The First Rule of Troubleshooting: Everything is True

The First Rule of Troubleshooting: Everything is True
One of my phone calls last week was from an engineer that had last seen an operational amplifier in some long forgotten analog class at the university. He opened the conversation with the exclamation, "OK...how did you guys do it? I mean, make an op amp that does not invert the signal."

Welcome to the world of the Analog Applications Engineer, the technical reference whose job is to guide customers in the use of the devices he supports. That's me: Every call I answer contains a nugget of troubleshooting technique, insight into the workings of a class of devices. Never a sarcastic outburst or statement of opinion.

Falling into my "lets get the details" mode, I asked for a description of the circuit he had constructed. The op amp circuit described to me seemed simple enough: A function generator, whose output goes through a 4.99 k-ohm resistor to the inverting input of the amplifier. (The non-inverting input is connected to ground.) The output of the op amp goes to a scope. (There is also a feedback resistor, 4.99 k-ohms, connected between the output of the op amp and the inverting input). Sounds like an inverter, right? But measurements taken at the output showed a perfect gain of +1 - where theory says the gain should be -1. Must be some new class of op amp, I thought.

The op amp circuit described to me seemed simple enough

The device was one of the precision bi-polar op amps that are powered from dual supply rails, 15V, which eliminated a whole list of single supply concerns. The circuit was classical inverting, so common mode voltage levels were not at issue. My first attack was the instrumentation. Is the invert signal switch off in the scope channel? That was confirmed as were checks of the scale factor and setting for probe sensitivity.

Finally I smelled the rat. I asked our puzzled engineer to take a DC voltmeter and measure the voltage from breadboard ground to each of the pins on the device. Please be sure to measure at the device pin, not on the socket or even a short jumper away. The reading on pins 1, 5, and 8 were zero, just as expected. They are not connected within the device. Pin 3 was at ground as was expected from the circuit diagram. The answer to the problem came at pins 4 and 7. They were also at ground. He had not connected the power supplies. With no load the input signal had just been connected to the scope through the feedback resistor.

This little exercise just illustrates the most common thread in all calls for applications assistance. The engineer is so close to the problem that some of the most basic assumptions turn out to be not valid. As I say quite often, "You have built a very fine contradiction here. One of the 'TRUTHS' that you are quoting is not. Our job is to find that non-truth."