V05N3 - AUGUST

LINEAR TECHNOLOGY
AUGUST 1995
IN THIS ISSUE . . .
COVER ARTICLE
The LT ®1304: Micropower DC/DC
Converter with Independent
Low-Battery Detector ................ 1
Steve Pietkiewicz
Editor's Page ............................ 2
Richard Markell
VOLUME V NUMBER 3
The LT1304: Micropower
DC/DC Converter
with Independent
Low-Battery Detector
LTC in the News ....................... 2
DESIGN FEATURES
LT1510 Battery Charger IC
Supports All Battery Types,
Including Lithium-Ion ............... 3
Chia Wei Liao
LT1118 Terminates
Active SCSI and More ................ 5
Gary Maulding
LT1580 Low Dropout Regulator
Uses New Approach to
Achieve High Performance ........ 7
Craig Varga
Power Factor Correction —
Part Three: Discontinuity ......... 9
Dale Eagar
The New LT1508/LT1509 Combines
Power Factor Correction and a
PWM in a Single Package ........ 11
Kurk Mathews
LTC ®1329 Micropower,
8-Bit, Current Output DAC ...... 16
K.S. Yap
DESIGN INFORMATION
RS232 Transceiver Protected to
±15kV IEC-801-2 ESD Standards
............................................... 20
Gary Maulding
V.35 Transceivers Allow
3-chip V.35 Port Solution ....... 22
Y.K. Sim
DESIGN IDEAS ................. 24–37
by Steve Pietkiewicz
Introduction
In the expanding world of low power
portable electronics, a 2- or 3-cell
battery remains a popular power
source. Designers have many options
for converting the 2V-4V battery voltage to 5V, 3.3V, and other required
system voltages using low voltage
DC/DC converter ICs. The LT1304
offers users a micropower step-up
DC/DC converter featuring Burst
ModeTM operation and a low-battery
detector that stays alive when the
converter is shut down. The device
consumes only 125µA when active,
yet can deliver 5V at up to 200mA
from a 2V input. High frequency
operation up to 300kHz allows the
use of tiny surface mount inductors
and capacitors. When the device is
shut down, the low-battery detector
draws only 10µA. An efficient internal
power NPN switch handles 1A switch
current with a drop of 500mV. Up to
85% efficiency is obtainable in
2-cell-to-5V converter applications.
The fixed-output LT1304-5 and
LT1304-3.3 versions have internal
resistor dividers that set the output
voltage to 5V or 3.3V, respectively.
(complete list on page 24)
Operation
New Device Cameos ................ 38
The LT1304’s operation can best
be understood by examining the
block diagram in Figure 1 (page 17).
Comparator A1 monitors the output
voltage via resistor-divider string R3/
R4 at the FB pin. When VFB is higher
than the 1.24V reference, A2 and the
timers are turned off. Only the reference, A1, and A3 consume current,
typically 120µA. As VFB drops below
1.24V plus A1’s hysteresis (about
6mV), A1 enables the rest of the circuit. Power switch Q1 is then cycled
on for 6µs, or until current comparator A2 turns off the on-timer,
whichever comes first. Off-time is fixed
at approximately 1.5µs. Q1’s switching causes current to alternately build
up in inductor L1 and discharge into
output capacitor C2 via D1, increasing the output voltage. As V FB
increases enough to overcome C1’s
hysteresis, switching action ceases.
C2 is left to supply current to the
load until VOUT decreases enough to
force A1’s output high, and the entire
cycle repeats.
If switch current reaches 1A, causing A2 to trip, switch on-time is
reduced. This allows continuousmode operation during bursts. A2
monitors the voltage across 7.2Ω
resistor R1, which is directly related
to the switch current. Q2’s collector
current is set by the emitter-area
ratio to 0.5% of Q1’s collector
, LTC and LT are registered trademarks of Linear Technology Corporation.
Burst Mode and Super Burst are trademarks of Linear Technology Corporation.
continued on page 17
EDITOR'S PAGE
Eggs on the Wall?
I was in the lunch room several
weeks ago calmly eating my lunch,
when suddenly a large “boom”
sounded behind me. (No, it wasn’t an
underrated tantalum capacitor.) I
turned around (slowly) to find that
one of our physicists had tried to
hard boil an egg in the microwave
without first putting a vent hole in
the eggshell. My first reaction was to
exclaim that any physicist should
know that when most things get hot,
they expand (go boom!). Sure, we
could plot the size of the vent hole
versus the “time to go boom.” The
conservative engineer might do this.
Both the analytical approach and trying things to see what happens are
useful, but maybe we need more
“booms.” It seems that, in today’s
business atmosphere, there is too
much conservatism and too few
people just trying things to see what
happens. At LTC we try to encourage
both approaches.
The featured article for this issue
highlights the LT1304 micropower
DC/DC converter, which includes a
low-battery detector that stays alive
when the converter is shut down. The
LT1304 is a micropower step-up
DC/DC converter tailored for operation from two or three cells and
featuring Burst Mode operation.
The device consumes only 125µ A
when operating and can deliver 5V at
up to 200mA from a 2V input. The
device features 300kHz operation,
allowing the use of tiny surface mount
inductors.
A second article features the
LT1118 family of fixed voltage regulators that can both source and sink
current. These devices are ideal for
use as active SCSI terminators. The
LT1118 family of regulators features
excellent stability and load transient
settling with most capacitive loads.
The devices can source up to 800mA
and sink up to 400mA. Quiescent
current is only 600µA unloaded.
The LT1510 is a battery charger IC
that allows fast charging of all types
of batteries, including Li-Ion. The
2
LT1510 allows users to precisely set
battery current and voltage, and to
choose the best charging algorithm
for a given application. For many
situations, such as Li-Ion, no external monitoring of the battery is
required. The LT1510 is a current
mode PWM battery charger that
allows constant-current and/or constant-voltage charging. The internal
switch is capable of delivering 1.5A
DC current (2A peak current).
Another featured product is the
LT1508/LT1509 “combo” power
factor correction circuit and PWM
controller in a single package. The
L T1508/LT1509 integrates the
functionality of an LT1248 power
factor controller with a 50% maximum duty cycle PWM in a single
20-pin DIP or SOIC.
“Power Factor Correction: Part
Three,” continues the series that began two issues ago. This installment
discusses discontinuity, and the grief
it causes for those trying to understand the operation of the PFC.
The LTC1329, a micropower, 8-bit,
current output DAC, is the subject of
a short article. The LTC1329 is designed to digitally control a power
supply’s output voltage or to serve as
a trimmer pot replacement. This is a
versatile part with a 1, 2 or 3 wire
interface. Additional information on
applying LTC’s DAC product line will
be found in a feature on 12-bit voltage
mode DACs. This article highlights
circuits ranging from a computercontrolled 4-to-20mA current loop to
an opto-isolated serial interface.
In our “Design Information” section, we present information on
electrostatic discharge (ESD) and the
IEC-801-2 specification. Our popular
RS232 transceiver, the LT1137A, has
been upgraded to include on-chip
ESD protection compliant with level
four of the IEC-801-2 standard. This
upgrade allows full ESD protection
with no energy absorbing components
external to the LT1137A.
Our “Design Ideas” section is chock
full of useful circuits. These include
by Richard Markell
circuits for driving diode-ring mixers,
a fully differential 8-channel, 12-bit
A/D system, and a circuit for
touchscreen interface.
LTC in the News...
Linear Technology reports record
annual and quarter sales!!!
Thanks to all of you for your hard
work and constant drive to make Linear
Technology the best analog company
world wide, the company’s annual
sales once again increased to a record
$265,023,000 for fiscal year end July
2, 1995, an increase of 32% over net
sales from the previous year. The
company also reported record net
income for the year of $84,696,000 or
$2.22 per share, an increase of 49%
over the $56,827,000, or $1.51 per
share, reported for fiscal 1994.
Linear Technology also announced
that the board of directors approved
both an increase in the quarterly
dividend to $0.08 from $0.07 per
share and a two-for-one stock
split of its common stock. A cash
dividend will be paid on August 23,
1995 to shareholders of record on
August 4, 1995. Shareholders of
record on August 11, 1995 will be
issued certificates reflecting the stock
split; these shares will be distributed
on September 1, 1995.
Chief Executive Magazine’s May,
1995 issue features the prestigious
‘Charting The Chief Executive 100
Index.’ Robert Swanson Jr. is among
the top ten Chief Executives for a third
year in a row. In order to make the
CE100, the following criteria had to
be met. Each CEO had to have been in
their position throughout 1992-1994,
the company had to be public for that
time frame and the market value of the
firms public stock had to exceed $500
million at the end of 1994.
Financial World’s July 18, 1995
issue included Linear Technology
Corporation on its list of “America’s 50
Best Mid-Cap Companies.” LTC was
ranked number 5 up from 21 the
previous year.
Investor’s Business Daily June
28,1995 issue featured an interview
with Paul Coghlan, titled ANALOG
LIFE, Linear Technology Keeps Pulse
Of Light, Heat, Sound, Speed. Mr.
Coghlan attributes the company’s
success to being in the right market at
the right time with a good strategy.
Linear Technology Magazine • August 1995
DESIGN FEATURES
LT1510 Battery Charger
IC Supports All Battery Types,
Including Lithium-Ion
by Chia Wei Liao
Introduction
The LT1510 is a current mode PWM
step-down (buck) switcher (see Figure 1). The battery DC charging
current is programmed by a resistor
RPROG (or a DAC output current) at
the PROG pin. Amplifier CA1 converts the charging current through
RS1 to a much lower current (IPROG =
500µA/A), which is then fed into the
PROG pin. Amplifier CA2 compares
the output of CA1 with the programmed current and drives the PWM
loop to force them to be equal. High
DC accuracy is achieved with averaging capacitor CPROG. Note that IPROG
has both AC and DC components.
IPROG flows through R1 and generates
a ramp signal that is fed to the PWM
200kHz
OSCILLATOR
+
SHUTDOWN
0.7V
+
Circuit Operation
VCC
–
VSW
S
BOOST
–
VCC
R
+
QSW
R
SLOPE COMPENSATION
SW
+
1.5V
GND
PWM
VBAT
C1
–
+
B1
R2
R1
1k
R3
+
IPROG
+
VC
RS1
–
BAT
+
0VP
IPROG
= 500µA/A
IBAT
VA
CA2
+
60k
SENSE
IBAT
CA1
–
Linear Technology Magazine • August 1995
The LT1510 can charge batteries
ranging from 1V to 20V; groundsensing of current is not required.
The battery negative terminal can be
tied directly to ground. A saturating
switch running at 200kHz gives high
charging efficiency and small inductor size. A blocking diode between
the chip and the battery is not required because the chip goes into
sleep mode and drains only 3µA when
the wall-adapter is unplugged. Softstart and shutdown features are also
provided. The LT1510S8 comes in an
8-pin SOIC and the LT1510S16 comes
in a 16-pin, fused-lead, power SO
package with a thermal resistance of
45°C/W.
–
A sure way to start an argument
among electronic engineers is to bring
up the subject of battery charging.
This relatively simple problem has
spawned more solutions than the
proverbial mouse trap. This situation
seems to be the result of misinformation about the basic electrochemistry
of the batteries themselves, the difficulty of quantifying results, and the
fertile imaginations of engineers who
like to create. At times, battery charging seems to have as much to do with
religion as with science.
The LT1510 was created with this
situation in mind. It is a basic building block for transferring energy
efficiently and accurately from one
source to another. Battery current
and voltage are precisely defined by
the user, who can decide which charging algorithm is best and set up the
LT1510 in whatever programmed
current or voltage combination is
required. For many situations, especially lithium-ion, no extra external
battery monitoring is needed.
The LT1510 current mode PWM
battery charger is the simplest, most
efficient solution to fast charge modern rechargeable batteries, including
lithium-ion, nickel-metal-hydride
(NiMH), and nickel-cadmium (NiCd),
that require constant-current and/
or constant-voltage charging. The internal switch is capable of delivering
1.5A DC current (2A peak current).
The onboard current sense resistor
(0.1Ω) makes programming the charging current very simple. One resistor
(or a programming current from a
DAC) is required to set the full charging current (1.5A max.) to within 5%
or the trickle-charge current (150mA)
to 10% accuracy. With 1% referencevoltage accuracy, the LT1510S16
meets the critical constant-voltage
charging requirement for lithium cells.
gm = 0.64
VREF
PROG
CPROG
RPROG
IPROG
Ω
–
VREF
2.465V
CHARGING CURRENT IBAT
= (IPROG)(2000)
1510 BD
( )
= 2.465V (2000)
RPROG
Figure 1. LT1510 block diagram
3
DESIGN FEATURES
control comparator C1 through buffer
B1 and level-shift resistors R2 and
R3, forming the current mode inner
loop. The BOOST pin drives the NPN
switch (QSW) into saturation and reduces power loss. For batteries such
as lithium, which require both constant-current and constant-voltage
charging, the 1% 2.5V reference and
the amplifier VA reduce the charging
current when battery voltage reaches
the preset level. For NiMH and NiCd,
VA can be used for overvoltage protection. When an input voltage is not
present, the charger goes into low
current (3µA typical) sleep mode as
input drops to 0.7V below the battery
voltage. To shut down the charger,
pull the VC pin low with a transistor.
Typical Charging Algorithms
The following algorithms are
representative of current techniques:
Lithium-Ion — charge at constant
voltage with current limiting set to
protect components and to avoid overloading the charging source. When
the battery voltage reaches the programmed voltage limit, current will
automatically decay to float levels.
The accuracy of the float voltage is
critical for long battery life. Be aware
that lithium ion batteries in series
suffer from “walk away,” because of
the required constant-float-voltage
charging technique. “Walk away” is a
condition where the batteries in the
series string wind up in different states
of the charge/discharge cycle. They
may need to be balanced by redistrib-
30µH
VCC
SW
10µF
C1
0.22µF
BOOST
0.1µF
GND
BAT
+
R2
5.6K
1K
+
–
ON: IBAT = 1A
OFF: IBAT = 0.1A
IBAT
2V TO 20V
1510_2. eps
Figure 2. Charging NiCd or NiMH batteries
uting charge from one battery to another. This phenomenon is minimized
by carefully matching the batteries
within a pack.
Nickel-Cadmium — charge at a
constant current, determined by the
power available or by a maximum
specified by the manufacturer. Monitor battery charge state using voltage
change with time (dV/dt), second derivative of voltage (d2V/dt), battery
pressure, or some combination of
these parameters. When the battery
approaches full charge, reduce the
charging current to a value (top-off)
that can be maintained for a long time
without harming the battery. After
the top-off period, usually set by a
simple time out, reduce the current
further to a trickle level that can be
maintained indefinitely, typically 1/10
to 1/20 of the battery capacity.
Nickel-Metal-Hydride — same as
NiCd, except that some NiMH batteries cannot tolerate a continuous low
level trickle charge. Instead, they
require a pulsed current of moderate
VCC
+
11V TO 25V
DC WALL
ADAPTOR
1µF
D1
1N914
0.1µF
300Ω
3.83kΩ
1K
VC
BAT
+
+
22µF
4.2V
Q3**
VN2222
–
+
4.2V
–
**OPTIONAL
Figure 3. Charging Li-Ion batteries
4
R1
50K
Q1
VN2222
22µF
BOOST
PROG
LT1510S16
GND
SENSE
300Ω
VC
SENSE
10µF
OVP
1µF
LT1510S8
D1
1N914
1N5819
C1
0.22µF
WALL
ADAPTOR
PROG
30µH
SW
+
1N5819
R3
59K R4
25K
1510_3. eps
+
–
value, with a low duty cycle, so that
the average current represents a
trickle level. A typical scenario would
be one second “on” and thirty seconds “off,” with the current set to
thirty times desired trickle level.
Recharging NiMH
or NiCd Batteries
The circuit in Figure 2 will charge
battery cells with voltages up to 20V
at a full charge current of 1A (when
Q1 is on) and a trickle charge current
of 100mA (when Q1 is off). If the third
charging level is needed, simply add
a resistor and a switch. The basic
formula for charging current is:
2.465
× 2000 ( when Q1 on)
R1 R2
2.465
× 2000 ( when Q1 off )
R1
For NiMH batteries, a pulsed trickle
charge can be easily implemented
with a switch in series with R1; switch
Q1 at the desired rate and duty cycle.
