Very low DC gain of pstb simulation

Dear NeoTED,

NeoTED said:

But, pss+pstb shows different DC gain, which is only 20 dB.

Neither pss nor pstb analyses simulate the response to DC. Are you sure the 20 dB value you are observing is not the gain at some non-zero frequency? In other words, the gain of your amplifier may fall off quickly with increasing frequency - especially when the signal amplitude increases from the zero value of a small-signal analysis. This would also explain why the GBW and high frequency performance of your pss/pstb and ac/stb analyses are consistent.

Shawn

Dear Logan,

Thanks for your reply.

The observed 'DC gain' is the low-frequency gain before dominant pole (eg, 10Hz). The GBW are almost same in pac/pstb (with chopper) and ac/stb (without chooper).

While pstb shows very low gain at low frequency, but the same GBW to other simulators.

The method utilizing huge cap and ind to break the feedback loop, cannot converge in pss+pac simulation. Without pac comparison data, I am not sure the correction of the gain result from pstb.

Best regards.

TED

Dear Logan,

Thanks for your reply.

The observed 'DC gain' is the low-frequency gain before dominant pole (eg, 10Hz). The GBW are almost same in pac/pstb (with chopper) and ac/stb (without chooper).

While pstb shows very low gain at low frequency, but the same GBW to other simulators.

The method utilizing huge cap and ind to break the feedback loop, cannot converge in pss+pac simulation. Without pac comparison data, I am not sure the correction of the gain result from pstb.

Best regards.

TED

Dear NeoTED,

NeoTED said:

The observed 'DC gain' is the low-frequency gain before dominant pole (eg, 10Hz). The GBW are almost same in pac/pstb (with chopper) and ac/stb (without chooper).

While pstb shows very low gain at low frequency, but the same GBW to other simulators.

Thank you for your clarification regarding your expression 'DC gain". I now understand you are referring to its low frequency gain before the dominant pole. In fact, your observation from the pstb result now makes sense to me given my lack of understanding of your exact circuit. A chopper amplifier by its nature will attenuate frequencies below the chopper frequency - that is what it is designed to do. Hence, the fact that the pstb analysis shows a much lower gain at 10 Hz than the gain at 10 Hz you observe in your AC/STB simulation without the chopping elements present is expected.

As an example of the expected frequency response of a chopper amplifier, Figure 1 illustrates the frequency response for a chopper amplifier where the corner frequency is varied from 0.50 Hz to 4.0 Hz by changing the clock frequency of the switched capacitor integrator.

NeoTED said:

When I use pss+pac or pss-tran to simulate or derive the open-loop gain, the result is also about 70dB.

NeoTED said:

The method utilizing huge cap and ind to break the feedback loop, cannot converge in pss+pac simulation

I am not sure what methodology you used to estimate the open loop gain using a pss/transient analysis, but if you used a huge capacitor and inductor to break the feedback loop, I would be skeptical in the accuracy of the result without knowing all the details and simulator settings.

In summary, my opinion is that the lower low frequency gain at 10 Hz using the pstb analysis relative to the 10 Hz gain from a conventional ac/stb result appears reasonable and expected for a chopper based amplifier.

Shawn

Figure 1

Figure 8 from reference [1]

[1] Liu Y, Zhou Z, Zhou Y, Li W, Wang Z. A Low-Noise Chopper Amplifier with Offset and Low-Frequency Noise Compensation DC Servo Loop. Electronics. 2020; 9(11):1797.