V06N3 - AUGUST

LINEAR TECHNOLOGY
AUGUST 1996
IN THIS ISSUE . . .
COVER ARTICLE
Safe Hot Swapping
Using the LTC®1421 ........ 1
Robert Reay and James Herr
VOLUME VI NUMBER 3
Safe Hot Swapping
Using the LTC1421
by Robert Reay and James Herr
Issue Highlights .............. 2
DESIGN FEATURES
The Care and Feeding of
High Performance ADCs:
Get All the Bits You Paid For
........................................ 8
William C. Rempfer
LTC1433/LTC1434:
High Efficiency,
Constant-Frequency
Monolithic Buck Converter
...................................... 13
San-Hwa Chee
The LTC1343 and LTC1344
Form a Software-Selectable
Multiple-Protocol Interface
Port Using a DB-25 Connector
...................................... 15
Robert Reay
DESIGN IDEAS
Li-Ion Battery Charger
Does Not Require Precision
Resistors ....................... 25
Arie Ravid
Constant-Voltage Load Box
for Battery Simulation
...................................... 26
Jon Dutra
LT1510 Charger with
–∆V Termination ............ 27
Arie Ravid
(more Design Ideas on pages 28–41.
Complete list on page 25)
Design Tools .................. 43
Sales Offices ................. 44
When a circuit board is inserted
into a live backplane, the large bypass capacitors on the board can
draw huge inrush currents from the
backplane power bus as they charge.
The inrush current, on the order of 10
to 100 amps, can destroy the board’s
bypass capacitors, metal traces or
connector pins. The inrush current
can also cause a glitch on the backplane power bus, which may force all
of the other boards in the system to
reset. At the same time, the system
data bus can be disrupted when the
board’s data pins make or break
contact.
The LTC1421 can turn on up to
three board supply voltages at a programmable rate, allowing boards to
be safely inserted in or removed from
a live backplane. The chip also provides board connection sensing, a
method for halting the system data
bus during insertion or removal, flexible supply voltage monitoring,
power-on reset outputs, short-circuit
protection and digital input or pushbutton power cycling control.
Typical Application
Figure 1 shows a typical application
using the LTC1421. The power supplies on the board are controlled by
placing external N-channel pass transistors Q1, Q2 and Q3 in the power
path for VCC, VDD and VEE, where VCC
and VDD can range from 3.0V to
12.0V. By ramping the gates of the
pass transistors up or down at a
controlled rate, the transient surge
current (I = C × dv/dt) drawn from
the main backplane supply will be
limited to a safe value.
The LTC1421 is designed for use
with a staggered 3-level connector.
Ground should make connection first
to discharge any static buildup. VCC,
VDD, VEE and DISABLE should make
connection second and the data bus
and all other pins last. The connection sense pins CON1 and CON2 must
be located on opposite ends of the
connector because most people will
rock the board back and forth during
insertion.
The system timing is shown in
Figure 2. When the supply pins make
contact, (Figure 2, time point 1), the
LTC1421 prevents transistors Q1 and
Q2 from turning on by holding their
gates (GATELO and GATEHI) at
ground, while C3 and R4 keep Q3 off
by pulling its gate to –12V. The
Schottky diode in the output stage of
CPON allows the pin to be pulled
below ground. The two connection
sense pins, CON1 and CON2, are
initially pulled to VCCLO by 10k pullup resistors. PWRGD and RESET are
held low while VCCLO and VCCHI are
pulled to ground by internal transistors N1 and N2. At the same time,
DISABLE is pulled high, turning on
transistor Q4, which will halt traffic
on the backplane data bus before the
data pins make contact. CON1 and
CON2 are the last pins to make contact and are shorted to ground on the
backplane side of the connector. After
continued on page 3
, LTC and LT are registered trademarks of Linear Technology Corporation. Adaptive Power, C-Load, LinearView,
Micropower SwitcherCAD and Burst Mode are trademarks of Linear Technology Corporation. Other product names may
be trademarks of the companies that manufacture the products.
EDITOR'S PAGE
Issue Highlights
This issue of Linear Technology features more exciting new products from
Linear Technology’s design team. Our
lead article explains “safe hot swapping” using the LTC1421 hot swap
controller. This device prevents inrush
currents that can destroy components
and traces. It also eliminates glitches
that disrupt the system bus when a
circuit board is plugged into a “live”
backplane. The LTC1421 can turn on
up to three board supply voltages at
programmable rates. The part provides board connection sensing, which
can be used to halt the system bus
during board insertion or removal,
power-on reset outputs, short-circuit
protection and digital input or pushbutton power cycling.
Also in this issue, we have an
in-depth examination of the “Care
and Feeding of High Performance
ADCs,” such as the 1.25Msps, 12-bit
LTC1410. Designers will be finding
many new applications for these parts
because of their excellent specifications and attractive pricing. This
article provides the layout, bypassing
and other design techniques that will
allow you to obtain maximum performance from these precision parts.
In the area of power products, we
introduce the LTC1433 and LTC1434
high efficiency, constant-frequency
monolithic buck converters. These
devices achieve high efficiency and
constant frequency at low load currents using the Adaptive Power™
mode first introduced in the LTC1435–
LTC1439 family of DC/DC controllers.
Also new are a pair of interface
products, the LTC1343 and LTC1344.
The LTC1343 contains four drivers
and four receivers and the LTC1344
contains six switchable resistive terminators. Together, these two devices
can implement a software-selectable,
multiple-protocol serial port that supports a wide range of protocols.
This month, we have a generous
helping of Design Ideas, including
two battery chargers, plus a constant-voltage load box for battery
simulation; some unusual amplifier
2
LTC in the News...
“Technology stocks have fallen so
far in recent months that they have
become tantalizing to bottom-fishing value investors, who rarely get a
chance to buy normally pricey computer and semiconductor makers,”
wrote E.S. Browning, staff reporter
of the Wall Street Journal, in an
article in the Dow Jones Newswire.
“These companies now look like real
bargains,” he writes.
Chicago Corp. analyst David Wu
said, “Linear is a good company in a
great business with high barriers to
entry.” He added that he looks for a
resumption of double-digit growth
after a period of adjustment, which
could take a year.
Even more bullish on Linear is Ed
Jamieson, manager of the Franklin
Small Cap Growth fund, who told
Barron’s Financial Weekly in July
that he believes LTC can produce
25%–50% returns in the next 12
months. “Historically, its market
grew 35% a year…. It’s crazy not to
think this company will reemerge.”
The perceptions of these shrewd
industry observers are turning out
to be prophetic. On July 23, LTC
announced that net sales for its
fiscal year ended June 30 were a
record $377,771,000, an increase
of 43% over the previous year. The
company also reported record net
income for the year of $133,964,000
or $1.72 per share, an increase of
58% over $84,696,000 or $1.11 per
share reported for fiscal 1995.
According to Robert H. Swanson,
Jr., president and CEO of Linear
Technology Corp., “1996 was a very
strong year for us in sales, profits
and cash growth. Our rate of sales
growth was higher than our historical average in response to a very
robust market early in the year. The
current environment is balancing
out some of this accelerated growth
and our sales for the fourth quarter
were less than the previous quarter
for the first time in ten years,”
Swanson said.
“Net income as a percentage of
sales continued to be the strongest
in the industry. The long term prospects for our business are excellent
and we continue to invest in the
plant infrastructure and technical
talent to maximize our opportunities. In the short term, however,
reduced backlog and shorter lead
times have caused the business to
be more dependent on orders that
are received and shipped in the same
quarter,” Swanson said.
applications featuring the LT1210 and
the LT1336; a micropower voltage-tofrequency converter; a circuit for
measuring small capacitance changes
and several power supply designs.
Also new from Linear Technology
this month are the LinearView™
CD-ROM and the Micropower
SwitcherCAD™ design software.
The LinearView CD-ROM contains
all product data from Linear
Technology’s Databooks (volumes I–
IV) and applications information from
Linear’s Applications Handbooks (volumes I and II), plus the complete
collections of Design Notes and back
issues of Linear Technology magazine.
MicroPower SwitcherCAD is a powerful tool for designing DC/DC
converters based on Linear
Technology’s micropower switching
regulator ICs. Given basic design
parameters, MicropowerSCAD selects
a circuit topology and offers you a
selection of appropriate Linear Technology switching regulator ICs.
MicropowerSCAD also performs circuit simulations to select the other
components that surround the DC/
DC converter.
See page 43 for ordering information for either LinearView or
Micropower SwitcherCAD.
As always, we welcome your questions and comments. Call (408)
432-1900 to talk to the authors of
any of this issue’s articles about your
application needs.
Linear Technology Magazine • August 1996
DESIGN FEATURES
they have stopped bouncing (Figure
2, time point 2), the LTC1421 will
wait for 20ms.
At the end of 20ms, if VCCLO and
VCCHI have exceeded the undervoltage lockout threshold (2.45V), and
VOUTLO is less than 100mV above
ground, the LTC1421 is ready to turn
on the supplies (Figure 2, time point
3). A 20µA reference current is connected to the RAMP pin, the charge
R4
20k
5%
Q3
1/2 Si4936DY
C3
0.047µF
Q2
1/2 Si4936DY
R2
0.025
12V
R1
0.005
5V
VCC
+ 50mV
–
AUX VCC
20 VOUTLO
16 VOUTHI
17 GATEHI
CHARGE
PUMP
N2
C5
220µF
VEE
–12V
1A
VDD
12V
1A
VCC
5V
5A
C4
2200µF
N1
+
–
+
AUX VCC
21 GATELO
18 SETHI
+ 50mV
–
–
AUXVCC 24
+
C1
1.0µF
19 VCCHI
23 VCCLO
22 SETLO
+
R5
16k
5%
C6
220µF
+
CRAMP
0.1µF
Q1
MTB50N06E
10 RAMP
–12V
pacitors C4 and C5 is I = (CBYPASS ×
20µA)/CRAMP. The internal charge
pumps are designed to provide at
least 8V of gate drive to Q1 and Q2.
The negative supply voltage is controlled using the CPON pin. When the
board first makes connection, transistor Q3 is turned off by R4 and C3.
After the charge pump turns on, CPON
is pulled to VCCLO and the gate of Q3
ramp up with a time constant determined by R4, R5 and C3.
pumps turn on, CPON pulls high and
the voltage at GATEHI begins to rise
with a slope equal to 20µA/CRAMP.
The voltage at the GATELO pin is
clamped one Schottky diode drop below GATEHI. The ramp time for each
supply is t = (VSUPPLY × CRAMP)/20µA.
Because the N-channel transistors
Q1 and Q2 act as source followers,
the voltage at VOUTLO and VOUTHI have
the same ramp rate. Therefore, the
inrush current into the bypass ca-
+
Hot Swapping, continued from page 1
CP2
CP1
BOARD
73.5k
–
+ 100mV
–
CP3
71.5k
–
BACKPLANE
+
UNDER
VOLTAGE
LOCKOUT
CPON 9
11 FB
26.7k
+
CP4
D1
FAULT 4
20µA
20µA
FAULT
6 PWRGD
20µA
CON2 2
RESET
TIMING
POR 3
+
CP5
GND 12
LTC1421
GND
QS3384
1
BEA
13
BEB
µP
7 RESET
15 COMPOUT
–
DISABLE 5
DATA BUS
I/O
DIGITAL
CONTROL
CON1 1
POR
8 REF
1.232V
REFERENCE
R3
1k
14 COMP –
13 COMP +
RESET
R8
10k
5%
I/O
R6
107k
1%
R7
13.7k
1%
DATA BUS
VCC
Figure 1. LTC1421 typical application
Linear Technology Magazine • August 1996
3
DESIGN FEATURES
As soon as VCCLO reaches the 4.65V
reset threshold, (Figure 2, time point
4), the PWRGD signal immediately
pulls high. After a 200ms delay, the
RESET signal pulls high, (Figure 2,
time point 6), and the DISABLE pin
pulls low, thus enabling the system
data bus. In this application, the free
comparator, CP5, is used to monitor
the 12V supply. When VOUTHI reaches
1
10.8V, the COMPOUT signal pulls
high (Figure 2, time point 5).
Monitoring
The Supply Voltages
The LTC1421 features a 1.232V reference, internal resistor divider from
VCCLO and precision voltage comparators CP4 and CP5 (Figure 1) to monitor
the supply voltages. The reset thresh-
2
3
4
5
old voltage for VOUTLO is determined
by the FB pin connection as summarized in Table 1.
When the VOUTLO voltage rises above
its reset threshold voltage, the comparator (CP4) output goes low, and
PWRGD is immediately pulled high to
VCCLO by a weak pull-up current
source or external resistor. After a
200ms delay, RESET is pulled high.
6
200ms
20ms
VCCLO
VCCHI
DISABLE
CON1
CON2
5V
CPON
–12V
10.8V
VOUTHI
4.65V
VOUTLO
PWRGD
RESET
COMPOUT
Figure 2. Typical insertion timing
4
Linear Technology Magazine • August 1996
DESIGN FEATURES
5V
12V
20 VOUTLO
16 VOUTHI
VCCLO
73.5k
20µA
CP4
71.5k
–
PWRGD 6
11 FB
10k
5%
107k
1%
26.7k
+
15 COMPOUT
VCCLO
RESET
TIMING
+
CP5
20µA
RESET 7
13 COMP+
14 COMP–
–
8
1.232V
13.7k
1%
REF
LTC1421
Figure 3. Monitoring a 12V supply with a 10.8V threshold; the 5V supply generates a reset
when it dips below 4.65V.
The weak pull-up current sources to
VCCLO on PWRGD and RESET have a
series diode so the pins can be pulled
above VCCLO by an external pull-up
resistor without forcing current back
into VCCLO.
When VCCLO drops below its reset
threshold, the comparator (CP4)
output goes high, and PWRGD immediately pulls low. After a 64µs delay,
RESET is pulled low. The RESET delay allows the PWRGD signal to be
used as an early warning that a reset
is about to occur. If the PWRGD signal
is used as an interrupt input to a
microprocessor, a short power-down
routine can be run before the reset
occurs.
The uncommitted voltage comparator (CP5) can be used to monitor
output voltages other than VOUTLO.
Figure 3 shows how the comparator
can be used to monitor a 12V supply
with a 10.8V threshold, while the 5V
3.3V
5V
20 VOUTLO
16 VOUTHI
VCCLO
73.5k
107k
1%
20µA
CP4
71.5k
–
PWRGD 6
11 FB
26.7k
10k
5%
+
15 COMPOUT
supply generates a reset when it dips
below 4.65V. The FB pin is left floating.
Figure 4 shows how to monitor a
5V supply with a 4.65V threshold,
while the 3.3V supply generates a
reset when it dips below 2.9V. The FB
pin is tied to VCCLO.
Figure 5 shows how the comparator can be used to generate a reset
when the 12V supply drops below
10.8V. The 5V supply also generates
a reset when it dips below 4.65V.
When 12V dips below 10.8V,
COMPOUT will pull the FB pin low,
setting the internal threshold voltage
for CP5 to 5.88V. Since VOUTLO is less
than 5.88V, PWRGD immediately goes
low, followed by RESET 64µs later.
Figure 6 shows how comparator
CP5 can be used to override the internal reset voltage threshold. A 5k
resistor is tied from the FB pin to
VOUTLO, setting the internal threshold
to about 2.9V. The new reset threshold voltage is set by the external
resistor divider connected to CP5, in
this case 4.5V. When VOUTLO drops
below the new threshold voltage,
COMPOUT pulls FB to ground, changing the internal threshold to 5.88V
and generating a reset.
Finally, Figure 7 shows how CP5
can be used to monitor a negative
supply voltage.
Electronic Circuit Breaker
The LTC1421 features an electronic
circuit-breaker function that protects
against short circuits or excessive
current on the supplies by placing
sense resistors (R1 and R2) between
the supply input and sense pin of
either supply. The circuit breaker will
be tripped whenever the voltage across
the sense resistor is greater than
50mV for more than 20µs. When the
VCCLO
RESET 7
RESET
TIMING
+
CP5
20µA
14 COMP–
–
8
1.232V
Table 1. Reset voltage thresholds
13 COMP+
38.3k
1%
REF
LTC1421
Figure 4. Monitoring a 5V supply with a 4.65V threshold; the 3.3V supply generates a reset
when it dips below 2.9V.
Linear Technology Magazine • August 1996
Feedback Pin
VOUTLO
Reset Voltage
Floating
4.65V
VOUTLO
2.90V
GND
5.88V
5
DESIGN FEATURES
73.5k
20µA
CP4
PWRGD 6
71.5k
11 FB
107k
1%
26.7k
15 COMPOUT
VCCLO
CP5
20µA
RESET 7
RESET
TIMING
13 COMP+
14 COMP–
13.7k
1%
8 REF
1.232V
LTC1421
Figure 5. Generating a reset when the 12V supply drops below 10.8V; the 5V supply also
generates a reset when it dips below 4.65V.
Figure 9 shows the typical poweron reset cycle. After the POR pin is
held low for 20ms, internal transistors N1 and N2 are turned on to start
discharging VOUTLO and VOUTHI. At the
same time, GATEHI and GATELO are
also actively pulled down internally,
while CPON goes low. When VOUTHI
reaches the reset threshold, the
COMPOUT pin pulls low (time point
3). When VOUTLO reaches the reset
threshold, PWRGD immediately pulls
low, followed by RESET 64µs later.
When VOUTLO is discharged within
100mV of ground, the LTC1421 will
reset and start a normal power-up
sequence.
5V
20 VOUTLO
16 VOUTHI
12V
VCCLO
73.5k
20µA
CP4
PWRGD 6
71.5k
–
11 FB
5k
5%
102k
1%
26.7k
+
15 COMPOUT
VCCLO
CP5
20µA
RESET 7
RESET
TIMING
+
–
The POR can be used to completely
cycle the power supplies on the board
or to reset the electronic circuitbreaker feature. The POR pin can be
connected to a grounded push button
or toggle switch, or to a logic signal
from the backplane through the
connector.
VCCLO
–
Power-On Reset
16 VOUTHI
+
When a short circuit occurs on the
board, it is possible to draw enough
current to cause the backplane supply voltage to collapse to a low enough
voltage that the LTC1421 gate drive
circuitry is unable to shut off the
N-channel pass transistors (Q1 and
Q2). This could also happen if VCCLO
breaks contact but VCCHI remains
connected.