If a micro-controller is used to control
the charging, connect the DAC current sink output to the PROG pin.
Recharging Lithium-Ion
Batteries
The circuit in Figure 3 will charge
lithium ion batteries at a constant
current of 1.5A until battery voltage
reaches 8.4V, set by R3 and R4. It
then goes into constant-voltage charging and the current slowly tapers off
to zero. Q3 can be added to disconnect R3 and R4 so they will not drain
the battery when the wall-adapter is
unplugged.
Linear Technology Magazine • August 1995
DESIGN FEATURES
LT1118 Terminates Active SCSI
and More
by Gary Maulding
The new LT1118 family of positive,
fixed voltage regulators is unique in
maintaining output voltage regulation while both sinking and sourcing
current. Conventional positive voltage regulators lose regulation if the
load circuit attempts to return current to the regulator output. Capable
of sourcing up to 800mA and sinking
400mA, LT1118 family regulators
feature excellent stability and loadtransient settling with any capacitive
load greater than 0.22µF. Their insensitivity to capacitive loading makes
the regulators easy to use in any
application within their power level.
Quiescent current is only 600µA when
unloaded. 5V, 2.5V, and 2.85V versions of the LT1118 are available in
3-pin SOT-223 or fused-lead SO-8
packages. The SO-8-packaged devices
include an enable pin that allows
the device to be shut down with the
output at high impedance. Quiescent
current drops to zero when the enable
pin is low.
The ability to maintain regulation
while both sourcing and sinking current is valuable in several regulator
applications. Appropriate applications are data-bus termination,
power-supply splitting, and any application where fast settling of large
load transients is needed. The 2.85V
version is ideal for use as an active
termination for SCSI cables. The 2.5V
version makes a convenient low power
supply splitter in 5V systems. The 5V
version is ideal where fast settling of
load transients is required. The shutdown pin on the SO-8-packaged
devices is valuable for power management and supply sequencing.
Active SCSI Terminator
The SCSI parallel bus requires cable
terminations at each end of the cable.
The termination serves the dual purposes of impedance matching and
providing the pull-up to the 2.8V
negated logic level. The open-drain or
open-collector drivers on the bus establish the low level and make
multiplexing of equipment simple,
due to the “wired-OR” nature of the
bus signals.
Passive terminators consisting of a
resistor divider from the termpower
supply (Figure 1) were commonly used
on early SCSI networks. Passive terminators have mostly disappeared
from use in recent systems due to the
high power dissipation and lack of
power supply rejection of the resistor
divider. The Boulay or active termination has replaced the passive
terminations in today’s equipment. A
Boulay active termination uses a
2.85V voltage regulator to establish
the negated logic level and 110Ω resistors to provide the impedance
match to each data and control line
(Figure 2). Power consumption when
all lines are in the high (negated)
state is dominated by the low quiescent current of the regulator, a 0.225W
saving over resistor terminators in a
27-line wide SCSI bus.
SCSI-2 introduced the use of
active-negation drivers in SCSI bus
systems. Active-negation drivers have
three states: active (low), negated
(high), and high impedance. By raising the data line above the 2.85V
termination voltage when negated,
the effects of cable reflections and
crosstalk between lines are overcome,
improving the noise margin in high
speed data busses. Active negation
may be used on all but three SCSI
data and control lines (three lines,
BSY, SEL, and RST, must be wireOR’d and cannot use active negation
drivers). Each negated driver can
source up to 7mA when in the high
state. Most voltage regulators will
lose regulation if any current is returned to their outputs, causing the
termination voltage to rise.
The LT1118-2.85, with current
sinking, is the perfect solution for
this application. The 800mA sourcing
capability and 400mA current-sink
capability provide ample margin to
terminate a 27-line wide SCSI bus
in the fully asserted or negated state.
Even though SCSI standards restrict
the active negation drivers to a 3.24V
and 7mA negated signal level, many
driver circuits in use exceed the allowed SCSI drive levels. If the
maximum of 24 lines are negated in a
Wide-SCSI bus, 168mA of current is
forced back into the terminator. The
400mA sink capability of the LT1118
gives more than twice the required
TERMPOWER
5V
TERMPOWER
5V
220Ω
ONE PER
BUS LINE
0.1µF
VIN
18 OR
27 LINES
VOUT
2.85V
REGULATOR
1µF
330Ω
1118_2.eps
1118_1.eps
Authors can be contacted
at (408) 432-1900
Figure 1. SCSI passive terminator
Linear Technology Magazine • August 1995
Figure 2. Boulay active terminator
5
DESIGN FEATURES
LT1118-2.85
OUT
IN
5V
GND
+
ILOAD
1µF
CERAMIC
2.2µF
1118_3.eps
Figure 3. LT1118 typical application
current handling without damage or
loss of regulation.
The fast settling time and stability
of the LT1118 allow the regulator
output capacitor to be reduced from a
typical 10µF to a surface mount 1µF
ceramic capacitor saving board
space and cost. The low, 600µA
quiescent current of the regulator
allows the terminator to be connected
to the SCSI term power line when not
in use.
The maximum voltage spikes for
Figure 3’s circuit are less than 0.3V
and settle out in 5µs. These tests were
under maximum load transient conditions: 800mA sourcing to 400mA
sinking. This transient is far more
severe than would be seen in a SCSI
bus termination, yet the voltage
output is always within acceptable
SCSI voltage levels. The amplitude
of the output voltage spikes may
be reduced further by use of larger
output capacitors.
Several competitors offer active
SCSI terminators that use on-chip
resistors. These terminators, while
offering some convenience, are limited in performance and are not as
economical a solution as the LT11182.85. Capacitance on the output pins
of the terminators cause impedance
mismatches at high frequencies,
which increase cable reflections and
degrade the noise margin of the bus
signals. The capacitance inevitably
comes from the pin-to-pin capacitance of the IC package, the ESD
protection devices on the chip, and
the distributed capacitance of the diffused resistors on the chip. The
reduced noise margin of the SCSI bus
increases error rates and limits the
length of the bus and the number of
nodes that can reliably be connected.
The external surface mount resistors
used with the LT1118-2.85 reduce
the capacitance to the stray PC board
capacitance.
Power consumption further limits
the capabilities of SCSI terminators
with on-chip resistors. The worstcase power consumption for a SCSI
terminator occurs when all data and
control lines are active (low). For a
27-line wide SCSI bus, the regulator
dissipates 1.2W and the resistors dissipate 2.0W. The surface mount IC
packages used for SCSI terminators
cannot dissipate this total power of
3.2W. This fundamental limit restricts
the devices with on-chip resistors to
nine lines. Three terminator chips
with integral resistors must be used,
versus one LT1118-2.85 with inexpensive surface mount resistors. The
cost of three SCSI terminator ICs
with integral resistors is more than
three times the cost of an LT11182.85 and the external 110Ω resistors.
Power Supply Splitter
Often a system has only a single 5V
power supply, but has analog sections that require positive and negative
supplies. The current source and sink
capabilities of the LT1118-2.5 are
ideal for generating a virtual ground
5V
VOUT
VIN
0.1µF
LT1118-2.5
2.5V
VIRTUAL
GROUND
1µF
GND
1118_4.eps
Figure 4. Supply splitter
to allow the use of analog components without a negative supply.
Figure 4 shows the LT1118-2.5 used
as a 5V supply splitter. The LT1118
can absorb large DC or transient
ground currents that would force high
power consumption if a resistive supply splitter were used. The regulator’s
fast settling time allows the use of
small capacitors, and its stability with
any large capacitive load makes it
easy to apply to almost any splitter
application. Competing supply splitters are known to suffer from stability
problems with capacitive loading.
Power Management
The enable control on the SO-8
packaged LT1118 provides easy power
management solutions. A logic low
on the enable pin causes the
regulator’s output to go high impedance and its quiescent current to
drop to zero. A logic high brings the
regulator into normal regulation.
Intelligent power controllers can
make use of this capability to selectively power subcircuits as needed
(see Figure 5). The small output capacitor requirements of the LT1118
minimize turn-on time and the supply current spikes that result from
charging large filter capacitors.
continued on page 15
+V
IN
0.1µF
OFF
0.22µF
LT1118-5
10k
ON
VOUT 1
5V
(FIRST ON/LAST OFF)
OUT
EN
0.1µF
IN
GND
10k
OUT
0.1µF
LT1118-5
EN
VOUT 2
5V
0.22µF (LAST ON
FIRST OFF)
GND
0.01µF
1118_5.eps
Figure 5. Power sequencing using the LT1118
6
Linear Technology Magazine • August 1995
DESIGN FEATURES
LT1580 Low Dropout Regulator
Uses New Approach to
Achieve High Performance by Craig Varga
The LT1580 Brings
Many New Features
Introduction
Low dropout regulators have become more common in desktop
computer systems as microprocessor
manufacturers have moved away from
5V-only CPUs. A wide range of supply
requirements exist today, with new
voltages just over the horizon. In many
cases the input-output differential is
very small, effectively disqualifying
many of the low dropout regulators
on the market today. Several manufacturers have chosen to achieve lower
dropout by using PNP-based regulators. The drawbacks of this approach
include much larger die size, inferior
line rejection, and poor transient
response.
Enter the LT1580
The new LT1580 NPN regulator is
designed to make use of the higher
supply voltages already present in
most systems. The higher voltage
source is used to provide power for
the control circuitry and supply the
drive current to the NPN output transistor. This allows the NPN to be driven
into saturation, thereby reducing the
dropout voltage by a VBE compared to
a conventional design. Applications
for the LT1580 include 3.3V to 2.5V
conversion with a 5V control supply,
5V to 4.2V conversion with a 12V
control supply, or 5V to 3.6V conversion with a 12V control supply. It is
easy to obtain dropout voltages as
low as 0.4V at 4A, along with excellent static and dynamic specifications.
The LT1580 is capable of 7A maximum, with approximately 0.8V
input-to-output differential. The
current requirement for the control
voltage source is approximately 1/50
of the output load current, or about
140mA for a 7A load. The LT1580
presents no supply-sequencing issues. If the control voltage comes up
Linear Technology Magazine • August 1995
first, the regulator will not try to supply the full load demand from this
source. The control voltage must be
at least 1V greater than the output to
obtain optimum performance. For
adjustable regulators the adjust-pin
current is approximately 60µA and
varies directly with absolute temperature. In fixed regulators the ground
pin current is about 10mA and stays
essentially constant as a function of
load. Transient response performance
is similar to that of the LT1584
fast-transient-response regulator.
Maximum input voltage from the
main power source is 7V, and the
absolute maximum control voltage is
14V. The part is fully protected from
overcurrent and over-temperature
conditions. Both fixed voltage and
adjustable versions are available. The
adjustables are packaged in 5-pin
TO-220s, whereas the fixed voltage
parts are 7-pin TO-220s.
Why so many pins? The LT1580
includes several innovative features
that require additional pins. Both the
fixed and adjustable versions have
remote-sense pins, permitting very
accurate regulation of output voltage
at the load, where it counts, rather
than at the regulator. As a result, the
typical load regulation over a range of
100mA to 7A with a 2.5V output is
approximately 1mV. The sense pin
and the control-voltage pin, plus the
conventional three pins of an LDO
regulator, give a pin count of five for
the adjustable design. The fixedvoltage part adds a ground pin for
the bottom of the internal feedback
divider, bringing the pin count to six.
The seventh pin is a no-connect.
Note that the adjust pin is brought
out even on the fixed-voltage parts.
This allows the user to greatly
improve the dynamic response of the
regulator by bypassing the feedback
5V
5
3.3V
VCONT
VIN
U1 SENSE
LT1580
ADJ
VOUT
4
1
VOUT = 2.5V
3
VCC
2
C3
22µF
25V
+
+
C2
220µF
10V
C4
0.33µF
R1
110Ω
1%
C1
R2
110Ω 100µF
10V
1%
+
100µF
10V
2X
+
1µF
25V
10X
MICROPROCESSOR
SOCKET
VSS
RTN
1580_1.eps
Figure 1. LT1580 delivers 2.5V from 3.3V at up to 6A
7
DESIGN FEATURES
temperature, guaranteed, while
operating with an input/output differential of well under 1V.
In some cases a higher supply voltage for the control voltage will not be
available. If the control pin is tied to
the main supply, the regulator will
still function as a conventional LDO
and offer a dropout specification
approximately 70mV better than
conventional NPN-based LDOs. This
is the result of eliminating the voltage
drop of the on-die connection to the
control circuit that exists in older
designs. This connection is now made
externally, on the PC board, using
much larger conductors than are
possible on the die.
50mV/DIV
2A/DIV
50µs/DIV
Figure 2. Transient response of Figure 1’s
circuit with adjust-pin bypass capacitor. Load
step is from 200mA to 4A
divider with a capacitor. In the past,
using a fixed regulator meant suffering a loss of performance due to lack
of such a bypass. A capacitor value of
0.1µF to approximately 1µF will generally provide optimum transient
response. The value chosen depends
on the amount of output capacitance
in the system.
In addition to the enhancements
already mentioned, the reference
accuracy has been improved by a
factor of two, with a guaranteed 0.5%
tolerance. Temperature drift is also
very well controlled. When combined
with ratiometrically accurate internal divider resistors, the part can
easily hold 1% output accuracy over
Circuit Examples
Figure 1 shows a circuit designed
to deliver 2.5V from a 3.3V source
with 5V available for the control voltage. Figure 2 shows the response to a
load step of 200mA to 4.0A. The circuit is configured with a 0.33µF
adjust-pin bypass capacitor. The performance without this capacitor is
shown in Figure 3. This difference in
performance is the reason for providing the adjust pin on the fixed-voltage
50mV/DIV
2A/DIV
50µs/DIV
Figure 3. Transient response without adjustpin bypass capacitor. Otherwise, conditions
are the same as in Figure 2
devices. A substantial savings in
expensive output decoupling capacitance may be realized by adding a
small “1206-case” ceramic capacitor
at this pin.
Figure 4 shows an example of a
circuit with shutdown capability. By
switching the control voltage rather
than the main supply, the transistor
providing the switch function needs
only a small fraction of the current
handling ability that it would need if
it was switching the main supply.
Also, in most applications, it is not
necessary to hold the voltage drop
across the controlling switch to a very
low level to maintain low dropout
performance.
Q1
Si9407DY
5V
R3
10k
5
3.3V
VCONT
VIN
U1 SENSE
LT1580
SHUTDOWN
ADJ
VOUT
4
1
VOUT = 2.5V
3
2
C3
22µF
25V
+
+
C2
220µF
10V
R1
110Ω
1%
LOAD
C4
0.33µF
C1
R2
110Ω 100µF
10V
1%
+
RTN
1580_4.eps
Figure 4. Small FET adds shutdown capability to LT1580 circuit
8
Linear Technology Magazine • August 1995
DESIGN FEATURES
Power Factor Correction —
Part Three: Discontinuity
In Part One of this series (in Linear
Technology V:1, pp. 17–18, 21), we
investigated power factor correction
(PFC) by looking at its line-frequency
voltage, current, and power waveforms. We developed the concept of
a DC variac and demonstrated its
utility for performing PFC.
In Part Two (LT V:2, pp. 3–7), we
developed the boost converter and
investigated its properties when operating in the continuous-mode. We
showed that a continuous-mode boost
converter does a pretty good job of
emulating the DC variac.
Part Three continues our investigation of the boost converter by
focusing on its properties when operating in the discontinuous-mode. We
will see that a discontinuous-mode
boost converter doesn’t look anything
like a DC variac.
There is elegance and simplicity in
the concept of the DC variac. There is
beauty in its application to PFC, a
beauty that is manifest in the portions of line cycle where the boost
converter is operating in the continuous-mode.
Insubordination
Wreaks Havoc
In analog engineering, we like to
view things as continuous and
smooth, with all nonlinearities well
defined. We look at an AC sine wave
and see the voltage go from positive,
through zero, to negative with no
discontinuity. This explains a good
portion of the genuine perplexity experienced by the author when the
bank practices alchemy on perfectly
legitimate checks—changing the
paper into rubber. One would think
L1
I1
E
SW1
OFF
ON
LOAD
pfc3_1a.eps
Figure 1a. The boost converter from heaven
Linear Technology Magazine • August 1995
the bank would know that it is only
two days until payday and the check
book is still half full.
