In Electronics today we were playing around with some op-amps. Specifically, TL081 op-amps from ST Microelectronics.

We were only using them as comparators. The inverting input was connected to ground and the noninverting input to a potentiometer. The op-amps and potentiometer were powered from a ±15V supply. When the potentiometer was set at 15V, the op-amp had a 14V output as expected. When set to -7V, the op-amp had a -13.5V output. But when set to -15V, the op-amp went back to 14V...

What could explain this odd behaviour? It did not happen when the op-amp was in the reverse configuration with noninverting input tied to ground and inverting to the pot.

I suspect the input ranges were being exceeded causing odd behaviour.

 

It's topical behavior, google "phase reversal" or "phase inversion" and check in datasheet allowed input common-mode voltage for TL

 

It's also known as "latch-up" and "lock-up". Many more modern opamps have protections against this phenomenon.

Here's an extract from Linear Technology's LT1057/LT1058 datasheet. Note the input sine wave, a phase inversion output from an older opamp, and the LT1057/LT1058 output on the right.

 

It's also known as "latch-up" and "lock-up". Many more modern opamps have protections against this phenomenon.

What I know as latch/lock-up refers to when an amp enters this state and bringing the input voltage back in range does not fix it. That's much worse than what is shown in the 'scope traces.

The only way to fix it is to power-down and start over. I also seem to remember that they would start to get hot when in this state.