To prevent the system from freezing
up in a permanent short condition,
the gate-discharge circuitry inside the
LTC1421 is powered from the AUXVCC
pin, which, in turn, is powered from
VCCLO through an internal Schottky
diode and current limiting resistor
(Figure 8).
When VCCLO collapses, there is
enough energy stored on the 1.0µF
capacitor connected to AUXVCC to
keep the gate discharge circuitry alive
long enough to fully turn off the
external N-channels.
12V
20 VOUTLO
+
Auxiliary VCC
5.0V
–
circuit breaker trips, both N-channel
transistors (Q1 and Q2) are immediately turned off and the FAULT pin is
pulled low. When VCCLO drops below
the 4.65V threshold, PWRGD is pulled
low and RESET is pulled low 64µs
later. The chip will remain in the
tripped state until a power-on reset is
generated, (pulling the POR pin low
momentarily), or the power on VCCHI
and VCCLO is cycled. If the circuit
breaker feature is not used, VCCLO can
be shorted to SETLO and VCCHI to
SETHI.
13 COMP+
14 COMP–
8
1.232V
38.3k
1%
REF
LTC1421
Authors can be contacted
at (408) 432-1900
6
Figure 6. CP5 can be used to override the internal reset voltage threshold. A 5k resistor is tied
from the FB pin to VOUTLO, setting the internal threshold to about 2.9V. The new reset
threshold voltage is set by the external resistor divider connected to CP5.
Linear Technology Magazine • August 1996
DESIGN FEATURES
5V
VCCLO
20 VOUTLO
23
16 VOUTHI
12V
VCCLO
10k
10k
5%
73.5k
21
20µA
CP4
71.5k
–
PWRGD 6
AUXVCC
11 FB
24
GATE DRIVE
CIRCUITRY
17
GATELO
GATEHI
1.0µF
26.7k
+
LTC1421
15 COMPOUT
VCCLO
Figure 8. Auxiliary VCC circuitry
RESET
TIMING
+
CP5
20µA
RESET 7
13 COMP+
Conclusion
14 COMP–
With the explosion of systems requiring distributed power, the need for
products that can be safely inserted
into a live backplane has increased.
Up to now, the design of the protection circuitry has required the talents
of an analog guru, but with the
LTC1421, safe hot swapping becomes
as easy as hooking up an IC, a couple
of power FETs and a handful of resistors and capacitors.
–
8 REF
1.232V
13.7k
1%
LTC1421
107k
1%
–12V
Figure 7. CP5 can be used to monitor a negative supply voltage.
1
2
3
4
20ms
5
6
7
8
64µs
9
200ms
POR
CPON
VOUTHI
10.8V
10.8V
VOUTLO
4.65V
4.65V
100mV
PWRGD
RESET
COMPOUT
Figure 9. LTC1421 typical power-on reset cycle
Linear Technology Magazine • August 1996
7
DESIGN FEATURES
The Care and Feeding of
High Performance ADCs:
Get All the Bits You Paid For
by William C.
Rempfer
Introduction
A new generation of ADCs currently
appearing on the scene brings higher
performance and lower cost to new
markets. Figure 1 shows an example
of how high speed 12-bit converters
are becoming affordable for the first
time to a new range of applications. At
the same time, the new converters
achieve better dynamic performance
with high frequency input signals. All
this means that more system designers are facing the challenge of using
high performance ADCs. In this
article, we will talk about some of the
problems designers encounter, how
to recognize their symptoms and how
to avoid them. We will focus on the
particular case of the LTC1410, a
1.25Msps, 12-bitADC. The same considerations become important in
higher resolution ADCs at lower
speeds. Conversely, lower resolution
ADCs will need this same attention at
higher speeds.
input, but many. Ground pins, supply pins and reference pins also act as
inputs and must be given special care
to prevent noise and unwanted signals from corrupting the ADC output.
Grounding, bypassing of the supplies
and the reference and driving the
analog and clock inputs are the major
weapons in this battle against corruption.
Ground Planes
and Grounding
Designing a high speed ADC system
without using a proper ground is like
trying to play basketball on a huge
trampoline. No matter how well you
mount the baskets to the court, the
whole court will bounce and wobble
as the players jump and try to shoot.
To play the game, you must have a
solid floor. Similarly, to give a solid
Providing a clean analog input signal
to an ADC doesn’t always guarantee a
clean digital output signal. This is
because an ADC has not just one
LIST PRICE OF
12-BIT 1.25Msps
ADCs
5V
–5V
An ADC Has Many Inputs
10µF
10µF
0.1µF
0.1µF
AIN+
AIN–
X
X
AVDD
DVDD
VSS
REFCOMP
XX
XX
AGND
10µF 0.1µF
$100
ground for your data converter circuit, you must use an analog ground
plane. This will put your circuit on a
solid foundation.
Figure 2 shows grounding techniques for the LTC1410, a 1.25Msps,
12-bit ADC. This provides an example
that can be modified for the particular high performance converter used.
All bypass caps, reference caps and
ground connections for the ADC
should be tied to the analog ground
plane. Tie them as close together as
possible to reduce the sensitivity to
currents that may flow in the ground
plane. The input signal circuitry, filter caps and op amp bypass caps (not
shown) should also be grounded to
the ground plane near the ADC.
Noise from digital components in
the system must be kept out of the
analog ground. To do this, boards
DIGITAL
GROUND
LTC1410
SO-28
ANALOG
GROUND
X
OGND
DIGITAL
LOGIC
SUPPLY
D0-D3
V+
(LTC1410)
DGND
DIGITAL
LOGIC
X
D4-D11
DGND
1988 1989 1990 1991 1992 1993 1994 1995 1996
X = VIA TO GROUND PLANE
X
INTRODUCTION YEAR
Figure 1. High performance 1.25Mbps, 12-bit
ADCs are becoming affordable to a new range
of applications. More system designers will
need to know how to use them effectively.
8
Figure 2. High performance ADC layout must have separate analog and digital ground planes,
bypass caps with short connections and digital outputs routed away from the inputs.
Linear Technology Magazine • August 1996
DESIGN FEATURES
The high conversion rates of high
performance converters require
proper bypassing on the supply pins.
The key to good bypassing is low lead
inductance between the ADC and the
bypass capacitors. The goal is to force
AC currents to flow in the shortest
possible loop from the supply pin
through the bypass cap and back
through ground to the ground pin.
In Figure 2, the first components
placed around the ADC are the bypass caps, which are located as close
as possible to the supply pins. The
capacitors must have low inductance
and low equivalent series resistance
(ESR). Tantalum 10µF surface mount
devices are good if they are used in
conjunction with 0.1µF ceramics.
Even better are the new surface mount
ceramic capacitors, which can be used
alone. They come in values of 10µF or
more and have ESR values as low as
20mΩ.
Figure 3a shows the differential
nonlinearity (DNL) of the LTC1410
with good supply bypassing. Figure
3b shows the effects of 2 inches of
lead length (corresponding to roughly
60nH of inductance) in series with
the supply bypass caps. This is an
exaggerated case of poor bypassing
layout, which causes the DNL to deLinear Technology Magazine • August 1996
DNL ERROR (LSBs)
0.5
0.
–0.5
–1.0
0
512
1024
1536
2048
2560
OUTPUT CODE
3072
3584
4096
0
512
1024
1536
2048
2560
OUTPUT CODE
3072
3584
4096
0
512
1024
1536
2048
2560
OUTPUT CODE
3072
3584
4096
1.0
0.5
DNL ERROR (LSBs)
Supply Bypassing
1.0
0.
–0.5
–1.0
1.0
0.5
DNL ERROR (LSBs)
should be designed with separate analog and digital ground planes, as
shown in Figure 2. (The figure shows
a 2-layer board layout. If more layers
are available, separate layers may be
used for analog and digital ground
planes.) All noisy digital logic devices
must be on the digital ground plane.
All the grounds and bypass caps of
the ADC (even the digital ones) should
tie to the analog ground plane. Tie the
two ground planes together at only
one point to keep digital currents
from taking shortcuts through the
analog ground. The ideal connection
point is the ground pin for the ADC
output drivers (or the digital ground
pin). If that is not possible, a different
connection point can be used (for
example, at the power supply). In any
case, be sure to use only a single
connection point.
0.
–0.5
–1.0
Figure 3. Poor layout will degrade the differential nonlinearity (DNL) of fast ADCs: a. (top) a
clean LTC1410 layout with bypass cap wires of less than 0.5 inch; b. (middle) 2-inch wires to
supply-bypass caps; c. (bottom) a wire of more than 2 inches to the reference-bypass cap.
9
DESIGN FEATURES
grade beyond 1LSB, reducing the
accuracy to 11 bits. For best performance, use supply bypass leads of
less than one-half inch. A little care
pays off with excellent performance.
Reference Bypassing
The analog reference input provides
the scale factor for the conversion.
For a clean data output the reference
must be stable. Dynamic currents
pulled from the reference by the ADC
as it converts perturb the reference
unless it is properly bypassed. Surface mount tantalum or ceramic
capacitors provide good results. They
should be located near the reference
pin and should be grounded very
near the ADC analog ground pin, as
shown in Figure 2.
Figure 3c shows the easily recognizable signature of a reference
bypassing problem—a bow-tie shape
to the error curve. This occurs because
reference perturbations feed in with
full strength for inputs near plus or
minus full scale but have less effect
for inputs near zero scale. This degradation in DNL results from several
inches of lead length in series with
the reference bypass cap. Once again,
this is an exaggerated case to make
the consequences of poor bypassing
more visible. To maintain high accuracy, keep the lead lengths less than
half an inch.
Driving the Analog Input
Switched Capacitor Inputs
The inputs to switched capacitor ADCs
are easy to drive if you allow for the
fact that they draw a small inputcurrent transient at the end of each
conversion. This happens when the
internal sampling capacitors switch
back onto the input to acquire the
next sample. For accurate results,
the circuitry driving the analog input
must settle from this transient before
the next conversion is started.
There are two ways to accomplish
this. One is to drive the ADC with an
op amp that settles from a load transient in less than the acquisition time
of the ADC. Fortunately, most op
amps settle much more quickly from
10
a load transient than from an input
step, so meeting this requirement is
not too difficult. The LT1363, for example, is a good choice for driving the
LTC1410 input.
A second solution to handling the
input transient is to use an input RC
filter with a capacitor much larger
than the ADC input capacitance. This
larger capacitor provides the charge
for the sampling capacitor, which
eliminates the voltage transient altogether. Figure 4 shows such a filter
for the LTC1410. The 1000pF capacitor provides the input charge for the
ADC’s sampling capacitor. The
LT1363’s capacitive load driving capability makes it a good choice for use
with this filter.
Filter Wideband Noise
from the Input Signal
Many new converters have wide S/H
input bandwidths. This is great for
capturing high frequency input signals, but for lower input bandwidth
applications the converter will pick
up any wideband noise that may be in
the input signal. To avoid this, use a
filter at the ADC input to pass only
your desired signal bandwidth.
The simple filter in Figure 4
bandlimits the input signal to 3MHz
and still allows clean sampling up to
the Nyquist frequency (625kHz).
Figure 7a shows the Nyquist performance of the LTC1410 using this
filter. The signal to noise and distortion ratio (SINAD) is 71.5dB and total
harmonic distortion (THD) is –84dB.
Choosing an Op Amp
To drive high performance ADCs, you
will need a high performance op amp.
The noise and distortion of good ADCs
are now so low that they no longer
mask the performance of the op amp.
This adds another tradeoff to op amp
selection.
High speed, current feedback op
amps have lower DC precision and
don’t settle as well to high accuracy
(for example, 0.01%) as the voltage
feedback types. However, they have
the best distortion and drive for high
speed AC frequency domain applications. Figure 5a shows the FFT result
of an LT1227 current feedback amp
driving a 172kHz signal into the
LTC1410. The distortion (THD) of
–82dB is about 3dB worse than the
–85dB of the ADC alone.
High speed voltage feedback amplifiers have better precision and
settling. They work well in frequency
domain applications but are best
suited for high speed, time domain or
multiplexed applications where their
DC precision and settling are required.
Figure 5b shows the voltage feedback
LT1363’s 2dB further degradation in
distortion (to –80dB) under the same
conditions.
Slower op amps like the OP-27/
OP-37 are excellent in noise and precision but are simply not fast enough
for high frequency applications. They
distort as they are pushed beyond
their slewing capabilities (as shown
in the FFT plot of Figure 5c).
Driving the
Convert-Start Input
An improperly driven conversion-start
input can create conversion errors in
a couple of ways. First, if an ADC has
internal timing, the returning edge of
the convert signal (the opposite edge
LTC1410
±2.5V
INPUT
+
50Ω
AIN+
LT1363
–
1000pF
20MHz
S/H
AIN–
ADC
AGND
Figure 4. Many new ADCs have wide-bandwidth sample-and-holds. In lower bandwidth applications, a simple RC filter will remove wideband noise that may be present in the input signal.
Linear Technology Magazine • August 1996
DESIGN FEATURES
0
FSAMPLE = 1.25Msps
FIN = 172kHz
THD = 82.6dB
–20
AMPLITUDE (dB)
–40
–60
–80
–100
–120
0
100K
200K
300K
400K
INPUT FREQUENCY (Hz)
500K
600K
0
FSAMPLE = 1.25Msps
FIN = 172kHz
THD = 80.9dB
–20
AMPLITUDE (dB)
–40
–60
–80
–100
–120
0
100K
200K
300K
400K
INPUT FREQUENCY (Hz)
500K
600K
0
FSAMPLE = 1.25Msps
FIN = 172kHz
THD = 26.6dB
–20
AMPLITUDE (dB)
–40
–60
from the one that starts the conversion) can couple noise into the
converter if it occurs during the conversion time. To avoid this, use a
narrow pulse for convert-start instead
of a square wave. This ensures that it
either returns quickly (after the
sample is taken but before the conversion gets underway), or returns
after the conversion is over. (This
does not apply to those ADCs that
draw all their timing from a clock
input and require precise 50% dutycycle clock inputs.)
A convert-start signal that overshoots or rings can also degrade
performance. If it overshoots beyond
the supply rails it can turn on the
ADC’s input protection diodes and
couple noise into the converter. If it
rings, it may still be bouncing around
as the ADC’s sample-and-hold captures the input signal, which can
affect the conversion result. Normally,
overshoot and ringing are not a problem with high speed CMOS logic on a
well designed board but they are still
things to watch out for.
High frequency or high slew rate
input signals impose another requirement on the ADC: low aperture jitter.
Aperture jitter is the variation in the
ADC’s aperture delay from conversion to conversion and results in an
uncertainty in the time when the input sample is taken. Figure 6 shows
how this jitter causes an equivalent
input noise by working against the
slew rate of the analog input signal.
The faster the input signal slew rate,
the worse the noise for a given jitter.
The best possible SINAD for an ADC
is limited by the jitter according to the
formula:
SINAD(dB) ≤ 20log[1/(2π × tJITTER(RMS) × fINPUT)]
–80
where:
tJITTER(RMS) = the RMS jitter in seconds
–100
–120
fINPUT = the analog input frequency in Hz
0
100K
200K
300K
400K
INPUT FREQUENCY (Hz)
500K
600K
Figure 5. Op amp selection is important when an ADC has low distortion levels. a. (top) Current
feedback op amps such as the LT1227 (seen here driving the LTC1410) provide the lowest THD
in the FFT output; b. (middle) Fast voltage feedback op amps do nearly as well in THD as
current feedback amps and offer better precision; c. (bottom) slower op amps pushed beyond
their slew limits will severely distort fast signals.
Linear Technology Magazine • August 1996
The LTC1410’s 5ps (RMS) aperture jitter allows clean sampling of
inputs far beyond the Nyquist frequency. However, to achieve this
performance, the convert-start input
signal applied to the ADC must also
have low jitter. Figure 7a shows the
11
DESIGN FEATURES
0
FSAMPLE = 1.25Msps WITH NO JITTER
FIN = 600kHz
SINAD = 71.5dB
–20
–40
AMPLITUDE (dB)
ADC, driven from a low jitter source,
capturing a 600kHz input with 71.5dB
SINAD. As Figure 7b shows, adding
70ps of jitter to the conversion-start
input signal will raise the noise floor,
and reduce the SINAD, by 3dB.
If generating a lower jitter signal is
a problem, one trick is to start with a
higher frequency clock, which will
usually have lower jitter, and then
divide the frequency down with fast
logic (which retains the lower jitter) to
get the desired sample clock.
–60
–80
–100
Routing the Data Outputs
–120
0K
100K
200K
300K
400K
INPUT FREQUENCY (Hz)
500K
600K
500K
600K
0
FSAMPLE = 1.25Msps WITH 70ps (RMS) JITTER
FIN = 600kHz
SINAD = 68.5dB
–20
–40
AMPLITUDE (dB)
One of the worst potential sources of
digital noise and coupling in an ADC
is its output data bus. Fortunately,
the user can control this with proper
board layout. First, to prevent the
data outputs of the ADC from capacitively coupling to the analog input
circuitry, they should be routed in
the opposite direction. This will naturally occur if separate digital and
analog ground plane layouts are used,
as in Figure 2. Second, the digital
–60
–80
IDEAL SAMPLING
INSTANT
ADC CONVERSION
START SIGNAL
–100
–120
ADC ANALOG
INPUT SIGNAL
0
100K
200K
300K
400K
INPUT FREQUENCY (Hz)
Figure 7. Jitter in the conversion signal creates noise: a. (top) with a low jitter source, the
LTC1410 will give 71.5dB SINAD when sampling a Nyquist input signal; b. (bottom) adding 70ps
of jitter to the convert-start signal will raise the noise floor by 3dB to 68.5dB.
EFFECTIVE NOISE DUE
TO APERTURE JITTER
∆V
∆t
APERTURE JITTER
Figure 6. Aperture jitter in a sampling ADC or
jitter in the conversion-start signal applied to
the ADC can degrade its noise performance.
The time jitter works against the slope of the
analog input signal to generate an effective
noise voltage that appears in the ADC’s
output spectrum.
