When I was a kid I had a Radio Shack “Science Fair 200-in-1 electronic project kit”, similar to the 150-in-1 kit Fran Blanche recently talked about on her show. It was designed in 1981. I don’t know if I built 10 circuits with it or 100, but probably somewhere in that range.
According to the manual, it contained:
Not counting the wires, that’s 62 components, most of which cost a cent or so nowadays, although I think at the time the kit was more like US$100. The components were mounted on brightly printed cardboard with some extension springs mounted around them; these served to grab the stranded copper wire when you fingered them sideways. I don’t know what the advantage of this method was over jumper wires in a standard breadboard, except that I guess each component terminal has a unique identifying number, so the wiring instructions in the manual could say things like “1-81-84, 2-41-49-55-176, 26-44-46,...”, and you could be reasonably sure you’d hooked it up correctly.
The designs of the circuits are pretty interesting in that they are adapted to the very minimal resources and poor tolerances available in the kit; they include a few different single-transistor oscillators, for example. (I think they’re Hartley oscillators, often using the center tap on the audio output transformer for their tapped coil, but I’m not sure I understand them.)
The circuits include various kinds of AM radio transmitters and receivers, various kinds of audio oscillators, games that control audio oscillators etc. with light, a “strobe light” with an LED, push-pull amplifiers, RTL and DTL logic gates, a “door alarm”, random number generators, a divide-by-4 counter with decoded output, a VCO, a voltmeter, an ohmmeter, and so on. Many of the circuits use the speaker or piezo earphone as microphones.
It’s been 39 years since it was designed, and a few of the components
are obsolete (TTL logic, germanium diodes, and variable capacitors)
while others are harder to find (CdS cells, piezo earphones, galvos,
relays, incandescent bulbs). And nowadays, if you were designing
something similar to build out of new parts, you might take advantage
of some of the parts that are cheaper and more robust than they were
then: power MOSFETs, op-amps (maybe LM324s, TLC272s, and as Viper-7
suggests (see file notes/jellybeans.html in Dercuano), TL084s for
JFET input), Schottky diodes, Darlington arrays like the ULN2003,
zeners, colored LEDs, some 555s, phototransistors, but especially and
above all else, microcontrollers. If you’re going to have discrete
logic circuits, make them CMOS.
If we’re limited to parts we can salvage from discarded equipment, what could we patch together?
The easiest way to get wire is from discarded wire, especially power cords, but sometimes also things like telephone line and coax.
Batteries are right out, but there are lots of perfectly capable AC power supplies out there. Surprisingly, the power supply often is not the first thing that breaks; sometimes it’s the supply chain.
LEDs, silicon signal diodes, resistors, capacitors, buttons, and switches are abundant, and optointerruptors are found at times; most power supplies also contain transformers, inductors, silicon PN power diodes, and Schottky diodes. Speakers are reasonably common. Crystal resonators are also quite common (this VCR has nine of them), potentially permitting very high precision timing measurements. Potentiometers with knobs attached do occur occasionally, but trimpots are enormously more common.
Even this 12-watt LED lightbulb that burned out the other day in the bathroom has a little power-supply board in it containing two resistors, an MLCC capacitor, a diode, two electrolytic capacitors, and a transformer (a center-tapped coil, really), plus a couple of chips (one of which may be a bridge rectifier), plus 14 bright LEDs in series, two of which are burned out. Perhaps the power supply works fine and it was just the LEDs that overheated, in which case I have a non-isolated power supply the size of my fingertip designed to supply some 56 volts, 300 mA, from 240VAC. Or perhaps it would be more useful in pieces.
Transistors are a little messier. The VCR, from 1996, has apparently several hundred of them, but apart from half a dozen power transistors in its power supply, they’re mostly tiny surface-mount components. I more often find BJTs than MOSFETs, but in this case I haven’t looked them up yet.
Inductors are a sufficiently expensive component that the 200-in-1 kit didn’t have any except as part of its transformers and antenna. But they are straightforward to make by hand from wire, especially for low inductances, or to salvage from discarded equipment.
Connectors are another tricky question. The 200-in-1 kit had only 62 electronic components — including post lugs to attach wires to — but some 80 wires and 176 springs. The dude from Espacio de César demonstrated rigging up a solderless breadboard out of DIP sockets from old circuit boards — snip the two sides off and you have two rows of 2.54-mm-spaced socket holes you can plug pins into. Other connectors, such as DIMM slots or CPU sockets, may also work for this. Through-hole components are easy to slot into those, as long as the leads aren’t too short, but surface-mount components need to have pins added to them.
