I've designed a 6th-order filter using Analog Devices' online filter wizard. The filter has a cutoff frequency of 2MHz, and I'm aiming for 50dB attenuation at 5MHz, with a continued roll-off beyond that point. To achieve this, I've replaced the last stage with a multiple feedback configuration to extend the roll-off.
I have a couple of questions:
Is it a good idea to use a multiple feedback configuration in the last stage for a continued roll-off, or are there better alternatives?
When I perform noise analysis, the online tool indicates minimal noise in the noise spectral density graph beyond the filter's cutoff frequency. However, in LTSpice, I observed another peak of noise at 20MHz. I anticipated that the noise would be suppressed due to bandwidth limitations. Is my understanding correct, and if so, why does the LTSpice simulation yield different results from the online tool's noise graph?