-conduction. The
presence of cross-conduction or
“shoot-through” currents always saps
efficiency and erodes precious ther-
mal margins; in extreme cases it can
lead to catastrophic failure.
Figure 4. 50W high efficiency switching regulator illustrates the design ease afforded by adaptive deadtime generation
1160_3a.eps
V
IN
V
OUT
I
OUT
C
LOAD
=
1000 - 9000pF
+
Figure 3a. CMOS driver topology
1000pF
SWITCHING TIME (ns)
0
0V
V
OUT
15V
600 800 1000 1200 1400 1600400200
9000pF
1000pF
9000pF
SWITCHING TIME (ns)
0
–1.2A
I
OUT
0A
1.2A
1160_3b.eps
600 800 1000 1200 1400 1600400200
1000pF
1000pF
9000pF
9000pF
Figure 3b. CMOS driver switching time increases 400ns for an increase of 8,000pF in load
capacitance
1160_4.eps
1
2
3
4
5
6
7
1
2
3
4
5
6
7
8
16
15
14
13
12
11
10
9
14
1N4148
13
12
11
10
9
(2) IRFZ44
IRFZ34
L
47µHR
S
7m
5400µF
LOW ESR
BOTH GNDS
JOINED AT
ONE POINT
5V
MBR340
8
12V
60V
(MAX)
L: HURRICANE LABS HL-KM147U 801 6352003
R
S
: DALE TYPE LVR-3 ULTRONIX RCS01
+
+
+
2.2µF
1µF
4700pF
f = 40kHz
+
18k
0.1µF
0.1µF
2 x 2200µF
LOW ESR
10µF
6800pF
100pF
25k
1k
500k
2k
10k
18k
10k
1k
2N2222
5k
330k
++
+
1N4148
1N4148
LT1846
BOOST
T GATE DR
T GATE FB
T SOURCE
P V
+
B GATE DR
B GATE FB
S V
+
IN TOP
IN BOT
UV OUT
S GND
P GND
NC
LT1160
DESIGN FEATURES
16
Linear Technology Magazine • December 1995
losses. Figure 5 shows the operating
efficiency for Figure 4’s circuit.
Two Halves Equal a Whole
The LT1162 is functionally and
electrically identical to two LT1160s.
The LT1162 is ideal for motor-control
applications driving an N-channel
MOSFET H-bridge, as shown in Fig-
ure 6. The DC motor shown can be
driven in both directions. By applying
the appropriate control signals at the
inputs, the motor can be made to run
clockwise, counterclockwise, stop rap-
idly (“plugging” braking) or free run
(coast) to a stop. A very rapid stop can
be achieved by reversing the current,
though this requires more careful
design to stop the motor dead. In
practice, a closed-loop control sys-
tem with tachometric feedback is
usually necessary.
Designing In Ruggedness
The output of a bridge circuit driv-
ing multiple-ampere inductive loads
can be a very ugly signal to couple
back into a hapless IC. Although
MOSFETs contain integral body di-
odes which conduct during the
dead-time, slow turn-on times and
wiring inductances can result in
spikes of several volts below ground
Figure 6. Full-bridge motor control with the LT1162
components. As a result, more dead-
time than is normally needed must be
used, increasing the conduction time
in the lossy MOSFET body diodes.
The LT1160 and LT1162 (like the
LT1158) use an adaptive system that
maintains dead-time independent of
the type, the size and even the num-
ber of MOSFETs being driven. It does
this by monitoring the MOSFET gate
turn-off to see that it has fully dis-
charged before allowing the opposite
MOSFET to turn on. In this way,
cross-conduction is completely elimi-
nated as a design constraint.
The high efficiency 10A stepdown
(buck) switching regulator shown in
Figure 4 illustrates how different sized
MOSFETs can be used without hav-
ing to worry about shoot-through
currents. Since 40V is being dropped
down to 5V, the duty cycle for the
switch (top MOSFET) is only 5/40 or
12.5%. This means that the bottom
MOSFET will dominate the R
DS(ON)
efficiency losses, because it is turned
on seven times as long as the top.
Therefore a smaller MOSFET is used
on the top, and the bottom MOSFET
is doubled up, all without having to
worry about dead-time. The noncriti-
cal Schottky diode across the bottom
MOSFETs reduces reverse-recovery
OUTPUT CURRENT (A)
75
80
85
90
95
EFFICIENCY (%)
10
1160_5.eps
024 86
1160_6.eps
UV OUT A
BOOST
V+
60V MAX
12V
T GATE DR A
T GATE FB A
T SOURCE A
B GATE DR A
B GATE FB A
GND
IN TOP A
IN BOT A
LT1162
UV OUT B
BOOST
T GATE DR B
T GATE FB B
T SOURCE B
B GATE DR B
B GATE FB B
GND
IN TOP B
IN BOT B
+
+
10µF
IRFZ44
IRFZ44
1N4148
1000µF
PWM
0-100kHz
25k
CBOOTSTRAP
0.1µF
+
IRFZ44
IRFZ44
1N4148
1000µF
PWM
0-100kHz
25k
CBOOTSTRAP
0.1µF
V+V+V+
1 7 14 26
4
24
23
22
21
19
6
5
2
3
10
18
17
16
15
13
12
11
8
9
Figure 5. Operation efficiency for Figure 4’s
circuit. Current limit is set at 15A
or above supply. The LT1160 and
LT1162 source pins have been de-
signed to take repetitive output
transients 5V outside the supply rails
with no damage to the part. This
saves the expense of adding high cur-
rent Schottky diode clamps.
Conclusion
The LT1160 and LT1162 N-
channel power MOSFET drivers an-
ticipate all of the major pitfalls
associated with the design of high
efficiency bridge circuits. The de-
signed-in ruggedness and numerous
protection features make these driv-
ers the best solution for 10V to 60V
medium-to-high current synchronous
switching applications.
DESIGN FEATURES
Linear Technology Magazine • December 1995
17
LTC
®
1430 Provides Efficient GTL Supply
by Craig Varga
1.5V output. The initial current was
1A and the final current 5A. The
top trace is the output voltage
(100mV/div) and the bottom trace is
the regulator’s inductor current
(5A/div). In Figure 4, the load resistor
is switched from ground to a 3V
supply, resulting in bidirectional load-
ing. Current is switched from 5A to
4A. Again, the top trace is the output
of the 1.5V supply. The relatively long
recovery period is actually a response
to the perturbation that appears on
the input voltage as a result of the
large load change. Voltage mode con-
trol suffers from inherently poor line
rejection. In this case, the lab supply
used for the tests exhibits rather
poor dynamics. If a better input sup-
ply is used, the overall settling time
will be much faster. The LTC1430
also has a very accurate reference
and exhibits good static load regula-
tion, as can be seen from the very
small offset between the settled light
load and heavy load conditions on the
output-voltage trace.
