Op Amp Tests
Low Level Distortion
©2021 By Jack Orman
I designed and built a null tester to try to determine if I could detect a difference in op amp chips. The test signal goes into the null circuit and is split with one path going into the op amp being tested, which is set up as a unity gain inverter. The inverted signal from the device under test is combined with the direct signal from the input in a mixer circuit with +40db gain, and is then analyzed for distortion harmonics. The test signal is at 1k Hz at 0 dbv, and the load on the test chip is 5k ohms. You will notice that all 60 Hz interference and its harmonics are removed by the null circuit, as well as any impurities in the input sine signal. Only distortion produced by the op amp will pass through the null mixer to the analyzer.The first chip that I put in the test box was the highly regarded LM4562. As you can see by the graph below, there are no detectable distortion harmonics. This chip is about as distortion-free as any available. The noise floor is also lower than the other chips tested with this circuit.
The next step was to change out the op amp chip. I replaced the LM4562 with an RC4558 also from TI. It should be noted that because there is 40db of output gain, the peaks in the chart are actually 40db lower than they appear. For example, the second harmonic at 2k Hz is diplayed as around -96db but the actual value is -136db. The third harmonic is barely detectable above the noise floor.
A TL072 is shown below and is is similar to the 4558 but with the third harmonic distortion making an appearance. Even so, is is actually at -140db and definitely not audible.
To show how previous generation op amps compare, I found an old ua1458 dual chip and put it in the tester. The second, third and fourth harmonics produced by it are much higher than with the newer chips previously shown. The 1458's transistors are producing additional distortion that is not present in the more refined recent production designs. Again, remember that the peaks are 40db lower than they appear on the graph.
These few tests show that, while there are differences in op amp chips and their performances, the harmonics produced are very low, and far below the threshold of hearing.
This null tester was not cheap to make and used some high performance (expensive) active devices, as well as a 10-turn wirewound precision potentiometer. The results were similar to what I expected but a simpler design would probably do the job and give results that are closer to that found in stompboxes. I've designed a circuit and pc board for further evaluations, and will feature that op amp tester in a follow up article.
If you found this article interesting, then you might also like to read about the testing of the LM358 op amp.
Share this article with your Twitter followers: Tweet
AMZ-FX Home Page
Lab Notebook Main Page
Guitar Effects Blog
©2021 Jack Orman
All Rights Reserved
©2021 Jack Orman