12
output drivers in the ADC switch
quickly and will create large current
transients if they are loaded with too
much capacitance. Locating the
receiving buffers or latches close to
the ADC will minimize loading.
Although reduced, some capacitive currents still flow, and it is
important to control their return path
to the driver of the ADC. Starting from
the output drivers of the ADC, the
current goes through the output lines,
charges the input capacitance of the
receiving latches or buffers, and
returns through the digital ground
plane to the ADC’s output driver. For
a falling edge, this current returns
into the output driver ground pin. For
a rising edge, it returns to the ground
point of the output driver’s supplybypass cap. Tying the digital and
analog grounds together at the ADC
output driver ground pin (as in Figure
2) helps prevent this current from
flowing across the analog ground
plane. If the grounds must be tied at
the power supply instead of at the
ADC, the return currents will flow
through the analog ground plane. In
this case, it is especially important to
minimize these currents by minimizing the capacitance on the digital
outputs.
Linear Technology Magazine • August 1996
DESIGN FEATURES
LTC1433/LTC1434:
High Efficiency, Constant-Frequency
Monolithic Buck Converter by San-Hwa Chee
In portable communications products
where high efficiency and constant
frequency operation are prime requirements, the LTC1433 and LTC1434
are a perfect fit. These two new devices are packed with features but
still fit in a small footprint. The
LTC1433 comes in a 16-pin narrow
SSOP, whereas the LTC1434 comes
in a 20-pin narrow SSOP. The
LTC1434 provides an additional feature that allows the device to be
synchronized with an external clock
through its internal PLL.
High efficiency and constant frequency at low load current are
achieved by using the new Adaptive
Power™ output stage, first introduced
in the LTC1435–LTC1439 DC/DC
controllers (“New LTC1435–LT1439
DC/DC Controllers Feature Value and
Performance”; Linear Technology VI:1
(February 1996)). The LTC1433 and
LTC1434 combine Adaptive Power
operation with internal power MOSFETs for the first time, with operating
frequency programmable up to
500kHz by means of a single external
capacitor.
What is Adaptive Power
With no load, the devices require Mode Operation?
only 470µA of quiescent current,
which drops to 15µA in full shutdown. In dropout conditions, the
internal 0.6Ω (at an input supply of
10V) power P-channel MOSFET switch
is turned on continuously (DC),
thereby maximizing the life of the
battery source. In the event of an
output short circuit, the oscillator
frequency is reduced by a factor of 4.5
to prevent inductor-current runaway.
In addition, an internal sense resistor
limits the switch current to 1.2A.
Both devices contain a low-battery
detector and a power-on reset (POR)
timer that generates a signal delayed
by 65,536 oscillator clock cycles after
the output is within 5% of regulated
output voltage. A soft-start pin allows
the LTC1433/LTC1434 to power up
gently and also serves as a shutdown
pin. For maximum flexibility, internal
resistive feedback dividers are selectable via programming pins for 3.3V or
5V, or can be configured with an
adjustable output voltage to meet any
requirement. Both devices function
down to an input voltage of 3.5V and
up to an absolute maximum of 13.5V.
The LTC1433/LTC1434 have two
internal P-channel MOSFETs of different sizes, with each drain bonded
out separately. To maximize efficiency,
only the smaller sized P-channel
MOSFET is switched on and off at low
load currents. This reduces gatecharge losses without changing the
operating frequency. At higher
load currents, both MOSFETs are
switched, since losses due to drops
across the FETs are more significant
than gate-charge losses. The
100
VIN = 5V
90
EFFICIENCY (%)
Introduction
80
VIN = 12V
VIN = 9V
70
60
50
40
0.001
0.01
0.10
1.00
LOAD CURRENT (A)
Figure 2. Efficiency versus load current
for Figure 1’s circuit
BSW
L1
0.1µF
* MBRS130LT3
** COILCRAFT DO3316-104
†
AVX TPSD107M010R0100
††
AVX TPSE686M020R0150
LTC1433
LTC1434
68µF††
20V
SSW
+
D1
100µH
L1**
VOUT
3.3V
+
D1*
†
100µF
10V
1 SSW
2 NC
3 BSW
PWRVIN 16
PGND 15
SVIN 14
13
C
LTC1433 OSC
5 SGND
POR 12
6 RUN/SS
ITH 11
INPUT VOLTAGE
3.6V TO 12V
Figure 3. Single-inductor configuration
10kΩ
L1
4 NC
0.1µF
7 LBO
8 LBI
VOSENSE 10
VPROG 9
BSW
POWER ON RESET
680pF
5.1kΩ
47pF
6800pF
LTC1433
LTC1434
D1
L2
SSW
D2
Figure 1. Typical application using the LTC1433
Linear Technology Magazine • August 1996
Figure 4. Dual-inductor configuration
13
DESIGN FEATURES
100
3.4
3.3
VIN = 5V
5.0
3.2
70
VOUT = 3.3V
COSC = 47pF
VIN = 9V
60
ONE 22 µH INDUCTOR
ON SSW & BSW
100 µH ON SSW
22 µH ON BSW
SOLID LINE
50
DOTTED LINE
40
0.001
0.01
0.10
3.1
3.0
2.9 IOUT = 400mA
2.8
2.7
2.6
VPROG = 0V
COSC = 50pF
L = 20µH
IOUT = 500mA
2.5
2.4
3.2
1.00
IOUT = 200mA
4.9
OUTPUT VOLTAGE (V)
80
OUTPUT VOLTAGE (V)
EFFICIENCY (%)
90
5.1
IOUT = 300mA
4.8
4.7
IOUT = 300mA
4.6
4.5
4.4
VPROG = VIN
COSC = 50pF
L = 20µH
IOUT = 400mA
4.3
4.2
3.6
4.0
4.4
4.8
5.2
4.6
SUPPLY VOLTAGE (V)
LOAD CURRENT (A)
4.8
5.2
5.0
5.4
5.6
5.8
6.0
6.2
SUPPLY VOLTAGE (V)
Figure 5. Efficiency comparison between
single- and dual-inductor configurations
Figure 6. Dropout characteristics at different
load currents for VOUT = 3.3V
Figure 7. Dropout characteristics at different
load currents for VOUT = 5.0V
LTC1433/LTC1434 monitor two conditions to determine when to switch
to low current mode: inductor current and error amplifier output voltage
(on the ITH pin). If the peak current of
the inductor does not exceed 260mA
and the voltage at the ITH pin does not
exceed 0.6V, the small MOSFET will
be used. When either one of the conditions is exceeded, the large MOSFET
will be used on the next clock cycle.
works all the way down to 3.6V at a
load current of 250mA before dropping out and the oscillator frequency
is a constant 210kHz down to 20mA
load current.
current rating. As can be seen from
Figure 5, the average efficiency gain
over the region where the small
P-channel is ON is about 3%. Hence,
the dual inductor configuration is
good for applications that require
maximum efficiency at low load
currents, while retaining constantfrequency operation.
Efficiency
Figure 1 shows a practical LTC1433
circuit that can be used for cellular
telephone applications. Efficiency
curves for this circuit at various input
voltages are shown in Figure 2. Note
that the efficiency reaches 93% at a
supply voltage of 5V and a load current of about 150mA. This high
efficiency makes the LTC1433 and
LTC1434 attractive for all other powersensitive applications. The circuit
Efficiency Considerations
Since there are two separate pins for
the drains of the small and large
P-channel switches, we could use two
inductors to further enhance the efficiency of the regulator over the low
current range. Figures 3 and 4 show
the single-inductor and dual-inductor
circuit configurations, respectively.
To reduce core losses, a higher value
inductor can be used on the small
P-channel switch. Since this switch
only carries a small part of the overall
current, the user can still specify a
small physical size inductor without
sacrificing on copper losses. The
Schottky on the small P-channel drain
(SSW) can also be chosen with a lower
100% Duty Cycle in Dropout
When the input voltage decreases,
the inductor’s ripple current starts to
decrease and the duty cycle increases
to provide the required output current. Further decrease in input voltage
will eventually cause the ITH voltage to
be at its maximum limit. Any decrease in input voltage from this point
will result in the P-channel switch
being turned on continuously. The
dropout voltage, VIN – VOUT, is governed by the switch resistance, load
current and the voltage drop across
continued on page 42
* MOTOROLA MBRS130LT3
** COILCRAFT DO3316 SERIES
†
AVX TPSD107M010R0100
††
AVX TPSE107M016R0100
+
VOUT
–5.0V
VIN (V)
14
IOUT MAX (mA)
3.0
180
4.0
240
5.0
290
6.0
340
7.0
410
7.5
420
L1**
68µH
100µF†
10V
D1*
0.01µF
1
SSW
2
NC
3
BSW
4
NC
5
SGND
6
RUN/SS
7
LBO
8
LBI
LTC1433
PWRVIN
16
PGND
15
SVIN
14
COSC
13
POR
12
ITH
11
VOSENSE
10
VPROG
INPUT VOLTAGE
3V TO 7.5V
100µF††
16V
100pF
+
0.1µF
6800pF
9
680pF
5.1kΩ
Figure 8. Positive-to-negative (–5.0V) converter
Linear Technology Magazine • August 1996
DESIGN FEATURES
The LTC1343 and LTC1344
Form a Software-Selectable
Multiple-Protocol Interface Port
Using a DB-25 Connector
by Robert Reay
Introduction
With the explosive growth in data
networking equipment has come the
need to support many different serial
protocols using only one connector.
The problem facing interface designers is to make the circuitry for each
serial protocol share the same connector pins without introducing
conflicts. The main source of frustration is that each serial protocol
requires a different line termination
that is not easily or cheaply switched.
With the introduction of the
LTC1343 and LTC1344, a complete
software-selectable serial interface
port using an inexpensive DB-25 connector becomes possible. The chips
form a serial interface port that supports the V.28 (RS232), V.35, V.36,
RS449, EIA-530, EIA-530A or X.21
protocols in either DTE or DCE mode
CTS
DSR
DCD
and is both NET1 and NET2 compliant. The port runs from a single 5V
supply and supports an echoed clock
and loop-back configuration that
helps eliminate glue logic between
the serial controller and the line
transceivers.
A typical application is shown in
Figure 1. Two LTC1343s and one
LTC1344 form the interface port using a DB-25 connector, shown here in
DTE mode.
Each LTC1343 contains four drivers and four receivers and the
LTC1344 contains six switchable
resistive terminators. The first
LTC1343 is connected to the clock
and data signal lines along with the
diagnostic LL (local loop-back) and
TM (test mode) signals. The second
LTC1343 is connected to the control-
DTR
RTS
RL
D3
D2
D1
TM
RXD
RXC
TXC
D4
R3
R2
Review of
Interface Standards
The serial interface standards RS232,
EIA-530, EIA-530A, RS449, V.35, V.36
and X.21 specify the function of each
signal line, the electrical characteristics of each signal, the connector type,
SCTE
TXD
LL
D3
D2
D1
LTC1343
LTC1343
R4
signal lines along with the diagnostic
RL (remote loop-back) signal. The
single-ended driver and receiver could
be separated to support the RI (ringindicate) signal. The switchable line
terminators in the LTC1344 are connected only to the high speed clock
and data signals. When the interface
protocol is changed via the digital
mode selection pins (not shown), the
drivers and receivers are automatically reconfigured and the appropriate
line terminators are connected.
D4
R1
R3
R4
R1
R2
LTC1344
13 5
22 6
10 8
23 20 19
4
21
1
7
25
16 3
9
17 12 15
11 24 14 2
18
LL A (141)
TXD A (103)
TXD B
SCTE A (113)
SCTE B
TXC A (114)
TXC B
RXC A (115)
RXC B
RXD A (104)
RXD B
TM A (142)
SGND (102)
SHIELD (101)
RTS A (105)
RL A (140)
RTS B
DTR A (108)
DTR B
DCD A (109)
DCD B
DSR A (107)
CTS A (106)
DSR B
CTS B
DB-25 CONNECTOR
Figure 1. LTC1343/LTC1344 typical application
Linear Technology Magazine • August 1996
15
DESIGN FEATURES
A'
A
BALANCED
INTERCONNECTING
CABLE
GENERATOR
R1
51.5Ω
LOAD
CABLE
TERMINATION
S1
RECEIVER
LTC1343
R8
6k
R2
51.5Ω
B'
V.10 (RS423) Interface
A typical V.10 unbalanced interface
is shown in Figure 2. A V.10 singleended generator (output A with ground
C) is connected to a differential
receiver with input A' connected to A
and input B' connected to the signalreturn ground C. The receiver’s
R7
10k
GND
Figure 3. V.10 receiver configuration
Figure 2. Typical V.10 interface
the transmission rate and the data
exchange protocols. The RS422 (V.11)
and RS423 (V.10) standards merely
define electrical characteristics. The
RS232 (V.28) and V.35 standards also
specify their own electrical characteristics. In general, the US standards
start with RS or EIA, and the equivalent European standards start with V
or X. The characteristics of each interface are summarized in Table 1.
Table 1 shows only the most commonly used signal lines. Note that
each signal line must conform to only
one of four electrical standards, V.10,
V.11, V.28 or V.35.
RECEIVER
S4
C'
C'
R6
10k
R4
20k
B
B'
C
R5
20k
S3
R3
124Ω
S2
A'
A
LTC1344
ground C' is separate from the signal
return. Usually, no cable termination
between A' and B' is required for V.10
interfaces. The V.10 receiver configuration for the LTC1343 and LTC1344
is shown in Figure 3.
With the introduction of the
LTC1343 and LTC1344,
a complete softwareselectable serial interface
port using an inexpensive
DB-25 connector becomes
possible. The chips form
a serial interface port that
supports a wide variety of
standards.
In V.10 mode, switches S1 and S2
inside the LTC1344 and S3 inside the
LTC1343 are turned off. Switch S4
inside the LTC1343 shorts the noninverting receiver input to ground so
the B input at the connector can be
left floating. The cable termination is
then the 30k input impedance to the
ground of the LTC1343 V.10 receiver.
V.11 (RS422) Interface
A typical V.11 balanced interface is
shown in Figure 4. A V.11 differential
generator with outputs A and B and
ground C is connected to a differential receiver with ground C', input A'
connected to A and input B' connected to B. The V.11 interface has a
differential termination at the receiver
end with a minimum value of 100Ω.
The termination resistor is optional
in the V.11 specification, but for the
high speed clock and data lines, the
termination is required to prevent
reflections from corrupting the data.
In V.11 mode, all switches are off
except S1 inside the LTC1344, which
connects a 103Ω differential termination impedance to the cable, as
shown in Figure 5.
Table 1. Interface summary
Clock and Data Signals
Control Signals
Test Signals
TXD
SCTE
TXC
RXC
RXD
RTS
DTR
DSR
DCD
CTS
RI
LL
RL
TM
CCITT#
(103)
(113)
(114)
(115)
(104)
(105)
(108)
(107)
(109)
(106)
(125)
(141)
(140)
(142)
RS232
V.28
V.28
V.28
V.28
V.28
V.28
V.28
V.28
V.28
V.28
V.28
V.28
V.28
V.28
EIA-530
V.11
V.11
V.11
V.11
V.11
V.11
V.11
V.11
V.11
V.11
—
V.10
V.10
V.10
EIA-530A
V.11
V.11
V.11
V.11
V.11
V.11
V.10
V.10
V.11
V.11
V.10
V.10
V.10
V.10
RS449
V.11
V.11
V.11
V.11
V.11
V.11
V.11
V.11
V.11
V.11
V.10
V.10
V.10
V.10
V.35
V.35
V.35
V.35
V.35
V.35
V.28
—
V.28
V.28
V.28
—
—
—
—
V.36
V.11
V.11
V.11
V.11
V.11
V.11
—
V.11
V.11
V.11
—
V.10
V.10
V.10
X.21
V.11
V.11
V.11
V.11
V.11
V.11
—
—
V.11
—
—
—
—
—
16
Linear Technology Magazine • August 1996
DESIGN FEATURES
A'
BALANCED
INTERCONNECTING
CABLE
GENERATOR
A
LOAD
CABLE
TERMINATION
R1
51.5Ω
S1
100Ω
MIN
R2
51.5Ω
C'
C
R5
20k
R6
10k
S3
R3
124Ω
S2
B'
B
R8
6k
RECEIVER
A'
A
LTC1343
LTC1344
B'
Figure 4. Typical V.11 interface
R7
10k
R4
20k
B
RECEIVER
S4
C'
GND
Figure 7. V.28 receiver configuration
A'
A
R1
51.5Ω
LTC1344
R8
6k
S2
R2
51.5Ω
50Ω
R7
10k
R4
20k
B
B'
125Ω
125Ω
50Ω
GND
B
B'
C
C'
A'
A
R1
51.5Ω
BALANCED
INTERCONNECTING
CABLE
LTC1343
LTC1344
R8
6k
S1
S2
RECEIVER
R7
10k
R4
20k
B
B'
S4
C'
C'
RECEIVER
S3
R3
124Ω
R2
51.5Ω
A'
C
R5
20k
R6
10k
LOAD
CABLE
TERMINATION
A
50Ω
Figure 8. Typical V.35 interface
Figure 5. V.11 receiver configuration
GENERATOR
RECEIVER
50Ω
S4
C'
CABLE
TERMINATION
A'
A
RECEIVER
S3
R3
124Ω
BALANCED
INTERCONNECTING
CABLE
GENERATOR
R5
20k
R6
10k
S1
LOAD
LTC1343
GND
Figure 6. Typical V.28 interface
Figure 9. V.35 receiver configuration
Table 2. LTC1343/LTC1344 mode selection
M2
M1
M0
CTRL/
CLK
D1
D2
D3
D4
R1
R2
R3
R4
V.10/RS423
0
0
0
X
V.10
V.10
V.10
V.10
V.10
V.10
V.10
V.10
RS530A clock & data
0
0
1
0
V.10
V.11
V.11
V.11
V.11
V.11
V.11
V.10
RS530A control
0
0
1
1
V.10
V.11
V.10
V.11
V.11
V.10
V.11
V.10
Reserved
0
1
0
X
V.10
V.11
V.11
V.11
V.11
V.11
V.11
V.10
X.21
0
1
1
X
V.10
V.11
V.11
V.11
V.11
V.11
V.11
V.10
V.35 clock & data
1
0
0
0
V.28
V.35
V.35
V.35
V.35
V.35
V.35
V.28
V.35 control
1
0
0
1
V.28
V.28
V.28
V.28
V.28
V.28
V.28
V.28
RS530/RS449/V.36
1
0
1
X
V.10
V.11
V.11
V.11
V.11
V.11
V.11
V.10
V.28/RS232
1
1
0
X
V.28
V.28
V.28
V.28
V.28
V.28
V.28
V.28
No Cable
1
1
1
X
Z
Z
Z
Z
Z
Z
Z
Z
LTC1343 Mode Name
Linear Technology Magazine • August 1996
17
DESIGN FEATURES
A
51.5Ω
LATCH
S2
ON
LTC1344
S1
ON
DCE/
DTE M2
22
23
M1 M0 (DATA)
24
1
51.5Ω
CONNECTOR
(DATA)
B
C1
100pF
R1, 10k
LTC1343
M0
C
17
R2, 10k
20
Figure 10. V.35 driver using the LTC1344
M1
CTRL/CLK
18
R3, 10k
22
M2
LATCH
19
A typical V.28 unbalanced interface
is shown in Figure 6. A V.28 singleended generator (output A with ground
C) is connected to a single-ended
receiver with input A' connected to A
and ground C' connected via the signal return ground to C. In V.28 mode,
all switches are off except S3 inside
the LTC1343, which connects a 6k
impedance (R8) to ground in parallel
with 20k (R5) plus 10k (R6), for an
combined impedance of 5k, as shown
in Figure 7. The noninverting input is
disconnected inside the LTC1343
receiver and connected to a TTL level
reference voltage for a 1.4V receiver
trip point.
