Independent Low and High Pass Filters

[JFace]

I have found the low and high pass filters on my recording interface to be invaluable to shaping my sound in a very intuitive way. I find this much more natural and pleasant sounding than a parametric equilizer. I realize this can be implemented using a state variable filter (check out the Alembic Like State Variable Filter schematic). I was wondering if there is a smart way to implement an independent low and high pass module into a pedal or even on-board preamp without having to cascade two state variable filters one after the other, which would require at least 7 op amps (buffer + 3 SVF + 3 SVF). I believe a circuit like this would be very valuable as it lets you tune the resonant low and high frequencies of your instrument.

Having a Sallen-Key arrangement is fine except that controlling the resonance also fluctuates the gain. Ideally, there would be a frequency and resonance control for each filter, independent of one another and of gain.

Another arrangement that is of interest is the Fliege filter. It uses two opamps to control frequency and resonance, with the gain pre set. This implementation would require 5 op amps (buffer + 2 LP Fliege + 2 HP Fliege).

Any thoughts on this?

A good article for reference: http://sound.westhost.com/articles/active-filters.htm

[lungdart]

If you put a buffer before your filters, you actually would not be changing the resonance frequency of the instrument. To change the resonance you would need to modify the overall capacitance and inductive reactance of the pickup circuit (Tone controls, cables, and first devices buffer characteristics). Here is a good article explaining this. http://buildyourguitar.com/resources/lemme/

Now as for adjusting the tone with a series LP and HP filter, you might have to experiment on what is the best. There are many passive and active LP/HP filters to play with, if you could grab your the schematic of your recording interface, that would be the easiest way to emulate it. Another idea would be to step different frequencies through the filter, and make a table of the tone at various knob settings. It is tedious, but would give you a goal of what your looking to emulate. You could then fire up your favourite simulator, and play with the different circuits until you get something that closely matches your needs. Then you can move into breadboard territory.

Sorry I don't have more specific information on the different active LP/HP filters. I haven't played with a lot of these, so I do not know how they control or how they sound.

[PRR]

> more natural and pleasant sounding than a parametric equilizer

That's like going to a plastic surgeon to trim your nails.


> Sallen-Key ...controlling the resonance also fluctuates the gain

Not inevitably; there's a trick.

How much Q change do you really need?

And how much frequency change?

> at least 7 op amps

I hate excess pins to wire, but I must note that today opamps are 21 cents each and a whole pile of opamps cost less than the pots, nice knobs, and snazzy box. I hate overcomplication but if it serves a musical purpose, stuff the box.

[JFace]

> more natural and pleasant sounding than a parametric equilizer
Not inevitably; there's a trick.

How much Q change do you really need?

And how much frequency change?

Care to share the trick? I can only see putting the feedback through another op amp, which is exactly what the fliege arrangement does. I don't want anything crazy with the Q...perhaps 10dB. Frequency range for HPF would be 20 to 250 Hz and LPF would be 500 Hz up to 7kHz. Ballpark.

[moosapotamus]

I love the HP/LP combination as well. Did some comparisons of a crossover versus independent HP/LP filters a while back (for an effect loop with clean blend pedal for bass guitar) and concluded that you get a whole lot more tonal flexibility from the independent HP/LP filters (LP on the clean signal path and HP on the effect loop). So, in that pedal the HP and LP filters were actually in parallel. But with nothing plugged into the effect loop, the thing was a really awesome tone shaper.

If you are going to design from scratch, I would suggest making the frequency ranges overlap. Instead of leaving a gap between 250Hz and 500Hz, you might want to make that your overlap region. Or maybe even go with a wider overlap, like 250Hz to 2KHz. The reason I say that is because it would give you the ability to also create band-pass and band-reject (scooped) responses.

Very cool idea.

~ Charlie

[PRR]

> putting the feedback through

Yes; except by stretching there may be no need for another 19 cent opamp.



Drawbacks:

* The internal node has signal up-to 2.8 times higher than the Output node, so the output can clip well inside the supply rails.

* Output impedance is non-zero, up to 360 ohms. This will generally be acceptable driving another filter with 10K impedance.

* Peak boost is sensitive to pot tolerance. Which is often loose. If pot max value runs high, "10" will oscillate. If it runs low you won't get your whole 10dB. You may need to trim the "560" for sloppy pots.

The "100K" are dual gang 100K pot with 10K stoppers. The 10n set the base frequency and would probably be more like 30n. There's no good stock taper for these. Linear is fine for 4:1 range but cramped for 10:1 range. Regular Audio taper is wrong unless you number the dial backward. Dual Reverse Audio 100K is hard to source. You may consider switching caps to cover the wide range.

The 1K is conveniently a Linear. With these values, 500r or "5" gives a Butterworth max-flat alignment, sweet.

I don't see much point in a 20Hz setting because it won't do much to signals 50hz and up. With practical loudspeakers I see 50Hz as a practical low end.

[JFace]

Looks good, I'm going to breadboard it and compare it with the two op amp arrangement and see which will handle the q.

I chose 20Hz for the low end as a functionally "off" position in case I wanted all the low end intact and only adjust the high end filter. I realize there is basically nothing down in that range...Using 47nF, 100K pots with 10k stoppers I get a frequency range of 31 to 339 Hz...good range and standard parts.

In case anyone else cares, the dual gang reverse 100K pots are sold at smallbear, and that's the only place I've ever seen them. I bought a couple a few months ago for this exact purpose. However, with some resistors across the lugs, one could convert a standard linear dual gang pot into an anti log taper, albeit smaller resistance due to the parallel resistance.
http://electro-music.com/forum/phpbb-files/pot_taper_amz_168.pdf

[PRR]

> with some resistors across the lugs, one could convert a standard linear dual gang pot into an anti log taper

No.

Those plots are for POTENTIOMETER connection. This type filter uses RHEOSTAT connection. If you try to pull-down the middle you just get less range.

IMHO, a truly "off" position would switch-bypass the stage. Just going super-low wastes a lot of rotation and still allows some slight bending depending on the Boost knob.

[kingswayguitar]

struggled with this myself.
compromised on a fairly fixed, but deep mid scooping boost.
just saw this chip
http://www.farnell.com/datasheets/156813.pdf
anybody use it?

[PRR]

> just saw this chip

LM1036 is one of the oldest chips in the National book.

It's Delco/GM's take on a Baxandall bass/treble control.

If you are doing Stereo, it is glorious because you control two channels with one cheap pot.

Audio quality, levels, impedance is very much early 1980s car-radio.

There's some neat tricks possible. But I suspect it is not the action that Jface is looking for.

[JFace]

> with some resistors across the lugs, one could convert a standard linear dual gang pot into an anti log taper

No.

Those plots are for POTENTIOMETER connection. This type filter uses RHEOSTAT connection. If you try to pull-down the middle you just get less range.

Busted. Well I'm glad you caught that here rather than me spending an hour pondering why my homemade pot isn't performing as it should. I'm going to look into digital pots with the goal to control as many rheostats/potentiometers as you want with one pot.

IMHO, a truly "off" position would switch-bypass the stage. Just going super-low wastes a lot of rotation and still allows some slight bending depending on the Boost knob.

Point taken...that's why I'd go with the 31 Hz design. Not far from the bass guitar range. I don't want to have a separate switch for each filter (perhaps no switch at all), thus having a small portion of the frequency range outside the usable range of the instrument.