In the good old days of the 1960s -– when everyone was gainfully employed –- military equipment designers implemented amplifiers from discrete transistors, resistors, and capacitors from the approved list, rather than simply plugging in an op amp, as we would probably do today.
While prototyping a circuit for a Navy flight-line ECM tester, I encountered an unusual problem. An amplifier circuit that was too simple to fail was not working correctly. My prototyping technician reported that the stage couldn’t support the signal levels and was clipping badly.
I left my cube, sat in front of his bench, and checked the scope. Sure enough, the output was clipping. Yet there didn’t seem to be any obvious errors in the layout or connections. Nevertheless, the biasing of the transistor was way off. We prototyped using 1/4w 5 percent military RC05 resistors that were clearly marked. I checked the DC decoupling capacitor values and leakage, and there was nothing amiss. I pulled the transistor from the board, and it tested fine.
Then I measured the voltage at the base-bias divider network. It was very low, despite the divider consisting of 330k and 33k to give roughly 9 percent of the supply rail. Finally, I unsoldered the two resistors and measured them. The 33k was actually 330 ohms! Clearly, the 33k resistor had been mismarked orange-orange-orange by the manufacturer and shipped as 33k. In reality, it was 330 ohms. The third most significant orange band should have been brown.
So much for the minutia of the lesson learned many years ago. Yet it’s relevant today. When prototyping and when troubleshooting, we must never assume the components are the correct value, whether they be chip resistors, capacitors, or larger leaded components. Prototyping and test personnel should be instructed to measure each component before installing, as annoying and time-consuming as that may be. Each part should be verified to make sure it is correct per the parts list and schematic.
Then when problems arise, you will be certain that it is not a design issue, but rather a fabrication issue. It should be noted that wrong-value components apply equally to modern designs using op amps and chip components. Assume nothing is my mantra in troubleshooting, and it should be yours.
This entry was submitted by Michael Neidich and edited by Rob Spiegel.
Michael Neidich is a circuit and system designer in disciplines ranging from DC to Lightwaves, from discrete Germanium to GaAs, from ICs to VLSI, and from consumer to military equipment. He has worked at semiconductor firms such as Anadigics, Analog Devices and Zoran Corp., General Microwave Corp., and Sharp Electronics Corp.
Tell us your experience in solving a knotty engineering problem. Send to Rob Spiegel for Sherlock Ohms.