Interesting circuit ideas which we have checked but not built and tested. Contributions from readers are welcome and will be paid for at standard rates.
"Safe" oscillator for watch crystals
This circuit was developed to allow watch crystals to be used in an existing CMOS oscillator circuit that was to run from a 12V supply.
The problem is that these crystals only work up to a supply
voltage of about 6V. Any more than that and the crystal will be over-driven, causing it to shatter.
This circuit solves the problem by using LEDs 1 & 2 and a
470nF capacitor (C3) to limit the drive to the crystal to about 4V peak-to-peak. Note that it may be necessary to adjust C1 & C2 to ensure reliable start-up and stable oscillation with some crystals. However, the C1:C2 ratio should be maintained.
As a bonus, the two LEDs both glow, giving a visual indication
that the oscillator is working.
Hamilton, NZ. ($35)
Editor's note: the relatively high values used here for capacitors C1 & C2 will load the crystal, which means that the oscillator will run at less than the nominal crystal frequency (32.768kHz).
Internal resistance tester for batteries
This circuit is designed to check the condition of lead-acid and gel cell batteries with capacities greater than 20Ah. It switches a load of about 18A at a rate close to 50Hz so that the internal resistance of the battery can be measured using a digital multimeter across the battery terminals.
The measured AC voltage in millivolts divided by 10 (ie, a
shift of the decimal point) is approximately equal to the battery's internal resistance in milliohms.
As shown, the circuit is quite straightforward and is based on
two 555 timer ICs (IC1 & IC2) and power Mosfet Q1. IC1 operates as a monostable timer with a period of 10s. When switch S1 (Test) is pressed, IC1's pin 3 output goes high for 10s and this enables IC2 which operates as a 50Hz astable oscillator.
IC2 in turn drives power Mosfet Q1 which is connected across
the load in series with three 0.22W 50W resistors. IC2 then turns off again after 10s - ie, at the end of the monostable timing period.
LED1 provides power indication when the circuit is connected to
a battery, while LED2 (green) comes on during the test period. The thermostat is not necessary unless the unit is to be used repeatedly (the Jaycar ST-3823 70°C unit is suitable) and you want to protect the output circuit against overheating.
Note that the power Mosfet does not need cooling but the
thermostat and the 0.22W 50W resistors should all be mounted on an aluminium heatsink at least 2mm thick.
In practice, the internal resistance of car batteries can vary
from about 15mW down to about 3mW. Before testing the battery, check that the electrolyte level is correct and that the voltage across its posts exceeds 12.5V for a nominal 12V battery; ie, close to full charge.
That done, switch on the car's headlights and measure the DC
voltage between each battery post and its connecting terminal. It should be less than 10mV in both cases; if not, the terminals need cleaning.
Once you've done that, you can turn off the headlights, connect
the tester and proceed with the internal resistance test. Be sure to connect the multimeter's test probes directly to the battery posts, to read the internal resistance (not the battery terminals).
Here's how to build a pendulum-controlled clock which can be
made really accurate. Retro? - yes, but an interesting project all the same.
You'll need a spare quartz clock which must be adapted by first
isolating the two pads on the chip which lead to the coil. You then have to connect wires to these pads and feed them out through a hole in the case (see SILICON CHIP, December 1996, p38, for full instructions, or October 2001, p37, for brief notes.) You'll
also need a spare battery driven pendulum from another, or the same, clock.
As originally used, these pendulums are for appearance only and
play no role in timekeeping. The salvaged unit should be mounted on a substantial vertical backboard. You'll find that the pendulum swings pretty fast and it must be slowed down by adding weights near the lower end.
However, it's not the mass of a pendulum that controls its rate
- instead, it's the distance from the support to the centre of mass that counts.
The aim is to make the pendulum operate so that it takes
exactly 1s for a full "to and fro" swing - ie, 0.5s "beats". Fine adjustment on mine was made by adding an adjustable (up and down) weight to the pendulum rod. This consisted of a small G-clamp fabricated from a brass strip and held by a small screw.
At the bottom end of the pendulum attach an inverted T-shape
aluminium vane, about 10mm wide and as thin as possible. This should be painted black.
This vane is used to trigger a photo-interrupter which is
attached to the backboard. The lengths of the arms of the "T" are made so that when the pendulum swings one way, the interrupter triggers - ie, the light is no longer blocked. Conversely, when the pendulum swings the other way, the vane must continue to interrupt the light.
This means that, with the pendulum swinging in 0.5s beats, we
get a short pulse from the photointerrupter at 1s intervals. This pulse is inverted by IC1a and inverted again by IC1b which then clocks IC2, a 4013 flipflop.
IC2 alternately produces 1s-long pulses at its pin 12 & 13
outputs. These outputs are then fed to IC1c & IC1d respectively, where they are gated by the short pulses on pin 4 of IC1b. This produces two short pulses to drive the clock in alternate directions at 1s intervals. And that's all you need to drive the clock.
Alternatively, this circuit could be a master clock and could
be used to drive several slaves, all remaining in time. And model train enthusiasts could drill one or more holes in the vane to make their "railway" clocks run at what ever speed they need.
The circuit can be built on a small piece of strip board. Note
that the photo-interrupter should be mounted with the photocell facing the backboard. This minimises the risk of interference by ambient light. The photo-interrupter is available from Jaycar - Cat.ZD 1901.