A recent trend in computer bus
architecture is the move toward GTL
(Gunning Transition Logic) for rout-
ing high speed signals between the
CPU and chipset logic. Buses with
Figure 1. Schematic diagram of GTL supply
+
+
++
+
L1
2.7µH
V
CC
I
MAX
SHDN
COMP
SS
SGND
FSET
PV
CC2
1
16
3
5
7
6
13
2
14
12
8
10
9
4
11
15
G1
G2
PGND
SENSE
+SENSE
FB
I
FB
PV
CC1
U1
LTC1430
R6
10k
R2, 3
R1
1k
R3
3
R4
51
R5
16.5k
1%
R9
88.7k
1%
R7
15k
R10
3.3k R8
130k
C7
0.1µF
C8
22µF
35V
C12
330
6.3V
OSCON
C13
330
6.3V
OSCON
C4
330
6.3V
OSCON
1.5V OUT
C5
330
6.3V
OSCON
S/D
5V
C6
0.01µF
C16
0.1µF
C2
680pF
C10
1µF
C9
1µF, 25V
C11
0.068µF
C3
220pF
C1
0.1µFQ1
Si4410
Q2
Si4410
D2
D1
MBRS330T3
di1430_1.eps
330
6.3V
TANT
330
6.3V
TANT
ADDITIONAL
DECOUPLING
AT THE LOAD
+ +
MBR0530T1
DESIGN IDEAS...
LTC1430 Provides Efficient
GTL Supply ................... 17
Craig Varga
Battery Charger Sinks
Constant Current .......... 18
Mitchell Lee
Low Noise Charge Pump
Biases GaAsFET Gates .. 19
Mitchell Lee
C-Load
TM
Op Amps
Conquer Instabilities ..... 20
Frank Cox and Kevin R. Hoskins
Lowpass Filters with Rail-to-
Rail Input and Output ... 23
Philip Karantzalis and Jimmylee Lawson
speeds in excess of 66MHz are using
this technology. Several different GTL
supply voltage standards seem to be
evolving, with voltages from 1.2V to
1.5V being discussed most often. In
some implementations of the logic
family, the GTL supply must both
source and sink currents of 5A–10A.
This causes a great deal of difficulty
for a typical current mode control
power supply design that is incapable
of controlling currents less than zero.
The LTC1430 is a voltage mode con-
troller, and as such, does not exhibit
any problems with negative load cur-
rents.
Most GTL incarnations, however,
appear not to require current-sinking
supplies, and the LTC1430 is well
suited for these applications as well.
The circuit shown in Figure 1
switches at 300kHz, regulating a 5V
input down to a 1.5V output while
maintaining good efficiency. (See Fig-
ure 2. Figures 2–4 are on the bottom
of page 18.) Figure 3 shows the effects
of a fast 4A load transient on the
DESIGN IDEAS
18
Linear Technology Magazine • December 1995
Battery Charger Sinks
Constant Current by Mitchell Lee
the current available from the source.
The circuit shown in Figure 1 does
just that, while charging 2 NiCd cells.
Input current is sensed by a PNP
transistor and applied as feedback to
an LTC1174 stepdown converter. The
Burst Mode™ converter draws just
enough power to hold the input cur-
rent at 50mA, matching the abilities
of the wall adapter.
Output current into two cells is
nominally 180mA, but climbs to over
300mA when the batteries are com-
pletely discharged. Efficiency is about
78% at 3V output. If the batteries are
removed from the charging circuit,
the output voltage could climb to
levels destructive to the load. To
DC-to-DC converters make very
efficient constant current sources for
charging NiCd batteries. Unfortu-
nately, the input of a switcher exhibits
a negative impedance, and this can
cause problems in systems where the
source is power or current limited.
Various schemes are used to preclude
input foldback. Undervoltage lockout
works well with power-limited
sources, keeping the converter OFF
until the input voltage has risen to
the point where it can develop ad-
equate power.
Undervoltage lockout does not work
with current-limited sources. A bet-
ter approach is to control the
converter’s input current to match
prevent this, a second feedback
path is applied to Pin 1 via a diode,
limiting the output voltage to approxi-
mately 3.7V.
Another condition that can spell
trouble for simple circuits such as
this one is the loss of input power
while the batteries are still connected
in circuit. An extra Schottky diode
placed in series with the switch (Pin 5)
blocks reverse current flow into the
LTC1174 and prevents damage. The
output can also be short circuited
without damage to the device. Shut-
down (Pin 8) allows logic control of the
charger, so the charging current can
be interrupted without disconnecting
the source supply.
Figure 2. Efficiency plot of the GTL supply
LOAD CURRENT (A)
55
65
60
70
75
80
85
90
95
EFFICIENCY (%)
7
di1430_2.eps
0123 654
LTC1430, continued from page 17
di1430_3.eps
INDUCTOR
CURRENT
5A/DIV
OUTPUT
VOLTAGE
100mV/DIV
OUTPUT
VOLTAGE
100mV/DIV
INDUCTOR
CURRENT
5A/DIV
di1430_4.eps
Figure 4. Oscillograph showing effects of
bidirectional loading on the GTL supply
Figure 3. Oscillograph showing effects of
4A load step on the 1.5V output
Figure 1. Schematic diagram of constant-current battery charger
+
+
+ +
di1174_1.eps
50mA
INPUT
2×
100µF
16V
OS-CON
100µF
16V
OS-CON
47µF
6.3V
OS-CON
2 NiCd
CELLS
11100
5
62N3906
1
1N5818
1N5818
LOAD
100pF *COILTRONICS
(407) 241-7876
7234
8
1N4148
CTX100-1*
SHDN SW
FB
IPGM GND
V
IN
LTC1174HV-3.3
Authors can be contacted
at (408) 432-1900
DESIGN IDEAS
Linear Technology Magazine • December 1995
19
Low Noise Charge Pump Biases
GaAsFET Gates by Mitchell Lee
ure more typical of generic charge
pumps. A plot of spot noise is shown
in Figure 2. Above 20kHz, broadband
output noise is gradually eliminated
by the 10µF output capacitor.
Throughout the audio range the noise
is 600–700nV/Hz.
Figure 3 shows the spectral re-
sponse of the LTC1550 over a range of
100kHz to 10MHz. Switching noise at
900kHz and related harmonics are
less than 10µV. The ferrite bead used
during these tests attenuates har-
monics above 5MHz. Depending on
the application, a series resistor of 1
to 10 may be just as effective at
attenuating noise.
