You did what?

Got a good "goof up" you'd like to share? Tell us what you did wrong, what happened, and what you learned. We'll send a PCB clipboard to contributors whose stories we publish. E-mail us at goofs@tmworld.com.

Years ago, we built a system console that included many displays for sequences, times, and process variables. In this era, before LEDs and LCDs, we used Nixie tubes—neon-filled tubes with cathodes that formed numerals. We used a standard driver IC for each Nixie tube, and the IC's binary inputs let us select one numeral at a time. The driver and power lines all ran from the displays to the control logic in a single large bundle. Our design included a current-limiting resistor for each Nixie tube. Ionized-gas lights such as Nixie tubes and neon bulbs require resistors to limit current flow, because after a high voltage (about 100 V) ionizes the gas, resistance approaches 0 Ù.

When we applied power, there was a huge inrush of current through the large cable running to the displays. The current spike coupled to the display-driver ICs and burned them all out in a few milliseconds. Our investigation showed someone forgot to include current-limiting resistors for the individual neon bulbs used to form colons in our time displays. So, these bulbs acted like short circuits. We ended up replacing a substantial number of ICs, and the lab that funded our project was not pleased with the increased cost for new drivers.

T. R. Clem, Sr., Ivanhoe, NC

One configuration of the photomultiplier tubes (PMTs) we use connects the cathode to ground and the anode to a positive high voltage. A coupling capacitor links a tube's signal to a low-voltage instrument, often an oscilloscope. For tube testing, we set the oscilloscope input to high impedance and use a 50-Ù external terminator at the scope's input. A student trying to test a PMT plugged the tube's output into the scope, unknowingly blowing out the channel's amplifier. He saw no signal, so he moved the cable to the next scope channel, and eventually blew out all four channels. When he went to find another scope, someone discovered his problem—improper scope settings—and prevented further disaster.

Months later, a physicist was running a similar test, but he had a problem and asked me for help. He said he couldn't see a signal on the scope. As he pulled the cable from the first scope input to try the next input, I quickly realized what he'd done and grabbed his hand. He was using the positive high-voltage configuration and had set the scope to an internal 50-V termination. That configuration provided an expensive way to learn that leakage currents really do exist and that capacitors can store a charge.

John Segal, Newport News, VA