DCE/DTE
VCC
VCC
NC
R4, 10k
V.28 (RS232) Interface
VCC
21
VCC
NC
CABLE
LTC1343
DCE/DTE
M2
VCC
20
22
CTRL/CLK
M1
LATCH
M0
21
19
18
17
(DATA)
Figure 11. Mode selection by cable
PORT #1
M0
M1
V.35 Interface
A typical V.35 balanced interface is
shown in Figure 8. A V.35 differential
generator with outputs A and B and
ground C is connected to a differential receiver with ground C', input A'
connected to A and input B' connected to B. The V.35 interface
requires T or delta network termination at the receiver end and the
generator end. The receiver differential impedance measured at the
connector must be 100 ±10Ω, and
the impedance between shorted terminals (A' and B') and ground (C') is
150 ±15Ω.
In V.35 mode, both switches S1
and S2 inside the LTC1344 are on,
connecting the T-network impedance,
as shown in Figure 9. Both switches
in the LTC1343 are off. The 30k input
impedance of the receiver is placed in
parallel with the T-network termination, but does not affect the overall
input impedance significantly.
18
M2
DCE/DTE
CONNECTOR #1
124Ω
21
LATCH
PORT #2
M0
M1
M2
DCE/DTE
CONTROLLER
CONNECTOR #2
V.35
DRIVER
LATCH
PORT #3
M0
M0
M1
M1
M2
M2
DCE/DTE
DCE/DTE
LATCH 1
CONNECTOR #3
LTC1344
LATCH 2
LATCH 3
LATCH
Figure 12. Mode selection by controller
Linear Technology Magazine • August 1996
DESIGN FEATURES
SERIAL
CONTROLLER
LL
LTC1343
LTC1344
D1
R4
SERIAL
CONTROLLER
LL
LTC1344
LL
LL
LTC1343
TXD
D2
TXD
TXD
103Ω
R3
TXD
SCTE
D3
SCTE
SCTE
103Ω
R2
SCTE
D4
R1
RXD
D2
RXD
TM
R4
TM
TM
D1
TM
1
0
1
0
0
0
1
0
1 0 1 0 0
1 0 1 1 0
1
0
1
LTC1343
LATCH
RXD
EC
103Ω
LB
R3
DCE/DTE
RXD
CTRL/CLK
RXC
M2
D3
M1
RXC
M0
RXC
M0
M1
M2
DCE/DTE
LATCH
103Ω
M0
M1
M2
DCE/DTE
LATCH
R2
LATCH
RXC
EC
TXC
LB
D4
DCE/DTE
TXC
CTRL/CLK
TXC
M2
103Ω
M1
R1
M0
TXC
0
1
0
1
0
LTC1343
RL
RL
D1
RTS
D2
RTS
DTR
D3
DTR
RL
R4
RL
RTS
R3
RTS
DTR
R2
DTR
D4
R1
D2
CTS
RI
R4
RI
RI
D1
RI
LATCH
CTS
EC
CTS
LB
R3
DCE/DTE
CTS
CTRL/CLK
DSR
M2
D3
M1
DSR
M0
DSR
LATCH
R2
EC
DSR
LB
DCD
DCE/DTE
D4
CTRL/CLK
DCD
M2
DCD
M1
R1
M0
DCD
1
0
1
1
0
0
1
0
1
0
1
1
1
0
1
0
Figure 13. Normal DTE Loop-back
The generator differential impedance must be 50Ω to 150Ω, and the
impedance between shorted terminals (A and B) and ground (C) is 150Ω
±15Ω. For the generator termination,
switches S1 and S2 are both on and
the top side of the center resistor is
brought out to a pin so it can be
bypassed with an external capacitor
to reduce common mode noise, as
shown in Figure 10.
Linear Technology Magazine • August 1996
Figure 14. Normal DCE loop-back
Any mismatch in the driver rise
and fall times or skew in driver propagation delays will force current
through the center termination resistor to ground, causing a high
frequency common mode spike on
the A and B terminals. This spike can
cause EMI problems that are reduced
by capacitor C1, which shunts much
of the common mode energy to ground
rather than down the cable.
LTC1343/LTC1344
Mode Selection
The interface protocol is selected using the mode select pins M0, M1, M2
and CTRL/CLK, as summarized in
Table 2. The CTRL/CLK pin should
be pulled high if the LTC1343 is being
used to generate control signals and
pulled low if used to generate clock
and data signals.
19
DESIGN FEATURES
For example, if the port is configured as a V.35 interface, the mode
selection pins should be M2 = 1, M1 =
0, M0 = 0. For the control signals,
CTRL/CLK = 1 and the drivers and
receivers will operate in RS232 (V.28)
electrical mode. For the clock and
data signals, CTRL/CLK = 0 and the
drivers and receivers will operate in
V.35 electrical mode, except for the
single-ended driver and receiver,
which will operate in the RS232 (V.28)
electrical mode. The DCE/DTE pin
will configure the port for DCE mode
when high, and DTE when low.
The interface protocol may be selected by simply plugging the
appropriate interface cable into the
connector. The mode pins are routed
to the connector and are left unconnected (1) or wired to ground (0) in the
cable, as shown in Figure 11.
The pull-up resistors R1–R4 ensure
a binary 1 when a pin is left unconnected and also ensure that the two
LTC1343s and the LTC1344 enter
the no-cable mode when the cable is
removed. In the no-cable mode, the
LTC1343 power supply current drops
to less than 200µA and all LTC1343
driver outputs and LTC1344 resistive
terminators are forced into a high
impedance state. Note that the data
latch pin, LATCH, is shorted to ground
for all chips.
The interface protocol may also be
selected by the serial controller or
host microprocessor, as shown in
Figure 12.
The mode selection pins M0, M1,
M2 and DCE/DTE can be shared
among multiple interface ports, while
each port has a unique data-latch
signal that acts as a write enable.
When the LATCH pin is low, the buffers on the MO, M1, M2, CTRL/CLK,
DCE/DTE, LB and EC pins are transparent. When the LATCH pin is pulled
high, the buffers latch the data, and
changes on the input pins will no
longer affect the chip.
The mode selection may also be
accomplished by using jumpers to
connect the mode pins to ground or
VCC.
20
LTC1343
5
DCE/DTE
D1
21
16
VCC
39
20
24
R4
26
CTRL/CLK
EC
Figure 15. Single-ended driver and receiver
enable
Loop-Back
The LTC1343 contains logic for placing the interface into a loop-back
configuration for testing. Both DTE
and DCE loop-back configurations
are supported. Figure 13 shows a
complete DTE interface in the loopback configuration and Figure 14 the
DCE loop-back configuration. The
loop-back configuration is selected
by pulling the LB pin low.
Enabling the Single-Ended
Driver and Receiver
When the LTC1343 is being used to
generate the control signals (CTRL/
CLK = high) and the EC pin is pulled
low, the DCE/DTE pin becomes an
enable for driver 1 and receiver 4 so
their inputs and outputs can be tied
together, as shown in Figure 15.
The EC pin has no affect on the
configuration when CTRL/CLK is high
except to allow the DCE/DTE pin to
become an enable. When DCE/DTE
is low, the driver 1 output is enabled.
The receiver 4 output goes into threestate, and the input presents a 30k
load to ground.
When DCE/DTE is high, the driver
1 output goes into three-state, and
the receiver 4 output is enabled. The
receiver 4 input presents a 30k load
to ground in all modes except when
configured for RS232 operation, when
the input impedance is 5k to ground.
Multiprotocol Interface
with DB-25 or µDB-26
Connectors
A multiprotocol serial interface with a
standard DB-25 connector EIA-530
pin configuration is shown in Figure
16. [Figures 16–19 follow on pp. 21–
24]. The signal lines must be reversed
in the cable when switching between
DTE and DCE using the same connector. For example, in DTE mode,
the RXD signal is routed to receiver 3,
but in DCE mode, the TXD signal is
routed to receiver 3. The interface
mode is selected by logic outputs
from the controller or from jumpers to
either VCC or GND on the mode-select
pins. The single-ended driver 1 and
receiver 4 of the control chip share
the RL signal on connector pin 21.
With EC low and CTRL/CLK high, the
DCE/DTE pin becomes an enable
signal.
Single-ended receiver 4 can be connected to pin 22 to implement the RI
(ring indicate) signal in RS232 mode
(see Figure 17). In all other modes,
pin 22 carries the DSR(B) signal.
A cable selectable multiprotocol
interface is shown in Figure 18. Control signals LL, RL and TM are not
implemented. The VCC supply and
select lines M0 and M1 are brought
out to the connector. The mode is
selected in the cable by wiring M0
(connector pin 18) and M1 (connector
pin 21) and DCE/DTE (connector pin
25) to ground (connector pin 7) or
letting them float. If M0, M1 or DCE/
DTE are floating, pull-up resistors
R3, R4 and R5 will pull the signals to
VCC. The select bit M1 is hard wired to
VCC. When the cable is pulled out, the
interface goes into the no-cable mode.
A cable-selectable multiprotocol
interface found in many popular data
routers is shown in Figure 19. The
entire interface, including the LL signal, can be implemented using the
tiny µDB-26 connector.
Conclusion
The LTC1343 and LTC1344 allow the
designer of a multiprotocol serial interface to spend all of his time on the
software rather than the hardware.
Simply drop the chips down on the
board, hook them up to the connector
and a serial controller, apply the 5V
supply voltage and you’re off and
running. In addition, the chip set’s
small size and unique termination
topology allow many ports to be placed
on a board using inexpensive connectors and cables.
Linear Technology Magazine • August 1996
DESIGN FEATURES
C6
100pF
C7
100pF
3
C8
100pF
8
11
12 13
LTC1344
VCC
5V
14
+
C3
1µF
C1 +
1µF
1
44
2
43
42
4
CHARGE
PUMP
3
+
C5
1µF
41
8
R1
100k
+
C11
1µF
C9
1µF
VCC
+
C12
1µF
40
GND
23
LB
R2
100k
LB
7
1
+ C10
43
42
1µF
4
CHARGE
PUMP
DCE
TM A
RXD A
RXD B
RXC A
RXC B
TXC A
TXC B
RXC A
RXC B
RXD A
RXD B
TXC A
TXC B
SCTE A
SCTE B
TXD A
TXD B
TM A
LL A
SGND
SHIELD
C13
41
8
+ 3.3µF
LTC1343
39
D1
38
37
36
35
34
33
D2
D3
D4
R2
32
31
30
29
R3
28
27
R4
20
CTRL
22
LATCH
11
INVERT
25
423 SET
21
DCE
19
M2
18
M1
17
M0
R1
16
LATCH
VCC
44
15
VCC
24
1
14
DTE_CTS/DCE_RTS
EC
2
9
DTE_DSR/DCE_DTR
DTE
LL A
25
10
12
13
DTE_DCD/DCE_DCD
18 17 19 20 22 23 24 1
26
21
DCE
19
M2
18
M1
17
M0
7
DTE_DTR/DCE_DSR
16 15
R3
R2
6
DTE_RTS/DCE_CTS
10
16
R4
20
CTRL
22
LATCH
11
INVERT
25
423 SET
R1
5
DTE_RL/DCE_RL
9
15
12
17
9
3
16
D4
3
+
7
32
31
30
29
28
27
D3
15
DTE_TM/DCE_LL
6
2
TXD A
14
TXD B
24
SCTE A
11
SCTE B
14
DTE_RXD/DCE_TXD
4
38
37
36
35
34
33
10
12
13
DTE_RXC/DCE_SCTE
DB-25 CONNECTOR
DCE/
DTE M2 M1 M0
5
D2
9
DTE_TXC/DCE_TXC
2
VEE
C4
+ 3.3µF
18
7
DTE_SCTE/DEC_RXC
1µF
39
D1
6
DTE_TXD/DCE_RXD
+ C2
LTC1343
5
DTE_LL/DCE_TM
VCC
40
GND
23
LB
21
RL A
4 RTS A
19 RTS B
20
DTR A
23
DTR B
8
DCD A
10
DCD B
6
DSR A
22
DSR B
5
CTS A
13
CTS B
RL A
CTS A
CTS B
DSR A
DSR B
DCD A
DCD B
DTR A
DTR B
RTS A
RTS B
26
EC
24
DCE/DTE
M2
M1
M0
Figure 16. Controller-selectable multiprotocol DTE/DCE port with DB-25 connector
Linear Technology Magazine • August 1996
21
DESIGN FEATURES
C6
100pF
C7
100pF
3
C8
100pF
8
11
12 13
LTC1344
VCC
5V
14
+
C3
1µF
C1 +
1µF
1
44
2
43
42
4
CHARGE
PUMP
3
+
C5
1µF
D2
D3
9
TXC
D4
10
12
13
VCC
LL
R1
100k
+
VCC
C11
1µF C9
1µF
+
+
C12
1µF
VCC
10
12
13
VCC
LATCH
R2
100k
SHIELD (101)
+ 3.3µF
D2
D3
D4
R2
R3
28
27
R4
20
CTRL
22
LATCH
11
INVERT
25
423 SET
21
DCE
19
M2
18
M1
17
M0
40
GND
23
LB
5
13
6
22
8
10
38
37
36
35
34
33
R1
16
RL
1
SGND (102)
39
D1
32
31
30
29
15
CTX
RXD A (104)
RXD B
RXC A (115)
RXC B
TXC A (114)
TXC B
LTC1343
14
DTR
TM A (142)
C13
41
8
9
DCD
7
1µF
7
DSR
16 15 18 17 19 20 22 23 24 1
VCC
+ C10
6
CTS
24
43
42
5
RI
10
VCC
44
CHARGE
PUMP
9
24
SCTE A (113)
11
SCTE B
2
TXD A (103)
14
TXD B
18
LL A (141)
2
3
7
VCC
1
4
6
3
16
17
9
15
12
26
21
DCE
19
M2
18
M1
17
M0
EC
4
25
16
R4
20
CTRL
22
LATCH
11
INVERT
25
423 SET
40
GND
23
LB
DB-25 FEMA;E
CONNECTOR
DCE/
DTE M2 M1 M0
38
37
36
35
34
33
R3
R2
15
TXD
VEE
39
32
31
30
29
28
27
R1
14
SCTE
2
+ 3.3µF
5
D1
7
RXC
1µF
LTC1343
6
RXD
LB
41
5
TM
+ C2
C4
8
VCC
20
DTR A (108)
23
DTR B
4
RTS A (105)
19
RTS B
21
RL A (140)
26
EC
CTS A (106)
CTS B
DSR A (107)
DSR B/RI A (125)
DCD A (109)
DCD B
RIEN = RS232
24
M2
M1
M0
Figure 17. Controller-selectable multiprotocol DCE port with ring-indicate and DB-25 connector
22
Linear Technology Magazine • August 1996
DESIGN FEATURES
C6
100pF
C7
100pF
3
C8
100pF
8
12 13
11
LTC1344
VCC
5V
14
+
C3
1µF
C1 +
1µF
1
44
2
43
42
4
CHARGE
PUMP
3
+
C5
1µF
41
D4
10
12
13
R3
32
31
30
29
28
27
16
R4
20
CTRL
22
LATCH
11
INVERT
25
423 SET
26
21
DCE
19
M2
18
M1
17
M0
R1
14
DTE_RXC/DCE_SCTE
R2
15
DTE_RXD/DCE_TXD
R1
100k
+
C11
1µF
C9
1µF
VCC
+
40
GND
23
LB
+
C12
1µF
8
23 24 1
DTE
2
TXD A
14
TXD B
24
SCTE A
11
SCTE B
TXC A
TXC B
RXC A
RXC B
RXD A
RXD B
TXC A
TXC B
SCTE A
SCTE B
TXD A
TXD B
SHIELD
VCC
VCC
R3
10k
VCC
R4
10k
VCC
R5
10k
25
+ 3.3µF
21
18
DCE/DTE
M1
M0
39
D1
4 RTS A
19 RTS B
20
DTR A
23
DTR B
R2
32
31
30
29
R3
28
27
8
DCD A
10
DCD B
6
DSR A
22
DSR B
5
CTS A
13
CTS B
R4
20
CTRL
22
LATCH
11
INVERT
25
423 SET
26
21
DCE
19
M2
18
M1
17
M0
40
GND
23
LB
DCE
RXD A
RXD B
RXC A
RXC B
SGND
38
37
36
35
34
33
16
LB
18 17 19 20 22
LTC1343
15
R2
100k
16 15
C13
41
14
VCC
24
1µF
10
12
13
DTE_CTS/ DCE_RTS
EC
CHARGE
PUMP
9
DTE_DSR/DCE_DTR
10
1
+ C10
7
DTE_DCD/DCE_DCD
9
7
43
42
6
DTE_DTR/DCE_DSR
7
VCC
44
5
DTE_RTS/DCE_CTS
6
15
12
17
9
3
16
1
3
4
VCC
2
4
DB-25 CONNECTOR
DCE/
DTE M2 M1 M0
38
37
36
35
34
33
D3
9
VEE
5
D2
7
DTE_TXC/DCE_TXC
2
+ 3.3µF
39
D1
6
DTE_SCTE/DEC_RXC
1µF
LTC1343
5
DTE_TXD/DCE_RXD
+ C2
C4
8
VCC
D2
D3
D4
R1
24
EC
CTS A
CTS B
DSR A
DSR B
DCD A
DCD B
DTR A
DTR B
RTS A
RTS B
VCC
CABLE WIRING FOR MODE SELECTION
MODE
PIN 18
PIN 21
V.35
PIN 7
PIN 7
EIA-530, RS449,
NC
PIN 7
V.36, X.21
RS232
PIN 7
NC
CABLE WIRING FOR DTE/DCE
SELECTION
MODE
PIN 25
DTE
PIN 7
DCE
NC
Figure 18. Cable-selectable multiprotocol DTE/DCE port with DB-25 connector
Linear Technology Magazine • August 1996
23
DESIGN FEATURES
C6
100pF
C7
100pF
3
C8
100pF
8
12 13
11
LTC1344
VCC
5V
14
+
C3
1µF
C1
1µF
+
1
44
2
43
42
4
CHARGE
PUMP
3
+
C5
1µF
41
D4
10
12
13
R3
32
31
30
29
28
27
16
R4
20
CTRL
22
LATCH
11
INVERT
25
423 SET
26
21
DCE
19
M2
18
M1
17
M0
R1
14
DTE_RXC/DCE_SCTE
R2
15
DTE_RXD/DCE_TXD
R1
100k
+
C11
1µF
C9
1µF
VCC
+
40
GND
23
LB
+
C12
1µF
CHARGE
PUMP
LB
DTE
2
TXD A
14
TXD B
24
SCTE A
11
SCTE B
TXC A
TXC B
RXC A
RXC B
RXD A
RXD B
TXC A
TXC B
SCTE A
SCTE B
TXD A
TXD B
SHIELD
VCC
VCC
R3
10k
VCC
R4
10k
VCC
R5
10k
25
+ 3.3µF
21
18
DCE/DTE
M1
M0
39
D1
4 RTS A
19 RTS B
20
DTR A
23
DTR B
R2
32
31
30
29
R3
28
27
R4
20
CTRL
22
LATCH
11
INVERT
25
423 SET
26
21
DCE
19
M2
18
M1
17
M0
8
DCD A
10
DCD B
6
DSR A
22
DSR B
5
CTS A
13
CTS B
26
LL B
40
GND
23
LB
DCE
RXD A
RXD B
RXC A
RXC B
SGND
38
37
36
35
34
33
16
R2
100k
18 17 19 20 22 23 24 1
LTC1343
15
VCC
16 15
C13
8
14
DTE_CTS/DCE_RTS
24
41
10
12
13
DTE_DSR/DCE_DTR
EC
1µF
9
DTE_DCD/DCE_DCD
10
1
+ C10
7
DTE_DTR/DCE_DSR
9
7
43
42
6
DTE_RTS/DCE_CTS
7
VCC
44
5
DTE_LL/DCE_LL
6
15
12
17
9
3
16
1
3
4
VCC
2
4
µDB-26 CONNECTOR
DCE/
DTE M2 M1 M0
38
37
36
35
34
33
D3
9
VEE
5
D2
7
DTE_TXC/DCE_TXC
2
+ 3.3µF
39
D1
6
DTE_SCTE/DEC_RXC
1µF
LTC1343
5
DTE_TXD/DCE_RXD
+ C2
C4
8
VCC
D2
D3
D4
R1
24
EC
CTS A
CTS B
DSR A
DSR B
DCD A
DCD B
DTR A
DTR B
RTS A
RTS B
LL B
VCC
CABLE WIRING FOR MODE SELECTION
MODE
PIN 18
PIN 21
V.35
PIN 7
PIN 7
EIA-530, RS449,
NC
PIN 7
V.36, X.21
RS232
PIN 7
NC
CABLE WIRING FOR DTE/DCE
SELECTION
MODE
PIN 25
DTE
PIN 7
DCE
NC
Figure 19. Cable-selectable multiprotocol DTE/DCE port with µDB-26 connector
24
Linear Technology Magazine • August 1996
DESIGN IDEAS
Li-Ion Battery Charger Does Not
Require Precision Resistors
by Arie Ravid
In constant-voltage mode charging, a Li-Ion cell requires 4.1V ±50mV.