The LTC1550 is part of a larger
family of devices including multiple
versions and the LTC1551. The only
difference between the LTC1550 and
LTC1551 is shutdown polarity (the
LTC1550 has inverting shutdown, and
the LTC1551 has noninverting shut-
down). Packages with higher pin
counts include features such as ad-
justable output and a “power good”
pin for controlling power to the
GaAsFET drain. All devices exhibit
the same characteristic low output
noise. See the LTC1550/LTC1551
data sheet for a complete description
of all available options.
Figure 2. Spot noise plot of the circuit shown in Figure 1
tra, rendering the usual charge pump
unsuitable for this application.
A new generation of charge pumps
with on-chip post regulators has been
developed to satisfy this application.
Although the charge pump is retained
as the most economical and compact
method of generating a negative volt-
age from a positive input, a linear
post regulator is added to “purify” the
otherwise noisy charge-pump output.
The result is a negative bias supply so
quiet you’ll need to build a low noise
preamp in order to observe the noise
on a spectrum analyzer.
Figure 1 shows the LTC1550 low
noise, regulated, switched-capacitor
voltage inverter. Total noise over a
40MHz bandwidth is approximately
200µV
RMS
with a 1mA load. Compare
this with 10–100mV
RMS
noise, a fig-
Gallium arsenide MOSFETs are
widely used in wireless applications
as RF output amplifiers, especially in
cellular telephones. Although they
provide excellent power gain, output
power and efficiency on low voltage
battery supplies, GaAsFETs are deple-
tion-mode devices and require a
negative gate voltage for proper bias.
Since only a small current (less than
10mA) is required, a charge pump is
the best solution for generating a
negative gate-bias supply.
Charge pumps normally conjure
visions of noise, and rightly so. Their
very operation relies on transferring
energy from input to output by alter-
nately charging and discharging a
transfer capacitor—a decidedly noisy
process. Noise on the gate-bias line
translates directly to the output spec-
di1550_1.eps
FERRITE BEAD*
C1
0.1µF
C2
0.1µF
C
L
0.1µF
C
OUT
10µF
5V
4.7µF
1
2
3
4
8
7
6
5
*DALE ILB-1206-19
3900
V
OUT
= –4.1V
+
LTC1550-4.1
SHDN
V
CC
C1+
V
OUT
SENSE
CP
OUT
GND
C1–
+
Figure 1. Circuit diagram: LTC1550 low noise, regulated, switched-capacitor voltage inverter
FREQUENCY (Hz)
1n
10n
100n
1µ
10µ
100µ
NOISE (V/Hz)
100k
di1550_2.eps
1k 10k
FREQUENCY (Hz)
–40
0
–10
20
30
50
–20
–30
10
40
60
NOISE (dBµV)
10M
di1550_3.eps
1k 10k 100k 1M
BW = 100Hz
Figure 3. Spectral response of the LTC1550, 100kHz to 10MHz
DESIGN IDEAS
20
Linear Technology Magazine • December 1995
C-Load
Op Amps
Conquer Instabilities
Introduction
LTC has taken advantage of pro-
cess technology advances and circuit
innovations to create a series of
C-Load operational amplifiers. All of
these amplifiers are tolerant of ca-
pacitive loading, and a majority of
them remain stable driving any ca-
pacitive load. This series of amplifiers
has a bandwidth that ranges from
160kHz to 140MHz. These amplifiers
are appropriate for a wide range of
applications such as coaxial cable
drivers, video amplifiers, analog-to-
digital converter (ADC) input buffer/
amplifiers and digital-to-analog con-
verter (DAC) output amplifiers.
This article looks at a few of the
many applications for which the
C-Load amplifiers are appropriate,
including overcoming the large junc-
tion capacitance of surge protection
diodes on the outputs of video
amplifiers, driving the dynamically
changing capacitive loads of high
speed ADC converters’ analog inputs,
and driving an ADC’s reference with
varying voltages.
Table 1 lists LTC’s unconditionally
stable voltage-feedback C-Load
amplifiers. Table 2 lists other voltage-
feedback C-Load amplifiers that are
stable with loads up to 10,000pF.
by Frank Cox and
Kevin R. Hoskins
current of many amplifiers causes
slew limiting. The excellent output
drive (70mA minimum at ±15V
supplies) and the C-Load stability
features of the LT1363 make the
video performance of this circuit
almost indistinguishable from the
same driver without the added ESD
protection. Of course, the 75 series
matching resistor on the output
helps to isolate the capacitive load.
For the best high speed performance,
we recommend that the coax cable
be matched in this way. However,
there are cases where the maximum
output swing is required and this
resistor cannot be used (of course, for
quality video a termination resistor
is always used on the end of the
cable), so tests were run with com-
posite video to quantify the effect of
this component. Table 3 gives differ-
ential gain and phase measurements
for the amplifier circuit with and
without the output series resistor.
Video Performance
with ESD Protection
The circuit shown in Figure 1 is a
simple gain-of-2 line driver that is
capable of withstanding high levels of
electrostatic discharge (ESD) while
retaining its excellent video perfor-
mance. The output of the LT1363
amplifier is protected with transient
voltage suppressors. These suppres-
sors are Zener diodes designed for
circuit protection and specified for
high peak power dissipation, stand-
off voltage and maximum surge
current. The P6KE15A diodes used in
this circuit have a peak power dissi-
pation of 600W, a minimum stand-off
voltage of 12.8V and a maximum surge
current of 28A. This presents an ex-
traordinary level of protection for
the output of the amplifier, but the
large junction diodes present a prob-
lem for some video circuits.
The series connection of the two
P6KE15A diodes has 500pF of total
capacitance in the OFF state
(the normal operating condition).
Normally, large capacitive loads are
to be avoided in video amplifiers. Not
only can capacitive loading cause os-
cillation in amplifiers that are not
designed for it, but the limited output
Figure 1. Line driver circuit diagram
dicl_1.eps
+
+
+
15V P6KE20
4.7µF TANT
0.1µF CER
15V
1000
75
4.7µF TANT
0.1µF CER P6KE20
LT1363
–15V
–15V
1N571275
1N5712
510
510
P6KE15A ×2
OUT
IN
Table 1. Unity-gain stable C-Load amplifiers
stable with all capacitive loads
GBW IS/Amp
Singles Duals Quads (MHz) (mA)
LT1012 0.6 0.4
LT1112 LT1114 0.65 0.32
LT1097 0.7 0.35
LT1457 2 1.6
Table 2. Unity-gain stable C-Load amplifiers
stable with C
L
10,000pF
GBW IS/Amp
Singles Duals Quads (MHz) (mA)
LT1368 LT1369 0.16 0.375
LT1200 LT1201 LT1202 11 1
LT1220 45 8
LT1224 LT1208 LT1209 45 7
LT1354 LT1355 LT1356 12 1
LT1357 LT1358 LT1359 25 2
LT1360 LT1361 LT1362 50 4
LT1363 LT1364 LT1365 70 6
DESIGN IDEAS
Linear Technology Magazine • December 1995
21
of Figure 3. The LT1206 high current
CFA has no measurable performance
degradation for video signals with the
suppressor diodes on the output,
although the restriction of the input-
clamp network still holds. The 250mA
output drive (over 0°C to 70°C tem-
perature with ±5V to ±15V supplies)
capability and the C-Load feature of
the LT1206 make it a natural for
driving difficult loads.