Consumer electronics are by and large full of single-sided PCBs, which are full of jumper wires, which can be pressed into service as pins in a pinch, but a better alternative when possible is to rip apart male Molex-style connectos.
Connectors are also very valuable for a different reason: they permit modularity, and if you’re generating, say, an audio or video signal, you can use them to connect it to something external.
7-segment LED displays can still be found in things like discarded
clock radios or microwaves,
but a better option may be to build them out of
now-abundant LEDs and commonplace non-electronic materials like paper
and aluminum foil.
CdS cells are virtually unheard of in the last decades, but phototransistors are ubiquitous, though most often infrared, often with shielding. LEDs can sometimes serve as photodiodes, too, although they are poorly characterized for this use.
A soldering iron and soldering flux may be difficult to improvise.
The circuit cookbook probably can’t be as cut-and-dried as the Radio Shack cookbook was, because the available components will be more variable.
You need to start from basic tools. First you need a power supply with voltage in a reasonable range. But you need to be able to detect that its voltage is in a reasonable range. How do you do that without a multimeter?
See also the note on multimeter metrology.
A white illumination LED from a lightbulb can probably dissipate a whole watt, no problem, which is 300 mA or so, and it will probably light up visibly with any current above 0.1 mA. You probably want a couple of separate measuring instruments here, made of two such LEDs in antiparallel in series with a resistor: one to ensure that the voltage is not outrageously high, one to verify that there is some useful voltage.
The not-outrageously-high detector uses a resistor in the 100kΩ–1MΩ range, which should illuminate the LED and heat up the resistor noticeably, but probably not burn up, if placed across a circuit carrying hundreds of volts. Still, you want to make sure you’re using a through-hole kind of resistor for this to handle the heat, not a surface-mount. At 100V and 1MΩ you get 100μA, which should be visible on the LED, if barely. If both LEDs light up, you know it’s AC.
The some-useful-voltage detector is used after you’ve established that the circuit doesn’t have 100V or more on it, so it uses a resistor in the 330Ω–3.3kΩ range. So those same 100μA will appear, and the LED will start to light up, at 0.033–0.33 volts above the LED’s forward voltage drop (typically 3V). At 100V the LED will have 30–300mA running through it and will illuminate brightly. XXX the resistor will explode
XXX Hmm, I need to rethink this a bit. Even at 3.3kΩ the resistor dissipates 3 W at 100V.
The resistors can be pulled from broken or surplus power supplies, which commonly have large resistors in them, and identified using the resistor color code, without a need for a multimeter. It will need to be verified that they do conduct electricity.
By attaching the some-useful-voltage detector to one side of the output of a known-good power supply, you also get a diode and continuity tester.
Once you know a given regulated DC power supply works, you need to be able to derive other DC voltages from it. Suppose it’s 12V, the highest-voltage rail on an ATX power supply (and typically provided with a lot of current). You can rig a 10kΩ potentiometer across it to get a variable voltage reference, then feed that into the emitter (or gate) of a power transistor whose collector (or drain) is connected to the appropriate power-supply rail, thus giving you an emitter (or source) follower.
This allows you to get whatever regulated output voltage you want, up to a diode drop below the input voltage. But how do you know what voltage you’re getting if you don’t have a multimeter?
Three or four LEDs in series to ground, ideally a 1.5-volt indicator type rather than a 3V illumination type, can provide some kind of indication of how high the input voltage is. At below 1.5 V, no LEDs will light. At 1.5 V, the bottom one will light, fed by a string of resistors to it from the voltage input. Successive resistors in parallel with the other LEDs will develop enough voltage to light those LEDs as the current rises; this requires them to have lower and lower resistances.
On one side of the bridge we use a potentiometer (presumed linear) with a knob glued to it; the other side pits the unknown resistance against a known resistance. Rather than Wheatstone’s galvanometer across the middle, we use a pair of antiparallel LEDs in series with a small protective resistance. This may require that the input voltage be rather high, tens of volts, to get good precision.
With an AC source, I think this setup also works to measure ratios of capacitances or inductances.
Then, it should be possible to replace the crude LED pair with a delicate differential pair of NPN transistors.
These detectors of voltage differences can also be used to directly compare voltages, for example to calibrate positions on the potentiometer knob on the linear power supply against known regulated voltages, either from a multi-voltage power supply or from a 7805 or something.
There are lots of circuits for this but I don’t know which ones are simple, free of soakage, thermal coefficients, and whatnot. But if you build one you can hook it up to a speaker to listen to your signals; one of the 200-in-1 projects does this.