Actual simulation of real load? Are you playing with us? A simple spice simulation with dubious parameters and graphs does not remotely make it "real world." Nothing in your simulation used the type of AC mains spectrum and noise I am showing. Nor do you have any idea what you are simulating. As I keep explaining, you are confusing the backward effect of a power supply on a load, than the effect of load harmonics and noise on the output of the power supply. We care about the latter, not the former. Nothing you have shown in your simulations makes any sense or back your backward point of view.
It is not meant to simulate what you did. It is to show that you can't filter to anywhere near 120Hz as that will cause massive voltage droop under load at the DC side.
There is nothing remotely dubious about any of the values I have used nor the simplicity of my circuit. If anything a more complex model will further prove my point.
I will state again the obvious. The existence of the diodes means the circuit is not linear so linear circuit analysis does not work.
I am not confusing anything. I have a deeper knowledge of this and while my simulations are simple my understanding is not. Simplistic would be testing a product purely on differential noise, making a claim about the need for just about 120hz filtering, and not even testing for common mode noise rejection (it's primary function) and then declaring based on wrong assumptions and inadequate testing the product does not work, which may be true, but the tests do not prove that conclusively.
My simulations show clearly and frankly irrefutably that attempting to filter down to 120hz, is a fools errand because the actual filter frequency is load dependent and you need those higher frequency harmonics to maintain the DC rail voltage. The voltage is only a sign wave at the source. The diodes modulate that voltage. That is unavoidable unless your power supply is power factor corrected in which the diodes commutate near 0V.