Driving ADCs
Most contemporary analog-to-digi-
tal converters (ADCs) incorporate a
sample-and-hold (S/H) using a topol-
ogy exemplified by the circuit in Figure
4. The hold capacitor’s (C1) size var-
ies with the ADC’s resolution, but is
generally in the range of 5pF–20pF,
10pF–30pF or 10pF–50pF for 8-, 10-
or 12-bit ADCs, respectively.
At the beginning of a conversion
cycle, this circuit samples the applied
signal’s voltage magnitude and stores
it on its hold capacitor. Each time the
switch opens or closes, the amplifier
driving the S/H’s input faces a
Differential Differential
Phase Gain
LT1363 with 0.11°0.03%
output clamp and
75 output R
LT1363 with 0.11°0.3%
output clamp and no R
LT1206 with 0.1°0.1%
output clamp and
75 output R
LT1206 with 0.11°0.1%
output clamp and no R
dicl_2.eps
15V
20k
–15V
1N5819
75
1N5819
NONINVERTING
INPUT OF LT1363
>> –AV
C1
AIN
GND
dicl_4.eps
TO 
SAR
FROM INTERNAL DAC
dynamically changing capacitive load.
This condition generates current
spikes on the input signal. These
spikes, along with the capacitive load,
are a very challenging load that can
potentially produce instabilities in an
amplifier driving the ADC’s input.
These instabilities make it difficult
for an amplifier to settle quickly. If the
output of an amplifier has not settled
to a value that falls within the error
band of the ADC, conversion errors
can result. That is, unless the ampli-
fier is designed to gracefully and
accurately drive capacitive loads, as
are Linear Technology’s C-Load line
of monolithic amplifiers.
As can be seen in Figure 5a, an
amplifier whose design is not opti-
mized for handling large capacitive
loads has some trouble driving the
hold capacitor of the LTC1410’s S/H.
Although the LT1006 has other very
Except for a slight rise in the LT1363’s
differential phase, there is no mea-
surable degradation. The output and
power supply pins of the driver circuit
withstand repeated air discharges of
17kV (the limit of the test equipment)
without failure. The input is sensitive
to the loading effects of the diode and
resistor network because the junc-
tion capacitance of the clamp diodes
works against the impedance of the
cable-termination resistors. For this
reason, low capacitance clamp di-
odes were used. The input of the
driver circuit still stood off 10kV.
If some reduction in the frequency
response (approximately 3dB at
5MHz) can be tolerated, the network
shown in Figure 2 can be added to the
input to increase the standoff voltage
to 17kV.
For even more demanding applica-
tions, we recommend the driver circuit
Figure 3. 250mA output line driver
Table 3. Differential gain and phase
measurement summary
Figure 2. Circuit with modified clamp for
higher voltage Figure 4. Typical ADC input stage showing
input capacitors
CONVST
1V/DIV
INPUT
200mV/DIV
dicl_5a.eps
200ns/DIV
dicl_5b.eps
200ns/DIV
INPUT
50mV/DIV
CONVST
1V/DIV
Figure 5a. Input signal applied to an
LTC1410 driven by an LT1006
Figure 5b. Input signal applied to an
LTC1410 driven by an LT1360
dicl_3.eps
+
+
+
15V P6KE20
4.7µF TANT
0.1µF CER
0.01µF
V
+
1k
75
4.7µF TANT
0.1µF CER P6KE20
LT1206
V
–15V
1N571275
1N5712
1200
1200P6KE15A ×2
DESIGN IDEAS
22
Linear Technology Magazine • December 1995
desirable characteristics, such low
V
OS
, high slew rate and low power
dissipation, it has difficulty accurately
responding to dynamically changing
capacitive loads and the current
glitches and transients they produce,
as indicated by the instabilities that
appear in the lower oscilloscope trace
in Figure 5a.
By contrast, Figure 5b shows the
LTC1360 C-Load op amp driving the
same LTC1410 input. The photo
continued on page 25
+
+
+
+
–5V
–5V
5V
0.1µF
5V
dicl_6.eps
0.1µF
INPUT
0.1µF10µF
+A
IN
12-BIT DATA
–A
IN
AV
DD
DV
DD
V
SS
AGND
REFCOMP
V
REF
10µF
OGND CONTROL LINES
LTC1410
DGND
0.1µF
0.1µF 10µF
LT1006 OR
LT1360
CONVST
Figure 6. Test circuit used to measure LTC1410 input signal waveform
Figure 7. The 12-bit LTC8043 DAC and the C-Load LT1220 op amp create variable reference voltages, enhancing the 12-bit LTC1410 ADC’s
input range flexibility
AV
DD
DV
DD
V
SS
BUSY
CS
CONVST
RD
SHDN
NAP/SLP
OGND
D0
D1
D2
D3
+A
IN
–A
IN
V
REF
COMP
AGND
D11 (MSB)
D10
D09
D08
D07
D06
D05
D04
DGND
1
2
3
4
5
6
7
8
9
10
11
12
13
14
28
27
26
25
24
23
22
21
20
19
18
17
16
15
U4
LTC1410
dicl_7.eps
5V
–5V
–5V
C8
0.1µF
BUSY
CHIP
SELECT
CONV START
READ
SHUTDOWN
SHUTDOWN MODE
DATA BIT 0
DATA BIT 1
DATA BIT 2
DATA BIT 3
DATA BIT 4
DATA BIT 5
DATA BIT 6
DATA BIT 7
DATA BIT 8
DATA BIT 9
DATA BIT 10
DATA BIT 11
SERIAL CLOCK
SERIAL DATA
DAC LOAD
R1
5.1k
C5
0.1µF
±2.5V ANALOG
INPUT RANGE
C4
0.1µF
5V
C2
0.1µF
C3
0.1µF
C1
0.1µF
C9
0.1µF
+
+
V
DD
CLK
SRI
LD
V
REF
FB
I
OUT
GND
1
2
3
4
8
7
6
5
U2
LTC8043
U3
LT1220
U1
LT1004-2.5
C7
0.1µF
C6
10µF
2
3
6
5V
–5V
7
4
shows that the LTC1360 is an ideal
solution for driving the ADC’s input
capacitor quickly and cleanly with
excellent stability. Its wide 50MHz
gain bandwidth and 800V/µs slew
rate very adequately complement the
LTC1410’s 20MHz full-power band-
width. The LTC1360 is specified for
±5V operation.
