Designing a State Variable Filter for Low Impedance Pickups

[Gus]

What about Alembic electronics? 

[HD Evans]

Whoa, that six-pack project is wild.  I'll definitely be following along, and the creator of it might be interesting to bounce some ideas around with.  As neat as hexaphonic pickups are, my personal philosophy is to at least stick with traditional coil geometry. 

I'm aware that alembic has had several state variable designs over the years.  While the basic idea is solid, I'm not as interested in making a really flexible tone control as I am in shaping the character of each pickup...  That's why my system calls for up to one SVF filter per coil, or as few as one per pickup.  Switching is the other challenge I'm facing.  While I'm on the topic of alembic, I know they still have at least one rack unit with a couple of state variable filters and perhaps some EQ...  Have you looked at the price?  I've already done the math for my design, and it can be sold at a similar price point to other active pickups on the market while making enough profit to cover operating costs and future development overhead.  By their standards I should plan to sell a set of pickups for at least $500.  There are far more bedroom guitarists than rockstars, and I would like for all of them to afford my pickups and guitars. 

Since no one has mentioned it, I should acknowledge a wonderful reference I picked up about a decade ago (right after I first had the idea for this system...  which is the reason I was looking for texts on the topic):  "The Active Filter Cookbook" by Don Lancaster.  It dates back to the mid-70s, and I'm pretty sure the copy I got from the engineering library was an original print.  That's where I learned all of the basics I needed to start developing a SVF.  If you want to know all about active filters, yet still be provided with a "ripoff section" in almost every chapter, it's hard to go wrong. 

[HD Evans]

Since there seems to be some interest, this is the basic form I've been working from without any values.  I can add in some values and explain the math, as well as my logic, sometime soon.  While it may be somewhat antiquated, my common practice consists of scribbling a block diagram on whatever paper is nearby with additional details added as necessary.  Rather than uploading a scan or picture from my cell phone, I drew this up so anyone following this thread can be on the same page with me. 

For the sake of simplicity almost all of the resistors are of the same value (I've used both 10k and 100k on different occasions.)  Among other things, this makes what would otherwise be ugly math, extremely simple.  The dual ganged frequency setting pot is chosen by establishing the desired frequency range, in the actual circuit a resistor is added in series going to the inverting input to restrict the range.  Q is a similar deal although it can oscillate if things are pushed too far, although this can be addressed with the addition of a capacitor or by designing within reasonable limits.  The gain setting pot is nothing special, working off of a simple relation with a feedback resistor.  Not all of the resistors going to ground from the non-inverting inputs are absolutely necessary, and there are some guidelines I will post for their values.  Other features I left out include a separate bipolar power supply I built to operate from AC.  It could run on batteries as well.  The input is rectified, passes through a couple of adjustable voltage regulators up to ±18V, and finally some filter caps for good measure.  Also notably absent are any decoupling caps, which never hurt but might not be absolutely necessary...  So knowing myself, they will be included for good measure in the final product.  I anticipate a goodly amount of real estate will be available on the board when the design shifts over to surface mount.   

Again, my apologies for the ultra simple layout.  But it is much better than a drawing on the back of an envelope (and I do have a few sitting around.)  Actually, I think creating a decent design guide for state variable filters is a good idea.  There's not much out there for most designers to work from, and I've picked up more than a few helpful hints after developing this system on and off for the past decade.  I promise to write up a comprehensive guide.

For the record, unless I come up with something really clever, the collection of designs created to make this system possible will be made available under an open source license.  This includes a CNC coil winder, wire tensioner, and other goodies such as pickup measurements that can be as inputs to model the character of classic designs.