Research into the character of
banking employees reveals the startling truth that the local bank is
staffed by the genetic precursor of the
digital engineer, a human so vile, so
base, as to actually enjoy inserting
discontinuities into an otherwise
continuous system. In the boost converter it is the output diode that acts
as the banker.
The boost converter shown in Figure 1a is the boost converter from
heaven. Its glory is directly attributed
to its having a switch, rather than a
vile diode, in its output. The switch
allows current flow in both directions, and thus inserts no additional
discontinuities. The boost converter
from heaven always operates in the
continuous-mode. Back here on earth
we must live with discontinuities.
In a boost converter operating in
continuous-mode, the volt-seconds,
and thus the current in the inductor,
is entirely controlled by:
1. The input voltage
2. The output voltage
3. The duty factor.
The boost converter of Figure 1b is
the mortal version. The inductor current (I1) in the boost converter, driven
by the three factors above, goes to
zero when the input line voltage
swings through zero. Figure 2 shows
that the input line current is actually
the average of the inductor current
over a full switching-frequency cycle.
Figure 2a is the inductor current when
the input line voltage has reached its
peak with the average value representing the line current at the plug.
(The ripple is filtered out by the input
EMI filter.) Similarly, Figure 2b shows
the inductor current and input current when the instantaneous input
line voltage is at 40% of its peak
value. Finally, Figure 2c shows the
by Dale Eagar
inductor current and input line current as the instantaneous input line
voltage is at 10% of its peak value.
Notice that on each switchingfrequency cycle, the inductor current
goes to zero and then tries to reverse.
When the inductor current attempts
to reverse, the diode commits the
disgraceful act of insubordination,
destroying the beauty, simplicity, and
continuity of the control function of
the continuous-mode boost converter.
Figure 2c also shows the area gained
due to the clipping of the negative
portion of the inductor current. The
net effect is that the average current
is higher than the current that the
continuous-mode’s duty factor control of this DC variac would have
requested. The very act of preventing
the inductor current from reversing,
a hideous act indeed, necessitates a
new set of models to describe the
discontinuous-mode operation of the
boost converter. Further, a new control law is required to control the
discontinuous-mode boost converter.
The Boundary
is a Moving Target
The boost converter could be
likened to a fish tank, with the surface of the water representing the
boundary of continuity, above which
the boost converter operates in the
continuous-mode and below which it
operates in the discontinuous-mode.
Folks living outside the fish tank tend
to obtain their oxygen from the air,
whereas folks living inside the fish
tank tend to get their oxygen from the
water. Air folk and water folk have
great trouble communicating, as
L1
D1
I1
E
SW1
LOAD
pfc3_1b.eps
Figure 1b. Mortal boost converter
9
DESIGN FEATURES
ACTIVE
OFF
TIME
INACTIVE
OFF
TIME
SWITCH
ON
SWITCH
OFF
AVE
I1
AVE
AVE
0
0
0
DIODE
COMMUTATES
(2a)
(2b)
AREA GAINED
(2c)
pfc3_2.eps
Figure 2. Choke-current waveforms
neither is equipped to speak, let alone
breathe, in the other’s medium. This
lack of communication has had a
significant effect on the way the two
groups think.
When air folk look at controlling
their boost converters, they speak of
duty-factor, the all important controlling entity. Water folk control their
boost converters with on-time. We,
looking at the whole system from the
outside, know that the switchingfrequency is constant, and thus ontime and duty-factor are describing
the same thing.
When air folk plot duty-factor, they
see a smooth curve extending down to
the surface of the water, and then
being refracted at the surface to a
different slope as it goes further down
to the bottom of the tank. The air folk
mathematicians have developed
equations that predict the dutyfactor plot, equations that exactly
describe everything it does in the air.
The air folk equations are incapable
of predicting what happens when the
plot is below the water. The water folk
mathematicians have the exact same
problem in reverse. We the onlookers
10
can inspect the equations, and when
we do, we see that the respective
equations have no hope of being
consolidated into one. The air folk
equation states that the duty-factor
is a function of instantaneous input
and output voltage and the rate of
change in the current flow, but is
altogether independent of the steadystate value of the current flow,
the operating frequency, and the
inductor’s inductance value.
The water folk equation states that
the on-time is a function of the steadystate current flow, the inductance of
the inductor, the switching frequency,
and the input voltage, and is independent of the output voltage and the
rate of change in current flow. Fortunately for all of us, both equations
converge on the same point at the
water’s surface.
Air folks use volt-seconds to describe the stretch in their inductors,
and water folk use joules to describe
the energy stored in their inductors.
As onlookers, we can convert between
Joules and volt-seconds by using the
following equations:
2
J = 1 (V • S)
2L
V•S=
2LJ
J = Energy in Joules
L = Inductance in Henries
V • S = Stretch in Volt Seconds
To further complicate matters, the
index of refraction of the water
changes and is directly related to the
water level. The water level changes
due to input voltage, output voltage,
switching frequency, output current,
and inductance of the inductor.
Memory Lost
In continuous-mode, the choke
current, and thus power intercepted
from the input line, depends on what
the choke current was during the
previous cycle. Evaluation of continuous-mode boost is best performed
by looking at the overall action of an
integral number of switching cycles
of a running converter.
In discontinuous-mode, the choke
current goes to zero during each
continued on page 15
Linear Technology Magazine • August 1995
DESIGN FEATURES
The New LT1508/LT1509 Combines
Power Factor Correction and a PWM
in a Single Package
by Kurk Mathews
Introduction
Hand-Holding
When a supply requires power factor correction (PFC), designers
commonly add a PFC controller, such
as the LT1248, to the front end of a
traditional off-line DC/DC converter.
Although this is a viable solution,
additional circuitry is usually required to ensure that the two
controllers work well together. The
LT1508/LT1509 eliminate the need
for the additional circuitry and PWM
by integrating the functionality of a
LT1248 power factor controller and a
50% maximum-duty-cycle PWM in a
single 20-pin DIP or SOIC. The LT1508
retains all the pins of the LT1248
except the EN/SYNC, and adds quietground, soft-start, control-voltage,
current-limit and gate-driver pins
to form a voltage-mode PWM (see
Figure 1). In the LT1509, the current
limit becomes the ramp pin (PWM
current sense), forming a currentmode PWM (see Figure 2).
Whether the application involves
voltage mode or current mode control, smooth interaction between the
PFC and PWM stages requires some
hand-holding. The first thing to
consider is start-up. The PFC preregulator turns on with soft start as
usual. During this period the PWM’s
soft-start pin is held low, preventing
the second stage converter from turning on until the intermediate bus
voltage reaches its preset value (typically 380VDC to 400VDC for a
universal input). In addition to delaying the PWM soft start during start-up,
the same comparator employs hysteresis to shut down the PWM stage in
the event the intermediate bus voltage drops below 73% of the preset
voltage (typically 280VDC). Once enabled, the PWM runs somewhat
independently of the PFC controller,
with a couple of exceptions. The oscillator frequency is set by the values of
the capacitor and resistor to ground
on the CSET and RSET pins. Both the
PFC and PWM stages are synchronized to this oscillator and operate at
the same switching frequency. Synchronization is accomplished by
delaying the turn-on of the PWM stage
through 50% of the switching period.
In addition to making oscilloscope
measurements easier, these features
ensure that both switches don’t turn
on simultaneously, causing bus voltage spikes, noise and problems with
current mode in the PWM stage.
Single frequency and synchronization also ensure the absence of beat
frequencies in the conducted emissions. A preset duty cycle limit of
50% and leading-edge blanking further ease the design of the PWM
section. The PWM stage is quickly
disabled by pulling the control
voltage low. Finally, the LT1509’s
Full Featured PFC
All features of the LT1248 remain,
requiring no compromises when
implementing PFC with the LT1508/
LT1509. Continuous-boost average
current mode reduces magnetics size
while operating at a fixed frequency,
without the need for slope compensation. Built-in overvoltage protection
reduces the external parts count.
Current amplifier offset voltage is dealt
with internally, minimizing crossover
distortion. Full differential input current sense is available, as are 1.5A
gate drivers for both the PFC and
PWM stages. In addition to these
familiar performance features there
are a few new ones incorporated in
the PWM section.
Linear Technology Magazine • August 1995
current mode comparator has a 1.2
VDC offset to accommodate opto-feedback, and both the LT1508 and
LT1509 contain independent overcurrent comparators.
Typical Application
Figure 3 shows a 24VDC, 300W
power-factor-corrected, universalinput supply. The continuous,
current mode boost PFC preregulator
minimizes the differential-mode input filter size required to meet
European low-frequency conducted
emission standards while providing
a high power factor. The 2-transistor
forward converter offers many benefits including low peak currents, a
non-dissipative snubber, 500VDC
switches and automatic core reset
guaranteed by the LT1509’s 50%
maximum duty-cycle limitation.
An LT1431 and inexpensive optoisolation is used to close the loop
conservatively at 3kHz with excess
margin (see Figure 4). Figure 5 shows
the output voltage’s response to a 2A
to almost 10A current step. Regulation is maintained to within 0.5V.
Efficiency curves for output powers
of 30, 150, and 300W are shown
in Figure 6. The PFC preregulator
alone has efficiency numbers of between about 87% and 97% over line
and load.
Start up of the circuit begins with
the LT1509’s VCC bypass capacitors
trickle charging through 91kΩ
to 16VDC, overcoming the chip’s
.25mA typical start-up current (VCC ≤
lockout voltage). PFC soft start is
then released, bringing up the
382VDC bus with minimal overshoot.
As the bus voltage reaches its final
value, the forward converter comes
up, powering the LT1431, and
closing the feedback loop. A 3-turn
11
12
13
SS2
7.5V
16
SS1
11
OVP
9
IAC
14
12µA
12µA
7.9V
–
+
+
–
PWMOK
–
16V TO 10V
VSENSE
+
VCC
+
–
EA
7.0V
4.7V
2.2V
25k
12
10
IB
IA
IA2IB
200µA2
M1
IM =
–
+
RUN
7.5V
VREF
VREF
VAOUT
IM
7µA
2R
R
8
VC
18
–
+
7
1V
19
CA
ILIMIT
+
–
50µA
MOUT ISENSE
–
+
+
–
CL
4
CSET
–
+
6
0.7V
+
–
5
CAOUT PKLIM
+
–
RSET
15
OSC
S
R
55%
DELAY
RUN
R
R
S
R
Q
150ns
BLANKING
3
GND1
Q
16V
16V
17
VCC
1508_1.eps
20
GTDR2
2
GND2
1
GTDR1
DESIGN FEATURES
Authors can be contacted
at (408) 432-1900
Figure 1. Block diagram: LT1508
Linear Technology Magazine • August 1995
Linear Technology Magazine • August 1995
13
SS2
7.5V
16
SS1
11
OVP
9
IAC
14
12µA
12µA
7.9V
–
+
+
–
PWM0K
–
16V TO 10V
VSENSE
+
VCC
+
–
EA
7.0V
4.7V
2.2V
–
+
25k
12
10
1V
IB
IA
IA2IB
VC
RAMP
+
18
–
1.2V
CL
IM
7µA
8
7
+
–
50µA
MOUT ISENSE
19
+
–
200µA2
M1
IM =
–
+
RUN
7.5V
VREF
VREF
VAOUT
CA
–
+
15
OSC
S
R
R
55%
DELAY
RUN
RSET
+
–
4
0.7V
+
–
5
CSET
–
+
6
CAOUT PKLIM
R
S
R
Q
150ns
BLANKING
3
GND1
Q
16V
16V
17
VCC
1508_2.eps
20
GTDR2
2
GND2
1
GTDR1
DESIGN FEATURES
Figure 2. Block diagram: LT1509
13
14
VIN
0.1
“X”
20k
1%
499k
1%
9
11
20k
1%
2
3
91k
2W
PKLIM
13
1µF
0.001µF
15
15k
1%
RSET
4.02k
1%
MOUT
8
200µF
15V
0.047µF
16
SS1
1.8k
5
ERA82-004
ERA82-004
+
+
–
ISENSE
7
T1
17
VCC
1µF
–
+
CAOUT
20k
RAMP
19
20
0.022µF
GTDR2
18
VREF
1k
VC
1
GTDR1
1.2V
6
1N5819
20k
330µF
35V
GNDF
6
2.2k
2
COMP
LT1431
2.2k
330Ω
470µ
IRF840
COLL
1
MURH860CT
1µF
400V
ELECTRONIC
CONCEPTS
5MP12J105K
CNY17-3
2N2222A
100pF
1N5819
FUJI
ERA82-004
0.6AMP/40VR
+
15V
20Ω
2.2µF
50V
2.2µF
50V
T1
COILTRONICS
CTX02-12378-2
100pF
0.001µF
4.02k
1%
0.15, 3W
FUKUSHIMA MPC71
LT1509
10k
0.001µF
SS2
VREF
10 12
VREF
4
GND2 GND1 CSET
+
–
VAOUT
330k
0.01µF
KETEMA
SG57
SURGE-GARD
0.1
“X”
GI
KBU4J
NOTE: UNLESS OTHERWISE SPECIFIED
1. ALL RESISTORS 1/4W, 5%
2. ALL CAPACITANCE VALUES IN MICROFARADS
IAC
OVP
0.1
“X”
0.047µF
14 VSENSE
0.47µF
4700pF
“Y”
4700pF
“Y”
0.0047µF
499k
1%
499k
1%
499k
1%
382VBUS
90 TO
264 VAC 1M
1/2W
6A
FAST
VIN
V+
GNDS
1k
220Ω
5
3
2N2222A
220Ω
100pF
+
–
+
20k
RTOP
7
REF 8
RMD
4
30.1k
1%
2N2907
20Ω
15V,
1N965
(× 2)
20Ω
+
+
+
2000pF
10Ω
2W
IRF840
1508_3.eps
T2 7 TURNS
0.9" × 0.005" CU
ETD44-P
LPRI = 3.1mH
GI
FEP
30DP
1000pF
10Ω
2W
24VOUT,
12.5 AMPS
T2
35 TURNS
29AWG × 4
T2
35 TURNS
29AWG × 4
MUR450
0.51, 2W
RG ALLEN
RFS2
(× 2)
4700pF 4700pF
“Y”
“Y”
1000pF
“Y”
20k
MUR450
20k
IRF840
470µF, 50V
67µH
NICHICON 39 T 12AWG
PL12.5X25
T150-52
(× 3)
OUTPUT COM
3.4k
1%
1µF
63V
FILM
10k
1%
24.9k
1%
VREF
10 : 15
TURNS
28AWG
846T500-3E2A
1µF
10Ω
0.1µF
180µF
450V
382VBUS
DESIGN FEATURES
Figure 3. Schematic diagram of 300W, 24V DC output, power-factor corrected, universal-input supply
Linear Technology Magazine • August 1995
DESIGN FEATURES
80
90
60
75
40
60
20
45
0
30
–20
15
90
5A/DIV
85
EFFICIENCY (%)
PHASE MARGIN (DEGREES)
LOOP GAIN (dB)
300W
0.5V/DIV
150W
80
75
30W
100µs/DIV
–40
10
100
1K
10K
FREQUENCY (Hz)
0
100K
100µs/DIV
70
100
150
200
VIN (AC)
250
300
1508_6.eps
1508_4.eps
Figure 4. Bode plot of the circuit shown in
Figure 3
Figure 5. Figure 3’s response to a 2A to
approximately 10A load step
Figure 6. Efficiency curves for Figure 3’s
circuit
secondary added to the 70-turn primary of T1, bootstraps VCC to about
15VDC, supplying the chip’s 13mA
requirement as well as about 25mA to
cover the gate current of the three
FETs and high-side transformer
losses. A 0.15Ω sense resistor senses
input current and compares it to a
reference current (IM) created by the
outer voltage loop and multiplier. Thus
the input current follows the input
line voltage and changes, as necessary, in order to maintain a constant
bank voltage. The forward converter
sees a voltage input of 382VDC unless the line voltage drops out, in
which case the 180µF main capacitor
discharges to 250VDC before the PWM
stage is shut down. Compared to a
typical off-line converter, the effective
input voltage range of the forward
converter is smaller, simplifying the
design. Additionally, the higher bus
voltage provides greater hold-up times
for a given capacitor size. The highside transformer effectively delays the
turn-on spike to the end of the builtin blanking time, necessitating the
external blanking transistor.
lator turns on only after the first
output voltage is stable. Turn-off is in
the reverse order, due to the rapid
discharge of capacitor C3 through
the diode.