Figure 6 shows the circuit used to
test the performance of op amps driv-
ing the LTC1410’s input and measure
the ADC’s input waveforms.
Adjustable ADC
Reference Voltage
An ADC’s reference voltage should
always be set to a value that maxi-
mizes the number of codes generated
during input signal conversion. If the
reference voltage is set for input sig-
nals with the widest full-scale range,
this might be too large for input sig-
nals with a much smaller range. For
example, an ADC’s full-scale input
DESIGN IDEAS
Linear Technology Magazine • December 1995
23
Lowpass Filters with Rail-to-Rail Input
and Output
The proliferation of systems oper-
ating exclusively with single 5V
supplies has created a need for ana-
log signal conditioning circuits with
rail-to-rail input and output. Op
amps, A/D converters, and D/A con-
verters with rail-to-rail inputs and
outputs are readily available, but not
high-order lowpass filters. High-
order lowpass filters are used to
provide antialiasing protection for
A/D converters or to bandlimit the
outputs of D/A converters.
Rail-to-Rail, DC-Accurate,
Butterworth and Bessel 7th-
Order Lowpass Filters.
Ther circuit of Figure 1 uses a
dual rail-to-rail op amp and a DC-
accurate switched-capacitor filter to
provide a 7th-order lowpass filter.
The rail-to-rail op amps accommo-
date the nonrail-to-rail output
operation of the switched capacitor
filter. Overall rail-to-rail operation is
achieved by attenuating and level
shifting the input signal and by level
shifting and amplifying the output of
the switched-capacitor filter.
difilt_1.eps
+
R1
10k R3
10k
R4
49.9k 0.1µF
R2
20k
C3*
0.039µF
IN
5V
5V
1
2
3
4
8
7
6
5
LTC1063
OR
LTC1065
R
OS
*
200k
1µF
TANT
4.99k
4.53k 5V
0.1µF
C
OS
510pF
R8
20k
R9
10k
OUT
C1
0.1µF
R10
20k
R7
20k 5V
C2
0.1µF
R6*
11.3k
R5*
5.62k
*THE VALUES FOR THESE COMPONENTS DETERMINE 
THE FILTERS –3dB FREQUENCY (SEE TABLE 1).
NOTE: ALL RESISTORS ARE 1% METAL FILM AND
ALL CAPACITORS ARE 5%.
+
1/2
LT1366
+
1/2
LT1366
The circuit of Figure 1 operates
with a single 5V supply and attenu-
ates the input signal by a factor of
two. A 0-to-5V
P–P
input signal ap-
pears as 1.25–3.75V at the input of
the filter. The attenuation factor of
two is chosen as a convenient num-
ber, and does not fully exploit the
3.5V output-voltage range of the
switched-capacitor filter. The penalty
is a mere 2dB loss of signal-to-noise
ratio, yet the circuit is still usable for
12- to 13-bit applications.
The input op amp of Figure 1
provides buffering. A single pole
antialiasing filter is added in front of
the switched capacitor filter through
capacitor C3; its cutoff frequency is
3.3 times the cutoff frequency of the
switched-capacitor filter.
The output op amp of Figure 1
provides an additional 2-pole, con-
tinuous-time filter and restores the
rail-to-rail output swing. Table 1
shows the component values of Fig-
ure 1 for three different cutoff
frequencies. The passive components
were carefully selected to work for
both filter types, Butterworth and
Bessel.
Other cutoff frequencies can be
obtained by scaling the component
values of Table 1. The 7th-order am-
plitude response of the circuit is shown
in Figure 2. Figure 2 was traced with
a 1.7V
RMS
input (4.76V
P–P
), and the
measurement bandwidth was 22kHz.
The filter cutoff frequency was set to
1kHz. Attenuation as high as 80dB is
obtained with the linear-phase
LTC1065 filter. Figure 3 shows the
dynamic range (S/N + THD) of the
filter of Figure 1.
by Philip Karantzalis and Jimmylee Lawson
FREQUENCY (Hz)
–90
–40
–50
–60
–70
–80
10
0
–10
–20
–30
GAIN (dB)
10k
difilt_2.eps
100 1k
VS = 5V
LTC1063
LTC1065
Figure 1. 100Hz 7th-Order Butterworth or Bessel lowpass filter with rail-to-rail input and output
Figure 2. Amplitude response of Figure 1’s
circuit for both LTC1063 and LTC1065
DESIGN IDEAS
24
Linear Technology Magazine • December 1995
3dB Freq C3 ROS COS R5 R6
10Hz 0.39µF 2M 510pF 56.2k 113k
100Hz 0.039µF 200k 510pF 5.62k 11.3k
1kHz 0.0039µF 19.1k 510pF 5621.13k
Table 1. Component values for Figure 1's circuit for three
different cutoff frequencies
20Hz Rail-to-Rail
Linear-Phase Lowpass
Filter with 60Hz Notch.
The circuit of Figure 5 is a 20Hz
Bessel filter with an additional 60Hz
notch. This type of circuit is useful in
applications where excessive 60Hz
noise is present and, on top of
bandlimiting, additional attenuation
of 60Hz is required. The same scheme
of rail-to-rail op amps is used for
input and output buffering, except
the output of the first op amp is fed
forward to the input of the Sallen-Key
lowpass filter. The output of the
LTC1065 is connected to the nonin-
verting input of the Sallen-Key filter
through R11. When the output of the
first op amp is 180 degrees out of
phase from the output of the LTC1065,
a notch depth of 50dB or greater can
be achieved. In order to achieve this
notch depth for 60Hz, the resistors
R5, R6, R11 and R
OS
must be 0.1%
and the C
OS
capacitor must be 1%
tolerance. Figure 6 shows the gain
response of the circuit in Figure 5.
The 76dB maximum dynamic range
occurs around 4.75V
P–P
and this
should be the full scale of the system.
Figure 4 shows the transient response
of both Butterworth and Bessel 7th-
order filters.
The measured DC gain error of
Figure 1’s circuit is less than 0.5%.
This depends on the matching of re-
sistors R1–R4 and R7–R10. The
measured DC gain nonlinearity of
this circuit is 0.1% (–80dB).
difilt_5.eps
+
1
2
3
4
8
7
6
5
R1
10k R3
10k
R4
49.9k 0.1µF
R2
20k
C3
0.039µF
IN
5V
5V
LTC1065
R
OS
200k
0.1%
1µF
TANT
4.99k
4.53k 5V
0.1µF
C
OS
510pF
1%
R8
20k
R9
10k
OUT
C1
0.1µF
R10
20k
R7
20k 5V
C2
0.1µF
R6
150k
0.1%
R5
49.9k
0.1%
NOTE: ALL RESISTORS ARE 1% METAL FILM AND
ALL CAPACITORS ARE 5% UNLESS SPECIFIED
OTHERWISE.
