... is made around a hex-buffer. A handful of cheap 100 nF ceramic capacitors, various resistors and a single transistor make do for the rest. The construction is suitable for further experiments and add-ons, so for that end some testpins and a diode (to give some protection) are employed too.
The power consumption is rather modest and batteries (6 - 12 V.) as well as any reasonably stable powersource can be used.
... is vastly dependent on the size and shape of the actual components. Here a couple of SIL-resistor-arrays were employed. They may, however, be replaced by small, conventional resistors without changing the basic lay-out significantly.
In case you decide to try the "crawling light", you may get some help from the illustrations. Please observe, that a couple of the connecting wires (not drawn in red!) run on top of the board -- and that my solution does not have to be the best one!
|The circuit diagram of the master
Please notice the minus-/common-ground-rail runs from the capacitor top, left and down right.
Except for the hex buffer-IC (= 4050) there are no special requirements to the components.
|Add-ons: more light
If the 'crawling light' is to be part of a toy or a display, higher intensities are needed. For that purpose an external driver will be useful. I have tried a simple approach like the one in the diagram. As you will see, I could not stand the temptation to use the first driver as an inverter for a second one. The first trio of LEDs does not, in this configuration, extinguish completely.
The drivers (6 of them) run readily on a separate powersupply and a plain rectifier will suffice. Resistance values are calculated for 12 V. and 3 yellow or green LEDs. In case you are bent on red LEDs, just add an extra one, making it to 4 diodes.
The drivers, when connected, will prevent the LEDs in the mastercircuit from lighting up.
|Add-ons: sounds instead of light.
Small, adjustable LF-oscillators are readily made around CMOS-Schmitt-trigger-NAND-gates. These come in packages of 4 (= 4093), so 2 of the kind were required for this purpose.
The value of the capacitor 'C' has to be determined experimentally because the internal parameters of the 4093 affect the frequency a lot. 1 nF is a good place to start. It is not adviceable to go lower than 220 pF.
Another peculiarity was discovered during my experiments: apparently there is a smudging effect amongst gates in the same housing, making two oscillators, running at nearly the same frequency, 'lock' to one another. The problem was partly solved by involving the vacant two NAND-gates as stand-ins for the troublesome ones.
The 'logic' outputs from the mastercircuit do not match the controlling input in the oscillators and some modifications have to be made. You may choose from two possibilities: either remove (or omit) the LEDs or establish an extra connection going right to the bufferoutputs.
The oscillators are best run from the same powersupply as the mastercircuit.
Since the LF-oscillators generate square-waves, the sound does appear rather harsh and vulgar. Fortunately there is plenty of it offering ample room for filtering and curveshaping.