differential elliptic active filter

[speedracer]

Hi! All,

I'm new to this. Could someone help me in designing a differential elliptic active filter? If there is any software can do this, I would like to know whether it exists.

Thanx!

 

[flatulent]

differential?

I suspect you mean to convert from differential to single ended with one op amp. Then elliptic filter and convert back to push pull.

The Nuhertz filters **broken link removed** are available for a 20 day free trial. This program can design passive and active filters.

 

[speedracer]

thanx for the reply

The filter needs to be active, because the cutoff frequency is about 1 MHz. The components for the passive filter, namely inductors, would be too large. The filter needs to be differential, because both the signal source and the output interface device are balanced (differential).

It's easy to implement a such filter on IC, but I am having the most difficult time to do this off chip. The solutions that I am getting so far are extremely ugly. Am I the first person facing this issue?

HELP!

p.s. filter solutions came up with a single-ended active filter design, but I am not sure how to convert it into differential mode.

 

[flatulent]

use an input amplifier

You will have to convert from differential input to single ended with an op amp. Us the classical method where the two inputs are connected to the inputs of the op amp with resistors and there are resistors from the - input to the output and from the + input to ground. Make all 4 of them the same value. To go back to differential output, use the filter output as one side and an inverting gain of -1 amplifier circuit to make the other output.

At this frequency range, the passive filter components start becoming more ideal (less resistance) and you sould consider such filters. To make a differential filter take the single ended one and copy the series parts on both lines and put half the value of the shunt parts between the lines.

What is amusing is that filter theory was originally invented by people working for telephone company laboratories and the filters were differential at the beginning.

 

[RegUser_2]

You should bear in mind the gain product ratio of the OP used. Depend on the filter topology U may need to select OP with a very high gain at 1 MHz.