There’s always a need for a multimeter when working with electronic circuits. Especially when it comes to semiconductors. Using the internal diode test, which is now a part of every good meter, we are able to double-check the proper function of every diode, bipolar transistor and the like, before using them in our projects. However, there are still certain parts, for which the standard checking voltage of a classic multimeter (2 to 3 volts) is just not enough; let me name a few examples such as special rectifier blocks with a higher voltage drop (selenium diodes, high voltage rectifiers), or e.g. neon glow lamps, Geiger-Müller tubes, breakdown voltages of Zener diodes, transils, capacitors and the list goes on and on. And this is the reason why I have chosen to construct this machine. Behold:
This is nothing more than a simple, portable, battery powered inverter with a single transistor and a ferrite transformer inside. It is a flyback converter with self oscillating design; i.e. there is no external circuitry which provides the input signal to the transistor, these necessary pulses are obtained from the feedback winding on the transformer. The transistor switches current to the primary winding and high frequency voltage spikes appear on the secondary winding. These are then rectified with a fast diode (UF400x/FR15x series), a few nanofarad capacitor is charged and connected directly to the output terminals, or through a special glow lamp with a somewhat higher igniting voltage — you can find these types in fluorescent bulb starters, they look like this). These have a bimetallic strip and a different mixture of noble gases inside to achieve a higher voltage drop required to start a fluorescent bulb (120-180V DC), classic NE-2′s or derivatives have an igniting voltage of ca. 75-80V DC.
The tester can operate in two modes; the first mode shorts out the starter glow lamp, so the output terminals are connected directly to the charged capacitor. Voltage output is now at maximum, in my case 380-400 volts. The second mode adds the glow lamp and a 100k resistor in series prior to the output. This allows the glow lamp to light up in case something slightly/sufficiently conductive is connected to the output, thus indicating current flow. This also clamps and limits the terminal voltage output to glow lamp’s voltage drop (in my case, ca. 170-190V) and it limits the short circuit current, too, so it will not destroy more or less sensitive parts.
You might have asked why I am using a glow lamp instead of an LED, for example. Well, neon lamps start to light up at tens or hundreds of microamps. I can touch the output terminals in this mode directly, almost without feeling anything, and the current is sufficient enough for the lamp to glow. The same applies if I connect “myself” through a few megohm resistor! Try that with an LED in place of the neon.
As you can see, the circuit is pretty straightforward. You can use any small, generic signal PNP (or if you make proper changes, NPN) transistor with hFE over 120. To obtain similar results as I did, you need to experiment, though.
Try to fiddle with transformer’s air gap, or with the 4k8 trimpot to get a compromise between output voltage/current draw ratio. Or you can add a few stage multiplier without rectifying the output first. As a last resort, try increasing the input voltage or adding more turns to the primary/feedback/secondary. Just don’t go over 5V with supply voltage and be sure to wind all three coils in the same direction!
My setup: I’ve used a ferrite (EI-cored) transformer out from an old ATX power supply, glued those two cores together (no air gap) and made 15 primary, 15 feedback and 600 secondary turns. After tuning with the trimpot, it gave me 125V DC unloaded with a 1.2V NiMH; 380-410V DC unloaded with a LiIon 3.7V; from which it draws 50 mA when on, and 140mA at shorted output terminals. The frequency it runs on is 24 kHz, measured.
Finally, here is a YouTube video of the machine in action. Watch and enjoy