I want to implement this
analog front end with the op-amps in the PIC32MK1024GPD064-I/PT. I might buffer the inputs before R5 and R6 too, but for now let's assume that the source has just a 50k input resistance to the inverting input of U1. The datasheet gives the minimum stable closed-loop gain, OA4 and OA16, as 8. Open loop gain is 90dB and GBWP is 10MHz. Both op-amp circuits have a nominal closed-loop gain of 1, so the first op-amp pole should be around 10MHz as well.
I added C5 to form a 60kHz low-pass with R7 in the feedback loop, so U1's feedback network's gain should drop to -40dB by 10MHz. The R9-C3 lowpass in U2's feedback loop will bring the feedback network gain there close to -60dB by 10MHz. Each of these filters will give 90 degrees of phase shift well before 10MHz. So my question is, is there any chance the op-amp's poles will drop the open-loop gain 50dB by the time they add 90 degrees of phase shift? If so, this should result in a loop gain of less than 1 at 180 degrees phase shift, which will make each circuit stable. But it seems like filters will typically show a phase shift much more quickly than they'll attenuate to that degree.
Is there anything I'm missing in this analysis? How is dominant-pole frequency compensation of this sort viable outside of extremely low signal frequencies or extremely fast op-amps?
analog front end with the op-amps in the PIC32MK1024GPD064-I/PT. I might buffer the inputs before R5 and R6 too, but for now let's assume that the source has just a 50k input resistance to the inverting input of U1. The datasheet gives the minimum stable closed-loop gain, OA4 and OA16, as 8. Open loop gain is 90dB and GBWP is 10MHz. Both op-amp circuits have a nominal closed-loop gain of 1, so the first op-amp pole should be around 10MHz as well.
I added C5 to form a 60kHz low-pass with R7 in the feedback loop, so U1's feedback network's gain should drop to -40dB by 10MHz. The R9-C3 lowpass in U2's feedback loop will bring the feedback network gain there close to -60dB by 10MHz. Each of these filters will give 90 degrees of phase shift well before 10MHz. So my question is, is there any chance the op-amp's poles will drop the open-loop gain 50dB by the time they add 90 degrees of phase shift? If so, this should result in a loop gain of less than 1 at 180 degrees phase shift, which will make each circuit stable. But it seems like filters will typically show a phase shift much more quickly than they'll attenuate to that degree.
Is there anything I'm missing in this analysis? How is dominant-pole frequency compensation of this sort viable outside of extremely low signal frequencies or extremely fast op-amps?