These and other applications will
benefit from the LT1118 family’s
unique performance attributes. Any
application that requires fast transient load response, unconditional
stability under capacitive loading,
or logic control of the power output
is a potential application for the
LT1118 family.
The Energy Bucket
power line. Because this converter
is very efficient, essentially all of
the intercepted power is delivered to
the load.
This concludes part three of our
series introducing the power factor
correction conditioner.
Conclusion
The LT1508 and LT1509 offer
complete solutions for low-to-high
power, isolated, off-line power factor-corrected supplies. The many
built-in features ensure a simplified,
low-parts-count design for a variety
of applications.
1118, continued from page 6
The enable pin provides many other
useful capabilities to the resourceful
designer. Figure 5 shows two LT11185 regulators configured to provide
sequential turn-on and turn-off from
logic control. A logic high level turns
on the first regulator; the second regu-
PFC, continued from page 10
off-time. This allows us to model the
discontinuous-mode boost converter
as a sum of discrete events. Each
event consists of an on-time and an
off-time, with the off-time being longer
than the time it takes to empty all
stored energy from the inductor.
Figure 2c shows that the off-time
consists of two separate times:
❏ the time when the diode is
conducting, the active off-time.
❏ the time after the diode has
committed, yet before the next
on-time, the inactive-off-time.
Linear Technology Magazine • August 1995
We look at a discontinuous-mode
boost converter as an energy bucket
that gets filled and dumped with each
cycle. During each cycle the bucket
takes a certain amount of energy (J).
Over a given number of cycles the
bucket will intercept many bucketsful of energy and deliver them to the
load. A boost converter operating at a
switching frequency of f and intercepting an amount of energy of J
during each cycle, will be intercepting
f × J watts of power from the input
15
DESIGN FEATURES
LTC1329 Micropower,
8-Bit, Current Output DAC
The LTC1329 is a micropower,
8-bit, current output DAC that
provides precision current output over
a wide supply range. This part is ideal
for portable and battery-powered
applications.
Description
The LTC1329 can communicate
with external circuitry by means of
one of three interface modes: standard, 3-wire serial-data mode, 1-wire
pulse mode, and 2-wire pulse mode.
In serial-data mode, the system microprocessor can serially transfer the
8-bit data to and from the LTC1329.
In pulse mode, the upper 6 bits of
DAC output can be programmed for
increment-only (1-wire interface) or
increment/decrement (2-wire interface) operation. In increment-only
mode, when the count increases beyond full scale, the counter rolls over
and sets the DAC to zero. In increment/decrement mode, the counter
stops incrementing at full scale and
stops decrementing at zero, and will
not roll over.
The LTC1329 can operate over a
wide supply range from 2.5V to 6.5V,
with a very low quiescent current of
only 50µA, which drops to only 0.2µA
in shutdown mode. The DAC current
output can be biased from −20V to
+2.5V. The full-scale DAC current
output is 10µA ±2.5% and 50µA
±2.5% for the −10 and −50 versions
respectively. On power-up, the DAC
current output is automatically set
at mid-range.
The LTC1329’s micropower operation makes it ideal for portable and
battery-powered applications like
LCD contrast control, backlight
brightness control, power supply
voltage adjustment, battery charger
voltage/current adjustment, and trim
pot replacement. The LTC1329 is
available in 8-pin surface mount and
DIP packages.
16
Power Supply
Voltage Adjustment
Trimmer Pot Replacement
Figure 2 is a schematic of a digitally
controlled offset-voltage adjustment
circuit using a 1-wire interface. By
simply clocking the LTC1329, the DAC
current output is increased, causing
VR2 to increase accordingly. When the
DAC current output reaches full scale,
it will roll over to zero, causing VR2 to
change from the maximum offset trim
voltage to the minimum offset trim
voltage.
Figure 1 is a schematic of a digitally controlled power supply voltage
adjustment circuit using a 2-wire
interface. The LT1107 is configured
as a step-up DC/DC converter, with
the output voltage (VOUT) determined
by the values of the feedback resistors. The LTC1329’s DAC current
output is connected to the feedback
node of these resistors, and an 8051
microprocessor is used to interface to
the LTC1329. By simply clocking the
LTC1329, the DAC current output is
decreased or increased (decreased if
DIN = 0, increased if DIN = 1), causing
VOUT to change accordingly.
L1*
100µH
VIN
2V TO 3V
1N5817
VOUT
4V TO 6V
20mA
+
47Ω
ILIM
by K.S. Yap
100µF
VIN
39k
LTC1329-50
SW1
1
LT1107
GND
VCC = 3.3V
FB
SW2
2
+
3
10k
47µF
4
IOUT
DOUT
VCC
DIN
SHDN
CLK
GND
CS
8
7
6
MPU
(e.g. 8051)
5
P1.1
P1.0
*L1 = COILTRONICS CTX100-4
1329_1.eps
Figure 1. LTC1329 digitally controls the output voltage of a power supply
RF
RI
VIN
–
V–
(0.5) (IFULL SCALE)
TRIM RANGE = ±(0.5) (IFULL SCALE) (R2)
R1 =
VOUT
LT1006
+
1
VCC = 3.3V
R2
100Ω
R1
600k
2
3
V–
4
IOUT
DOUT
VCC
DIN
LTC1329-50
SHDN
CLK
GND
CS
8
7
6
MPU
(e.g. 8051)
5
P1.0
1329_2.eps
Figure 2. LTC1329 used to null op amp’s offset voltage
Linear Technology Magazine • August 1995
DESIGN FEATURES
VIN
+
C1
L1
D1
VOUT
2
3
LB0
SW
VIN
1.5V
UNDERVOLTAGE
LOCKOUT
+
4
C2
A 4-Cell-to-5V Converter
36mV
1
LBI
+
+
R2
1k
A2
–
–
R1
7.2Ω
A3
BIAS
~1V
1.2V
Q3
OFF
R3
8
FB
1k
–
ENABLE
A1
R4
+
TIMERS
6µs ON
1.5µs OFF
Q2
1x
DRIVER
1.24V
VREF
Q1
200x
SHUTDOWN
7
SHDN
6
ILIM
5
GND
1304_1.eps
Figure 1. LT1304 Block diagram. Independent low-battery detector A3 remains alive when
device is in shutdown
LT1304, continued from page 1
current. When R1’s voltage drop exceeds 36mV, corresponding to 1A
switch current, A2’s output goes high,
truncating the on-time part of the
switch cycle. The 1A peak current can
be reduced by tying a resistor between the ILIM pin and ground, causing
a voltage drop to appear across R2.
The drop offsets some of the 36mV
reference voltage, lowering peak
current. A 22k resistor limits current
to approximately 550mA. A capacitor
connected between ILIM and ground
provides soft start. Shutdown is
accomplished by grounding the
SHDN pin.
The low-battery detector A3 has its
own 1.2V reference and is always on.
The open collector output device can
sink up to 500µA. Approximately
35mV of hysteresis is built into A3 to
reduce “buzzing” as the battery voltage reaches the trip level.
A 2-Cell-to-5V Converter
A compact 2-cell-to-5V converter
can be constructed using the circuit
of Figure 2. Using the LT1304-5 fixedLinear Technology Magazine • August 1995
sag, possibly tripping the low-battery
detector. Output voltage reaches 5V
in approximately 1ms. The addition
of R1 and C1 to Figure 2’s circuit
limits inrush current at start-up, providing for a smoother turn-on, as
indicated in Figure 6.
output device eliminates the need for
external voltage-setting resistors, lowering component count. As the battery
voltage drops, the circuit continues
to function until the LT1304’s undervoltage lockout disables the part at
approximately VIN = 1.5V. 200mA is
available at a battery voltage of 2.0V.
As the battery voltage decreases below 2V, cell impedance starts to
quickly increase. End-of-life is usually assumed to be around 1.8V, or
0.9V per cell. Efficiency is detailed in
Figure 3. Burst Mode micropower
operation keeps efficiency above 70%,
even for load current below 1mA.
Efficiency reaches 85% for a 3.3V
input. Load transient response is illustrated in Figure 4. Since the
LT1304 uses a hysteretic comparator
in place of the traditional linear feedback loop, the circuit responds
immediately to changes in load current. Figure 5 details start-up behavior
without soft-start circuitry (R1 and
C1 in Figure 2). Input current rises to
1A as the device is turned on, which
can cause the input supply voltage to
A 4-cell-to-5V converter is more
complex than a simple boost converter because the input voltage can
be either above or below the output
voltage. The single-ended primary inductance converter (SEPIC) shown in
Figure 7 accomplishes this task, with
the additional benefit of output isolation. In shutdown conditions, the
converter’s output will go to zero,
unlike the simple boost converter,
where a DC path from input to output
through the inductor and diode remains. In this circuit, peak current is
limited to approximately 500mA by
the addition of 22k resistor R1. This
allows very small, low-profile components to be used. The 100µF
capacitors are D-case size, with a
height of 2.9mm, and the inductors
are 3.2mm high. The circuit can deliver 5V at up to 100mA. Efficiency is
relatively flat across the 1mA-to100mA load range.
Super BurstTM Mode Operation:
5V/100mA DC/DC with 15µA
Quiescent Current
The LT1304’s low-battery detector
can be used to control the DC/DC
converter. The result is a reduction in
quiescent current by almost an order
of magnitude. Figure 9 details this
Super Burst circuit. VOUT is monitored by the LT1304’s LBI pin via
resistor divider R1/R2. When LBI is
above 1.2V, LBO is high, forcing the
LT1304 into shutdown mode and reducing current drain from the battery
to 10µA. When VOUT decreases enough
to overcome the low battery detector’s
hysteresis (about 35mV), LBO goes
low. Q1 turns on, pulling SHDN high
and turning on the rest of the IC. R3
limits peak current to 500mA; it can
be removed for higher output power.
Efficiency is illustrated in Figure 10.
17
DESIGN FEATURES
MBRS130L
22µH*
90
VIN = 3.3V
SW
LBI
+
LT1304-5
100µF
2 CELLS
80
5V
200mA
SENSE
LB0
SHDN
IL
GND
+
R1
1M
100µF
VIN = 2.5V
70
VIN = 1.8V
60
50
+
C1
1µF
SHUTDOWN
40
0.1
*SUMIDA CD54-220
(708) 956-0666
1
10
100
LOAD CURRENT (mA)
1304_2.eps
500
1304_3.eps
Figure 2. 2-cell-to-5V/200mA boost converter takes four external parts. Components with
dashed lines are for soft-start (optional)
VOUT 100mV/DIV
AC COUPLED
ILOAD
EFFICIENCY (%)
VIN
Figure 3. 2-cell-to-5V converter efficiency
VOUT 2V/DIV
VOUT 2V/DIV
IIN
500mA/DIV
IIN
500mA/DIV
VSHDN
10V/DIV
200mA
0
VSHDN
10V/DIV
1ms/DIV
100µs/DIV
Figure 4. Boost converter load-transient
response with VIN = 2.2V
The converter is approximately 70%
efficient at a 100µA load, 20 points
higher than the circuit of Figure 2.
Even at a 10µA load, efficiency is in
the 40%-50% range, equivalent to
100µW–120µW total power drain from
the battery. In contrast, Figure 2’s
circuit consumes approximately
300µW–400µW unloaded.
Figure 5. Start-up response. Input current
rises quickly to 1A. VOUT reaches 5V in
approximately 1ms. Output drives 20mA load
Figure 6. Start-up response with 1µF/1MΩ
components in Figure 2 added. Input current
is more controlled. VOUT reaches 5V in 6ms.
Output drives 20mA load
An output capacitor charging cycle
or “burst” is shown in Figure 11, with
the circuit driving a 50mA load. The
slow response of the low-battery detector results in the high number of
individual switch cycles or “hits”
within the burst.
Figure 12 depicts output voltage at
the modest load of 100µA. The burst
repetition rate is around 4Hz. With
the load removed, the repetition rate
drops to approximately 0.2Hz, or one
burst every 5 seconds. Systems that
spend a high percentage of operating
time in sleep mode can benefit from
the greatly reduced quiescent power
drain of Figure 9’s circuit.
100µF
22µH*
3.5V TO 6.5V
1ms/DIV
MBR0530
85
+
80
LBI
+
4 CELLS
5V
100mA
SENSE
LT1304-5
47µF
VIN = 3V
SW
LB0
IL
SHDN
GND
+
22µH*
R1
22k
100µF
VIN = 4V
75
EFFICIENCY (%)
VIN
70
65
VIN = 6V
VIN = 5V
60
SHUTDOWN
55
50
1
*SUMIDA CD43-220
(708) 956-0666
1304_7.eps
Figure 7. 4-cell-to-5V step-up/step-down converter, also know as SEPIC (single-ended primary
inductance converter). Low profile components are used throughout
18
10
LOAD CURRENT (mA)
100
1304_8.eps
Figure 8. Efficiency plot of SEPIC converter
shown in Figure 7
Linear Technology Magazine • August 1995
DESIGN FEATURES
33µH*
IQ ≈ 15µA
MBR0530
5V
80mA
200k
0.01µF
47k
2
3
VIN
4
SW
LB0
80
LBI
+
330µF
FB
SHDN
GND
7
5
IL
6
R2
1.21M
+
2 CELLS
VIN = 3V
1
LT1304
2
90
R1
3.83M
EFFICIENCY (%)
Q1
2N3906
70
VIN = 2V
60
50
100µF
47k
22k
40
0.01
0.1
1.0
10
LOAD CURRENT (mA)
100
1304_9.eps
*SUMIDA CD54-330
(708) 956-0666
1304_10.eps
Figure 9. 2-cell-to-5V DC/DC converter using Super Burst Mode Operation draws only
15µA unloaded. Two AA Alkaline cells will last for years
Figure 10. DC/DC converter efficiency of
Figure 9
VOUT 100mV/DIV
AC COUPLED
VSW
5V/DIV
SHUTDOWN
IL
1A/DIV
50µs/DIV
1
Figure 11. Super Burst Mode Operation
in action. VIN = 2.5V, ILOAD = 50mA
2
VIN
VOUT 100mV/DIV
AC COUPLED
8
LT1304
7
3
6
4
5
+ CIN
VOUT
+COUT
50ms/DIV
Figure 12. Circuit using Super Burst
Mode Operation, 100µA load. Burst
occurs approximately once every 240ms
GND (BATTERY AND LOAD RETURN)
1304_13.eps
Figure 13. Suggested layout for best performance. Input capacitor placement as shown is
highly recommended. Switch trace (pin 4) copper area is minimized
VOUT 100mV/DIV
AC COUPLED
50ms/DIV
Figure 14. LT1304 Switch rise time is in the
5ns range. These types of edges emphasize
the need for proper PC board layout
VSW 1V/DIV
5ns/DIV
Figure 15. Switch fall time. Lower slope in
second and third graticules shows effect of
lead and bond wire inductance
Linear Technology Magazine • August 1995
Layout
The LT1304 switch turns on and
off very quickly. For best performance
we suggest the component placement
in Figure 13. Improper layouts will
result in poor load regulation, especially at heavy loads. Parasitic lead
inductance must be kept low for
proper operation. Switch turn-off is
detailed in Figure 141. A close look at
the rise time (5ns) will confirm the
need for good PC board layout. The
200MHz ringing of the switch voltage
is attributable to lead inductance,
switch and diode capacitance, and
diode turn-on time. Switch turn-on is
shown in Figure 15. Transition time
is similar to that of Figure 14. Adherence to the layout suggestions will
result in working DC/DC converters
with a minimum of trouble.
1. Instrumentation for oscillographs of Figures 14
and 15 include Tektronix P6032 active probe,
Type 1S1 sampling unit and type 547 mainframe.
19
DESIGN INFORMATION
RS232 Transceiver Protected to
±15kV IEC-801-2 ESD Standards
by Gary Maulding
ESD Effects on
Electronic Equipment
Electrostatic discharge (ESD) is a
well known killer of semiconductor
devices. Unless protective measures
are taken, ESD will result in various
forms of device degradation. Discharge damage ranges from increased
leakage currents on device inputs
and outputs, to increased offset or
shifted threshold voltages, to failure
of the affected circuit.
