TLV4112: Sallen Key Filter Stability

Hi Stanley,

An output capacitance of 10 uF to 20 uF is an exceedingly large load for an op amp to drive. The Sallen-Key filter should be stable alone, but adding the very high load capacitance will likely cause the op amp to become unstable. Sometimes the equivalent series resistance (ESR) of the large output capacitor work in conjunction with the capacitance re-stabilize the op amp, but in doing so the bandwidth of the circuit becomes unacceptably lower. Therefore, the op amp would have to rely on another stabilization technique to be stable.

Since the application is a Sallen-Key active filter that has a specific and defines response ideally, the ways to stabilize the stage are very limited. The TLV4112 datasheet in the "APPLICATION INFORMATION - driving a capacitive load" section discusses adding an RNULL in series with the output to isolate the capacitive load. This is a common stabilization technique, but RNULL and the CLOAD create an RC low-pass filter at the output. The result for your circuit would be the intended Sallen-Key active filter, followed by the RC low-pass filter. For example, if  RNULL is 10 Ohms and CLOAD is 10 uF, the -3 dB cutoff of the RC filter would be 1.59 kHz. Since your active filter has a cutoff frequency of 1 kHz following it with the 1.59 kHz filter will result in a filter response that is the product of the two responses. That may, or may not be acceptable.

If you have a schematic for your TLV4112 Sallen-Key filter and the proposed large capacitive load I could do a stability analysis using TINA-TI. Almost certainly the circuit will require some amount of RNULL resistance to attain a usable phase margin. If you provide the circuit schematic we can see what the circuit does in simulation and determine if it can be adequately stabilized.

Regards, Thomas

Precision Amplifiers Applications Engineering

Hi Stanley,

Additionally, is there going to be any other kind of load in parallel with the large load capacitance at the TLV4112 output? If there is that could increase, or may decrease the phase margin.

Thanks, Thomas

Precision Amplifiers Applications Engineering

Hi Stanley,

why not showing a schematic?

Kai

Hi Thomas,

Thanks for the quick response, it was very helpful. I won't be able to go too much more details on the design. One thing I noticed was that there are Op-Amps advertised as "unlimited capacitive" drive like the LM7332. What does that really mean? Figure 32 shows that the phase margin is down to <20° at 1nF 10MΩ. Technically it is still is stable, but it is not something I would like to use with low phase margin. What happens to this Op-Amp with a 10uF load?

Thanks,

Stanley

Hi Stanley,

in any case you must run a phase stability analysis. We could do this for you, if desired.

Kai

Hi Thomas,

Thanks for all the good responses. I have another question regarding the open and/or closed loop output impedance of the LM7332. The data sheet looks to have one data point for closed loop output impedance of 3Ω at 100kHz. Is there any data near DC? I am interested in the output error due to the closed loop output impedance when drawing a constant load of 50mA. In the simulations it looks to be quite significant.

Thanks,

Stanley

Hi Thomas,

I'm not sure how to navigate to their e2e forum. Can you comment on how to get there?

Thanks,

Stanley

Hey Stanley,

For the LM7332, the open loop output impedance was not something that was measured, however we can calculate the Z0 using figure 29. At 1kHz the Zo is 15kOhms, and probably increases as frequency lowers. Thomas is correct that the "unlimited" cap load comes at a bandwidth cost. Additionally, you can increase the cap load drive with compensation techniques, but 10-20uF is a significant cap load.

Do you have a circuit schematic we can use to simulate? Many of our newer devices have the Zo modeled, and will have effective stability simulations

Best,
Jerry

stanley_tlv4111.TSC

stanley_tlv4111_1.TSC

Kai

Kai

Kai