This 1.2% tolerance is tight. In a
regulation loop where a voltage divider
is compared against a reference, the
accuracy is achieved by selecting a
0.7% reference and a voltage divider
with 0.25% tolerance resistors.
Unfortunately, 0.25% precision resistors cost three times as much as
1% resistors and have very long lead
times.
One solution for moderate volume
production involves adding two 1%
resistors and two jumpers to the
charger circuit, as shown in Figure 1.
The jumpers are removed as necessary to bring the constant voltage to
the required accuracy of 1.2%.
The charger selected for this
example is the LT1510 and the number of Li-Ion cells in the battery is
three. Select a value for R4 (20k) and
calculate the values for resistors R1,
R2 and R3 using the equations in
Figure 1. K is the relative change
required for a circuit with all its tolerances in one direction. For example,
in the case of a 0.5% reference and
DESIGN IDEAS…
Li-Ion Battery Charger Does Not
Require Precision Resistors ... 25
Arie Ravid
Constant-Voltage Load Box for
Battery Simulation ................ 26
Jon Dutra
LT1510 Charger with
–∆V Termination .................... 27
Arie Ravid
Transparent Class-D Amplifiers
Featuring the LT1336 ........... 28
Dale Eagar
An Ultralow Quiescent Current,
5V Boost Regulator ................ 33
Sam Nork
24 Volt to 14 Volt Converter
Provides 15 Amps .................. 34
John Seago
The LT1210: High Power Op Amp
Yields Higher Voltage and Current
.............................................. 36
Dale Eagar
LTC1441-Based Micropower
Voltage-to-Frequency Converter
.............................................. 38
Jim Williams
Capacitive Charge Pump Powers
12V VPP from 5V Source ........ 40
Mitchell Lee
Bridge Measures Small
Capacitance in Presence of Large
Strays .................................... 41
two 1% resistors, the total tolerance
is 2.5%. In order to bring it back to
1.2%, the percentage change required
is 2.5% – 1.2% = 1.3% and K = 0.013.
The jumpers J1 or J2 need to be
opened based on the following:
If VOUT is K/2 below nominal,
remove J1.
If VOUT is K/2 above nominal,
remove J2.
The following values were calculated: R1 = 20k, R2 = 324Ω, R3 =
80.6Ω and R4 = 4.99k.
The voltage below which J1 should
be opened is 12.34V – 1.3%/2 =
12.22V.
The voltage above which J2 should
be opened is 12.34V + 1.3%/2 =
12.42V.
The complete schematic can be
seen in Figure 2. Q3 is off when the
charger is not powered, preventing
current drain from the battery through
the voltage divider. R5, a 100k resistor, isolates the OVP pin from any
high frequency noise on VIN. The
charger in Figure 2 is programmed for
1.3A constant current.
Jeff Witt
D3
1N5819
C1
D1
0.22µF 1N5819
SW
VCC
+
VIN
L1**
33µH
R1
OVP
R2
–
VREF
2.465V
J1
+ 3-CELL
Li-ION
R3
+
–
BOOST
VOUT = 12.3V
BAT
LT1510
CONSTANT VOLTAGE/
CONSTANT CURRENT
BATTERY CHARGER
J2
– BATTERY
GND
1µF
VC
+
R3 =
–+
CIN*
10µF
LT1510
0.1µF
300Ω
3.83k
1k
OVP
SENSE
BAT
+
COUT
22µF
TANT
3-CELL
Li-ION
BATTERY
J1
R4
R1 = R4 × VOUT – VREF
VREF
R2 = (R1+R2) × K
R5
100k
R1
R2
Q3
VN2222
*TOKIN OR MARCON CERAMIC SURFACE MOUNT
** COILTRONICS CTX33-2
J2
R3
R4 × K
V
1 – (1 – K) REF
VOUT
Figure 1. R2, R3, J1 and J2 eliminate the need for precision resistors.
Linear Technology Magazine • August 1996
D2
1N914
PROG
16V TO 28V
R4
Figure 2. 3-cell Li-Ion charger without precision resistors
25
DESIGN IDEAS
Constant-Voltage Load Box
for Battery Simulation
Linear Technology has developed
many new switcher-based battery
charger ICs. Testing accuracy, regulation and efficiency in the lab with a
battery load is inconvenient because
the terminal voltage of a battery constantly changes as it is being charged.
If much testing is to be done, a large
supply of dead batteries will be needed,
since one set of cells can quickly
become overcharged. This Design Idea
describes an active load circuit that
can be used to simulate a battery in
any state of charge. The battery simulator provides a constant-voltage load
for a battery-charging circuit, independent of applied charging current.
The simulator’s impedance is less
than 500mΩ at all reasonable input
frequencies. Best of all, the simulator
can never be overcharged, allowing
long-term testing and debugging of a
charger system without the possibility of battery damage.
by Jon Dutra
Circuit Operation
The simulator uses an LT1211 high
speed, single-supply op amp to drive
the base of a high gain PNP transistor-stage active load. Power for the
LT1211—a portion of the charging
current—is supplied through a diode
so the op amp and reference can
survive brief periods of zero charging
current. The op amp is configured for
a DC gain of four, so the voltage on its
noninverting input is one fourth of
the voltage that the load box is set to.
With S1 open, the load-voltage adjust
range will be from 10V to 20V, and
with S1 closed it will be approximately
3.5V–10V. Low voltage operation
could be improved by replacing the
top LT1004-2.5 with an LT1004-1.2
and reducing R1, the reference bias
resistor, to 1k. The 510Ω and 1.1k
resistors are required for high frequency stability; they suppress a
1MHz oscillation. The 1N5400 diode
and 4-amp fuse protect the circuit
from reverse voltages.
Results
The battery simulator circuit has been
tested “swallowing” currents from
30mA to 3A with the output voltage
essentially unchanged. When simulating a battery, the voltage adjust
can be increased until the charger
thinks the battery is fully charged
and reduces the current into the simulator. Conversely, as the voltage is
adjusted down, the battery charger
may think the battery is becoming
discharged and increase the current
into the simulator.
Figure 2 shows the circuit’s capacity for current absorption at two
voltages, 5V and 15V, from 50mA to 3
amps.
R1 10k
0.033µF
1N5817 OR
BAT-85
+
–
5.0V OR 2.5V
1/2
LT1211
100kΩ
10 TURN POT
+
LT1004-2.5
0.5Ω
5W
30k, 1%
100µF
25V
IN+
15.5
1.1kΩ
5.3
15.4
510Ω
+
5.2
15.3
Q1
2N6667
(SEE NOTE)
270µF
25V
2.5V OR 0V
0.033µF
LT1004-2.5
4A FUSE
VOLTAGE (V)
10k, 1%
1N5400
5.1
15.2
5V
15.1
5
15
15V
4.9
14.9
S1
4.8
14.8
IN–
4.7
14.7
0
0.5
1
1.5
2
CURRENT (A)
2.5
3
S1 CLOSED ≥ 0 TO 10V RANGE
S1 OPEN ≥ 10V TO 20V RANGE
ALL RESISTORS 5% UNLESS NOTED
Q1 DISSIPATES MOST OF THE POWER, MOUNT ON AN ADEQUATE HEAT SINK!
Figure 1. Schematic diagram of battery simulator
26
Figure 2. Current absorption capacity of the
battery simulator at 5V and 15V
Linear Technology Magazine • August 1996
DESIGN IDEAS
LT1510 Charger with –∆V Termination
by Arie Ravid
Any portable equipment that requires fast charge needs proper
charge termination. Commonly, a
LT1510 constant-voltage, constantcurrent type charger controlled by a
microcontroller is used. Sometimes,
however, a microcontroller is not
available or is not suitable for fastcharge termination.
When fast charging NiCd batteries
with constant current, the internal
battery temperature rises toward the
end of the charge. Since the temperature coefficient of NiCd is negative,
the temperature rise causes the battery voltage to drop. The drop can be
detected and used for termination
(called –∆V termination). The circuit
in Figure 1 is a solution for a 3-cell
(Panasonic P140-SCR) NiCd battery
charger with –∆V termination.
U1 in Figure 1 is programmed by
resistor R2 for a conservative charge
current of 0.8A, which is 0.57C. Typi-
cal fast-charge current is 1C. (The
boldfaced C represents a normalization concept used in the battery
industry. A C rate of 1 is equal to the
capacity of the cell in ampere-hours,
divided by 1 hour. Since the capacity
of the P140-SCR is 1.4 ampere-hours,
C is 1.4 amperes.)
To determine the voltage droop rate,
the battery was connected to an
LT1510 charger circuit programmed
for a 0.8A constant-current. The data
was plotted as voltage versus time
and the results are shown in Figure 2.
The voltage slope is calculated to be
–0.6mV/s. After the battery voltage
dropped 300mV from the peak of 4.93V
(100mV per cell), the charger was
disabled.
At the heart of the circuit in Figure
1 is U3, a sample-and-hold IC (LF398).
For every clock pulse at pin 8, the
output of U3 (pin 5) updates to the
input level on pin 3. When the battery
voltage drops, the input to U3 also
drops. If the update step at the output
of U3 is sufficiently negative, U2B
latches in the high state and Q1 turns
on. Q1 terminates the charge by pulling down the LT1510’s VC pin, and
thereby disabling it.
U2A and the associated passive
components smooth, amplify and level
shift the battery voltage. The timer
(U4) updates the hold capacitor (C8)
every fifteen seconds. The timer signal stays high for 7ms, sufficient time
for the hold capacitor to be charged to
the input level. U2B and the associated parts form a latch that requires
a momentary negative voltage at pin 6
to change state. R15 supplies the
negative feedback and Q2, R16, R17
and C10 reset the latch on turn-on.
U3’s output voltage droops at
a rate proportional to the hold
capacitor’s internal leakage and the
continued on page 32
12V
2
15
VCC1
VCC2 14
13
PROG
SW
C1
0.22µ
L1*
30µ
3 BOOST
U1
LT1510
4
GND
CR2
1N5819
CR1
1N914
6
VC
12
CR3
1N5819
C3 +
1µ
R1
300
R2
6.19k
C4
R3 0.1µ
1k
11
GND BAT
**1, 7, 8, 9, 10, 16
C2†
10µ
+
C5
22µ
25V
C7
0.1µ
R8
100k
–
C6
0.1µ
C12, 0.01µ
16
2
15
4
–
1
5
R20
100k
+
3
C9
0.1µ
OUTPUT
+
6
CLK
150Ω
8
U2B
5 LT1013
R13
10k
12
6
11
7
10
8
9
7
R15
100k
R16
100k
Q2
2N3904
R17
100k
4
Q1
2N3904
C10 +
22µ
–
INPUT
R10
30.1k
13
LT1029CZ
+
LOGIC
REFERENCE
14
U4
CD45368
5
R11
100k
7
1
3
5VREF
30k
LOGIC
R18, 100k
R19
100k
1
–
8
U2A
3 LT1013
+
4
R14
10k
U3
LF398
OFFSET
CLK
C11
0.22µ
N/C
R4
10k
2
B1***
R9
30.1k
2
R5
100k
R6
100k
SENSE
R7
10
5VREF
HOLD
CAPACITOR
6
C8
1µ
ECQV1HIOSJL
PANASONIC
R12
1k
R21
10k
NOTES: *L1 IS COILTRONICS CTX 33-2
**SOLDER TO GROUND PLANE FOR HEAT DISSIPATION
***B1 IS A NiCd 3 CELL PANASONIC P140-SCR
†C2 IS A TOKIN OR MARKON CERAMIC SURFACE MOUNT
Figure 1. Schematic diagram: 3-cell NiCd charger with –∆V termination
Linear Technology Magazine • August 1996
27
DESIGN IDEAS
Transparent Class-D Amplifiers
Featuring the LT1336
Introduction
+
+
HEFTY
WIRES
FROM CAR
BATTERY
9V-15V
+
C1
+
+
C2
C3
Class-D amplifiers can be simple or
complex, depending on what is required by the application. A simple
class-D amplifier is the thermostatic
switch in an electric heater. The thermostat controls the heater by turning
it on or off. The switch is essentially
lossless, dissipating practically no
power. This class-D amplifier is remarkably efficient, since even the
energy lost in the switch, power cord
and house wiring contributes to the
desired result. The duty factor, and
hence the average amount of power
delivered to the heater, can assume
an infinite number of values. This is
true even though a constant amount
of heat is delivered when the heater
is on.
PRi
20T
2x#14
SEC
4T
#26
MICROMETALS T150-52
R1
1Ω
C4
D3
MUR110
C5
22µF
20V
1200µF, 16V ×4
+
–
The Electric Heater—
a Simple Class-D Amplifier
ergy in both directions—input to output and output to input.
Class-D amplifiers also have a way
of ignoring reactive loads that can be
uncanny. A class-D amplifier operating with an AC output will draw very
little additional input power when a
sizable capacitive or inductive load is
placed at its output. This is because
the reactive load has AC voltage across
it and AC current flowing through it,
but the phase angle of the voltage and
current is such that no real power is
dissipated. The class-D amplifier ends
up shuttling power back and forth
between its input and its output, doing both with minimal loss. An ideal
class-D amplifier can be thought of as
having no place to dissipate power,
since all of its components are lossless;
that is, it contains no resistors.
Efficiency in the field of power conversion is like transparency in the
field of light transmission. It is no
wonder, then, that Class-D amplifiers
are often called transparent, since
they have no significant power losses.
