Continue to Site

# How Do I Change Gain on a 4 Pole LP Filter?

Status
Not open for further replies.

#### Hertzilla

##### Newbie level 4
I used TI's "WEBENCH Filter Designer" online to design a 4 pole low pass filter. The cutoff freq. is 10.5 kHz.

The pass band gain of the circuit is unity, but what I actually need is a gain of -6.5 dB @ 1 kHz. The software won't let me enter negative numbers, so I can't determine what I need to change.

I'm guessing that I may need to add a resistor in the negative feedback loop of one or both stages to reduce the gain(?) ...I don't want to change the specs of the existing circuit other than the gain.

-Thanks

Last edited:

You cannot change the gain within the filter itself, without seriously upsetting its operation.

If you need to change the overall gain, some outside gain or attenuation will be needed either ahead of the filter or after the filter.
As you need to reduce the amplitude by about half, a simple resistive voltage divider at the output would probably be the simplest solution.

Hertzilla

### Hertzilla

Points: 2
That circuit cannot readily give a gain of less than one in the passband so using an attenuator at the input or output as Warpspeed suggested would likely be the best way to go.

Why do you need a gain of less than 1?

The Sallen Key topology doesn't support gain below unity, adding a voltage divider will of course work, as suggested.

A multiple-feedback (MFB) topology gives the option of setting a lower gain, but I fear it's not supported by the Webbench designer.

A voltage divider at the output would compromise the output impedance. I'd replace either R1_S1 or R1_S2 with a voltage divider.

You cannot change R1_S2 that will screw up the relative gains within the filter.

But R1_S1 could be turned into an input attenuator, as long as the first stage is being fed from a source impedance of 6K8.

Closest I could quickly get with very common values would be with 14K and 13K in series.
The impedance at the junction works out to 6K74.
Attenuation of that would be 2.0769:1 or 6.35db

You could probably fudge that a bit with a third resistor, or even a potentiometer.

Hertzilla

Points: 2