To minimize the monetary loss due
to damaged IC’s and to improve the
reliability of equipment, the electronics industry has adopted many
standard safeguard procedures
against ESD. These safeguards include the use of grounded work
surfaces, equipment, and personnel;
330Ω
HIGH
VOLTAGE
SUPPLY
150pF
conductive plastic shipping materials, on-chip protection circuits at the
input and output pins, and testing of
the ESD sensitivity of circuits.
The “Human Body Model,” defined
by Mil-Std-883B Method 3015, is a
commonly used ESD test standard
for rating integrated circuits for ESDdamage immunity. This test method
is aimed directly at reducing the incidence of failed devices installed on
equipment circuit boards. The test
discharges a charged 100pF capacitor through a 1.5kΩ resistance directly
to the IC’s pins. Both positive and
negative voltages are tested. Rarely is
the device powered when testing is
conducted, and only permanent damage to the circuit is considered.
The widespread adoption of environmental static control and the
design of integrated circuits with
on-chip protection devices has dramatically reduced the incidence of
circuit failures due to ESD damage
during circuit manufacturing and
handling. Most semiconductor devices enter a much safer environment
once installed in a PC board. Installation of the PC board into a chassis
further shields the IC from the environment. No longer are most ICs’
input and output pins exposed to
direct ESD strikes, but ESD is still an
operational and reliability hazard for
electronic equipment. Everyone has
experienced the shock caused by
walking across a nylon carpet on a
dry winter day and touching a piece of
electrical equipment. This
static discharge can damage internal circuitry of
the equipment if the
manufacturer does not
DEVICE
provide a safe discharge
UNDER
path for the energy. InTEST
put/output ports on
personal computers are
especially vulnerable. The
ports are available so that users can
attach cabling to printers, mice, or
networks. Portable computers are
particularly vulnerable since the port
connections are made and broken
much more often than for desktop
machines. Most consumers would not
like to wear static control ground
straps when using their personal computers. Hence, it is important that
ESD protection be built in to the end
user equipment.
What is IEC-801-2?
The International Electrotechnical
Commission (IEC) has defined a test
and rating standard for electronic
equipment exposed to ESD transients.
IEC-801-2 defines two alternate test
methodologies. The preferred method
is a contact-mode test, but the alternate air-gap test is more easily
implemented and is more commonly
used. Equipment certified for IEC801-2 testing consists of a charging
circuit to apply the desired test voltage to a 150pF capacitor. A discharge
switch discharges the energy into the
equipment under test through a 330Ω
resistance (see Figure 1). IEC-801-2
certified test equipment must be able
to demonstrate that the applied test
32
24
0.7ns < TRISE < 1.0ns
AMPS
Linear Technology’s popular
LT1137A RS232 transceiver has
been upgraded to include on-chip
ESD protection compliant with
IEC-801-2 level 4 standards. Device
pins connected to the RS232 line can
absorb up to ±15kV air-gap or ±8kV
contact-mode transients using the
IEC-801-2 ESD model, with no
damage or interruption of operation.
The 28-lead SOIC or SSOP packaged
circuit and five inexpensive ceramic
capacitors are the only components
required to implement a 5V powered
AT serial port that is fully compliant
with RS232 or V.28 electrical
specifications and IEC-801-2 staticdischarge standards.
16
IEC-801-2
HUMAN BODY MODEL
8
0
–20
20
60
100 140
TIME (ns)
180
220
ESD_2.eps
ESD_1.eps
Figure 1. IEC-801-2 ESD test system
20
Figure 2. ESD Transient current waveforms
Linear Technology Magazine • August 1995
DESIGN INFORMATION
voltage and current waveform meet
the prescribed rise and decay waveshape shown in Figure 2. Linear
Technology uses a Keytek MiniZap
for IEC-801-2 ESD testing.
The test results on equipment are
classified by the voltage level successfully applied to the equipment
and the affect of test discharges on
the equipment. The IEC-801-2
standard defines the following performance criteria:
1. Normal performance within
specification limits
2. Temporary degradation or loss of
function or performance that is
self-recoverable
3. Temporary degradation or loss of
function or performance that
requires operator intervention or
system reset
4. Degradation or loss of function
that is not recoverable, due to
damage of equipment (components) or software, or loss of data
Table 1 details the rated test levels
for both contact and air-gap tests.
The contact-mode test is the
preferred methodology due to its
superior repeatability. The air-gap
test may be influenced by the test
environment’s relative humidity, as
well as by the approach speed of the
tester tip to the device under test.
LT1137A ESD
Test Performance
The LT1137A is an RS232 interface transceiver with an integral
charge pump for single 5V operation.
The complement of three drivers and
five receivers matches the requirements for AT serial ports on personal
computers. The eight RS232 lines are
directly connected to the socket on
the computer’s case, and are therefore directly exposed to possible ESD
transients when cables are installed
or removed from the machine. The
inclusion of rugged ESD protection
on the chip eliminates the need for
the user to install costly and spaceconsuming transient suppression
diodes on the RS232 lines.
The new, upgraded LT1137A has
been tested to level 4 IEC-801-2 testing. In both the ±8kV contact-mode
test and the ±15kV air-gap test, the
transceiver suffered no damage and
no interruption of normal operation.
Linear Technology provides this
enhanced ESD protection on the
LT1137A at no extra cost to the user.
The improved level of protection requires no increase in die area over the
older ±10kV protected devices. The
user must only follow the data sheet
and use low ESR capacitors (ceramic
chip caps are perfect) located close to
the device pins to obtain the benefits
of the on-chip protection (see Figure
3). The capacitors and the length of
PC board traces are important, since
the on-chip protection devices shunt
the ESD transient energy through
the capacitors and away from active
circuitry. The user should take care
to provide a short ground return path
for the energy in order not to cause
secondary damage to other PC board
components caused by the highspeed, high-current transients that
will flow during an ESD discharge.
Is ESD Protection Effective?
In 1992 Linear Technology introduced the first integrated circuits
capable of withstanding ±10kV ESD
transients as tested per Mil-Std.-883B
Method 3015 (Human Body Model).
Since their introduction, millions of
units have been installed in equipment, but not one single unit has
been returned as a field failure due to
ESD damage. Prior to the on-chip
10kV ESD protection, interface circuits would routinely be found to
have suffered ESD damage when they
were used in systems that did not use
a TransZorb for ESD protection. The
on-chip ESD protection has proven to
be a great enhancement to equipment reliability and a cost saving to
computer manufacturers, who no
longer need to add protective devices
to their I/O ports. The enhanced protection achieved by meeting the
IEC-801-2 test standards will prolong this perfect record.
TransZorb is a registered trademark of General Instruments, GSI.
+
0.1µF
+
0.2µF
DRIVER 1 OUT
RX1 IN
DRIVER 2 OUT
RS232
LINE PINS
PROTECTED
TO ±15kV
RX2 IN
RX3 IN
RX4 IN
DRIVER 3 OUT
RX5 IN
ON/OFF
28
LT1137A
V–
27
3
26
4
25
5
24
6
23
7
22
8
21
9
20
10
19
11
18
12
17
13
16
14
15
+
0.1µF
0.2µF
+
+
0.1µF
1
V+
5V VCC 2
Table 1. IEC-801-2 severity levels
DRIVER 1 IN
RX1 OUT
DRIVER 2 IN
RX2 OUT
RX3 OUT
RX4 OUT
DRIVER 3 IN
RX5 OUT
Level
1
2
3
4
Test Voltage
Contact Discharge (kV)
Test Voltage
Air Discharge (kV)
2
4
6
8
2
4
8
15
GND
DRIVER DISABLE
ESD-3.eps
Authors can be contacted
at (408) 432-1900
Figure 3. LT1137A ESD test circuit
Linear Technology Magazine • August 1995
21
DESIGN INFORMATION
V.35 Transceivers Allow
3-chip V.35 Port Solution
Two new LTC interface devices, the
LTC1345 and the LTC1346, provide
the differential drivers and receivers
needed to implement a V.35 interface. When used in conjunction with
an RS232 transceiver like the
LT1134A, they allow a complete V.35
interface to be implemented with just
two transceiver chips and one resistor-termination chip. The signals in
a typical V.35 interface are divided
into two groups, with five high-speed
differential signals for clock and data
transfer and five or eight slower,
single-ended V.28 (RS232) signals for
handshaking. The LTC1345 and
LTC1346 provide the three differential drivers and receivers necessary to
implement the high-speed path, and
the LT1134A provides the four RS232
drivers and receivers required for the
handshaking interface. Both the
LTC1345 and the LT1134A provide
onboard charge pump power supplies, allowing a complete V.35
interface to be powered from a single
5V supply. For systems where ±5V
supplies are present, the LTC1346 is
offered without charge pumps, representing a 30% power savings.
Appendix II of the CCITT V.35
specification describes the differential signal requirements for the
high-speed interface lines. The specification stipulates a differential
transmitter of 100Ω source termination impedance driving a 100Ω
twisted-pair cable and a differential
50Ω
125Ω
50Ω
A
300Ω
120Ω
300Ω
B
1345_2.eps
Figure 1. Y and ∆ termination topologies
22
receiver with 100Ω as load termination. The ground potential between
the source and load is allowed to vary
up to ±4V, and the common-mode
resistance at both source and load is
150Ω. The specification calls for differential transmitter signals of 0.55V
±20% and no more than ±0.6V
comon mode, exclusive of groundpotential differences. Early V.35
implementations used constant-voltage-output drivers for the differential
signals, and resistors in series with
the outputs to provide the required
±0.55V output swing. This technique
suffered from a strong dependence
on power supply voltage and driver
output impedance to control the output voltage. Further, common-mode
voltages present at the driver output
had a strong effect on output swing,
significantly limiting the driver-output common mode range.
The LTC1345/LTC1346 transceivers avoid these problems by using a
constant-current, high impedance
output driver. Designed to be used
with matching external resistor networks to meet the Appendix II
specifications, the transmitter output
structure consists of two complementary switched current sources that
maintain a 0.55V signal magnitude
over ±4V common mode range. These
current sources are trimmed to source
and sink 11mA at the complementary outputs. The differential output
voltage of the transmitter is set by the
external source and load termination
resistor networks. The typical differential source/load resistance of 100Ω
gives an output voltage of 0.55V (11mA
× 100Ω/2). The common mode voltage at the transmitter is typically
less than 0.2V with zero ground potential difference. The LTC1345 and
LTC1346 require 5% tolerance in the
termination resistors to meet the
20% differential voltage variation
specification. A termination resistor
by Y.K. Sim
network such as BI Technologies part
number 627T500/1250 (SOIC) or
899TR50/125 (DIP) meets this specification and is recommended for
proper cable termination. Discrete
5% resistors can also be used to terminate the drivers and receivers;
either a Y (wye, Figure 1A) or ∆ (delta,
Figure 1B) connection can be used to
give the required 100Ω differential
and 150Ω common mode impedances.
The LTC1345/LTC1346 differential input receivers have 30kΩ input
impedance and a typical hysteresis of
40mV. This allows the receivers to
meet the V.35 specification with the
specified termination resistor network. Additionally, they can be used
to detect RS422 signals if the 125Ω
V.35 common mode resistor is removed. Both the LTC1345 and
LTC1346 transceivers feature four
digitally-selectable operating modes:
DTE, with two drivers and three receivers active; DCE, with three drivers
and two receivers active; all three
drivers and receivers active, and shutdown. In shutdown, supply current
drops to 1µA typically. The differential transceivers are capable of
operating at data rates above
10MBaud in non-return-to-zero
(NRZ) format.
The RS232 handshaking lines can
be implemented with standard RS232
transceivers. The LT1134A provides
four RS232 drivers and four receivers, enough to implement the extended
8-line handshaking protocol specified in V.35. The LT1134A also
includes an onboard charge pump to
generate the higher voltages
required by RS232 from a single 5V
supply, making it an ideal companion
to the LTC1345. These two chips, together with the BI Technologies
termination resistor network chip,
provide a complete, surface mountable, 5V-only V.35 data port. Systems
that have multiple power supply
Linear Technology Magazine • August 1995
DESIGN INFORMATION
1µF
50Ω
1µF
+ +
VCC1
5V
4
T
1
2
27
LTC1345
6
+
1µF
DX
7
DX
11
RX
12
RX
13
RX
9
10
1µF
26
1
25
2
24
3
23
4
20
14
19
13
18
12
17
11
16
10
15
9
5
7
4
P
TXD (103)
S
S
U
T
SCTE (113)
T
TXC (114)
X
T
RXC (115)
T
T
RXD (104)
T
T
R
R
T
X
V
V
T
AA
Y
Y
T
U
W
W
AA
P
BI
627T500/
1250
(SOIC)
B
GND (102)
B
A
CABLE SHIELD
A
T
1
3
22
2
12
16
11
15
10
14
9
13
1
24
2
23
3
22
4
21
5
20
6
19
7
3
10
11
RX
4
DX
5
DX
6
DX
7
8
12
VCC2
0.2µF
23
4
LT1134A
0.2µF
3
22
1
0.1µF
0.1µF
19
20
18
OPTIONAL SIGNALS
16
14
17
15
23
24
LT1134A
2
21
0.1µF
RX
8
24
+
LTC1346
0.1µF +
0.2µF
1
0.1µF
T
8
14
VCC1
0.2µF
BI
627T500/
1250
(SOIC)
VCC2
5V
VEE2
50Ω
28
3
+
125Ω
=
0.1µF
2
DX
DX
RX
RX
RX
RX
DX
DX
5
H
DTR (108)
7
C
RTS (105)
6
E
DSR (107)
8
D
CTS (106)
10
F
DCD (109)
12
NN
TM (142)
9
N
RDL (140)
11
L
LLB (141)
13
H
6
C
8
E
5
D
7
F
9
NN
11
N
10
L
12
20
RX
18
RX
DX
DX
DX
DX
21
19
17
15
16
RX
14
RX
13
ISO 2593
ISO 2593
34-PIN DTE/DCE
34-PIN DTE/DCE
INTERFACE CONNECTOR INTERFACE CONNECTOR
1345_1.eps
Figure 2. Typical V.35 implementation using LTC1345 and LTC1346
voltages available and use only the
simpler 5-signal V.35 handshaking
protocol can use the LTC1346 with
the LT1135A or LT1039 RS232 transLinear Technology Magazine • August 1995
ceivers; this combination provides a
complete port while saving board
space and complexity. Figure 2 shows
a typical LTC1345/LTC1134A V.35
implementation with five differential
signals and five basic handshaking
signals, with an option for three additional handshaking signals.
23
DESIGN IDEAS
1MHz Step-Down Converter Ends
455kHz IF Woes
by Mitchell Lee
There can be no doubt that switching power supplies and radio IFs don’t
mix. One-chip converters typically
operate in the range of 20kHz–100kHz,
placing troublesome harmonics right
in the middle of the 455kHz band.
This contributes to adverse effects
such as “de-sensing” and outright
blocking of the intended signals.
A new class of switching converter
makes it possible to mix highefficiency power supply techniques
and 455kHz radio IFs without fear of
interference.
The circuit shown in Figure 1 uses
an LT1377 boost converter, operating
at 1MHz, to implement a highefficiency buck-topology switching
regulator. The switch is internally
grounded, calling for the floating
DESIGN IDEAS
1MHz Step-Down Converter Ends
455kHz IF Woes ...................... 24
Mitchell Lee
A Circuit That Smoothly Switches
Between 3.3V and 5V .............. 25
Doug La Porte
12-Bit Voltage-Mode DACs Pack
Versatility into 8-Pin SOs ....... 26
Kevin R. Hoskins
Wideband RMS Noise Meter ..... 28
Mitchell Lee
LTC1477: 70mW Protected
High-Side Switch Eliminates
“Hot Swap” Glitching ............. 32
Tim Skovmand
Driving a High-Level Diode-Ring
Mixer with an Operational
Amplifier ................................ 33
supply arrangement shown (D1 and
C1). The circuit converts inputs of
8V–30V to a 5V/1A output.
