| |
|
April 23, 1998
Nelson Nguyen, Anritsu Corp, Morgan Hill, CA
A project using Xilinx FPGAs brought
an interesting problem to light. When you turn on the board, one FPGA in three succumbs to
this problem. A lot of frustration and testing uncovered a negative-going spike (Figure 1) in the 5V line from the dc/dc converter. The system uses a dc/dc converter to convert 48V to 5V and
other voltages. The spike occurs before the converter delivers its intended 5V. Spikes
greater than 5V would kill the FPGA with the shortest path to the converter. The circuit
in Figure 2 solves the problem.
Because the spike occurs before the 5V supply line turns on, to prevent
the spike from destroying the FPGA, you should open the 5V path when you turn on the power
switch and then close the path when the 5V supply voltage is present. The R19-C13
RC network provides a delay in turning relay K1 on. The turn-on voltage for K1
is approximately 3.7V. The voltage divider comprising R19 and K1's
coil resistance (approximately 780 ohms for an NEC EA2-5NU) provides a voltage at the
junction of C13 and R19 sufficient to turn K1 on. The
value of C13 sets the delay at approximately 2 msec. (DI #2181)
Tom Gross, Linear Technology, Milpitas, CA
Many circuits, such as those that use
batteries or solar cells, must operate in the face of decreasing supply voltages. The
circuit in Figure 1 maintains the maximum load current as
the supply voltage drops. The regulator boosts a 2.5 to 4.2V input to 5V and provides 2A
load current, for 10W of output power. The circuit is a bootstrapped synchronous boost
regulator that uses an LTC1266 synchronous-regulator controller. Diodes D1
through D5 allow the circuit to start up using the low input voltage and then
to receive its power from the higher output voltage during normal operation. The crucial
elements in the circuit are the switches: two IRF7401 n-channel MOSFETs. The MOSFETs
receive full enhancement at low gate-source voltages. (At VGS=2V, the peak
drain current is 15A.) The low enhancement voltages allow the circuit to start with low
input voltages.
This low-voltage capability is
important for low-series-cell-count, battery-powered systems. Diodes D3 and D4,
along with capacitor C2, form a charge-pump circuit, which the controller uses
for the MOSFETs' gate drive. Because the circuit receives its power from the 5V output
voltage, the circuit still operates if the input supply voltage drops below the IC's
minimum input voltage. This bootstrapping allows the circuit to start up when the input
voltage is below the IC's 3.5V minimum input spec. With a 1A load, the regulator operates
with inputs as low as 1.8V. Figure 2 shows the regulator's
efficiency vs the input voltage with three load currents. With 2A load current, the
efficiency drops as the input voltage decreases, because of the higher power losses in the
inductor. A larger inductor would provide in-creased efficiency or allow for greater load
currents. (DI #2185)
Nicholas Gray and Terrance Smith, National Semiconductor, Santa Clara, CA
Many video-circuit clamps operate
well in the presence of a composite-video signal but cannot achieve a clamp level with
signals other than composite video or in the absence of an input signal. The circuit in Figure 1, developed for the ADC1175 (a popular and inexpensive,
high-performance, 8-bit, 20M-sample/sec ADC), provides the normal back-porch clamp
function to the input of the ADC in the presence of a composite-video signal. The circuit
further ensures that the voltage presented to the ADC is within its correct operating
range in the absence of an input signal and forces any signal other than composite video
to be within the ADC's input common-mode range.
The circuit accomplishes video clamping by building a control loop that
forces the dc voltage at IC2's output to a desired level during the blanking
period. This level, approximately 25% of full scale for a composite-video signal, forces
the ADC's output-pedestal (blanking) level to an 8-bit code of approximately 64. The
simple filter comprising R3 and C3 bandlimits the signal at the
output of IC2. This high-frequency attenuation is necessary to prevent noise
spikes from upsetting the operation of the LM1881. The LM1881 is a video sync-separator
chip that produces burst-gate pulses at its Pin 5 when a composite-video signal is present
at Pin 2.
The burst-gate output of the LM1881 serves to sample the blanking level of
the video signal. Potentiometer VR1 and R5 produce an adjustable
offset in the signal path when Q1 gates on. During the blanking period, the
ac-coupled burst-gate signal pulls Q2's base low (to approximately 4V), thus
pulling Q1's gate high, thereby sampling and storing the sum of the
video-blanking level and the dc offset from VR1 onto C7. At times
other than the back-porch interval, Pin 5 of the LM1881 is high, and Q2 is off,
thereby turning Q1 off. Divider R14-R15 attenuates the
voltage on C7 to ensure sufficient phase margin in the clamp loop. IC4
is an integrator that averages the attenuated dc value over many samples. This average
sums with the input signal in IC1.
If the integration time is too small,
the result could be shading across the display. A long integration results in slow,
perceptible adjustments when switching between fields with large differences in average
brightness. The dc feedback path for IC4 is through IC1 and IC2.
If no video signal exists or if the input signal has no sync, R11 holds Q2
on, thus holding the video output of the circuit within the ADC's operating range. With VR1
centered, the level halfway between the positive and negative peaks of the input signal
clamps at approximately 1.6V, or approximately halfway between the high and low reference
voltages (2.6 and 0.6V, respectively) of the ADC1175. The circuit achieves an effective
number of bits of 7.5, corresponding to a signal-to-noise and distortion of 47 dB. Figure 2 shows the offset at Point B in Figure
1, relative to the voltage at Point A. (DI #2184)
Hugh Adams, Fort Walton Beach, FL
Have you ever wondered which way a
blower motor is going to turn when you plug it into another socket, or have you ever
inherited the task of modifying three-phase wiring in your plant? The circuit in Figure 1 is a simple, approximately $5 phase sequencer that you
can probably build from parts in your junk box
and save approximately $50 to boot. The component values reflect 60-Hz operation, but the
design equations in Figures 2 and 3 allow you to
select values for other frequencies. The equations are in MathCAD spreadsheet format, but almost any other spreadsheet would do.
