Filter issue

[ljcox]

The filter shown below is from an old National Application Note. The AN states that the Fo = 10 kHz.

My analysis gives a transfer function Vo/Vi = 1/(SCR + 1)^2 which leads to the -3dB point at Fo = 0.2152/CR = 9538 Hz.

This agrees roughly with the AN. But when I did the SwitcherCADII analysis (attached), the -3 dB point is at 4563 kHz.

So either National and I have made the same mistake in our maths or I have done somthing wrong in the SwitcherCAD simulation.

Can someone please enlighten me?

 

[audioguru]

Hi Len, You don't have a Butterworth filter, it is a Bessel.
Calculating from the RC values, its response is down -6dB at 7092Hz.
The difference between a Bessel's -6dB response and a -3dB response for a 2nd order filter is 0.6434 so the -3dB frequency is 4563Hz. SwitcherCad is correct. :lol:

EDIT: Numbers corrected

 

[Roff]

For the schematic posted, Q=0.5. For Bessel, Q=0.5771.
That circuit is a filter, but I don't think it has a name, other than 2 pole lowpass.

 

[audioguru]

It is the filter used in a Linkwitz-Riley crossover network for a speaker system. Since it has an even number of poles, it causes a null at the crossover frequency so one driver is inverted. If a Butterworth filter is used then with a driver inverted the frequency response has a +3dB peak at the crossover frequency.

 

[ljcox]

Thanks for the responses. I'll absorb the info & post again if appropriate.