The chip’s internal oscillator operates at 1MHz for load currents of
greater than 50mA, with a guaranteed
tolerance of 12% over temperature.
Even wideband 455kHz IFs are
unaffected, as the converter’s operating frequency is well over one octave
distant.
Figure 2 shows the efficiency of
Figure 1’s circuit. You can expect
80%–90% efficiency over an 8V–16V
input range with loads of 200mA or
more. This makes the circuit suitable
for 12V battery inputs (that’s how I’m
using it), but no special considerations are necessary with adapter
inputs of up to 30V.
Mitchell Lee
Fully Differential, 8-Channel,
12-Bit A/D System Using the
LTC1390 and LTC1410 .......... 36
Kevin R. Hoskins
8V TO 30V
INPUT
+
D1
1N5818
100µF
100
VIN = 8V
5
8
V+
N.C.
+
C1
2.2µF
3
VSW
SHDN
LT1377
PFB
3.57k
2
EFFICIENCY (%)
4
90
1N4148 10Ω
NFB
VC
SG
6
PG
1
1.24k
100nF
80
12V
70
16V
VO = 5V
7
60
2k
50
200
4.7nF
47nF
CTX20-2P*
5V
1A
+
MBRS130
150µF
6.3V
OSCON**
*CTX20-2P, COILTRONICS 20µH
**OSCON, SANYO VIDEO COMPONENTS
400
600
IOUT (mA)
800
1000
dI1377_2.eps
Figure 2. Efficiency graph of the circuit
shown in Figure 1
dI1377_1.eps
Figure 1. Schematic diagram: 1MHz LT1377-based boost converter
24
Linear Technology Magazine • August 1995
DESIGN IDEAS
A Circuit That Smoothly Switches
Between 3.3V and 5V
by Doug La Porte
Many subsystems require supply
switching between 3.3V and 5V to
support both low-power and highspeed modes. This back-and-forth
voltage switching can cause havoc to
the main 3.3V and 5V supply buses.
If done improperly, switching the subsystem from 5V to 3.3V can cause a
momentary jump on the 3.3V bus,
damaging other 3.3V devices. When
switching the subsystem from 3.3V
to 5V, the 5V supply bus can be
pulled down while charging the
subsystem’s capacitors, and may inadvertently cause a reset.
The circuit shown in Figure 1 allows smooth voltage switching
between 3.3V and 5V, with added
protection features to ensure safe
operation. IC1 is an LTC1470 switchmatrix device. This part has on-chip
charge pumps running from the 5V
supply to fully enhance the internal
N-channel MOSFETs. The LTC1472
also has guaranteed break-beforemake switching to prevent crossconduction between buses. It also
features current limiting and
thermal shutdown.
When switching the subsystem
from 5V to 3.3V, the holding capacitor and the load capacitance are
initially charged up to 5V. Connecting these capacitors directly to the
main 3.3V bus causes a momentary
step to 5V. This transient is so fast
that the power supply cannot react in
time. Switching power supplies have
a particularly difficult time coping
with this jump. Switching supplies
source current to raise the supply
voltage and require the load to sink
current to lower the voltage. A switching supply will be unable to react to
counter the large positive voltage step.
This jump will cause damage to many
low-voltage devices.
The circuit in Figure 1 employs a
comparator (IC2) and utilizes the high
impedance state of the LTC1470 to
allow switching with minimal effect
on the supply. When the 3.3V output
is selected, IC1’s output will go into a
high impedance state until its output
falls below the 3.3V bus. The output
capacitors will slowly discharge to
3.3V, with the rate of discharge depending on the current being pulled
by the subsystem and the size of the
1µF
5V 3.3V
+
holding capacitor. The example shown
in Figure 1 is for a 250mA subsystem.
The discharge time constant should
be about 4ms. Once the subsystem
supply has dropped below the 3.3V
supply, the comparator will trip, turning on the 3.3V switch. The
comparator has some hysteresis to
avoid instabilities. The subsystem
supply will reach a low point of about
3V before the 3.3V switch is fully
enhanced.
When switching from 3.3V to 5V,
IC1’s current limiting prevents the
main 5V bus from being dragged down
while charging the holding capacitor
and the subsystem’s capacitance.
Without current limiting, the inrush
current to charge these capacitors
could cause a droop in the main 5V
supply.
If done improperly, supply voltage
switching leads to disastrous system
consequences. The voltage switch
should monitor the output voltage
and have current limiting to prevent
main supply transient problems. A
correctly designed supply voltage
switch avoids the pitfalls and results
in a safe, reliable system.
1µF
+
2
0 = 5V
1 = 3.3V
1k
3.3V
5V
3
+
8
IC2
LT1011
–
VOUT
IC1
V
LTC1470 OUT
0.1µF
3
2
7
7
4
4
1
TO
SUBSCRIBER
8
EN1
+
EN0
220µF
TANTALUM
HOLDING
CAPACITOR
5
500mV/DIV
3.3V
5V
5V/DIV
0V
➛
dI1470_1.eps
➛
5V
6
5VIN 3VIN 3VIN
51k
2ms/DIV
100µs/DIV
Figure 2. Oscillograph of the switchover
waveform showing smooth transitions
Figure 1. Schematic diagram of 3.3V and 5V switchover circuit
Linear Technology Magazine • August 1995
25
DESIGN IDEAS
12-Bit Voltage-Mode DACs Pack
Versatility into 8-Pin SOs
by Kevin R. Hoskins
Introduction
Let’s face it, microprocessors and
microcontrollers cannot control mechanical potentiometers. They have
no fingers, just pins. And they have
neither eyes with which to set a wiper’s
position nor the intuition to know
when it’s set correctly. They need the
organs that provide these functions
transplanted into their “operational
environment.” The “organs” that fill
these roles are analog-to-digital converters (ADCs) and the sensors
connected to their inputs, and digital-to-analog converters (DACs) and
the manipulative circuits or mechanics they drive.
This article concentrates on the
“fingers” and “hands” capabilities that
DACs bestow upon microprocessors
and microcontrollers. Linear Technology has recently introduced 12-bit,
serially interfaced, voltage-output
DACs packaged in small 8-pin SOs.
These parts include the LTC1257,
LTC1451, LTC1452, and LTC1453.
All of these DACs are micropower,
have 12-bit monotonic transfer functions, and operate on 4.75V to 15.75V
(LTC1257), 5V (LTC1451), or 3V to 5V
(LTC1452 and LTC1453) supply voltages. Table 1 captures the important
specifications of the LTC1257,
LTC1451, LTC1452, and LTC1453.
Figures 1 and 2 show the block
diagrams of the LTC1257 and the
LTC1451, LTC1452, and LTC1453.
These devices share many common
circuit elements. With a logic low
applied to the CS input, serial data is
applied to a 12-bit serial shift register. Each bit is latched on the rising
LOGIC
SUPPLY
CLK
DIN
5V REGULATOR
VCC
12-BIT
SHIFT REGISTER
DOUT
12
LOAD
GND
12-BIT LATCH
12
REF
2.048V REFERENCE
+
DAC
VOUT
–
dIDAC_1.eps
Figure 1. LTC1257 block diagram
edge of the CLK input signal. The
shift register’s contents are transferred to the DAC register on the
rising edge of CS, which in turn updates the output voltage’s magnitude.
Each of these DACs also has a data
output, DOUT, making it easy to
daisy-chain multiple DACs on a single
serial data line. The LTC1257 and
LTC1451 have an on-chip 2.048V
bandgap reference. The LTC1453 has
a 1.22V bandgap reference. The
LTC1257 has an internal 5V regulator that powers all onboard digital
circuitry.
System Autoranging
System auto-ranging, adjusting an
ADC’s full-scale range, is an application area for which these DACs are
appropriate. Auto-ranging is especially useful when using an ADC with
multiplexed inputs. Without autoranging, only two reference values
are used: one to set the full-scale
magnitude and another to set the
zero-scale magnitude. Since it is common to have input signals with
different zero-scale and full-scale
magnitude requirements, fixed reference voltages present a problem.
Although the ranges selected for some
of the inputs may take advantage of
the full range of ADC output codes,
inputs that do not span the same
range will not generate all codes, reducing the ADC’s effective resolution.
One possible solution is to match the
reference voltage span to the need of
each multiplexer input.
The circuit shown in Figure 3 uses
two LTC1257s to set the full-scale
and zero operating points of the
LTC1296 12-bit, 8-channel ADC. The
ADC shares its serial interface with
the DACs. To further simplify bus
connections, the DACs’ data is daisychained. Two chip selects are used,
one to select the LTC1296 when programming its multiplexer, and the
other to select the DACs when setting
their output voltages.
Table 1. Key specifications of the LTC1257, LTC1451, LTC1452, and LTC1453 8-pin, SO packaged, voltage-out DACs
Part Number
LTC1257
LTC1451
LTC1452
LTC1453
26
Supply
Voltage Range
4.75V to 15.75V
4.5V to 5.5V
2.7V to 5.5V
2.7V to 5.5V
Supply Current (Typ)
350µA at 5V Supply
400µA at 4.5V ≤ VCC ≤ 5.5V
225µA at 2.7V ≤ VCC ≤ 5.5V
250µA at 2.7V ≤ VCC ≤ 5.5V
Differential
Nonlinearity (Max)
±0.5 LSB
±0.5 LSB
±0.5 LSB
±0.5 LSB
Integral Linearity
Error (Max)
±3.5 LSB
±3.5 LSB
±3.5 LSB
±3.5 LSB
Linear Technology Magazine • August 1995
DESIGN IDEAS
8 VCC
CLK 1
LD
DIN 2
12-BIT
SHIFT
REGISTER
12-BIT DAC
DAC
REGISTER
+
7 VOUT
–
CS/LD 3
REFERENCE
LTC1451: 2.048V
LTC1453: 1.22V
POWER-ON
RESET
6 REF
DOUT 4
5 GND
dIDAC_2.eps
Figure 2. Block diagram of the LTC1451, LTC1452, and LTC1453
to channel 2 from the LTC1296. As
other multiplexer channels are selected, the DAC outputs are changed
to match their spans.
During the conversion process, U2
and U3 receive the full- and zeroscale codes, respectively, that
correspond to a selected multiplexer
channel. For example, let channel 2’s
span begin at 2V and end at 4.5V.
When a host processor wants a conversion of channel 2’s input signal, it
first sends code that sets the output
of U2 to 2V and U3 to 4.5V, fixing the
span to 2.5V. The processor sends
data to the LTC1296, selecting channel 2. The processor next reads data
that was generated during the conversion of the 3.5VP–P signal applied
Computer Controlled
4mA-to-20mA Current Loop
A common and useful circuit is the
4mA-to-20mA current loop. It is used
to transmit information over long distances using varying current levels.
The advantage of using current over
voltage is the absence of IR losses and
the transmissions errors and signal
losses they can create.
+22µF
5V
VCC
CH0
8 ANALOG
INPUT CHANNELS
•
•
•
U1
LTC1296
CH7
COM
REF + REF –
CS
DOUT
CLK
µP
DIN
SSO
50k
50k
5V
74HC04
0.1µF
VCC
100Ω
VOUT
0.1µF
VOUT
0.1µF
Opto-Isolated Serial Interface
CLK
U2
LTC1257 LOAD
GND VREF
VCC
100Ω
DIN
DOUT
DIN
CLK
U3
LTC1257 LOAD
GND VREF
DOUT
dIDAC_3.eps
Figure 3. Using two LTC1257 12-bit voltage-output DACs to set the input span of the 12-bit,
8-channel LTC1296
Linear Technology Magazine • August 1995
The circuit in Figure 4 is a computer controlled 4mA-to-20mA
current loop. It is designed to operate
on a single supply over a range of 3.3V
to 30V. The circuit’s zero output reference signal, 4mA, is set by R1 and
calibrated using R2, and its full-scale
output current is set by R3 and calibrated using R4. The zero- and
full-scale output currents are set as
follows: with a zero input code applied
to the LTC1453, the output current,
IOUT, is set to 4 mA by adjusting R2;
next, with a full-scale code applied to
the DAC, the full-scale output current is set to 20mA by adjusting R4.
The circuit is self regulating, forcing the output current to remain
stable for a fixed DAC output voltage.
This self regulation works as follows:
starting at t = 0, the LTC1453’s fixed
output (in this example, 2.5V) is
applied to the left side of R3; instantaneously, the voltage applied to the
LT1077’s input is 1.25V; this turns
on Q1 and the voltage across RS starts
increasing beginning from 1.25V; as
the voltage across RS increases, it
lifts the LTC1453’s GND pin above
0V; the voltage across RS continues to
increase until it equals the DAC’s
output voltage.
Once the circuit reaches this stable
condition, the constant DAC output
voltage sets a constant current
through R3 + R4 and R5. This constant current fixes a constant voltage
across R5 that is also applied to the
LT1077’s noninverting input. Feedback from the top of RS is applied to
the inverting input. As the op amp
forces its inputs to the same voltage,
it will fix the voltage at the top of RS.
This in turn fixes the output current
to a constant value.
The serial interface of the LTC1451
family and the LTC1257 make
opto-isolated interfaces very easy and
cost effective. Only three opto-isolators are needed for serial data
communications. Since the inputs of
the LTC1451, LTC1452, and LTC1453
have generous hysteresis, the switching speed of the opto-isolators is not
continued on page 31
27
DESIGN IDEAS
Wideband RMS Noise Meter
by Mitchell Lee
Recently, I needed to measure and
optimize the wideband RMS noise of
a power supply over about a 40MHz
bandwidth. A quick calculation
showed that the 12nV–15nV/√Hz
noise floor of my spectrum analyzer
would come up short—my circuit was
predicted to exhibit a spot noise of
perhaps 8nV–10nV/√Hz. In fact,
I didn’t have a single instrument in
my lab that would measure 50µV–
60µV RMS.
For the 40MHz bandwidth, the
HP3403C RMS voltmeter is a good
choice, but its most sensitive range is
100mV, about 66dB shy of my requirement. This obsolete instrument
today carries a hefty price on the used
market. The fact that here in the
Silicon Valley HP3403Cs are a common sight at flea markets is of little
consolation to most customers wishing to reproduce my measurements.
We have several of these meters in the
LTC design lab, but they are in constant use and closely guarded by “The
Keepers of the Secret RMS Knowledge.” I resolved to build my own
meter using an LT1088 thermal RMS
converter.
Full scale on the LT1088 is 4.25V
RMS. To measure 50µV full-scale, I’d
need an amplifier with a gain of
100,000. At 40MHz bandwidth, this
didn’t sound like it would have a good
chance of working first-time—built
by hand and without benefit of a
custom casting.
Rather than build a circuit with
40MHz bandwidth and a gain of
100dB, I decided to use just enough
gain to put my desired noise performance around twice minimum scale.
Aside from gain, this amplifier would
also need less than 5nV/√Hz input
noise, and the output stage would
have to drive the 50 ohm load presented by the LT1088.
It wasn’t hard to find an appropriate output stage. The LT1206 (see
Figure 1) can easily drive the required
120mA peak current into the LT1088
converter, and there’s plenty left over
for handling noise spikes. To preserve 40MHz bandwidth, the LT1206
was set to run at a gain of 2.
The front end was harder to solve.
I needed a low noise, high speed
amplifier that could give me plenty of
gain. Here I selected the LT1226. This
is a 1GHz GBW op amp with only
2.6nV/√Hz input noise. It has a minimum stable gain of 25, but in this
circuit, high gain is an advantage.
Cascading two LT1226s on the
front end gives a gain of 625, a little
shy of the 5,000 to 10,000 required.
Another gain of 5, plus the gain of 2 in
the LT1206 adds up to a gain of
6,250—just about right.
There are several ways to get 40MHz
bandwidth at a gain of 5, including
the LT1223 and LT1227 current-feedback amplifiers, but I settled on the
LT1192 voltage amplifier because it
is the lowest cost solution. This brings
the gain up to 6250, for a minimum
scale sensitivity of 34µV RMS, and a
full-scale sensitivity of 680µV RMS.
My advice-filled coworkers assured
me that there was no way I could
build a wideband amplifier with a
gain of 6250, and make it stable.
