# Help frequency to dB relationship

1. Nov 11, 2005

### frogdogbb

Hello all, I am having trouble figures out the following filter I am designing. I to use a filter with a 24db/oct slope I want 0dB@90hz. So if I want the knee of the cuttof curve to start at 90Hz how do I find the -6dB point so I can design the crossover, by the way this filter uses -6dB as the crossover point as opposed to the usual -3dB.
Thanks

2. Nov 12, 2005

### Staff: Mentor

It depends on the topology of your filter. Cheby, Butterworth, etc. Just put a generic form of the candidate polynomials into Mathematica or even Excel, and plot out the frequency response. That will start to show you how close you can get a 0dB flat spot to the -6dB point with each polynomial. It's way different for the different filter topologies. Keep in mind that in the real world, there can be some disadvantages to the sharper edged filters. Quiz question -- what is generally the main disadvangate of the sharper filters...?

3. Nov 13, 2005

### frogdogbb

I dunno what is the main disadvantage? Size, added complexity? It is an active filter so componet losses are not really an issue.

4. Nov 13, 2005

### Staff: Mentor

Nope. Hint -- you need to order the resistors and capacitors from Digikey to build your filter. Read the datasheets from digikey for the real components, and resimulate the filter response, based on the published tolerances. See any problems?

5. Nov 14, 2005

### Staff: Mentor

BTW, keep in mind that the different topology filters have different characteristics in the passbands and stop bands, as well as different knee sharpness characteristics. The Butterworth is flat and slow, the Cheby is better if you can tolerate some ripple in the passband (not 0dB all across), and the Elliptical is better yet in terms of a sharp knee, if you can tolerate gain ripple in both the passband and stopband.

But the quiz question that I posed above would also factor into your decision for a real world filter (another hint here), especially if you plan on building them in the millions.....