In contrast to class-D amplifiers’
nearly lossless switching, class-A
through class-C amplifiers are throttling devices that waste significant
energy. Amplifiers of the “lower
classes” (A–C) are modeled as
rheostats (variable resistors), whereas
class-D amplifiers are modeled as
variacs (variable transformers). The
ideal resistor dissipates power,
whereas the ideal transformer does
not. Like transformers (variacs), many
class-D amplifiers can transfer en-
by Dale Eagar
12V
D4
MUR110
SEC
T1
5
C6
1500pF
R2
100k
1
2
8
C7
0.1
Q1
2N3904
R7
1k
6
+
–
4
7
U2A
LT1215
4
R6
2.4k
FB
VREF
R5
15k
4
C9
0.15
VC
VIN
8
+
C11
33µF
16V
+
C12
33µF
16V
D2
MBR1060
R4
2.49k
R3
51k
7
PRi
36µH 20
30A
U1 GTDR
LT1243
R16
20Ω
6
D1
1N5819
+
C15
C19
1000µF, 63V ×6
R/C
C8
1µF
3
+
C14
R17
20Ω
C13
2200µF
25V
55V, 3.3A
+
Q3
IRFZ44
Q2
IRFZ44
+
R12
0.01Ω
1W
ISEN
GND
5
R8
1k
+
1
U2B
LT1215
–
R9
1k
3
2
R13
0.01Ω
1W
R14
0.01Ω
1W
R15
0.01Ω
1W
R11
1k
C10
220pF
R10
100Ω
Figure 1. 200W, 12V to 60V front end for automotive applications
28
Linear Technology Magazine • August 1996
DESIGN IDEAS
T1
COILTRONICS
CTX100-P
12V 55V
D3
1N4148
IN
TOP
OUT
C1
1µF
IN
BOTTOM
C2
0.1
R1
6.2k
12V
D5
1N4148
R2
2Ω
2
V+
1
16
10
SW
D1
1A
60V
IN TOP
IN BOTTOM
BOOST
13
TGD
12
TGF
ISEN
TSOU
4
BGD
IN TOP
15
S
GND
6
Q1
IRFZ44
11
9
Q2
IRFZ44
OUT
BGF
R6, 10Ω
D2
1A
60V
IN BOTTOM
SW
GND
R4, 10Ω
R5, 10Ω
U1
LT1336
3
55V
R3, 10Ω
14
V+
4-Quadrant
Class-D Amplifier
(SYMBOLIC REPRESENTATION)
C3
1µF
D4
1N4148
Q3
IRFZ44
Q4
IRFZ44
8
PGND
7
Figure 2. Half-bridge driver subcircuit and symbolic representation
Quadrants of Energy Transfer
Class-D amplifiers have a property
that requires new terminology, a property that generally isn’t considered in
lower-class amplifiers. This property,
quadrants of energy transfer, describes the output characteristics of
the class-D amplifier. The output
characteristics are plotted on a imaginary X-Y plot (I’ve yet to see someone
actually do one on paper), one axis
representing output voltage and the
other axis representing output current, with the intersection of the axes
representing zero volts and zero amps.
A simple switcher that can only provide a positive output current into a
positive output voltage can be
described as a 1-quadrant device.
This 1-quadrant device could be a
computer power supply, a battery
charger or any supply that delivers a
positive voltage into a device that can
only consume power.
The 2-quadrant converter can be
one of two different things: 1) A positive output voltage that can both
source and sink current, or 2) A posiLinear Technology Magazine • August 1996
plifier. Figure 2 details the LT1336
driving power MOSFETs and shows
the symbolic representation of this
subcircuit that will appear in subsequent figures. Table 1 shows the
logical states of this half-bridge power
driver.
tive current that can comply both
positive and negative output voltage.
Finally, the 4-quadrant converter can
both source and sink current into
both positive and negitive output
voltages.
1-Quadrant
Class-D Converter
To illustrate the 1-quadrant class-D
amplifier, we will focus on the boost
mode converter detailed in Figure 1
This circuit removes power from the
source (12V automotive battery) and
delivers it to the load (some as-yetunknown 55V device) This circuit is
classified as “1 quadrant” because it
can only regulate output voltage in
one polarity (positive) and it can output current in only one polarity
(positive).
Introducing the
LT1336 Half-bridge Driver
Taking a side step from our main
discussion, we will introduce a
component, the half-bridge power am-
Class-D amplifiers are commonly used
in subwoofer drivers. This is because
subwoofers require a great deal of
power. A class AB amplifier driving a
subwoofer will put about half of its
input power into its heat sink. Driving the same subwoofer at the same
volume with the same music, a classD amplifier will put about five percent
of its input power into the heat sink.
The difference is ten to one on the
heatsink size and two to one on the
input power supply. Figure 3 is the
200W class-D subwoofer driver. This
circuit uses the 200W front end developed in Figure 1 as its power source.
The circuit in Figure 3 performs as
follows: U1a, R1–R4 and C7 implement a 75kHz pseudosawtooth
oscillator. U1d is the input amplifier/
filter, with a gain of 6.1 and 200Hz
Butterworth lowpass response. U1b
and U1c are comparators that compare the sawtooth and the amplified/
filtered input signal to form two complimentary, pulse-width modulated
square waves. X1 and X2 are two
half-bridge power drivers and M1 is
the subwoofer driver.
One of the properties of Class-D,
4-quadrant amplification is the ability to transfer power both to and
from the load. In our subwoofer
driver, this happens when the driver
reaches the end of any given excurTable 1. Half-bridge power driver truth table
In Top
In Bottom
Output
L
L
Floating
L
H
Ground
H
L
55V
H
H
Floating
29
DESIGN IDEAS
12V
R1
15k
12V
R3
1.8k
+
8
U1A
LT1365
–
R4
15k
10
5
9
6
18" SUBWOOFER
DRIVER
1mH, 6.5Ω
C8
0.1
R2
15k
12V 55V
4
+
U1B
LT1365
7
U1C
LT1365
1
55V 12V
F1
10A
M1
–
C7
220pF
R5
15k
75kHz
3
+
2
C4
0.015
C1
2.2µF
R6
1.8k
R7
18k
–
11
R8
100Ω
R9
18k
3
INPUT
C3
0.1
C2
0.47
C5
0.022
C6
2.2µF
2
+
U1D
LT1365
1
–
R10
51k
R11
10k
Figure 3. 200W-powered subwoofer
12V
15k
12V
1.8k
+
8
0.1
15k
U1A
LT1365
–
10
+
9
55V 12V
L1
1mH
U1B
LT1365
–
M
220pF
15k
3
2
+
U1C
LT1365
–
11
1
3
–
12V
MOTOR
SPEED
AND
DIRECTION
LOAD
12V 55V
4
2
U1D
LT1365
1
+
POT 1
Figure 4. Class-D motor drive
30
Linear Technology Magazine • August 1996
DESIGN IDEAS
+
12V BATTERY
+
R1
1Ω
–
+
C1
+
C2
PRi
20T
2x#14
SEC
4T
#26
MICROMETALS T150-52
+
C3
C4
D1
MUR110
1200µF, 16V ×4
+
C5
22µF
20V
5
6
C6
1.5µF
R2
100k
1
2
8
C7
0.1
C9
0.15
R7
1k
R6
2.4k
VC
20
4
VIN
VREF
VN2222
C11
33µF
16V
+
C12
33µF
16V
+
C13
2200µF
25V
55V
+
C14
U1 GTDR
LT1243
+
+
C15
C19
6
1000µF, 63V ×6
R/C
C8
1µF
3
+
2k
4
FB
R5
15k
4
Q1
2N3904
T1
U2A
LT1215
–
12V
D2
MUR110
7
R4
2.49k
R3
51k
SEC
8
+
ISEN
R12
0.01Ω
1W
GND
5
R8
1k
+
1
U2B
LT1215
–
R9
1k
3
2
R13
0.01Ω
1W
R14
0.01Ω
1W
R15
0.01Ω
1W
R11
1k
C10
220pF
R10
200Ω
R16
49.9k
Figure 5. 200W, 2-quadrant front end for automotive applications
sion and the combination of the driver
spring and the acoustic spring drive
the cone back to center. During this
time, energy is transferred from the
driver back to the input of the classD amplifier stage. In the case shown
in Figure 3, the energy ends up on
the 55V bus, where the bus voltage
climbs during these periods of “negative energy delivered to the load.”
Fortunately, C14–C19 of Figure 1 can
store this energy; otherwise the 55V
bus would subject to excessive voltage until someplace was found for
the energy to go.
but when one wants to slow the motor
down by turning pot 1 back toward its
center, disaster strikes. Rotational
energy stored in the inertia of the
motor is converted back into electrical energy by the motor and is
presented to the output of the class-D
amplifier. L1, X1 and X2 do their job
by transferring the energy back into
the 55V bus. The energy goes into
C14–C19 of Figure 1, charging them
to some voltage significantly above
55V, and something breaks. The problem here is that the circuit in Figure 1
is only a 1-quadrant class-D amplifier.
Managing the
Substituting a motor and an inductor Negative Energy Flow
Class-D for Motor Drives
for the subwoofer in Figure 3 and
simplifying the control, we arrive at
the circuit shown in Figure 4. Connecting this circuit to the front end
shown in Figure 1 and then getting
the motor up to speed is no problem,
Linear Technology Magazine • August 1996
Sound like a course in management?
The negative energy transferred
through the class-D amplifier needs a
home. One simple home is the 62V
power Zener diode strapped across
the 55V bus and bolted to a massive
heat sink. One could easily imagine
the heat sink as the brake shoes
heating up as the electric vehicle winds
down the mountain road. Another
place to put the energy is back into
the 12V battery. This will require
upgrading the 12V to 55V front-end
power converter from 1 quadrant to 2
quadrants.
The 2-Quadrant
Class-D Converter
Converting Figure 1 to two quadrants
involves replacing D2 with a switch
and activating the switch out of phase
with the switch formed by Q2 and Q3.
The half-bridge power driver shown
in Figure 2 is just such a switch. Refer
to Figure 5. The ISENSE signal (U1, pin
3) needs to be offset to accommodate
negative current (add R16, Figure 5)
The ISENSE signal needs to be scaled
for twice the range (–30A to 30A rather
31
DESIGN IDEAS
A Trip Over the Great Divide
than 0A to 30A); this is done by changing R10.
Now we are happily winding down
the mountain road, watching the scenery unfold before us. We are happy in
knowing that we are recycling the
energy released from the descent by
charging our batteries, while watching the mountain bikers burn their
descent energy off in brake linings.
Once again technology wins over sweat
and brawn.
Climbing the great divide in an electric vehicle requires some planning.
Stops to recharge are necessary. Once
on top, the whole scheme changes:
descending the hill, charging our battery, all goes well until the battery is
fully charged; then we have to stop.
Further descent would overcharge our
battery, boiling out the electrolyte.
Not only would this ruin our battery,
VIN
R1
49.9k
R2
95k
U1
FIG 5
PIN 8
6
VREF
5.00V
Conclusion
R3
5M
8
R5
1Ω
200W
Class-D has been around for a long
time: the venerable electric heater
with its bang-bang controller is a
remarkably efficient and reliable
class-D amplifier. Class-D drives have
been used for decades in golf carts,
fork lifts, cranes and industry. The
advent of the half-bridge driver greatly
simplifies the Class-D Amplifier. Here
at Linear Technology we have a family
of half/full bridge MOSFET drivers.
For further information, contact us at
the factory or refer to the LT1158,
LT1160, LT1162 or LT1336 data
sheets.
R4
100Ω
+
U2A
LT1215
7
IRFZ40
–
14A
0A
14.3V 14.6V
Figure 6. Wolf Creek Pass adapter
–∆V Termination, continued from page 27
–dV/dT = VTRIG/(TCLK × GU2A)
where:
VTRIG is the trigger voltage of U2B,
VTRIG = VREF × R12/(R11 + R12)
= 5 × 1/101 = 49.5mV
VREF = 5V
TCLK is the clock period, 15 seconds,
GU2A is the gain of the first stage,
= R8/(R4 || R5)
= 11
The circuit in Figure 1 was built
and connected to a system that discharges the battery to 3V after
termination, at constant current of
32
5.0
NEGATIVE
VOLTAGE SLOPE
4.9
BATTERY VOLTAGE (V)
leakage current at pin 6 (10pA typical). This droop is very low and does
not affect the operation of the circuit.
The minimum negative battery
voltage slope required to trigger termination (–dV/dT) is 0.3mV/s. It can
be calculated from:
in the end we would have no place to
put the energy and our class-D
amplifier would find some way to fail.
We need to stop and drain off some
charge, trade batteries with someone
climbing the other side or put a power
Zener on our battery. Figure 6 details
the active Zener circuit. Using the
reference in U1 of Figure 5 and the
unused half of U2 we are able to make
a hysteretic clamp that puts all of the
heat into a resistor, R5. This circuit
will save the battery from destruction
and drop our level of smugness back
to that of the mountain bikers.
END OF
0.8A CHARGE
4.8
4.7
4.6
4.5
4.4
4.3
4.2
7:30
15:00
22:30
TIME (MIN)
30:00
Figure 2. Voltage-droop rate, 3-cell
NiCd battery
0.8A. Once the battery drops to 3V,
the system reenables charging, and
thus the complete system repeats
charge/discharge cycles indefinitely.
The duration of 70 charge discharge
cycles was recorded. The following is
condensed data from the test:
1. Average Charge Time:
2:00:55 Hours
2. Standard Deviation of Charge
Time: 5:37 Minutes
3. Average Discharge Time:
1:59:14 Hours
4. Standard Deviation of Discharge
Time: 48 Seconds.
The ratio of standard deviation of
charge time to average charge time
proves that the charger has good
repeatability. However, the ratio of
standard deviation of discharge time
to average discharge time shows that
the charge level at the time of termination is very consistent because the
discharge time at constant current is
a better measure of charge level than
charge time. A secondary termination method, such as time, battery
temperature, or the like, is also
recommended.
Linear Technology Magazine • August 1996
DESIGN IDEAS
An Ultralow Quiescent Current,
5V Boost Regulator
Many battery-powered applications
require an auxiliary 5V supply to
power infrequently used circuitry,
such as smart card readers, wireless
i.d. tags, or the like. Keeping the 5V
supply permanently active is desirable, since this eliminates timing
delays and inrush currents due to
supply start-up. The downside is that
most 5V boost converters consume
an unacceptable amount of quiescent
current under no-load conditions.
This problem is addressed by the
SHDN features of the LTC1516
VIN
10µF
micropower, charge-pump DC/DC
converter. Toggling the SHDN pin of
the LTC1516 allows the 5V supply to
remain in regulation with a typical
no-load input current of less than
5µA. When the 5V output load is
enabled, the part can supply up to
50mA of load current.
The LTC1516 produces a regulated 5V output from a 2V to 5V input
(refer to Figure 1). In normal operation, the part regulates by sensing
the output through a resistive divider
and enabling a switched-capacitor
SHDN
+
S2A
VOUT
+
S1A
C2 +
10µF
0.22µF
S2B
COMP1
C2 –
S1B
CLOCK 1
COMP2
CONTROL
LOGIC
S1C
C1 +
0.22µF
S2C
C1
–
CLOCK 2
COMP3
VOS
S1D
S3
VREF
CHARGE PUMP
CHARGE PUMP SHOWN IN TRIPLER MODE, DISCHARGE CYCLE
Figure 1. LTC1516 simplified block diagram
charge pump when the output droops
below the trip point of the sense comparator (COMP2). When the output
has been boosted above COMP2’s
upper trip point, the charge pump is
turned off. In shutdown mode, the
output load is disconnected from VIN
and the quiescent current drops below 1µA.
When the output is in regulation,
the internal sense resistor draws only
1.5µA (typical) from VOUT. During
no-load conditions, this internal load
causes a droop rate of only 150mV
per second on VOUT with COUT = 10µF.
Applying a 5Hz–100Hz, 95%–98%
duty-cycle signal to the SHDN pin
ensures that the circuit in Figure 2
comes out of shutdown frequently
enough to maintain regulation during no-load (or low-load) conditions.
Since the part is kept in shutdown
mode for the majority of the time, the
no-load quiescent current (see Figure
3) is approximately equal to (VOUT ×
(1.5µA + ILOAD))/(VIN × efficiency).
The LTC1516 must be taken out of
shutdown mode for a minimum of
200µs to allow the internal sense
circuitry to start up and keep the
output in regulation. As the VOUT load
current increases, the frequency with
which the part is taken out of shutdown must also be increased to
prevent VOUT from drooping below
4.8V during the OFF phase (see Figure 4). A 100Hz, 98% duty cycle signal
continued on page 35
0.22µF
1
VIN = 2V TO 5V
2
+
10µF
3
+
10µF 4
C1+
C1–
VIN
SHDN
LTC1516
VOUT
GND
C2+
C2 –
0.22µF
by Sam Nork
8
7
FROM MPU
SHDN PIN WAVEFORMS:
6
5
LOW IQ MODE (5Hz TO 100Hz, 95% TO 98% DUTY CYCLE) VOUT LOAD ENABLE MODE
IOUT ≤ 100µA
(IOUT = 100µA TO 50mA)
VOUT = 5V ±4%
Figure 2. Ultralow quiescent current (<5µA) regulated supply
Linear Technology Magazine • August 1996
33
DESIGN IDEAS
24 Volt to 14 Volt Converter
Provides 15 Amps
The LTC1435 is an extremely
versatile voltage controller. Most applications take advantage of its ease
of design and the high efficiency of its
synchronous regulator topology for
microprocessor-level output voltages.
The LTC1435 can also be used as a
conventional buck regulator, with very
high efficiency, in circuits requiring
hundreds of watts and higher-thanlogic-level output voltages. As a
constant-frequency, current mode,
step-down switching regulator, it controls external N-channel MOSFETs
for very efficient, low noise operation.
The current mode architecture provides a tightly controlled output
voltage with excellent load and line
regulation. Internal slope compensation eliminates subharmonic
oscillations. The 1% reference ensures good initial set-point accuracy.
The switching frequency can be set
between 50kHz and 400kHz, so circuit efficiency, component size and
transient response can be properly
balanced. The LTC1435 also features
both logic-level on/off control and
output current soft-start. When the
controller is turned off, voltage is
removed from the load and quiescent
input current drops to a mere 15µA.
