Voice bandwidth filter
This circuit passes frequencies in the 300Hz - 3.1kHz range, as
present in human speech. The circuit consists of cascaded high-pass and low-pass
filters, which together form a complete band-pass filter.
One half of a TL072 dual op amp (IC1a) together with two
capacitors and two resistors make up a second-order Sallen-Key high-pass filter.
With the values shown, the cut-off frequency (3dB point) is around 300Hz.
As the op amp is powered from a single supply rail, two
and a 10μF
decoupling capacitor are used to bias the input (pin 5) to one-half supply rail
The output of IC1a is fed into the second half of the op amp
(IC1b), also configured as a Sallen-Key filter. However, this time a low-pass
function is performed, with a cut-off frequency of about 3.1kHz.
The filter component values were chosen for Butterworth
response characteristics, providing maximum pass-band flatness. Overall voltage
gain in the pass-band is unity (0dB), with maximum input signal level before
clipping being approximately 3.5V RMS.
The 560Ω resistor at IC1b's output provides short-circuit protection.
M. Sharp, Berala, NSW. ($35)
El-cheapo fluoro ballast
This simple circuit can start small (15W or less) fluorescent
tubes such as those used in PC board exposure and EPROM ultraviolet erasure boxes.
As you can see, the tube’s filament heaters are not used.
Instead, ignition is provided by a voltage tripler formed by diodes D1-D3 and the two 6.8nF 2kV capacitors.
At switch on, C1 charges up via R1 until the gas in the tube
breaks down (around 700V). C1 then discharges through the tube, lowering the resistance enough to sustain continual AC current flow.
C1 then continues to act as the ballast, with R1 included to
prevent the three diodes shunting the tube on positive mains half-cycles.
Adrian Kerwitz, via email. ($30)
Surveillance camera recorder
This idea originated from the need to record video from a
surveillance camera on a standard VCR, without wasting hours of tape. The circuit waits for a trigger signal before starting the VCR, which then runs for a predefined period. Virtually any mechanism could be used to trigger the circuit, including the output from a PIR (passive infrared) sensor, door switch, alarm panel, etc.
The circuit is based on the popular Picaxe-08 microcontroller
and is programmed to simulate the normal key presses used to set a VCR into record and stop modes. Control of the VCR is achieved by connecting the normally open contacts of two relays (RLY1 & RLY2) across the "record" and "stop" buttons. Some VCRs also need the "play" button to be pressed at the same time as the "record" button to start recording. In this case, a double-pole relay is
needed for RLY1, as indicated on the circuit.
The trigger input is interfaced via an optocoupler (OPTO1) to
give complete isolation. This allows the entire circuit to be powered from an internal VCR supply rail, if available. Alternatively, it can be powered from an external 12-18V DC source.
When the Picaxe (IC1) detects a trigger input, it switches on
transistor Q1 to energise relay RLY1. The relay is held on for a second or so to simulate finger operation of the button(s). After the programmed time has elapsed (10 minutes), transistor Q2 is switched on to energise RLY2, stopping the VCR.
The necessary Picaxe program is shown in Listing 1. It is
easily modified for different times and applications.
|'picaxe 08 |
|'PORT 0 = NOT CONNECTED|
|'PORT 1 = START RELAY OUTPUT|
|'PORT 2 = STOP RELAY OUTPUT|
|'PORT 3 = TRIGGER INPUT|
|'PORT 4 = LED OUTPUT|
|if pin3=0 then start
||'detect input goto start|
||'flash LED1 (indicates running)|
||'start relay on|
||'start relay off'this sets VCR into record mode|
|for b0 = 1 to 60
||'sets 60x1 sec =1min |
||'waits 1 sec |
||'flash working led|
||'loop until 60 reached|
|let b3 = b3 + 1
||'count up 1 min|
|if b3>=10 then stop
||'goto stop sequence|
||'stop relay on|
||'stop relay off|
||'resets time variable|
Important: do not attempt to modify your VCR unless you know
exactly what you’re doing. Always use relays to connect to the record/play/stop switches, as these switches are "floating" in most VCR’s (ie. not connected to either the positive or ground rails). Also, keep switch lead lengths as short as
Darren Michell, Coraki, NSW.
Experimental pendulum clock
Using this design, you can construct an electromagnetically
impulsed pendulum clock with a 1-second beat.
On the prototype, the pendulum rod is 115cm long with a bob
adjusted to make it beat every second. It is suspended on a short piece of mainspring from a watch, which is attached to a vertical backboard with a 6mm screw. The rod extends some 15cm below the bob and is fitted with large washes at the lower end.
