2nd order HP & LP filter with variable passband gain

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

VacuumVoodoo

Well-known member
Joined
Jun 3, 2004
Messages
218
Location
Gothenburg, Sweden
Hi,

it' about time i contributed with more than a comment now and then....so here is my first "real" contribution:

A standard 2nd order active filter with Sallen-Key topology has a fixed gain but you can change cut-off frequency easily. Things get screwy when you want to change the passband gain.

Here is the idea which I originally developed for current feedback opamp, it works just as well with our normal opamps.

In principle: since the feedback loop around the opamp forces the inverting input to assume same potential as the noninverting input then the signal path between these two inputs can be looked upon as a virtual voltage follower. So we build the filter around this virtual voltage follower leaving the rest of the opamp to provide variable gain in the passband by varying the feedback resisitor's (R2) value. You can use either a logpot or a switched resistor ladder.

A much clearer PDF with both HP and LP filter curves is here:
http://anacon-tech.com/IPdocs/VARIGAIN_HP_LP_FILTER.pdf

Simulation using Microcap8 and 5532 opamp model:
VARIGAIN%20FILTER.gif


This might save you one opamp in a signal chain since the filter provides gain too. Hope you can find use for this.

I used it originally directly on the input stage in a RIAA preamp to provide effective HP at 30Hz.

Alex
 
Comment: Equal C (1,2) unequal R (3,4) works fine too. I've used the fixed-gain version of this in some low-cost designs when I needed more gain than the Q desired would yield, and it works well. There's a feedthrough term that makes the maximum attenuation in the stopband a little less than the perfect two-pole highpass, but this can be small if the feedback voltage divider impedance is low relative to the rest of the values.

For really large gains you might want to put a big C in series with the bottom of R1 to limit the overall d.c. gain to unity, so you don't develop a large output offset voltage.

The counterpart to this circuit is the fixed-gain variable Q version, where the unity-gain output is taken from the inverting input and the rest of the circuit is the standard equal R equal C Sallen-Key with gain less than times 3. The loading of the following stage has to be taken into account due to its appearing as a parallel load across the feedback divider R.

Brad
 
> idea which I originally developed for current feedback opamp

Just FYI: the unity-gain 2-pole filter is very old. The equal-value filter is popular because you can bulk-order the precision Rs and Cs and not worry which goes where, but the unequal-value unity-gain was old when the 741 appeared.

The modification where you take the inverting input of a gain-amp as your unity-gain reference for the RC is also old. Heck, I was doing it on tube amps when tube amps were not exotic.

> directly on the input stage in a RIAA preamp to provide effective HP at 30Hz.

Check your subsonic noise. With BJT input, that small input cap (plus 1/f noise) can lead to astonishing subsonic noise. I got an inexpensive phono preamp which used a smallish input cap to roll-off subsonic record-warp. But even without the needle in the groove, the subwoofer cone flapped, and at realistic levels (system gain set so recorded piano approximated the grand piano in the room) the power amp would shut-down. It thought any long subsonic noise-peak was DC and tripped the protection. I ended up swapping caps around to get the subsonic roll-off on the output, with a big cap at the input to suck-out some of the 1/f noise.

Still, of course, the idea is very useful in many situations.
 
Hi PRR,

thanks for the history lesson, i'm not going to argue with that :razz:

It happens more often than not that same ideas are developed simultaneusly or over and over again at different places and times.

What I presented in my post is basically the same thing as I had published in Electronic Design, Ideas for Design, Septemeber 1990 p 105.

I think i need another history lesson, who said "There's nothing new under the sun" ?

Alex
 
At an AES convention in '97 there was a long session of amplifier papers. The pattern that emerged was (1) the presentation of a 'new" idea (2) the question period, in which a cleared throat would lead to a query of the form "Are you aware of the work of X...", (3) the presenting author's negative response and expression of general regret.

A person quipped that in the future they would secure the patent before presenting the paper.

The challenges to the papers at the convention were not reported by columnists in Stereophile, only the gushing (highly patriotic in nature, following the affiliation of the columnist) about our lives being surely changed forever by this new stroke of genius.

It is hard but not impossible to do something new. It is even more difficult to do something new of true utility and sufficient novelty to warrant protection. Of course this hasn't stopped B*s* from patenting blatantly prior art and intimidating anyone who dares to criticize.

When I met Gerald Stanley, the amplifier guru at Crown, I had a book with me that seemed to have the basic design idea of his class D amp invention. The book predated the application by a few years. I showed it to him and he looked a bit worried, but after reading the section said that the Crown topology had one more inductor and that the book's modulation scheme was slightly but crucially different. He went on to say that he had presented the idea (after the filing) to a room full of power electronics experts at another convention, and had figured he was pretty safe after none approached the microphone and cleared his throat...
 
Back
Top