The LTC1435 is available in the popular 16-pin SO package.
by John Seago
Combining the LTC1435 with a
large geometry power MOSFET and
good PCB layout allows large currents to be processed easily and
efficiently. With the use of a current
sense transformer, output voltages
greater than 10V can be implemented.
The circuit in Figure 1 shows an
LTC1435 configured as a conventional
buck regulator using a single N-channel MOSFET to control an output
voltage greater than 10V with load
current exceeding 15 amps. The efficiency of the breadboard measured
94% with a 24V input, 14V output
and 15A of load current. If maximum
efficiency is required, adding a sec-
D5
1N4148
INPUT
18V TO 28V
+
C10
1000µF
35V
+
C11
1000µF
35V
T1
R3
10Ω
D1
1T
100T
C15
100pF
D2 1N758
Q1
2N3904
C1
120pF
Q3
1
C2, 0.1µF
2
R1, 10k
3
C3, 330pF
C4, 47pF
4
5
C5, 100pF
6
R2
11.8k
COSC
TG
RUN/SS
ITH
SFB
BOOST
SW
U1
LTC1435
VIN
INT VCC
SGND
BG
VOSENS
7
SENSE –
PGND
8
SENSE +
EXT VCC
15
C7, 0.1µF
C12
470pF
D3
C8, 0.1µF
12
11
Q4
2N3906
R9
0.62Ω
R10
100Ω
D6
C17
0.001µF
Q5
VN2222LL
R11
100k
L1
10µH
14
13
C16
0.001µF
D7
1N751
Q2
2N3906
16
R8
R7 1.2Ω
1K
+
C9
4.7µF
D4
14V
AT 15A
R6
127k
C13
470µF
25V
+
10
+
C14
470µF
25V
GND
9
R4, 100Ω
D8
1N4148
C6, 0.001µF
R5, 100Ω
R12, 100Ω
+
C18
1µF
R14
430Ω
R13
2.2k
R16
16k
Q6
2N3904
C19
0.01µF
R15
470Ω
C20
0.001µF
SPECIAL PARTS
C10 = C11 =NICHICON, UPL1V102MHH6
C13 = C14 =NICHICON, UPL1E471MHH6
D1 = D3 = D6 =MOTOROLA, MBRS0540
D4 =MOTOROLA MBR2045 WITH
THERMALLOY #7020 HEATSINK
Q3 =INTERNAL RECTIFIER, IRL3803 WITH
THERMALLOY #6299 HEATSINK
U1 =LINEAR TECHNOLOGY, LTC1435CS
L1: CORE =MAGNETICS, 55930-AZ
WINDING = 8T #14 BIFILAR
T1: CORE =MAGNETICS W-41406-TC
WINDING = PRI = 1T #18 SEC = 100T #32
Figure 1. 14V, 15A buck regulator
34
Linear Technology Magazine • August 1996
DESIGN IDEAS
ond power MOSFET for synchronous
switching will improve efficiency by
about 1%.
This circuit’s 100kHz switching frequency was selected to reduce
switching losses so that PCB mounted
heat sinks could be used without
requiring additional air flow. The
switching frequency can be set from
50kHz to 400kHz by selecting an
appropriate value for C1. The current
sense transformer T1 uses a 1:100
turns ratio to scale down the buck
inductor input current and develop
the voltage across R9, used by the
±SENSE inputs for regulation. Shortcircuit protection is provided by Q4
and Q5. When the current transformer secondary voltage developed
across R8 and R9 is enough to turn
on Q4, Q5 temporarily pulls the
RUN/SS pin low, turning off the regulator. Output current soft-starts when
Q5 releases the RUN/SS pin. This
results in frequent attempts to establish output voltage if a short exists,
without high current continuously
flowing through the power elements.
The power elements consist of input
capacitors C10 and C11, Current
sense transformer T1, buck inductor
L1, power MOSFET Q3, commutating
diode D4 and output capacitors C13
and C14.
Although the wide 3.6V–36V input
voltage range and 99% duty cycle
operation of the LTC1435 are ideal for
A = Q3 SWITCH VOLTAGE
20V/DIV
B = L1 CURRENT
10A/DIV
0.0V
0.0A
C = T1 PRIMARY CURRENT
10A/DIV
0.0A
D = OUTPUT VOLTAGE RIPPLE
14VDC
0.2V/DIV
2µs/DIV
Figure 2. Buck regulator circuit waveforms
battery/wall adapter input applications, operating above 95% duty cycle
causes problems for the current sense
transformer. To avoid transformer
saturation, the Q6 stage limits duty
cycle to approximately 90%. Current
through R16 tries to charge C20 to
the 3V base voltage of Q6. If the
switch cycle terminates at less than a
90% duty cycle, C20 is reset by D8. If
the duty cycle exceeds 90%, C20
charges until Q6 turns on, ending the
switch cycle.
Switch voltage, inductor current,
T1 primary current, and output voltage ripple waveforms are shown in
Figure 2. These waveforms were measured with a 24V input, 14V output,
and 15A load current. When MOS-
FET Q3 turns on, the switch voltage
(Trace A) goes high, the inductor current (Trace B) increases, as does the
T1 primary current (Trace C) and the
output ripple voltage (Trace D). When
Q3 turns off, the switch voltage goes
low, inductor current decreases as its
stored energy supplies load current
through D4, T1 primary current goes
to zero and the output voltage
decreases slightly.
In addition to its role as a microprocessor voltage controller, the
LTC1435 and current sense transformer can be very useful in higher
voltage and higher current application where high efficiency and
ease-of-design are important.
5V Boost Regulator, continued from page 33
on the SHDN pin ensures proper regulation with load currents as high as
100µA. When load current greater
than 100µA is needed, the SHDN pin
must be forced low, as in normal
operation. The typical no-load supply
current for this circuit with VIN = 3V is
only 3.2µA.
ICC (µA)
4.0
2.0
0.0
2.0
1000
SHDN ON PULSE WIDTH = 200µs
COUT = 10µF
MAX SHDN OFF TIME (ms)
6.0
100
10
1
3.0
4.0
5.0
1
VIN (V)
Figure 3. No-load ICC versus input voltage for
Figure 2’s circuit
Linear Technology Magazine • August 1996
10
100
1000
IOUT (µA)
Authors can be contacted
at (408) 432-1900
Figure 4. Maximum SHDN OFF time versus
output load current for ultralow IQ operation
35
DESIGN IDEAS
The LT1210: High Power Op Amp
Yields Higher Voltage and Current
by Dale Eagar
Introduction
30
The LT1210, a 1 amp current feedback operational amplifier, opens up
new frontiers. With 30MHz bandwidth, operation on ±15V supplies,
thermal shutdown and 1 amp of output current, this amplifier single
handedly tackles many tough applications. But can it handle output
voltages higher than ±15V or currents greater than 1 ampere? This
Design Idea features a collection of
circuits that open the door to high
voltage and high current for the
LT1210.
25
20
50Ω LOAD
GAIN (dB)
15
10
FIGURE 1 CKT
+10dBM INPUT
50Ω LOAD
5
0
–5
–10
–15
–20
10K
Fast and Sassy—
Telescoping Amplifiers
100K
1M
FREQUENCY (Hz)
input signal. This telescoping arrangement can be cascaded with additional
stages to get more than ±30V. This
amplifier is stable into capacitive
loads, is short-circuit protected
and thermally shuts down when
overheated.
1k
30V
TIP 29
15V
1kΩ
1µF
1k
15V
1
1µF
300Ω
1
LT1210
60Ω
2
+
6
3
1k
4
–
4
–
7
LT1210
7
2
5
5
+
6
3
0.01
0.01
6.2k
1µF
6.2k
–15V
60M
Figure 2. Gain versus frequency plot of telescoping amplifier
Need ±30V? Cascading LT1210’s will
get you there. This circuit (Figure 1)
will provide the ±30V at ±1A and has
13MHz of full-power bandwidth (see
Figure 2). How does it work? The first
LT1210 drives the “ground” of the
second LT1210 subcircuit, effectively
raising and lowering it while the second LT1210 further amplifies the
INPUT
10M
15V
1µF
TIP 30
1k
–30V
OUTPUT
Extending
Power Supply Voltages
Another method of getting high voltage
from an amplifier is the extendedsupply mode (see “Extending Op Amp
Supplies to Get More Voltage”; Linear
Technology Volume IV Number 2
(June 1994), pp. 20–22). This involves
steering two external regulators with
the power supply pins of an op amp
to get a high voltage amplifier.
Figure 3 shows the LT1210 connected in the extended-supply mode.
Placing an amplifier in the extendedsupply mode requires changing the
return of the compensation node from
the power supply pins to system
ground. R9 and C5 are selected for
clean step response. The process of
relocating the return of the compensation node slows the amplifier down
to approximately 1MHz (see Figure
4). Figure 3’s circuit will provide ±1A
at ±100V, is stable into capacitive
loads and is short-circuit protected.
The two external MOSFETs need heat
sinking.
Figure 1. Telescoping amplifiers
36
Linear Technology Magazine • August 1996
DESIGN IDEAS
100V
100V
Boosting Both
Current and Voltage
100k
100Ω
R8
300Ω
0.01
R7
10k
INPUT
IRF640
1k
1
P6KE
15A
4
–
C5
8pF
R9
9.1k
+
5
6
0.01
3
15k
AV = −
(
R8 R9 += R10
)
220Ω
7
LT1210
2
15V
LOAD
P6KE
15A
15V
0.01
R10 300Ω
R8 R9 − R7 R10
100Ω
IRF9640
100k
–100V
Figure 3. ±100V, ±1A power driver
–100V
Gateway to the Stars
Boosting Output Current
The circuit of Figure 3 can be expanded
to yield much higher voltages; the
first and most obvious way is to use
higher voltage MOSFETs. This causes
two problems: first, high voltage Pchannel MOSFETs are hard to get;
second, and more importantly, at ±1A
the power dissipated by the MOSFETs is too high for single packages.
The solution is to build telescoping
regulators, as shown in Figure 5. This
circuit can provide ±1A of current at
±200V and has the additional powerdissipation ability of four MOSFETs.
The current booster detailed in Figure 6 illustrates a technique for
amplifying the output current capability of an op amp while maintaining
speed. Among the many niceties of
this topology is the fact that both Q1
and Q2 are normally off and thus
consume no quiescent current. Once
the load current reaches approximately 100mA, Q1 or Q2 turns on,
providing additional drive to the output. This transition is seamless to the
outside world and takes advantage of
the full speed of Q1 and Q2. This
circuit’s small-signal bandwidth and
full-power bandwidth are shown in
Figure 7.
The current-boosted amplifier shown
in Figure 6 can be used to replace the
amplifiers in Figure 1, yielding ±10A
at ±30V. Placing the boosted amplifier
in the circuits shown in Figures 3 or
5 will yield peak powers into the
kilowatts.
Thermal Management
When the LT1210 is used with external transistors to increase its output
voltage and/or current range an
additional benefit can often be realized: system thermal shutdown.
Careful analysis of the thermal design of the system can coordinate the
overtemperature shutdown of the
LT1210 with the junction temperatures of the external transistors. This
essentially extends the umbrella of
protection of the LT1210’s thermal
continued on page 39
200V
0.47µF
250V
10k
1W
IRF640
15V
10k
1W
IRF640
0.1
15V
220Ω
300Ω
10k
5W
INPUT
1k
1
30
25
7
LT1210
2
90Vp-p INTO 50Ω
R9
9.1k
+
C5
8pF
±1A
±200V
4
–
5
LOAD
6
3
0.01
20
300Ω
GAIN (dB)
15
15V
10
0.1
100Ω
IRF9640
5
10k
1W
FIGURE 3 CKT
15V
0
–5
IRF9640
–10
10k
1W
–15
–20
10K
0.47µF
250V
100K
1M
FREQUENCY (Hz)
10M
Figure 4. Gain versus frequency plot of extended-supply amplifier
Linear Technology Magazine • August 1996
60M
–200V
Figure 5. Cascode power amplifier
37
DESIGN IDEAS
LTC1441-Based Micropower
Voltage-to-Frequency Converter
by Jim Williams
Figure 1 is a voltage-to-frequency
converter. A 0V–5V input produces a
0–10kHz output, with a linearity of
0.02%. Gain drift is 60ppm/°C. Maximum current consumption is only
26µA, 100 times lower than currently
available units.
To understand the circuit’s operation, assume that C1’s negative input
is slightly below its positive input
(C2’s output is low). The input voltage
causes a positive-going ramp at C1’s
input (trace A, Figure 2). C1’s output
is high, allowing current flow from
Q1’s emitter, through C1’s output
stage to the 100pF capacitor. The
2.2µF capacitor provides high frequency bypass, maintaining low
impedance at Q1’s emitter. Diode connected Q6 provides a path to ground.
The voltage to which the 100pF unit
charges is a function of Q1’s emitter
potential and Q6’s drop. C1’s CMOS
output, purely ohmic, contributes no
voltage error. When the ramp at C1’s
negative input goes high enough, C1’s
output goes low (trace B) and the
inverter switches high (trace C). This
action pulls current from C1’s negative input capacitor via the Q5 route
(trace D). This current removal resets
C1’s negative input ramp to a potential slightly below ground. The 50pF
capacitor furnishes AC positive feedback (C1’s positive input is trace E)
ensuring that C1’s output remains
negative long enough for a complete
discharge of the 100pF capacitor. The
Schottky diode prevents C1’s input
from being driven outside its negative
common mode limit. When the 50pF
unit’s feedback decays, C1 again
switches high and the entire cycle
repeats. The oscillation frequency
depends directly on the input-voltage-derived current.
Q1’s emitter voltage must be carefully controlled to get low drift. Q3
and Q4 temperature compensate Q5
and Q6 while Q2 compensates Q1’s
VBE. The three LT1004s are the actual
voltage reference and the LM334 current source provides 12µA bias to the
stack. The current drive provides
excellent supply immunity (better
than 40ppm/V) and also aids circuit
temperature coefficient. It does this
by using the LM334’s 0.3%/°C tempco
to slightly temperature modulate the
voltage drop in the Q2–Q4 trio. This
correction’s sign and magnitude
directly oppose the –120ppm/°C
100pF polystyrene capacitor’s drift,
+V = 6.2 → 12V
LM334
INPUT
0–5V
10kHz
TRIM
1.2M* 200k
6.04k*
+
2.2µF
Q8
Q1
–
C1
1/2 LTC1441
0.01
LT1004
1.2V
x3
+
+
0.47
50pF
Q2
100k
Q3
100Hz TRIM
3M TYP
15k
Q5
Q4
Q7
†
100pF
74C14
OUTPUT
= HP5082-2810
Q6
10M
2.7M
= 1N4148
0.1
Q1, Q2, Q8 = 2N5089
ALL OTHER = 2N2222
†
= POLYSTYRENE
* = 1% METAL FILM
GROUND ALL UNUSED 74C14 INPUTS
–
C2
1/2 LTC1441
+
Figure 1. 0.02% V/F converter requires only 26µA supply current
38
DIVF_01.eps
Linear Technology Magazine • August 1996
DESIGN IDEAS
A = 50mV/DIV
B = 5V/DIV
C = 5V/DIV
D = 1mA/DIV
E = 5V/DIV
HORIZ = 20µs/DIV
Figure 2. Waveforms for the micropower V/F
converter: charge-based feedback provides
precision operation with extremely low power
consumption.
racy permits, draws only small transient currents during its charge and
discharge cycles. The 50pF–100k
positive feedback combination draws
insignificantly small switching currents. Figure 3, a plot of supply
current versus operating frequency,
reflects the low power design. At zero
frequency, comparator quiescent current and the 12µA reference stack
bias account for all current drain.
There are no other paths for loss. As
frequency scales up, the 100pF
capacitor’s charge-discharge cycle
introduces the 1.1µA/kHz increase
shown. A smaller value capacitor
would cut power, but effects of stray
capacitance and charge imbalance
would introduce accuracy errors.
Circuit start-up or overdrive can
cause the circuit’s AC-coupled feedback to latch. If this occurs, C1’s
output goes low; C2, detecting this
via the 2.7M–0.1µF lag, goes high.
This lifts C1’s positive input and
LT1210, continued from page 37
Summary
shutdown to cover the external transistors. The thermal shutdown of the
LT1210 activates when the junction
temperature reaches 150˚C and has
about 10˚C hysteresis. The thermal
resistance RθJC of the TO-220 package (LT1210CY) is 5˚C/Watt).
The LT1210 is a great part; its performance in terms of speed, output
current and output voltage is unsurpassed. Its C-Load™ output drive
and thermal shutdown allow it to
take its place in the real world—no
kid gloves are required here. If the
generous output specification of the
grounds the negative input with Q7,
initiating normal circuit action.
To calibrate this circuit, apply
50mV and select the indicated resistor at C1’s positive input for a 100Hz
output. Complete the calibration by
applying 5V and trimming the input
potentiometer for a 10kHz output.
35
CURRENT CONSUMPTION (µA)
aiding overall circuit stability. Q8’s
isolated 100pF drive to the CMOS
inverter prevents output loading from
influencing Q1’s operating point. This
makes circuit accuracy independent
of loading.
The Q1 emitter-follower delivers
charge to the 100pF capacitor efficiently. Both base and collector
current end up in the capacitor. The
100pF capacitor, as small as accu-
30
25
20
SLOPE = 1.1µA/kHz
15
10
5
0
0
1
2
3
4 5 6 7 8 9 10 11 12
FREQUENCY (kHz)
DIVF_03.eps
Figure 3. Current consumption versus
frequency for the V/F converter: charge/
discharge cycles account for 1.1µA/kHz
current drain increase.
LT1210 isn’t big enough for your
needs, just add a couple of transistors to dissipate the additional power
and you are on your way. Only the
worldwide supply of transistors limits the amount of power you could
command with one of these parts.