Referring to Figure 1 and the equations, you
can see that the neon bulb that glows brighter indicates the phase sequence, or
phase-rotation order, ABC or CBA. The bulb glows brighter because it carries more current
because of the phase shift the 1.5-µF capacitor provides. You can verify this assertion
by examining the two sets of equations. Note that the two sets of equations have different
expressions for IB and IC. In one, IB lags IA
by 2pi/3; in the other, it lags by
4pi/3, and vice versa for IC.
The equations provide the mathematical way of reversing the phase sequence, and, as you
can see, the two currents IB and IC reverse their relative
magnitudes as the phase rotation reverses. (DI #2180)
Piezo device generates buzz, beep, or chime
Dennis Eichenberg, Parma Heights, OH
Piezoelectric buzzers, such as the
Murata (Smyrna, GA) PKB5-3A in Figure 1, make excellent
alarms. They're compact, lightweight, efficient, and reliable. However, a piezo alarm is a
dc device; it requires additional circuitry to operate from an ac source. The circuits in Figure 1 provide a simple and inexpensive way to obtain the dc
drive. The W04G full-wave bridge rectifier produces a full-wave dc waveform from the 120V
ac line. The 100 ohm resistor protects the circuit from surges when you first apply power.
The 5.5V 1N4733 zener diode protects the buzzer against high-voltage excursions. The 1-µF
capacitor provides filtering for the buzzer.
The circuit in Figure 1a produces a true
buzzer sound. The addition of an F336HD flashing LED (part number 276-036 at Radio Shack)
in Figure 1b changes the alarm to a beeper, and it also
provides a visual alarm. The LED produces a constant pulse of light at approximately 1 Hz
without the addition of a time-constant capacitor. The LED starts immediately when you
apply power, and it's insensitive to temperature variations. The addition of a 35-µF
capacitor in parallel with the buzzer (Figure 1c) changes
the audible alarm to a pleasing chime. The value of the capacitor is not critical; you can
obtain various sound effects by varying it. (DI #2194)
Robert Lindsey, Hansvedt EDM, Urbana, IL
Transformer-core saturation can cause
inexplicable fuse blowing, system crashes, or premature switch and relay failure. When a
core saturates, it loses its inductive characteristics; primary winding current can then
reach extremely high values for several ac cycles. Turning on a transformer may seem
fundamental, but in some power-supply designs and control applications, it can be a game
of Russian roulette. Because transformers remain polarized when turned off, saturation
occurrence is a function of the polarity and phase angle of the ac cycle when you switch
the circuit on and off. The smart-switch circuit in Figure 1
eliminates saturation, improves relay reliability, and provides a tool for determining
transformer and relay performance.
The circuit goes beyond typical
configurations using zero-crossing or peak-switching relays, by using the polarity of the
ac cycle, known phase angles, and soft-starting techniques. Figure 2
shows that the primary turn-on current of a 220-VA transformer can be disastrous when you
use a zero-crossing relay. Trace R1 shows 46A peak with a saturated core. Trace 1 shows
only a few amps with use of the smart-switch circuit. This large difference in current
demonstrates the value of the smart switch in controlling transformer magnetization.
Switching on during a positive half cycle and off during a negative half cycle or vice
versa prevents most core saturation.
Peak switching of the ac voltage
during turn-on and -off further reduces the susceptibility to core saturation, regardless
of ac polarity. This reduction is an important consideration in the event of an
uncontrolled power outage. Figure 3, trace R, shows the
primary current with peak and same-polarity switching. The vertical scale in Figures 2 and 3 is 10A per division, and Trace 2 is the
relay control voltage. The primary current in Figure 3
causes some core saturation (note that the current is not bipolar), but the saturation is
much lower than that in Figure 2. Trace 1 shows the reduced
primary current with the use of peak and opposite-polarity switching. Note that
transformer designs vary widely; some may favor particular phase angles.
Inrush current from power-supply
filter capacitors is also an important design consideration. By using a resistor, an
inrush device, or an inductive input filter in the secondary winding, you can reduce this
inrush surge. Another solution is to soft-start the transformer by using a resistor in the
primary to limit inrush and saturation currents to an acceptable level. After a brief
delay, a second solid-state relay shunts the resistor. The Microchip 12C508 µC uses its
internal 4-MHz RC oscillator for all timing. The chip is simple, inexpensive, reliable,
and well-suited for this application. For wide temperature variations, you can obtain more
accurate timing by using a 32-kHz crystal. To download Listing 1,
the source code for the µC's operation, from EDN's Web site, click here: Download
DI-SIG, #2170.
You can use either zero-crossing or random relays, but the random type
works better for transformers. Set Pin 4 high for zero-crossing relays and low for
random-turn-on relays. The HCPL-3760 optocoupler determines the polarity and phase of the
ac line. The coupler is configured as a near-zero detector. Its output is set to switch on
at 50V ac and off at 25V ac. One internal diode in the optocoupler rectifies the ac signal
to indicate the positive half cycle. The µC has two solid-state-relay outputs: SSR1 and
SSR2. When the Transformer Enable input goes high, the µC waits 250 msec, detects the
next positive edge from the optocoupler, waits 12 msec, and then turns off SSR1. SSR2 has
a 250-msec delay from SSR1 and operates as a last-on, first-off output to shunt a
soft-start resistor. Pin 3 is an optional output for a power-supply bleeder switch or a
status indicator. (DI #2170)
|