Nevertheless, I built my amplifier on a
+5V
10nF
10nF
5V
3
SIGNAL
INPUT
+
100nF
–
10nF
10nF
7
6
LT1226
2
5V
10nF
10nF
3
+
100nF
4
7
5
6
LT1226
2
10nF
–
3
100nF
4
7
–
TO
LT1206
4
10nF
1200Ω
–5V
6
LT1192
2
10nF
1200Ω
+
1200Ω
–5V
–5V
1000Ω
100Ω
100Ω
1000Ω
100Ω
5V
100Ω
1000Ω
300Ω
–5V
100nF
100nF
dI1226_1.eps
Authors can be contacted
at (408) 432-1900
Figure 1. Noise meter gain stage
28
Linear Technology Magazine • August 1995
DESIGN IDEAS
1.5" × 6" copperclad board, taking
care to maintain a linear layout. The
finished circuit was stable, provided
that a coaxial connection was made
to the input. The amplifier was flat,
with 3dB points at 4kHz and 43MHz,
and some peaking at high frequencies.
The performance of the amplifier
and thermal converter can be optimized by adjusting the value of the
feedback and gain setting resistors
around the LT1206. Slightly more
bandwidth can be achieved at the
expense of higher peaking by reducing the resistor values 10%. Reducing
the resistor values will decrease
peaking effects at the expense of
bandwidth. A good compromise value
is 680Ω.
I’ve shown the LT1226 amplifiers
operating from ±5 supplies, which
puts their bandwidth on the edge at
15V
10nF
10nF
100nF
24kΩ
SHUTDOWN CONTROL
100nF
FROM
LT1192
2
4
+
3
LT1206
1000Ω
1
–
7
5
6
10nF
TO LT1088
10nF
100nF
620Ω
–15V
620Ω
dI1226_2.eps
Figure 2. LT1206 buffer/driver section
3.3nF
15V
500Ω
27.4k
1%
10nF
FROM
LT1206
2
1k
1%
27.4k
1%
1k
1%
10nF
12
6
5
9
5
3
1
14
SUB
13
6
2
7
8
–15V
7
1k
2N2219
+
+
U4
1/4
LT1014
1
–
5V
33k
4
2N3904
U1
1/4
LT1014
15V
SHUTDOWN CONTROL
30k
–
15V
9.09M
1%
LT1088
SUB
9.09M
1%
+
8
12
10
10k
U3
1/4
LT1014 9
22nF
–
12k
13
+
U2
1/4
LT1014
–
14
OUTPUT
10k
11
5V
10k
–5V
100nF
100nF
dI1226_3.eps
Figure 3. LT1088 RMS detector section
Linear Technology Magazine • August 1995
29
DESIGN IDEAS
Coaxial Measurements
When measuring low-level signals, it is difficult to get a clean,
accurate result. Scope probes have
two problems. 10× probes attenuate
the already small signal, and both
1× and 10× probes suffer from circuitous grounds. Coaxial adapters
are a partial solution, but these are
expensive. They make for a lot of
wear and tear on the probes, and,
without a little forethought, they
can be a bear to attach to the circuit
under test. My favorite way to get
clean measurements of small signals is to directly attach a short
length of coaxial cable as shown in
Figure 4.
I use the good part of a damaged
BNC cable, cutting away the shorter
portion to leave at least 18" of RG58/U and one good connector. At
the cut, or as I call it, “real world”
end of the cable, I unbraid, twist,
and tin a very small amount of outer
conductor to form a stub 1/4" to
3/8" long. Next, I cut away the dielectric, exposing a similar length of
center conductor, which I also tin.
Now the probe is ready for use. It can
be soldered directly to a circuit or
breadboard, eliminating any lead
length that might otherwise pick up
stray noise, or worse, act as an antenna in a sensitive, high gain circuit.
Small signals aren’t the only beneficiaries of this technique. This works
great for looking at ripple on the outputs of switching supplies. Ripple
measurements are simplified because
the large voltage swings associated
with the switch node are completely
isolated and no loop is formed where
di/dt could inject magnetically
coupled noise.
In some instances, I’ve found it
important to retain the 50Ω termination impedance on the cable, but it is
rarely possible to place a terminator
at the “BNC” end of the cable, since
this creates a DC path directly across
the circuit under test. There is, however, another way as shown in Figure
5. Here, a technique known as backtermination is used. No termination
is used at the far end of the cable, but
a 51Ω resistor is connected in series
with the measurement end. Signals
sent down the cable reach the BNC
connector without attenuation, and
fast edges that bounce off the
unterminated end are absorbed back
into the 51Ω source resistor. I’ve found
this especially useful for measuring
fast switch signals, or, when measuring the RMS value of small signals,
for ensuring that the amplifier input
sees a properly terminated source.
The resistor trick does not work if the
node under test is high impedance;
an FET probe is a better choice for
high-impedance measurements.
While I’m at it, I might as well
give away my only other secret. We’ve
all encountered ground loop problems, giving rise to 60Hz (50Hz for
my friends overseas) injection into
sensitive circuits. Every lab is replete with isolation transformers and
“controlled substance” line cords
with missing ground prongs for battling ground loops. There are similar
problems at high frequencies, but
the victim is wave fidelity, not AC
pick up. To determine whether or
not high frequency grounding,
ground loops, or common-mode rejection is a problem for your
oscilloscope, simply clamp a small
ferrite E core around the probe lead
while observing any effects on the
waveform (see Figure 6). Sometimes
the news is bad; the waveform really
is messed up and there is some work
to be done on the circuit. But occasionally the circuit is exonerated,
the unexplainable aberrations disappear, proving that high frequency
gremlins are at work. If necessary,
several passes of the probe cable
can be made through the E-Core,
and it can be taped together for as
long as needed.
51Ω
dicoax_2.eps
dicoax_1.eps
Figure 4. BNC cable used as a “probe”
40MHz. Their bandwidth can be improved by operating at ±15V.
Because the LT1206 operates on
15V rails, it is possible to overdrive
the LT1088 and possibly cause permanent damage. One section of the
30
Figure 5. BNC cable probe back terminated
LT1014 (U3) is used to sense an overdrive condition on the LT1088 and
shut down the LT1206. Sensing the
feedback heater instead of the input
heater allows the LT1088 to accommodate high-crest-factor waveforms,
shutting down only when the average
input exceeds maximum ratings.
By the way, my power supply noise
measured 200µV; filtering brought it
down to less than 60µV.
Linear Technology Magazine • August 1995
DESIGN IDEAS
dicoax_3.eps
Figure 6. An E-Core serves to choke off high frequency common mode currents from a ’scope probe
VLOOP
3.3V TO 30V
12-Bit DACs, continued from page 27
LT1121-3.3
IN
OUT
critical. Further, because each of these
DACs can be daisy-chained to others,
only three opto-isolators are required.
Linear Technology Magazine • August 1995
R2
5k
R3
45k
R4
5k
1µF
CLK
FROM
OPTOISOLATED
INPUTS
Control Right
Where It’s Needed
Two major benefits of these DACs
are their small SO-8 size and serial
interface. These features make it very
easy to place an LTC1257 or an
LTC1451, LTC1452, or LTC1453 on a
PCB exactly where needed, next to its
associated circuitry. Unlike the compromises that exist with placing a
multi-DAC package and the difficulty
in optimizing analog-signal-trace
length and routing, the placement
flexibility of the LTC1257 and
LTC1451, LTC1452, or LTC1453 reduces analog-signal-path lengths to
a minimum.
R1
90k
OPTO-ISOLATORS
DIN
500Ω
VREF
10k
4N28
+
LTC1453 VOUT
CS/LD
3.3V
CLK
DIN
CS/LD
VCC
R5
3k
CLK
DIN
CS/LD
LT1077
–
1k
Q1
2N3440
RS
10Ω
IOUT
dIDAC_4.eps
Figure 4. The LTC1453 forms the heart of this isolated 4mA-to-20mA current loop
Conclusion
The LTC1257 and the LTC1451,
LTC1452, and LTC1453 serially programmed 12-bit DACs are ideal
solutions for applications that
require 12-bit functionality in small
SO-8 packages and the ease and
economy of a serial interface.
31
DESIGN IDEAS
LTC1477: 70mΩ Protected
High-Side Switch Eliminates
“Hot Swap” Glitching
When a printed circuit board is
“hot swapped” into a live 5V socket, a
number of bad things can happen.
First the instantaneous connection of a large, discharged,
supply-bypass capacitor may cause
a glitch to appear on the power bus.
The current flowing into the capacitor is limited only by the socket
resistance, the card-trace resistance,
and the equivalent series resistance
(ESR) of the supply bypass capacitor.
This supply glitch can create real
havoc if the other boards in the system have power-on RESET circuitry
with thresholds set at 4.65V.
Second, the card itself may be damaged due to the large inrush of current
into the card. This current is sometimes inadvertently diverted to
sensitive (and expensive) integrated
circuits that cannot tolerate either
over-voltage or over-current conditions even for short periods of time.
by Tim Skovmand
As the card is inserted, the LTC699
power-on RESET circuit holds the
enable pin of the LTC1477 low for
approximately 200ms. When the enable pin is asserted high, the output
is ramped on in approximately 1ms.
Even if a very large supply bypass
capacitor (for example, over 100µF) is
used, the LTC1477 will limit the inrush current to 2A and ramp the
capacitor at an even slower rate. Further, the board is protected against
short-circuit conditions by limiting
the switch current to 2A.
The 5V card supply can be disabled via Q1. The only current flowing
is the standby quiescent current
of the LTC1477, which drops below
1µA, the 600µA quiescent current of
the LTC699, and the 10µA consumed
by R1.
Third, if the card is removed and
then reinserted in a few milliseconds,
the glitching of the supply may “confuse” the microprocessor or peripheral
ICs on the card, generating erroneous data in memory or forcing the
card into an inappropriate state.
Fourth, a card may be shorted,
and insertion may either grossly glitch
the 5V supply or cause severe physical damage to the card.
Figure 1 is a schematic diagram
showing how an LTC1477 protected
high-side switch and an LTC699
power-on RESET circuit reduce the
chance of glitching or damaging the
socket or card during “hot swapping.”
The LTC1477 protected high-side
switch provides extremely low RDS(ON)
switching (typically 0.07Ω) with
built-in 2A current limiting and
thermal shutdown, all in an 8-pin
SO package.
5V
C1
1µF
1
VOUT
2
NC
CLOAD
100µF
VIN2
LTC1477CS8
VIN2
VIN3
7
LTC699CS8
+
VIN1
R1
510k
6
VOUT
VOUT,
ISC = 2A
EN
GND
DISABLE
3
4
5
8
Q1
2N7002
dI1477_1.eps
Figure 1. “Hot swap” circuit featuring LTC1477 and LTC699
Authors can be contacted
at (408) 432-1900
32
Linear Technology Magazine • August 1995
DESIGN IDEAS
Driving a High-Level Diode-Ring
Mixer with an Operational Amplifier
by Mitchell Lee
15V
15V
10nF
100nF
1kΩ
10nF
PN4416
24k
62pF
100nF
SHUTDOWN
120k
10MHz
2
110nF
10k
+
4
LT1206
1
–
6
3
OPERATE
7
51Ω, 2W
OPTIONAL DOUBLETERMINATION RESISTOR
RF
LO
IF
5 10nF
620Ω
10nF
5k
100nF
–15V
–15V
OSCILLATOR (TYPICAL)
BUFFER/AMPLIFIER
DIODE RING MIXER
dI1206_1.eps
Figure 1. Oscillator buffer drives +17 to +27dBm double balanced mixers
One of the most popular RF building blocks is the diode-ring mixer.
Consisting of a diode ring and two
coupling transformers, this simple
device is a favorite with RF designers
anywhere a quick multiplication is
required, as in frequency conversion,
frequency synthesis, or phase detection. In many applications these
mixers are driven from an oscillator.
Rarely does anyone try building an
oscillator capable of delivering +7dBm
for a “minimum geometry” mixer, let
alone one of higher level. One or more
stages of amplification are added to
achieve the drive level required by the
mixer. The new LT1206 high-speed
amplifier makes it possible to amplify
an oscillator to +27dBm in one stage.
Figure 1 shows the complete circuit diagram for a crystal oscillator,
LT1206 op amp/buffer, and diodering mixer. Most of the components
Linear Technology Magazine • August 1995
+
30dBm
LT1206
–
50Ω
620Ω
dI1206_2.eps
Figure 2. Spectrum plot of Figure 1’s circuit driving +30dBm into a 50Ω load (single
termination)
33
DESIGN IDEAS
+
are used in the oscillator itself, which
is of the Colpitts class. Borrowing
from a technique I first discovered in
Hewlett Packard’s Unit Oscillator, the
current of the crystal is amplified,
rather than the voltage. There are
several advantages to this method,
the most important of which is low
distortion. Although the voltages
present in this circuit have poor wave
shape and are sensitive to loading,
the crystal current represents essentially a filtered version of the voltage
waveform and is relatively tolerant of
loading effects.
27dBm
LT1206
–
50Ω
620Ω
dI1206_3.eps
Figure 3. Spectrum plot of Figure 1’s circuit driving +27dBm into a 50Ω load (single termination)
+
The impedance, and therefore the
voltage at the bottom of the crystal is
kept low by injecting the current into
the summing node of an LT1206 current-feedback amplifier. Loop gain
reduces the input impedance to well
under one ohm. Oscillator bias is
adjustable, allowing control of the
mixer drive. This also provides a convenient point for closing an output
power servo loop.
23dBm
LT1206
–
One of the most popular
RF building blocks is the
diode-ring mixer.
Consisting of a diode
ring and two coupling
transformers, this simple
device is a favorite with
RF designers anywhere a
quick multiplication is
required, as in frequency
conversion, frequency
synthesis, or phase
detection
50Ω
620Ω
dI1206_4.eps
Figure 4. Spectrum plot of Figure 1’s circuit driving +23dBm into a 50Ω load (single
termination)
34
Linear Technology Magazine • August 1995
DESIGN IDEAS
Operating from ±15V supplies, the
LT1206 can deliver +32dBm to a 50Ω
load, and, with a little extra headroom (the absolute maximum supply
voltage is ±18), it can reach 2W output power into 50Ω. Peak guaranteed
output current is 250mA.
Shown in Figures 2–6 are spectral
plots for various combinations of
single and double termination at
power levels ranging from +17 to
+27dBm—not bad for an inductorless circuit. Double termination may
be used to present a 50Ω source impedance to the mixer, or to isolate two
or more mixers driven simultaneously
from one LT1206 amplifier.
Although a 10MHz example has
been presented here, the LT1206’s
65MHz bandwidth makes it useful in
circuits up to 30MHz. In addition, the
shutdown feature can be used to gate
drive to the mixer. When the LT1206
is shut down, the oscillator will likely
stop, since the crystal then sees a
series impedance of 620Ω and the
mixer itself. Upon re-enabling the
LT1206 there will be some time delay
before the oscillator returns to full
power. The circuit works equally well
with an LC version of the oscillator.
Note that the current feedback topology is inherently tolerant of stray
capacitive effects at the summing
node, making it ideal for this application. Another nice feature is the
LT1206’s ability to drive heavy capacitive loads while remaining stable
and free of spurious oscillations.
For mixers below +17dBm, the
LT1227 is a lower cost alternative,
featuring 140MHz bandwidth in combination with the shutdown feature
of the LT1206.
+
50Ω
27dBm
LT1206
–
50Ω
620Ω
dI1206_5.eps
Figure 5. Spectrum plot of Figure 1’s circuit driving +27dBm into a 100Ω load (double
termination)
+
50Ω
17dBm
LT1206
–
50Ω
620Ω
dI1206_6.eps
Figure 6. Spectrum plot of Figure 1’s circuit driving +17dBm into a 100Ω load (double
termination)
Linear Technology Magazine • August 1995
35
DESIGN IDEAS
Fully Differential, 8-Channel, 12-Bit
A/D System Using the LTC1390
and LTC1410
by Kevin R. Hoskins
The LTC1410’s fast 1.25Msps conversion rate and differential ±2.5V
input range make it ideal for applications that require multichannel
acquisition of fast, wide-bandwidth
signals. These applications include
multitransducer vibration analysis,
race-vehicle telemetry data acquisi+5V
tion, and multichannel telecommunications. The LTC1410 can
be combined with the LTC1390
8-channel serial-interfaced analog
multiplexer to create a differential
A/D system with conversion throughput rates up to 625ksps. This rate
applies to situations where the
selected channel changes with each
conversion. The conversion rate increases to 1.25Msps if the same
channel is used for consecutive conversions.