+
1/2
LT1366
+
1/2
LT1366
R11
113k
0.1%
Figure 5. Schematic diagram of 20Hz Bessel lowpass filter with 60Hz notch
Figure 3. Dynamic range (THD + Noise) versus
input voltage plot for the circuit of Figure 1 Figure 6. Amplitude response of the circuit
detailed in Figure 5
Figure 4. Oscillograph of square wave
response; top trace is the LTC1065 Bessel
filter, bottom trace is the LTC1063 Butter-
worth filter
BUTTERWORTH
BESSEL
INPUT VOLTAGE (VP-P)
RLOAD = 100k
f–3db = 100Hz
fin = 50Hz
0.5
–90
THD + NOISE (dB)
–40
–45
5.55.0
difilt_3.eps
4.54.03.53.02.52.01.0 1.5
–80
–85
–70
–75
–60
–65
–50
–55
RLOAD = 5k
VS = 5V
FREQUENCY (Hz)
–90
–40
–50
–60
–70
–80
10
0
–10
–20
–30
GAIN (dB)
1k
difilt_6.eps
10 100
DESIGN IDEAS
Linear Technology Magazine • December 1995
25
C-load, continued from page 22
span is set to 5V. Input signals with
magnitudes that cover this span will
use all of the ADC’s codes. However,
an input signal that only spans 2V
will use only 2/5 of the codes, result-
ing in a loss of resolution.
When the reference voltage is re-
duced to 2V, the 2V signal will now
use all of the ADC’s codes. But with
the 2V reference, an input signal
having a 4V span is likely to exceed
the input span, resulting in clipped
signals and missed information. In
cases such as this, an adjustable
reference is needed to adjust the
ADC’s full-scale gain.
input bypass capacitor C5. The
LTC1410 can easily handle a refer-
ence voltage as low as 1V (±2V
P-P
input range) with only a 0.5 LSB
change in linearity error.
Conclusion
Using advanced process technol-
ogy and innovative circuit design,
LTC has created a series of C-Load
operational amplifiers that not only
tolerate capacitive loading, but re-
main stable driving any capacitive
load. C-Load amplifiers meet the
challenging and difficult capacitive
loads by remaining stable and
settling quickly.
The circuit in Figure 7 uses the
LTC8043 12-bit CMOS 4-quadrant
DAC driving the LT1220 C-Load
amplifier to generate variable ADC
reference voltages that allow the user
to adjust the ADC’s gain. An LT1004-
2.5 micropower 2.5V reference is used
as the LTC8043’s reference source,
setting its output span to 0V–2.5V.
The LT1220’s output is applied to the
LTC1410’s V
REF
input. The LT1220
C-Load amplifier ensures that, when
changing the LTC1410’s reference
voltage, the amplifier does not oscil-
late even while driving the reference
PFC, continued from page 11 (Design Features)
probe to see the PFC in action. Scope
probes don’t read duty factor very
well, and probing on the FET drain
reveals a waveform that doesn’t look
like anything you have ever seen be-
fore, and may awaken you in the
middle of the night. It’s not that the
waveform isn’t pretty, it’s that the
waveform is very busy, and it is hard
to see anything like a sine wave in it.
The input current waveform is far
easier to enjoy, it’s just harder to get
to without a good current probe.
Conclusion
PFC is fun. It offers technical chal-
lenges not because it is complicated
but because there are so many little
weirdnesses, all happening together
in concert to provide a beautiful func-
tion. One of the biggest challenges is
to find a place to hook your scope
CONTINUATIONS
Authors can be contacted
at (408) 432-1900
Included with this issue is a
complete index of the circuits
featured in Linear Technology
Magazine, issues I:1 through V:4.
This index evolved as a result of a
request from one of our Field
Applications Engineers, who was
unable to locate a circuit he
believed had appeared in Linear
Technology Magazine. We set out
to create an index in which the
reader could “browse” for a circuit
that fit his or her design needs or
search for a specific Linear
Technology part; we believe that
the table design we used here
thoroughly fulfills this mission. The
reader can page to the desired
circuit category, scan the columns
for the circuit type and/or part
number, and easily find the
appropriate article, issue, page
number, and figure number(s).
Over 250 circuits from the first
five years of Linear Technology
magazine are included in this
index. Hereafter, we will update the
index yearly, so new circuits won’t
get lost. Pleasant Reading!
Index of Circuits in Linear Technology, Issues I:1 Through V:4
26
Linear Technology Magazine • December 1995
New Device Cameos
LTC1409 and LTC1415:
High Speed, Low Power,
12-bit A/D Converters
LTC has two new 12-bit A/D con-
verters that set a new standard for
high performance at low power. The
LTC1415 is a 12-bit, 1.2Msps A/D
converter that operates on a single
5V supply and draws just 60mW of
power. The LTC1409 is a 12-bit,
800ksps A/D converter with out-
standing dynamic performance
(SINAD = 71dB at Nyquist) that oper-
ates on a ±5V supply.
The LTC1409 and LTC1415 both
have differential input sample-and-
hold circuits that reduce unwanted
high frequency common mode noise.
The LTC1415 has a unipolar input
range of 0V to 4.096V (1mV per LSB).
The LTC1409 has a bipolar input
range of ±2.5V. The input ranges of
both devices are set with an on-chip
2.5V precision voltage reference.
Both devices have parallel output
interfaces that allow easy connection
to microprocessors, DSPs or FIFOs.
Two power-shutdown modes allow
flexibility in further reducing the
power during inactive periods. Both
devices come in 28-pin SO packages.
The LTC1446/LTC1447:
World’s Only Dual
12-Bit, Rail-to-Rail DACs
in SO-8 Packages
The LTC1446 and LTC1447 are
complete dual, 12-bit micropower
DACs with rail-to-rail buffer amplifi-
ers and an onboard reference. Both
parts have a maximum DNL error of
0.5LSB. They have an easy-to-use
cascadable serial interface that is SPI
compatible. Applications for these
parts include industrial process
control, digital calibration, automatic
test equipment, cellular telephones
and portable battery-powered
applications where low supply
current is essential.
The LTC1446 operates on a single
supply ranging from 2.7V to 5.5V,
dissipating a mere 1.35mW from a 3V
supply. It has a wide output swing of
0V to 2.5V. The LTC1447 operates on
a single supply of 4.5V to 5.5V, dissi-
pating 3.5mW. It has an output swing
of 0V to 4.095V. Both parts have an
onboard reference that is connected
internally to each DAC. A D
OUT
pin
allows several parts to be daisy-
chained, and a built-in power-on reset
clears the output to zero-scale at
power-up.
