Discussed at the time:

The 101 of analog signal filtering - https://news.ycombinator.com/item?id=40946165 - July 2024 (20 comments)

(p.s. reposts are fine after a year or so; links to past threads are just to satisfy extra-curious readers!)

Thanks for reposting this. I missed it last year, and as an EE with a misspent youth (I originally got a BA in media), I have frequently struggled with the calculus. One of the few regrets is that I didn’t learn advanced math back when my brain was primed for it. Learning calculus in mid-life is tough.

>One of the few regrets is that I didn’t learn advanced math back when my brain was primed for it. Learning calculus in mid-life is tough

in the same position. tried to get into deep learning last year and found my calculus knowledge lacking. If you're looking to get back into it efficently, checkout mathacademy.com. i'm currently doing it myself and I'm on track to make another stab at getting into deeplearning next year.

On the other hand it's literally never been easier. Seriously, don't so much spend the time learning a discipline as learning to learn with any of the free online LLM tools for any subject and if you've actually got the gumption you'll go far. The 2nd qualification in life is often the toughest but the 3rd and 4th are often the easiest.

I'm partial to MathAcademy for learning math as an adult learner. Not cheap but so so so worth it.

TI App Notes and their design app, LT Spice, and the late Bob Pease are your friends

Really great intro to this topic -- rc filtering is rarely explained this intuitively.

I wonder if there's a blog post in the three levels of complexity of calculating Sallen-Key filters?

https://upload.wikimedia.org/wikipedia/commons/3/3f/Sallen-K...

Super simple, really stable, and very easy to calculate. The cutoff frequency and Q can be set independently, and while making the cutoff variable can be tricky (as seen in various analogue synthesizers) you can vary the Q by varying the gain of the amp. If you assume it's just a unity gain buffer (in some designs it's just an emitter follower) getting the component values is easy.

I don't know why nobody seems to ever begin with: A filter is a voltage divider with frequency-dependent complex impedance.

If you can calculate what a resistor network does, or get a computer to do it, you can also calculate what a filter does (and you'll definitely want a computer for that).

(Going backwards, from a desired frequency response to a filter design, takes more learning. Luckily some very smart people already designed filter templates where you can just plug in the frequency you want)

this is exactly how I intuitively approach filters as an applied engineer. Does it give a ground path to DC (low frequencies) and pass the higher frequencies, or vice versa. If we change the capacitance how does the frequency response of the divider change?