Note that for a pendulum to beat in seconds, there must be
99.4cm distance between the support and the centre of mass of the pendulum. Between the bob and the lower end is a 5mm wide white reflector facing back. Below the rod and 15mm to the left is the impulse solenoid, with a core but no actuator attached.
The circuit comprises of four parts: (1) the sensor; (2) the
counter and solenoid driver; (3) the clock driver; and (4) the clock.
The sensor is built on its own small piece of strip board and
is located on the centre line of the backboard behind the reflector. It utilises a Sharp IS471F infrared modulated detector (Farnell cat. 414-2860) to eliminate interference from external light sources. The infrared emitter (IRLED1) must be mounted near to the detector (IRDET1) but be masked from it.
The emitter radiates a coded signal toward the reflector. As
the pendulum passes the centre line it reflects the signal back to the detector, which then gives a negative-going output pulse on pin 2. This makes the surface-mount LED (LED1) flash once. It also sends a signal to the counter and clock driver circuits on the main circuit board.
Pulses from the sensor are fed into IC1, a 4020 14-stage ripple
counter. The counter’s output (pin 6) goes high every 128 counts (seconds). These long duration pulses are inverted by transistor Q1 and differentiated by the 10nF capacitor and 22kΩ resistor, providing a narrow trigger pulse for a 7555 CMOS timer (IC2).
The 7555 is wired as a monostable, driving the base of
transistor Q3 with a relatively short pulse width suitable for energising the impulse solenoid. LED2 flashes in unison with solenoid pulses, and can be mounted right on the solenoid as a visual aid.
Pushbutton switch S2 is used to provide gentle starting pulses
to get the pendulum swinging smoothly at the outset. Switch S1 resets the counter to zero. With this arrangement, the pendulum is set swinging and when it is to the left of centre, S1is pushed. Thus, the pendulum moves right to left on even numbered counts. At the 128th count, the solenoid gives a shot pull to the left just as the pendulum is passing through the centre line and moving right to
left. The distance of the solenoid below the pendulum is adjusted so that it does not jerk the pendulum but adds a gentle nudge.
The clock driver circuit also derives its timing from the
output of the sensor. Negative-going pulses from the sensor are inverted by Q4 before being fed into a 4013 flipflop. On the output side, pins 12 & 13 go high in turn for one second. These pulses are too long to directly drive the clock coil, so they’re logically "anded" with the short pulses from the sensor using two gates of a 4093 NAND Schmitt trigger (IC4). The outputs from these
gates then drive an adapted quartz clock movement.
A suitable clock can be made from a standard quartz movement by
isolating the coil and removing the battery. See SILICON CHIP, Dec. 1996, page 38 for full instructions or October 2001 page 37 for brief notes.
This is an experimental clock so you may have to try various
solenoids to find one that works for you. If necessary, the solenoid pulse duration can be changed by varying IC2’s timing components. If the suspension is too stiff, try impulsing at 64 beats from pin 4 of IC1, but note that the aim is to get the freest pendulum movement possible. The Synchronome and Hipp clocks were impulsed at 30-second intervals, so your clock could be even better.
In the prototype, the reflector was made from the back of an
adhesive cable clip snapped on to the pendulum rod. The white back was masked to give a 5mm wide central vertical strip, giving clean, short pulses as the pendulum passes.
Current drain is several milliamps, so the prototype was powered from an SLA battery fed from a float charger.
A pendulum beating in seconds is called a Royal pendulum. Its
length is the same as one in a typical long case (grandfather) clock.
A. J. Lowe, Bardon, Queensland, ($45)
Handy time delay with relay output
This circuit is designed to provide delayed relay switching
action at power on. The delay is a function of the time constant produced by the combination of R1 and C1.
At power on, C1 charges slowly via R1 and the coil of the
relay. When the voltage across C1 exceeds both the base-emitter voltage of Q1 and the gate trigger voltage of the SCR, gate current flows. This fires the SCR and switches on the relay.
At power off, diode D1 rapidly discharges C1 through the
100Ω resistor, so ensuring that every time the circuit is restarted, as in a temporary outage, the delay time is maintained.
Just about any NPN transistor can be used for Q1, since after
SCR1 fires, it is effectively out of the circuit. In fact, the only part that’s still active after SCR1 turns on is the relay. You can’t get much simpler than that!
This circuit can be used to delay speaker turn-on, so avoiding
the "thump" that occurs in some stereo systems at power on. A 5-second delay is enough for this application, requiring approximately 560kΩ for R1 and 10μF for C1. Another application
might be as a motor protector in a short power outage.
R. Besana, Henderson, New Zealand. ($30)