Both the LTC1446 and the
LTC1447 come in an SO-8 package or
a small 8-pin PDIP, allowing efficient
use of circuit board space. These dual
DACs offer excellent DNL, low power
and a convenient serial interface, all
in a tiny SO-8 package.
LT1464: First Micropower
JFET Op Amp for Driving
C-Loads
The LT1464 op amp combines fea-
tures that would normally be unique
to three different types of op amps.
The LT1464 is the first micropower
JFET op amp in our C-Load family of
op amps. The output can drive ca-
pacitive loads up to 10nF with any
load resistance. At the input side of
the op amp, input bias currents are
less than 1pA for demanding track-
and-hold circuits and active filters
that use high-value resistors. Input
common mode range includes the
positive supply. Power supply cur-
rent is typically 250µA, allowing
micropowered battery applications.
For single supply requirements, a full
set of specifications are included for
5V operation. The slew rate is 1V/µs
and the gain bandwidth is 1MHz.
The LTC1458 and LTC1459:
Quad 12-Bit, Rail-to-Rail
Micropower DACs
The LTC1458 and the LTC1459
are complete quad 12-bit micropower
DACs with rail-to-rail buffer amplifi-
ers and an onboard reference. Both of
these parts have a maximum DNL
error of 0.5LSB, and an easy-to-use,
SPI compatible, cascadable serial in-
terface. They also offer several
convenient features that make these
parts flexible and easy to use. Indus-
trial process control, automatic test
equipment and digital calibration are
just a few of the applications for
these parts.
The LTC1458 operates on a single
supply ranging from 2.7V to 5.5V and
dissipates a mere 2.5mW from a 3V
supply. It has a 1.22V onboard refer-
ence. The LTC1459 operates on a
single supply of 4.5V to 5.5V, dissi-
pating 6mW. Its onboard reference is
2.048V. Both of these parts have a
separate reference input pin for each
DAC, which can be driven by an ex-
ternal reference or connected to the
onboard reference output. The user
can also offset the zero-scale voltage
for each DAC by means of a REF
LO
pin. Each of the rail-to-rail buffer
amplifiers will swing to within a few
millivolts of either supply rail and can
be connected in a gain-of-one or a
gain-of-two configuration, giving an
output full-scale of 1×REF
IN
or
2×REF
IN
respectively. There is a built-
in power on reset and a CLR pin that
resets the output to zero-scale.
Both the LTC1458 and the
LTC1459 come in small 28-pin SSOP
or 28-pin SO packages. These quad
DACs offer the user flexibility with a
host of useful features, excellent DNL
and low power dissipation.
LT1537 5V Powered
RS232 Transceiver
The LT1537 is a new three-driver/
five-receiver RS232 interface trans-
ceiver designed to serve as a serial I/O
port for AT-compatible computers.
The LT1537 is pin compatible with
the LT1137A. Its integral charge-
pump power generator uses low
cost 0.1µF and 0.2µF capacitors to
NEW DEVICE CAMEOS
Linear Technology Magazine • December 1995
27
NEW DEVICE CAMEOS
generate RS232 levels from standard
5V power supplies. The device has
low power consumption: 8mA in
normal operation and 1µA in shut-
down mode.
The LT1537 is RS232/RS562 stan-
dard compatible. Operation at data
rates in excess of 250kBd for 1,000pF
loads and 120kBd for 2,500pF loads
is guaranteed. Slew rate with 3k,
2500pF loads is a minimum of 4V/µs.
When powered down or in shutdown
mode, the driver outputs remain
high impedance for line voltages to
15V. Both driver outputs and receiver
inputs can be forced to ±25V
without damage.
Whereas most other RS232 trans-
ceivers require external ESD
protection devices, the LT1537 is
protected against ±5kV ESD strikes
to the RS232 inputs and outputs.
This integrated ESD protection saves
the expense and space of external
protection devices. This level of ESD
protection is adequate in most appli-
cations. In applications where high
ESD level protection (IEC-801-2) is
required, our LT1137A provides that
solution. The LT1537 is available in
28-lead SO and SSOP packages.
The LTC1449/1450 Parallel
Input, 12-Bit Rail-to-Rail
Voltage Output Micropower
DACs
The LTC1449 and LTC1450 are
LTC’s first pair of parallel-input, volt-
age output DACs. These devices are
designed as complete DACs with in-
ternal references and buffered, true
rail-to-rail voltage outputs. The ana-
log specifications are equivalent to
those of the LTC1451 serial DAC,
with excellent DNL performance of
less than ±0.5LSBs.
The buffered true rail-to-rail volt-
age output can be configured for two
different full-scale settings. When the
internal reference is used and the
gain-setting pin is tied to ground, the
full-scale range is 2.500V for the
LTC1449 and 4.095V for the LTC1450.
When the gain-setting pin is con-
nected to V
OUT
, the full-scale ranges
The LTC1516 can be substituted
for the MAX619 with improved per-
formance. The part is available in
8-pin PDIP and 8-pin SO packages.
LT1307 1-Cell
Micropower 675kHz
PWM DC/DC Converter
The LT1307 is a micropower, fixed
frequency DC/DC converter that op-
erates from an input voltage as low as
1V. The first device to achieve true
PWM performance from a 1-cell sup-
ply, the LT1307 automatically shifts
to power-saving Burst Mode opera-
tion at light loads. High efficiency is
maintained over a broad 100µA to
100mA load range. The device in-
cludes a low-battery detector with a
200mV reference, and consumes less
than 5µA in shutdown mode. No-load
quiescent current is 50µA, and the
internal NPN power switch handles a
400mA current with a voltage drop of
just 240mV.
Unlike competing devices, the
LT1307 does not require large elec-
trolytic capacitors. Its high frequency
(675kHz) switching allows the use of
tiny surface mount multilayer ceramic
(MLC) capacitors and small surface
mount inductors. It works with just
10µF of output capacitance and re-
quires only 1µF of input bypassing.
The LT1307 is available in the
SO-8 package.
change to 1.220V for the LTC1449
and 2.047V for the LTC1450.
The LTC1449 and LTC1450 are
designed to run off a single supply
while consuming very little power.
The LTC1449 can be powered from a
single supply as low as 2.7V; when
powered from such a supply, it dissi-
pates only 0.75mW. The LTC1450 is
powered from a single 5V supply and
dissipates only 2mW of power. This
low power dissipation makes the
LTC1449 and LTC1450 ideal for bat-
tery-powered applications.
