I have an op-amp with a single order low pass filter on the output. A simple series resistor and shunt capacitor. This opamp is only supposed to work with DC so the filter has an extremely large time constant - somewhere around 1000s. About 1MΩ and 1000uF.
I would like to know anyones thoughts on keeping a system like this stable. Obviously the opamp will not be rated on the datasheet to be stable for much more than a few thousand pF since a DC precision opamp is required. The series resistor of 1MΩ I thought would be enough to isolate the capacitor but it appears this is not the case, and some capacitive peaking is occurring at about 1-10mHz, depending on the configuration of the rest of the system. This causes the system to oscillate at these very low frequencies when I add a small white noise source at the input.
The opamp and filter are connected in a control loop. The capacitive peaking (and oscillation) go away with removal of the filter. Unfortunately the filter is required to remove noise from the system above 1mHz, so I cannot really get rid of it unless someone else has an idea for attenuating super-low-freqency signals.
The load, as shown in the diagram, is a parallel resistor and capacitor. The load capacitance is also very large (0.3F supercap). It also causes problems with the stability, but eliminating the LPF fixes the system so I am hoping to see if I can fix the problem there and not change the feedback network. Sorry to black box it, but you can assume its input impedance is very high (like a FET amplifier buffer) so the LPF is not loaded down by more than a couple pA.
I would like to know anyones thoughts on keeping a system like this stable. Obviously the opamp will not be rated on the datasheet to be stable for much more than a few thousand pF since a DC precision opamp is required. The series resistor of 1MΩ I thought would be enough to isolate the capacitor but it appears this is not the case, and some capacitive peaking is occurring at about 1-10mHz, depending on the configuration of the rest of the system. This causes the system to oscillate at these very low frequencies when I add a small white noise source at the input.
The opamp and filter are connected in a control loop. The capacitive peaking (and oscillation) go away with removal of the filter. Unfortunately the filter is required to remove noise from the system above 1mHz, so I cannot really get rid of it unless someone else has an idea for attenuating super-low-freqency signals.
The load, as shown in the diagram, is a parallel resistor and capacitor. The load capacitance is also very large (0.3F supercap). It also causes problems with the stability, but eliminating the LPF fixes the system so I am hoping to see if I can fix the problem there and not change the feedback network. Sorry to black box it, but you can assume its input impedance is very high (like a FET amplifier buffer) so the LPF is not loaded down by more than a couple pA.