Manual amplifier nulling circuits are simple topologies, typically consisting of just a trimmer pot and a couple of fixed resistors intended to allow offset adjustment by a (usually small) symmetrical fraction of bipolar supply voltages. So, it’s surprising how many variations exist, some very good, some very not. Figure 1 is an example of the latter case.
 |
|
| Figure 1. | The bad: Attenuation of the supply voltages is done with subtraction instead of division, destroying the PSRR of the amplifier. |
This zero trim is a bad idea because attenuation of the supply voltages is done with (V+ – V–) subtraction instead of division. This virtually destroys the PSRR of the amplifier. That’s pretty bad.
Figure 2 corrects this serious defect, achieving attenuation with a proper (R3/R2) voltage divider instead of PSRR-robbing subtraction. But it still isn’t very pretty. Here’s why.
 |
|
| Figure 2. | The ugly: An attempt to correct for the destroyed PSRR can be done by achieving attenuation with a voltage divider instead; however, the supply rails must be symmetrical, leading us back to our PSRR problem. |
Figure 2 can only give the (usually) desirable symmetrical trim range if the supply rails are likewise symmetrical (and vice versa). You could add a series resistor between R1 and the larger rail voltage to fix the problem, but that would (at least partly) revive the PSRR shortcoming of Figure 1. Ugly.
Figure 3 fixes both problems.
 |
|
| Figure 3. | The good: Setting R2 = R3(–V+/ V–)/2 to get a symmetrical trim range. |
All you have to do is set
to get a symmetrical trim range regardless of the actual supply rail voltage ratio.
And I think that’s pretty good.