Separate pins are provided for both
ends of the voltage-scaling resistor
network to allow for versatile
connection in multiplying and
external-reference applications.
Double-buffered parallel digital in-
puts provide flexibility and easy
interface to popular DSPs and micro-
processors.
The LTC1449 and LTC1450 are
available in 24-lead PDIP and SSOP
packages.
The LTC1516: Micropower,
Regulated, 5V Charge-Pump
DC/DC Converter
The LTC1516 is a micropower
charge-pump DC/DC converter that
generates a regulated 5V output from
a 2V to 5V supply, without inductors.
Only four external capacitors (two
0.22µF and two 10µF) are required to
deliver a 5V ±4% output with up to
50mA of output load current. Ex-
tremely low supply current (12µA
typical with no load and less than 1µA
in shutdown conditions) and low ex-
ternal parts count make the LTC1516
ideal for small, battery-powered ap-
plications. The LTC1516 achieves
efficiency greater than 75% with load
currents as low as 100µA, due to its
ultralow quiescent current.
To further improve overall effi-
ciency, the LTC1516 operates as either
a doubler or tripler, depending on V
IN
and V
OUT
load conditions. The part
also has thermal shutdown protec-
tion and can survive an indefinite
short from V
OUT
to GND. In shutdown
conditions, the load is disconnected
from V
IN
.
Burst Mode, Bits-to-Nits and C-Load are
trademarks of Linear Technology Corporation.
, LTC and LT are registered trademarks
used only to identify products of Linear Tech-
nology Corp. Other product names may be
trademarks of the companies that manufac-
ture the products.
Information furnished by Linear
Technology Corporation is believed to be
accurate and reliable. However, Linear
Technology makes no representation that
the circuits described herein will not infringe
on existing patent rights.
For further information on the
above or any of the other devices
mentioned in this issue of Linear
Technology, use the reader service
card or call the LTC literature ser-
vice number: 1-800-4-LINEAR. Ask
for the pertinent data sheets and
application notes.
28
Linear Technology Magazine • December 1995
AppleTalk
is a registered trademark of Apple Computer, Inc.
© 1995 Linear Technology Corporation/ Printed in U.S.A./32K
LINEAR TECHNOLOGY CORPORATION
1630 McCarthy Boulevard
Milpitas, CA 95035-7417
(408) 432-1900
For Literature Only: 1-800-4-LINEAR
DESIGN TOOLS
Applications on Disk
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
PSPICE
TM
by MicroSim. Available at no charge.
Technical Books
1990 Linear Databook • Volume I — 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 commer-
cial 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
1994 Linear Databook • Volume III — This 1826 page supplement to the
1990 Linear Databook
and
1992 Linear Databook Supplement
is a collection
of all products introduced since 1992. A total of 152 product data sheets are
included with updated selection guides. The
1994 Linear Databook Volume III
is a supplement to the 1990 and 1992 Databooks, which should not be
discarded.
$10.00
1995 Linear Databook • Volume IV — This 1152 page supplement to the
1990, 1992 and 1994 Linear Databooks
is a collection of all products introduced
since 1994. A total of 80 product data sheets are included with updated
selection guides. The
1995 Linear Databook Vol IV
is a companion to the
1990,
1992 and 1994 Linear Databooks
, which should not be discarded. $10.00
Linear Applications Handbook • Volume I — 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, systems-oriented 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
Interface Product Handbook
— This 424 page handbook features LTC’s
complete line of line driver and receiver products for RS232, RS485,
RS423, RS422, V.35 and AppleTalk
applications. Linear’s particular
expertise in this area involves low power consumption, high numbers of
drivers and receivers in one package, mixed RS232 and RS485 devices, 10kV
ESD protection of RS232 devices and surface mount packages.
Available at no charge.
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
1995 Power Solutions Brochure,
First Edition
— This 64 page collection
of circuits contains real-life solutions for common power supply design
problems. There are over 45 circuits, including descriptions, graphs and
performance specifications. Topics covered include PCMCIA power manage-
ment, microprocessor power supplies, portable equipment power supplies,
micropower DC/DC, step-up and step-down switching regulators, off-line
switching regulators, linear regulators and switched capacitor conversion.
Available at no charge.
FRANCE
Linear Technology S.A.R.L.
Immeuble “Le Quartz”
58 Chemin de la Justice
92290 Chatenay Malabry
France
Phone: 33-1-41079555
FAX: 33-1-46314613
GERMANY
Linear Techonolgy GmbH
Oskar-Messter-Str. 24
85737 Ismaning
Germany
Phone: 49-89-962455-0
FAX: 49-89-963147
JAPAN
Linear Technology KK
5F NAO Bldg.
1-14 Shin-Ogawa-cho Shinjuku-ku
Tokyo, 162 Japan
Phone: 81-3-3267-7891
FAX: 81-3-3267-8510
KOREA
Linear Technology Korea Co., Ltd
Namsong Building, #403
Itaewon-Dong 260-199
Yongsan-Ku, Seoul 140-200
Korea
Phone: 82-2-792-1617
FAX: 82-2-792-1619
SINGAPORE
Linear Technology Pte. Ltd.
507 Yishun Industrial Park A
Singapore 2776
Phone: 65-753-2692
FAX: 65-754-4113
TAIWAN
Linear Technology Corporation
Rm. 602, 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: 44-1276-677676
FAX: 44-1276-64851
Linear Technology Corporation
1630 McCarthy Boulevard
Milpitas, CA 95035-7417
Phone: (408) 432-1900
FAX: (408) 434-0507
International
Sales Offices
World Headquarters
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
3220 Tillman Drive, Suite 120
Bensalem, PA 19020
Phone: (215) 638-9667
FAX: (215) 638-9764
Linear Technology Corporation
266 Lowell St., Suite B-8
Wilmington, MA 01887
Phone: (508) 658-3881
FAX: (508) 658-2701
NORTHWEST REGION
Linear Technology Corporation
782 Sycamore Dr.
Milpitas, CA 95035
Phone: (408) 428-2050
FAX: (408) 432-6331
SOUTHEAST REGION
Linear Technology Corporation
17000 Dallas Parkway
Suite 219
Dallas, TX 75248
Phone: (214) 733-3071
FAX: (214) 380-5138
Linear Technology Corporation
5510 Six Forks Road
Suite 102
Raleigh, NC 27609
Phone: (919) 870-7834
FAX: (919) 870-8831
SOUTHWEST REGION
Linear Technology Corporation
21243 Ventura Blvd.
Suite 227
Woodland Hills, CA 91364
Phone: (818) 703-0835
FAX: (818) 703-0517
Linear Technology Corporation
15375 Barranca Parkway
Suite A-211
Irvine, CA 92718
Phone: (714) 453-4650
FAX: (714) 453-4765