Figure 1 shows the complete differential, 8-channel A/D circuit. Two
LTC1390s, U1 and U2, are used
–5V
0.1µF
S0
+CH1
S1
V+
U1
LTC1390
D
+CH2
S2
V–
+CH3
S3
DATA2
+CH4
S4
DATA1
+CH5
S5
CS
+CH6
S6
CLK
+CH7
S7
0.1µF
+
10µF
GND
0.1µF
+AIN
AVDD
–AIN
DVDD
VREF
VSS
REFCOMP
D11
D10
CHIP SELECT MUX
SERIAL CLK
0.1µF
0.1µF
BUSY
AGND
SERIAL DATA
10µF
+
+CH0
+
10µF
CS
CONVST
U3
LTC1410
RD
D9
SHDN
D8
NAP/SLP
D7
OGND
D6
D0
D5
D1
D4
D2
DGND
D3
µP CONTROL
LINES
0.1µF
V+
–CH0
S0
–CH1
S1
–CH2
S2
V–
–CH3
S3
DATA2
–CH4
S4
DATA1
–CH5
S5
CS
–CH6
S6
CLK
–CH7
S7
GND
U2
LTC1390
D
0.1µF
12-BIT
PARALLEL
BUS
dI1390_1.eps
Figure 1. Fully differential 8-channel data acquisition system achieves 625ksps throughput
36
Linear Technology Magazine • August 1995
DESIGN IDEAS
as noninverting and inverting input
multiplexers. The outputs of the noninverting and inverting multiplexers
are applied to the LTC1410’s +AIN
and −AIN inputs, respectively. The
LTC1390 share the CHIP SELECT
MUX, SERIAL DATA and SERIAL
CLOCK control signals. This arrangement simultaneously selects the same
channel on each multiplexer: S0 for
both +CH0 and −CH0, S1 for both
+CH1 and −CH1, and so on.
As shown in the timing diagram
(Figure 2), MUX channel selection
and A/D conversion are pipelined to
maximize the converter’s throughput.
The conversion process begins with
selecting the desired multiplexer
channel-pair. With a logic high applied to the LTC1390’s CS input, the
channel-pair data is clocked into each
DATA 1 input on the rising edge of the
5MHz clock signal. CHIP SELECT
MUX is then pulled low, latching the
channel-pair selection data. The signals on the selected MUX inputs are
then applied to the LTC1410’s differential inputs. CHIP SELECT MUX is
pulled low 700ns before the LTC1410’s
conversion-start input, CONVST, is
pulled low. This accounts for the maximum time needed by the LTC1390’s
MUX switches to fully turn-on. This
ensures that the input signals are
fully settled before the LTC1410’s S/H
captures its sample.
The LTC1410’s S/H acquires the
input signal and begins conversion
on CONVST’s falling edge. During the
conversion, the LTC1390’s CS input
is pulled high and the data for the
next channel-pair is clocked into
DATA 1. This pipelined operation
continues until a conversion sequence
is completed. When a new channelpair is selected for each conversion,
the sampling rate of each channel is
78ksps, allowing an input-signal
bandwidth of 39kHz for each channel
of the LTC1390/LTC1410 system.
To maximize the throughput rate,
the LTC1410’s CS input is pulled low
at the beginning of a series of conversions. The LTC1410’s data output
drivers are controlled by the signal
applied to RD. The conversion’s results are available 20ns before the
rising edge of BUSY. The rising edge
of the BUSY output signal can be
used to notify a processor that a conversion has ended and data is ready
to read.
This circuit takes advantage of the
LTC1410’s very high 1.25Msps conversion rate and differential inputs
and the LTC1390’s ease of programming to create an A/D system
that maintains wide input-signal
bandwidth while sampling multiple
input signals.
SERIAL CLOCK LTC1390
5MHz, 200ns
CHIP SELECT MUX LTC1390
DATA LTC1390
EN
B2
B1
B0
EN
B2
B1
B0
EN
B2
B1
B0
EN
B2
B1
B0
CONVST LTC1410
BUSY LTC1410
700ns
CONVERSION 0
CONVERSION 1
CONVERSION 2
RD LTC1410
OUTPUT DATA LTC1410
DATA 0
DATA 1
DATA 2
dI1390_2.eps
A LOGIC LOW IS APPLIED TO THE LTC1410S CS INPUT DURING A CONVERSION SEQUENCE.
Figure 2. Timing diagram of circuit in Figure 1
Linear Technology Magazine • August 1995
37
NEW DEVICE CAMEOS
New Device Cameos
LTC1266/LTC1266-3.3/
LTC1266-5 Synchronous
Controllers Feature Precision
Output Regulation
The LTC1266, LTC1266-3.3, and
LTC1266-5 synchronous controllers
for N- or P-channel MOSFETs provide
a precision 1% internal reference for
tighter regulation of output voltage.
This precision reference, combined
with the LTC1266’s tight load regulation and ability to drive two N-channel
MOSFETs for high currents, provides
the accuracy and current capability
necessary to meet the stringent
demands of many Pentium microprocessor applications.
The LTC1266 is a synchronous,
current mode, switching regulator
controller with automatic Burst Mode
operation (Burst Mode operation can
be disabled), available in the 16-lead
narrow SO package. Two user-programmable pins make it possible to
drive an N-channel MOSFET for high
current or step-up operation, or a Pchannel MOSFET for low dropout
operation. With an N-channel topside MOSFET for higher efficiency at
high currents and Burst Mode operation enabled for high efficiencies at
low currents, the LTC1266 can provide 90% or better efficiency for loads
from 10mA to 10A.
Other features of the LTC1266 include a shutdown mode, which
reduces supply current to 40µA, and
an on-chip comparator for use as a
“power-good” monitor, which stays
active in shutdown. The maximum
operating frequency of 400kHz allows the use of small, surface mount
inductors and capacitors.
LT1368 and LT1369
Precision Rail-to-Rail Op
Amps with Excellent HighFrequency Supply Rejection
The LT1368 dual and LT1369 quad
are the latest members of the LT1366
rail-to-rail op amp family. Designed
to operate with a 0.1µF output-compensation capacitor, these devices
feature improved high frequency supply rejection and lower high frequency
output impedance. The output capacitance acts as a filter, reducing
noise pickup and ripple from the supply. This additional filtering is helpful
in mixed analog/digital systems with
common supplies, or in systems employing switching supplies.
Like the LT1366, the LT1368 and
LT1369 combine rail-to-rail input and
output operation with precision specifications. These devices maintain their
characteristics over a supply range of
1.8V to 36V. Operation is specified
with +3V, +5V, and ±15V supplies.
The input offset voltage is typically
150µV, with a minimum open loop
gain (AVOL) of 1 million while driving a
2k load. The typical input-bias and
offset currents are 10nA and 1nA
respectively. The supply current per
amplifier is typically 375µA.
Common mode rejection is typically 90dB over the full rail-to-rail
input range, and supply rejection is
110dB (DC). The filter formed with
the amplifier’s output impedance and
the external compensating capacitor
provides approximately 40dB of supply rejection above 30kHz.
The LT1368 dual is available in
plastic 8-pin DIP and 8-lead SO
packages. The LT1369 quad is available in a plastic 16-lead SO package.
LT1186 DAC-Programmable
CCFL Switching Regulator
(Bits-to-NitsTM)
The LT1186 is a fixed frequency,
current mode switching regulator that
provides the control function for coldcathode fluorescent lighting. The
LT1186 supports grounded-lamp and
floating-lamp configurations. The IC
includes an efficient, high-current
switch, an oscillator, output drive
logic, control circuitry, protection circuitry, and a micropower 8-bit, 50µA,
full-scale current output DAC.
The DAC provides simple “bitsto-lamp-current control,” and
communicates externally in two
interface modes: standard SPI mode
and pulse mode. On power-up, the
DAC counter is reset to half-scale
and the chip is configured to SPI or
pulse mode, depending on the CS
signal level. In SPI mode, the system
microprocessor serially transfers the
present 8-bit data and reads back the
previous 8-bit data from the DAC. In
pulse mode, the upper six bits of the
DAC implement increment-only
(1-wire interface) or increment/decrement (2-wire interface) operation.
The operational mode (increment-only
or increment/decrement) is selected
by the DIN signal level.
The LT1186 operates 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
typically reduces total supply current
to 35µA for standby operation while
the DAC retains its last setting. A
200kHz switching frequency minimizes the size of required magnetic
components and the use of currentmode switching techniques gives high
reliability and simple loop-frequency
compensation. The LT1186 is available in a 16-pin narrow body SO.
LTC1477/LTC1478: Single
and Dual 70mΩ, 2A,
Protected High-Side Switches
in SO Packaging
The LTC1477/LTC1478 single and
dual protected high-side switches
provide extremely low RDS(ON) switching with built in protection against
short-circuit and thermal-overload
conditions. The LTC1477 single is
available in 8-lead SO packaging and
the LTC1478 dual is available in
16-lead SO packaging. Both parts
operate from 2.7V to 5.5V.
The LTC1477/LTC1478 provide
two levels of protection. The first level
Bits-to-Nits is a trademark of Linear Technology Corporation.
Pentium is a registered trademark of Intel Corporation.
38
Linear Technology Magazine • August 1995
NEW DEVICE CAMEOS
is short-circuit current limiting, which
is set at 2A. The short-circuit current
can be reduced to as low as 0.85A by
disconnecting portions of the power
device. The second level of protection
is thermal-overload protection, which
limits the die temperature to approximately 130°C. A built in charge pump
produces gate drive from the 2.7V-to5.5V power supply input for
controlled ramping of the internal
ultra-low RDS(ON) NMOS switch. No
higher gate voltage power supply is
required. Rise time is set internally at
1ms, to reduce inrush currents to
negligible levels; therefore, extremely
large load capacitors can be ramped
without glitching the system power
supply. The switch resistance is typically 0.07Ω at 5V and only rises to
0.08Ω at 3.3V. Output leakage current drops to 0.01µA when the output
is switched off.
The NMOS switch has no parasitic
body diode; therefore, no current
flows through the switch when it is
turned off and the output is forced
above the input supply voltage. (DMOS
switches have parasitic body diodes
that become forward biased under
these conditions.)
All operating modes are micropower
and the quiescent current automatically drops below 1µA when the
switch is turned off.
The combination of micropower
operation and full protection make
these devices ideal for battery powered applications such as palmtop
computers, notebook computers, personal digital assistants, instruments,
handi-ter minals, and bar -code
readers.
LTC1550/LTC1551: LowNoise, Switched-Capacitor,
Regulated Voltage Inverters
The LTC1550 and LTC1551 are
switched-capacitor voltage inverters
with onboard linear post regulators,
designed to generate a low noise, regulated negative supply for use as bias
voltage for GaAs transmitters FETs in
portable-RF and cellular-telephone
applications. They operate from a
single 4.5V to 7V supply and provide
a fixed −4.1V or adjustable output
Linear Technology Magazine • August 1995
voltage, with low output ripple (1mV
typical) and a typical quiescent current of 5mA at VCC = 5V. An on-chip
oscillator sets the charge pump frequency at 900kHz, away from
sensitive 400kHz–600kHz IF bands.
The LTC1550-4.1 and the LTC15514.1 provide fixed −4.1V outputs,
whereas the LTC1550 offers adjustable output voltage. Both devices
include a TTL-compatible shutdown
pin that drops supply current to 0.2µA
typical. The LTC1550 shutdown pin
is active low (SD), and the LTC1551
shutdown pin is active high (SD).
Only three external components
are required for the fixed-voltage
parts: two 0.1µF charge pump capacitors and a 10µF filter capacitor at
the linear regulator output. Each version of the LTC1550/LTC1551 will
supply up to 20mA output current
with guaranteed output regulation of
±5%. All except the 8-pin versions of
the LTC1550/LTC1551 include an
open-drain REG output, which pulls
low to indicate that the output is
within 5% of the set value. This allows external circuitry to monitor the
output without additional components. The adjustable LTC1550
requires two additional resistors to
set the output voltage.
The LTC1550-4.1 and LTC15514.1 are available in 8-pin SO, 14-pin
SO, and 16-pin SSOP packages. The
LTC1550 adjustable-output voltage
version is available in 14-pin SO and
16-pin SSOP packages.
Low-Cost, High-Accuracy, 12Bit Multiplying DAC Family
Linear Technology has introduced
the LTC8043, LTC8143, LTC7543,
and LTC7541A, four 12-bit, current
output, 4-quadrant multiplying digital-to-analog converters (DACs). They
are superior pin-compatible replacements for the industry standard
DAC-8043, DAC-8143, AD7543, and
AD7541A. Low cost, tight accuracy
specs, low power, small size, and outstanding flexibility make these DACs
excel in process control, remote or
isolated applications, and softwareprogrammable gain, attenuation, and
filtering. Linear Technology’s parts
have improved accuracy and stability, tighter timing specs, reduced
sensitivity to external amplifier VOS,
lower output capacitance, and reduced cost.
The LTC8043 comes in a small SO8 or PDIP package and has an
easy-to-use serial interface. The
LTC8143 and LTC7543 come in 16Pin SO or PDIP packages and feature
a more flexible serial interface that
includes an asynchronous Clear pin.
The LTC8143 also has a serial data
output so users can daisy-chain
multiple DACs on a 3-wire bus. The
LTC7541A has a 12-bit wide parallel
interface and comes in 18-pin SO and
PDIP packages.
These DACs boast unbeatable accuracy and stability, with INL and
DNL of 1/2LSB MAX, 1/8LSB TYP
over temperature. Gain Error is 1LSB
MAX, eliminating adjustments
in most applications. Gain and
nonlinearity temperature coefficients
are typically better than 1 ppm/°C
and 0.1 ppm/°C respectively.
All four; parts are offered in the
extended industrial temperature
range and the LTC7541A is also available in commercial grades.
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 service number: 1-800-4-LINEAR. Ask
for the pertinent data sheets.
Burst Mode is a trademark of Linear
Technology Corporation.
, LTC and LT
are registered trademarks used only to
identify products of Linear Technology Corp.
Other product names may be trademarks
of the companies that manufacture the
products.
I nformation 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.
39
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
Available at no charge.
PSPICETM by MicroSim.
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 commercial and military grades. The catalog features well over 300 devices. $10.00
1992 Linear Databook Supplement — This 1248 page supplement to the
1990 Linear Databook is a collection of all products introduced since then.
The catalog contains full data sheets for over 140 devices. The 1992 Linear
Databook Supplement is a companion to the 1990 Linear Databook, which
should not be discarded.
$10.00
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
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 nonLTC 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.
World Headquarters
Linear Technology Corporation
1630 McCarthy Boulevard
Milpitas, CA 95035-7487
Phone: (408) 432-1900
FAX: (408) 434-0507
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
SOUTHWEST REGION
Linear Technology Corporation
21243 Ventura Blvd.
Suite 227
Woodland Hills, CA 91364
Phone: (818) 703-0835
FAX: (818) 703-0517
International
Sales Offices
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
Untere Hauptstr. 9
D-85386 Eching
Germany
Phone: 49-89-3197410
FAX: 49-89-3194821
JAPAN
Linear Technology KK
5F NAO Bldg.
1-14 Shin-Ogawa-Cho, Shinjuku-Ku
Tokyo, 162 Japan
Phone: 81-3-3267-7891
FAX: 81-3-3267-8010
KOREA
Linear Technology Korea Branch
Namsong Building, #505
Itaewon-Dong 260-199
Yongsan-Ku, Seoul
Korea
Phone: 82-2-792-1617
FAX: 82-2-792-1619
SINGAPORE
Linear Technology Pte. Ltd.
507 Yishun Industrial Park A
Singapore 2776
Phone: 65-753-2692
FAX: 65-541-4113
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
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 management, 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.
LINEAR TECHNOLOGY CORPORATION
1630 McCarthy Boulevard
Milpitas, CA 95035-7487
(408) 432-1900
For Literature Only: 1-800-4-LINEAR
AppleTalk is a registered trademark of Apple Computer, Inc.
©40
1995 Linear Technology Corporation/ Printed in U.S.A./27K
Linear Technology Magazine • August 1995