18V
8
0.01
R1
6.2Ω
4AP–P INTO 1Ω
0.033Ω
7
6
Q1, D45VH4
50Ω LOAD
5
3.6k
1
4
4
–
7
LT1210
2
OUT
5
+
GAIN (dB)
1.8k
IN
FIGURE 6 CKT
+10dBM INPUT
3
2
6
3
1
0.01
1Ω LOAD
Q2, D44VH4
0
0.01
R2
6.2Ω
0.033Ω
–1
–2
10K
–18V
Figure 6. ±10A/1MHz current-boosted power
op amp
Linear Technology Magazine • August 1996
100K
1M
10M
FREQUENCY (Hz)
Figure 7. Gain versus frequency response of current-boosted amplifier
39
DESIGN IDEAS
Capacitive Charge Pump Powers
12V VPP from 5V Source
by Mitchell Lee
+
10µF
CIN
C1+
C1
470nF
VCC
SHDN
OFF
ON
C1–
LTC1263
C2+
C2
470nF
C2–
12V AT 60mA
VOUT
+
10µF
COUT
VPP
FLASH
MEMORY
For further information on any
of the devices mentioned in this
issue of Linear Technology, use
the reader service card or call
the LTC literature service
number:
–7V OUTPUT (V)
VCC = 5V
12V LOAD = 3mA
7
6
1
10
–7V LOAD (mA)
Figure 3. Cross regulation with a
constant 12V load
40
100
4.75V–5.5V
10µF
C1+
VCC
SHDN
470nF
C1–
LTC1263
C2+
470nF
C2–
12V OUTPUT
VOUT
+
10µF
1µF
MBR0520L
MBR0520L
+
–7V OUTPUT
10µF
Figure 2. Split supply generator: cross
regulation is improved by driving the
inverting charge pump from C2+.
Figure 1. Programming two flash chips with
the LTC1263 charge pump. In shutdown
mode, the output is held at 5V.
8
Schemes like this one often suffer
from poor cross regulation. Although
the inverting output is not directly
regulated, the –7V load does affect
the 12V output, thereby improving
cross regulation (see Figure 3). The
regulation with a common load (such
as op amps) is shown in Figure 4.
16
VCC = 5V
12
+12V
OUTPUT (V)
4.75V-5.5V
C1 and C2 in parallel across 5V and
ground and then discharging them in
series across 5V and the output. In
theory, the output could reach 15V,
but an internal regulation loop maintains the output at a constant 12V.
SHUTDOWN reduces the quiescent
current of the LTC1263 to less than
1µA under logic control. In shutdown
mode, the output is held at 5V by an
internal 500Ω, VCC-to-VOUT switch.
Output-voltage fall time is guaranteed to be less than 15ms for the
component values shown. Output rise
time coming out of shutdown is guaranteed to be less than 800µs.
Designing a circuit to generate a
split supply from a single 5V source is
usually an unpleasant chore; one to
be avoided at all costs. If load current requirements are modest, the
LTC1263 can generate both 12V and
–7V for op amps and biasing needs.
Figure 2 shows how. The LTC1263 is
connected in the usual way to produce a regulated, 12V output, but a 2
diode, 2-capacitor charge pump is
added to the C2+ pin. This pin switches
between VCC and VOUT, swinging
approximately 7VP–P. The result is an
outboard charge pump inverter with
a –7V output.
+
The LTC1263, a regulating charge
pump tripler, converts a 5V input to a
regulated 12V, 60mA output. No
inductors are required; charge pumps
operate with capacitors only. Figure
1 shows the LTC1263 configured to
provide VPP for two flash memory
chips. The “flying” capacitors in the
charge pump, C1 and C2, are sized
well within the surface mount ceramic range. CIN and COUT, as shown,
are surface mount tantalum capacitors, such as Sprague 595D series. In
the 10µF capacitance range, tantalum capacitors cost less than ceramic
units. The chip operates by charging
8
–7V
1-800-4-LINEAR
4
Ask for the pertinent data sheets
and Application Notes.
0
0
20
40
60
COMMON LOAD CURRENT (mA)
80
Figure 4. Output regulation with a
common load
Linear Technology Magazine • August 1996
DESIGN IDEAS
Bridge Measures Small Capacitance
in Presence of Large Strays
by Jeff Witt
Capacitance sensors measure a
wide variety of physical quantities,
such as position, acceleration, pressure and fluid level. The capacitance
changes are often much smaller than
stray capacitances, especially if the
sensor is remotely placed. I needed to
make measurements with a 50pF
cryogenic fluid level detector, with
only 2pF full-scale change, hooked to
several hundred pF of varying cable
capacitance. This required a circuit
with high stability, sensitivity and
noise rejection, but one insensitive to
stray capacitance caused by cables
and shielding. I also wanted battery
operation and analog output for easy
interfacing to other instruments. Two
traditional circuit types have drawbacks: integrators are sensitive to
noise at the comparator and voltageto-frequency converters typically
measure stray as well as sensor
capacitance. The capacitance bridge
presented here measures small transducer capacitance changes, yet rejects
noise and cable capacitance.
The bridge, shown in Figure 1, is
designed around the LTC1043
switched-capacitor building block.
The circuit compares a capacitor, CX,
of unknown value, with a reference
VOUT = VREF CX
CREF
Balance is achieved by integrating
the current from node C using an op
amp (LT1413) and a third switch on
the LTC1043 for synchronous detection. With CREF = 500pF and VREF =
2.5V, this circuit has a gain of 5mV/
pF, and when measured with a DMM
achieves a resolution of 10fF for a
dynamic range of 100dB. It also rejects stray capacitance (shown as
ghosts in Figure 1) by 100dB. If this
rejection is not important, the switching frequency f can be increased to
extend the circuit’s bandwidth, which
is
BW = f
5V
VREF
5
+
CREF
COUT
V+
COSC
V–
1/2
LT1413
7
16
C1
0.01µF
17
COUT
2.2nF
7
11
–
A
1/4 LTC1043
8
CX
1µF
With CREF = 50pF, the circuit has a
gain of 5V/pF and can resolve 2fF.
Supply current is 1mA. The synchronous detection makes this circuit
insensitive to external noise sources
and in this respect shielding is not
terribly important. However, to achieve
high resolution and stability, care
should be taken to shield the capacitors being measured. I used this circuit
for the fluid level detector mentioned
above, putting a small trim cap in
parallel with CREF to adjust offset and
trimming R2 for proper gain.
1/2 LTC1043
4
6
VOUT – VREF ≈ VREF ∆CX (R1 + R2)
CREF R2
COUT should be larger than CREF.
The circuit operates from a single
5V supply and consumes 800µA. If
5V
100k
the capacitances at nodes A and C are
kept below 500pF, the LT1078
micropower dual op amp may be used
in place of the LT1413, reducing supply current to just 160µA.
If the relative capacitance change
is small, the circuit can be modified
for higher resolution, as shown in
Figure 2. A JFET input op amp
(LT1462) amplifies the signal before
demodulation for good noise performance, and the output of the
integrator is attenuated by R1 and R2
to increase the sensitivity of the circuit. If ∆CX << CX, and CREF ≈ CX, then
capacitor, CREF. The LTC1043, programmed with C1 to switch at 500Hz,
applies a square wave of amplitude
VREF to node A, and a square wave of
amplitude VOUT and opposite phase
to node B. When the bridge is balanced, the AC voltage at node C is
zero, and
5V
5
LT1004-2.5
C
2
6
13
3
CREF
12
14
2
+
8
1/2
LT1413
–
1
VOUT
4
B
VREF
NOTE: SHADED PARTS REPRESENT
PARASITIC CAPACITANCES
Figure 1. A simple, high performance capacitance bridge
Linear Technology Magazine • August 1996
41
CONTINUATIONS
VIN
5V
100k
VREF
1/2 LTC1043
4
6
5
V+
COSC
16
3
+
V–
1/2
LT1413
7
17
1/2
LT1462
10M 2
5V
1/4 LTC1043
VREF
11
–
10k
10k
1
–
7
6
10k
8
CX
1µF
+
0.01µF
5
2
100Ω
LT1004-2.5
–
VREF
5
+
8
1/2
LT1462
7
4
6
VREF
1µF
CREF
13
12
5V
14
R1
10.0Ω 1%
8
1
–
1/2
LT1413
+
R2
10.0k 1%
4
2
3
100K
VREF
VOUT
Figure 2. A bridge with increased sensitivity and noise performance
Bridge circuits are particularly
suitable for differential measurements. When CX and CREF are replaced
with two sensing capacitors, these
circuits measure differential capacitance changes, but reject common
mode changes. CMRR for the circuit
in Figure 2 exceeds 70dB. In this
case, however, the output is linear
only for small relative capacitance
changes.
LTC1433/44, continued from page 14
Power-On Reset
Monitor Included
An internal regulation monitor provided in the LTC1433 and LTC1434
continuously monitors the output.
When the device is out of regulation
or in shutdown mode, the POR open
drain output pulls low. At start-up,
once the output voltage has reached
5% of its final value, an internal timer
is started, after which the POR pin is
released. The timer counts 216 oscillator cycles, yielding a delay-to-release
reset of approximately 300ms in a
typical application. Once the output
is in regulation, it has to fall by 7.5%
before the POR pin is asserted.
42
100
90
EFFICIENCY (%)
the internal sense resistance. Figures 6 and 7 show the dropout
characteristic at different load
currents.
VIN = 7V
80
70
VIN = 3.5V
60
50
40
0.001
VOUT = –5.0V
COSC = 100pF
0.01
0.10
1.00
LOAD CURRENT (A)
Figure 9. Efficiency curve for Figure 8’s
positive-to-negative converter
Typical Application: Positiveto-Negative Converter
Both the LTC1433 and LTC1434 can
easily be set up for a negative output
voltage. Figure 8 shows the schematic using the LTC1433. The
efficiency curve is shown in Figure 9.
This circuit is set up so that the
output is taken from the device
ground. Components that are referenced back to the device
ground, Run/SS capacitor, oscillator
frequency capacitor and ITH compensation network, are connected to
the output instead of to the circuit
ground.
Conclusion
The LTC1433 and LTC1434, with their
low dropout, high efficiency and small
footprint, are ideal for battery-operated portable equipment applications.
In addition, their constant frequency
operation makes the devices suited
for applications that require the
switching regulator to operate at a
defined frequency.
Linear Technology Magazine • August 1996
DESIGN TOOLS
DESIGN TOOLS
Applications on Disk
Noise Disk — This IBM-PC (or compatible) program allows the user to calculate circuit noise using
LTC op amps, determine the best LTC op amp for a
low noise application, display the noise data for LTC
op amps, calculate resistor noise and calculate
noise using specs for any op amp.
Available at no charge.
SPICE Macromodel Disk — This IBM-PC (or compatible) high density diskette contains the library of
LTC op amp SPICE macromodels. The models can
be used with any version of SPICE for general
analog circuit simulations. The diskette also contains working circuit examples using the models
and a demonstration copy of PSPICE™ by MicroSim.
Available at no charge.
SwitcherCAD — SwitcherCAD is a powerful PC
software tool that 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. 144
page manual included.
$20.00
SwitcherCAD supports the following parts: LT1070
series: LT1070, LT1071, LT1072, LT1074 and
LT1076. LT1082. LT1170 series: LT1170, LT1171,
LT1172 and LT1176. It also supports: LT1268,
LT1269 and LT1507. LT1270 series: LT1270 and
LT1271. LT1371 series: LT1371, LT1372, LT1373,
LT1375, LT1376 and LT1377.
Micropower SwitcherCAD — MicropowerSCAD is
a powerful tool for designing DC/DC converters
based on Linear Technology’s micropower switching regulator ICs. Given basic design parameters,
MicropowerSCAD selects a circuit topology and
offers you a selection of appropriate Linear Technology switching regulator ICs. MicropowerSCAD
also performs circuit simulations to select the other
components which surround the DC/DC converter.
In the case of a battery supply, MicropowerSCAD
can perform a battery life simulation. 44 page manual
included.
$20.00
MicropowerSCAD supports the following LTC
micropower DC/DC converters: LT1073, LT1107,
LT1108, LT1109, LT1109A, LT1110, LT1111,
LT1173, LTC1174, LT1300, LT1301 and LT1303.
Technical Books
1990 Linear Databook, Vol 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, Vol II — This 1248 page
supplement to the 1990 Linear Databook is a collection of all products introduced in 1991 and 1992.
Linear Technology Magazine • August 1996
The catalog contains full data sheets for over 140
devices. The 1992 Linear Databook, Vol II is a
companion to the 1990 Linear Databook, which
should not be discarded.
$10.00
1994 Linear Databook, Vol III —This 1826 page
supplement to the 1990 and 1992 Linear Databooks
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
Vol III is a companion to the 1990 and 1992 Linear
Databooks, which should not be discarded. $10.00
1995 Linear Databook, Vol IV —This 1152 page
supplement to the 1990, 1992 and 1994 Linear
Databooks is a collection of all products introduced
since 1994. A total of 80 product data sheets are
included with updated selection guides. The 1995
Linear Databook Vol IV is a companion to the 1990,
1992 and 1994Linear Databooks, which should not
be discarded.
$10.00
1990 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 40 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
Power Solutions Brochure — This 80 page collection of circuits contains real-life solutions for
common power supply design problems. There are
over 79 circuits, including descriptions, graphs and
performance specifications. Topics covered include
battery chargers, 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.
High Speed Amplifier Solutions Brochure —
This 72 page collection of circuits contains real-life
solutions for problems that require high speed
amplifiers. There are 82 circuits including descriptions, graphs and performance specifications. Topics
covered include basic amplifiers, video-related applications circuits, instrumentation, DAC and
photodiode amplifiers, filters, variable gain, oscillators and current sources and other unusual
application circuits.
Available at no charge
Data Conversion Solutions Brochure — This 52
page collection of data conversion circuits, products and selection guides serves as excellent
reference for the data acquisition system designer.
Over 60 products are showcased, solving problems
in low power, small size and high performance data
conversion applications—with performance graphs
and specifications. Topics covered include ADCs,
DACs, voltage references and analog multiplexers.
A complete glossary defines data conversion specifications; a list of selected Application and Design
Notes is also included.
Available at no charge
CD-ROM
LinearView — LinearView™ is Linear Technology’s
interactive PC-based CD-ROM. LinearView allows
you to instantly access thousands of pages of
product and applications information, covering
Linear Technology’s complete line of high performance analog products, with easy-to-use search
tools.
The LinearView CD-ROM includes the complete
product specifications from Linear Technology’s
Databook library (Volumes I–IV) and the complete
Applications Handbook collection (Volumes I and
II). Our extensive collection of Design Notes and the
complete collection of Linear Technology magazine
are also included.
A powerful search engine built into the LinearView
CD-ROM enables you to select parts by various
criteria, such as device parameters, keywords or
part numbers. All product categories are represented: data conversion, references, amplifiers,
power products, filters and interface circuits. Upto-date versions of Linear Technology’s software
design tools, SwitcherCAD, FilterCAD, Noise Disk
and Spice Macromodel library, are also included.
Everything you need to know about Linear
Technology’s products and applications is readily
accessible via LinearView.
Available at no charge.
AppleTalk is a registered trademark of Apple Computer,
Inc.
Information furnished by Linear Technology Corporation
is believed to be accurate and reliable. However, Linear
Technology makes no representation that the circuits
described herein will not infringe on existing patent rights.
43
World Headquarters
Linear Technology Corporation
1630 McCarthy Boulevard
Milpitas, CA 95035-7417
Phone: (408) 432-1900
FAX: (408) 434-0507
U.S. Area
Sales Offices
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
1900 McCarthy Blvd., Suite 205
Milpitas, CA 95035
Phone: (408) 428-2050
FAX: (408) 432-6331
SOUTHEAST REGION
Linear Technology Corporation
17000 Dallas Parkway
Suite 219
Dallas, TX 75248
Phone: (214) 733-3071
FAX: (214) 380-5138
Linear Technology Corporation
5510 Six Forks Road
Suite 102
Raleigh, NC 27609
Phone: (919) 870-5106
FAX: (919) 870-8831
CENTRAL REGION
Linear Technology Corporation
Chesapeake Square
229 Mitchell Court, Suite A-25
Addison, IL 60101
Phone: (708) 620-6910
FAX: (708) 620-6977
SOUTHWEST REGION
Linear Technology Corporation
21243 Ventura Blvd.
Suite 227
Woodland Hills, CA 91364
Phone: (818) 703-0835
FAX: (818) 703-0517
Linear Technology Corporation
15375 Barranca Parkway
Suite A-211
Irvine, CA 92718
Phone: (714) 453-4650
FAX: (714) 453-4765
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 Technology GmbH
Oskar-Messter-Str. 24
85737 Ismaning
Germany
Phone: 49-89-962455-0
FAX: 49-89-963147
JAPAN
Linear Technology KK
5F NAO Bldg.
1-14 Shin-Ogawa-cho Shinjuku-ku
Tokyo, 162 Japan
Phone: 81-3-3267-7891
FAX: 81-3-3267-8510
KOREA
Linear Technology Korea Co., Ltd
Namsong Building, #403
Itaewon-Dong 260-199
Yongsan-Ku, Seoul 140-200
Korea
Phone: 82-2-792-1617
FAX: 82-2-792-1619
SINGAPORE
Linear Technology Pte. Ltd.
507 Yishun Industrial Park A
Singapore 2776
Phone: 65-753-2692
FAX: 65-754-4113
SWEDEN
Linear Technology AB
Sollentunavägen 63
S-191 40 Sollentuna
Sweden
Phone: 46-8-623-1600
FAX: 46-8-623-1650
TAIWAN
Linear Technology Corporation
Rm. 602, No. 46, Sec. 2
Chung Shan N. Rd.
Taipei, Taiwan, R.O.C.
Phone: 886-2-521-7575
FAX: 886-2-562-2285
UNITED KINGDOM
Linear Technology (UK) Ltd.
The Coliseum, Riverside Way
Camberley, Surrey GU15 3YL
United Kingdom
Phone: 44-1276-677676
FAX: 44-1276-64851
LINEAR TECHNOLOGY CORPORATION
1630 McCarthy Boulevard
Milpitas, CA 95035-7417
(408) 432-1900 FAX (408) 434-0507
For Literature Only: 1-800-4-LINEAR
44
© 1996 Linear Technology Corporation/ Printed in U.S.A./34K
Linear Technology Magazine • August 1996