A footnote for horologists - if you have a clock with a Hipp
butterfly escapement, you could rid yourself of the trailing arm and contact arrangement and replace it with a vane and photo-interrupter set so that as the arc of the swing becomes too small, a pulse is missed. This could then be detected by a 555 missing pulse detector circuit which would then energise the impulsing magnet.
Bardon, Qld. ($50)
Super light sensor circuit
Fig.1: light level fluctuations are detected by LDR1 and the resulting signal fed to comparator stage IC1. IC1 in turn triggers 7555 timer IC2 which is wired as a monostable and this drives transistor Q2 and a relay.
This "Super Light Sensor" responds to minute fluctuations in
light level, auto-adjusting over the range from about 200 lux up to 60,000 lux (ie, from a modestly lit room to direct sunlight). It has lots of potential uses - eg, detecting a car entering a driveway, a person moving in a room, or wind rustling the leaves of a tree. At the same time, it has a high level of rejection of natural light variations, such as sunrise, sunset and the movement of clouds.
While it is a "passive" system, it can also be used as an
"active" system - ie, used in conjunction with a light beam. Its great advantage here is that, since it responds to fluctuations in light level rather than the crossing of a specific light threshold, it is much more flexible than other typical "active" systems. It can be placed within the line-of-sight of almost any light source, including "vague" ambient light, and simply switched on.
As shown, the LDR is wired as part of a voltage divider so
that, between darkness and full sunlight, its output at "X" varies between about one-quarter and three-quarters of the supply voltage. A wide variety of sensors may be used in place of the LDR, including phototransistors, photodiodes and infrared and ultraviolet devices.
The signal from the sensor is fed to the inputs of comparator
IC1 via two 150kΩ resistors. However, any signal fluctuations will be slightly delayed on pin 3 compared to pin 2, due to the 220nF capacitor.
As a result, the pin 6 output of the comparator (IC1) switches
low during short-term signal fluctuations and this triggers monostable timer IC2. IC2 in turn switches on transistor Q2 which activates Relay 1. It also lights LED1 via a 1.5kΩ current-limiting resistor. Trimpot VR2 allows the monostable period to
be adjusted between about 3s and 30s.
As with all such circuits, the Super Light Sensor may not work
as well under AC lighting as under natural lighting. If AC lighting does prove a problem, a 16μF (16V) electrolytic capacitor can be connected between the sensor output and ground to filter the signal to the comparator.
When pin 3 of IC2 goes high, FET Q1 also turns on and pulls pin
2 of IC2 high. This transistor remains on for a very short period after pin 3 goes low again due to the 100nF capacitor on its gate. This "blanking" is done to allow the circuit time to settle again after the relay disengages (and stops drawing current).
The "blanking" also makes it possible to run external circuits
from the same power supply as the Super Light Sensor, without upsetting the circuit. The current consumption is less than 10mA on standby, so that battery operation (eg, 8 x AA batteries) is feasible.
Fig.2: the LDR should be installed inside a black tube, as shown here.
After building the circuit, switch on and wait for the circuit
to settle. It's then just a matter of adjusting VR1 so that the circuit has good sensitivity without false triggering.
With some experimentation, it's possible to set the circuit to
change seamlessly from natural to AC lighting. If maximum sensitivity under natural lighting false triggers the circuit under AC, then adjust VR1 to give maximum sensitivity under AC (and vice versa).
In daylight, the Super Light Sensor will typically detect a
single finger moving at a distance of 3m, without the use of any lenses. It will also detect a person crossing a path at a distance of more than 10m, again without lenses. And when used as an "active" system, it will typically detect a person walking in front of an ordinary light source (eg, a 60W incandescent light-bulb) at more than 10m.
Note that these ranges are achieved by placing the LDR (which
is used as the light sensor) in a black tube, as shown in Fig.2. A single lens will double these distances, while the use of two lenses in an "active" system will multiply the basic range by 6 or 7.
Capetown, South Africa. ($50)
LED lighting for dual-filament lamps
A number of readers have asked how the bayonet lamp described
in the "LED Lighting For You Car" project in March 2003 can be adapted to replace a dual filament lamp.
Contribute and win!
As you can see, we pay good money for each of the "Circuit Notebook" contributions published in "SILICON CHIP". But now there's an even better reason to send in your circuit idea: each month, the best contribution will win one of these superb Peak Atlas LCR Meters - valued at around $195.00.
So don't keep that brilliant circuit secret any more: send it to SILICON CHIP and you could be a winner!
Before we describe how it's done, note that we recommend that
the result be checked as having sufficient brightness for a stop & tail-light application. That's because the light output may be inadequate, depending on the tail-light lens and reflector assembly - so use any modified lamps with discretion!
As shown, an additional diode (D1) and resistor (68W) provide power from the "tail" circuit. Alternatively, when the "stop" circuit is powered, the resistor is bypassed by D2, thus increasing the LED current and the light output.
Modifications to the lamp assembly instructions are as
(1) After soldering in the copper tube but before soldering the
platform board to the bayonet lamp base, the three components inside the dotted box must be wired up inside the base.
(2) The anode leads of the diodes can be soldered directly into
the contacts ("bumps") on the base (a fine file or glass paper may be needed to get a nice round shape). Everything must be insulated (use heatshrink tubing).
The red wire from the Multidisc board is then soldered to the
junction of D2 and the resistor. The black wire is soldered directly the metal casing of the lamp.
We suggest testing the lamp before soldering the platform board
in place. It may be necessary to vary the value of the additional resistor to get the correct intensity change between stop & tail modes.