Pickup coils from multi-layer PC boards

[bbsailor]

You need lots of turns either on the pickup or on the secondary of the matching transformer, so either wind a pickup, or wind/buy a transformer. I am not sure it makes much difference.

Mike,

Very true. You need between 2,000 to 3,000 turns on the secondary of a matching transformer to match to the high Z input impedance of guitar amps. These turns are on a laminated transformer metal core so it takes less turns to get the same impedance as a pickup with closer to an air core or nearby metal using 6,700 to 10,000 turns.

Joseph Rogowski

[Joe Gwinn]

You need between 2,000 to 3,000 turns on the secondary of a matching transformer to match to the high Z input impedance of guitar amps. These turns are on a laminated transformer metal core so it takes less turns to get the same impedance as a pickup with closer to an air core or nearby metal using 6,700 to 10,000 turns.

In one of the patents, they are getting 8000 turns by stacking 40 thin circuit boards, each with a 200-turn spiral. So, no transformer needed.

[Mike Sulzer]

In one of the patents, they are getting 8000 turns by stacking 40 thin circuit boards, each with a 200-turn spiral. So, no transformer needed.

So this is a forty pc card high impedance pickup. I doubt that this is cheaper than winding a coil.

[bbsailor]

In one of the patents, they are getting 8000 turns by stacking 40 thin circuit boards, each with a 200-turn spiral. So, no transformer needed.

Joe,

That would mean 40 series connections or potential failure points to string each board together with in and out leads on the outer boards. Do you know how they do that to maintain the integrity of all those connections?

Joseph Rogowski

[Joe Gwinn]

That would mean 40 series connections or potential failure points to string each board together with in and out leads on the outer boards. Do you know how they do that to maintain the integrity of all those connections?

They appear to be plated buried vias.

[Hugh Evans]

I should have clarified that matching in this case would still be relatively low impedance... I was bouncing around the idea and nothing more. Ideally the coil will be between 200 and 1k ohms.

[ken]

That would mean 40 series connections or potential failure points to string each board together with in and out leads on the outer boards. Do you know how they do that to maintain the integrity of all those connections?

They appear to be plated buried vias.

I used to work many years ago in a circuit board plant in the plating department, and I can tell you from experience that plating a multilayer board with that many series connected vias is basically impractical. In fact, just thinking about it would drive me sane.

ken

[SonnyW]

I came across the planar transformer core not long ago, and it has made researching design parameters much easier. Ultimately, there's plenty of new ground to tread as they apply to guitar pickups. New concepts are popping up along the way, such as integrating a matching transformer before going into the preamp. Assuming everything works as I hope, this could prove to be a wonderfully flexible platform for anyone who doesn't mind deviating from traditional designs. Cost of fabrication will also likely be far below that of wound pickups. From the standpoint of IP, using PC boards opens another option: copyrighting trace patterns. Companies such as Mesa Boogie have done exactly that with their amps for years. It's not a strategy I plan to employ, but it's nice to have ideas in your back pocket.

Anyways, the weather should be warming up soon (when you start a company in a garage, lack of climate control can be a very annoying reality) at which point I can grab a bunch of PC boards and run them through my CNC router to quickly throw together a proof of concept.

I was a printed wiring board engineer for 5 years a long time ago at Texas Instruments. I am not 100% up to speed on the current technology but I would suppose that a 40 layer MLB with blind plated vias and probably something under 2 mil traces would get pretty expensive fast. (if the substrate is 1/2 inch wide, a 100 spiral would be about 1.2 mil traces/spaces by my rough math - that is half the currently producible standard, which I think is about 3 mils. I'd say a 200 turn spiral on a 1/2 inch substrate is beyond current PWB technology and getting into the IC/ thin film ballpark.) Not that it isn't doable but it would be awfully expensive. Even if the design was less challenging, they would have to be made in high volume on large panels to get the cost down. You might try getting a ball park quote from someone like this PCB Manufacturing | PCB Fabrication | PCB Fab | San Francisco Circuits They appear to have the necessary techniques.

[Steve Conner]

Yup. 40 layers is crazy. Multi-layer boards for consumer electronics usually have 4, 6 or 8 layers. Cell phone and PC motherboards might have a few more, and might have blind or buried vias, as they help to get the traces out from underneath big complex BGA chips.

[Hugh Evans]

Yeah, I want nothing to do with a 40 layer board and have no need for extremely fine traces. My only interest in exploring this construction style stems from experiments with low-z designs. I believe fabricating a traditional high impedance pickup from pc boards is nothing short of an exercise in futility. It is entirely possible that nothing useful will result from this plan, but as the saying goes: one test is worth a thousand expert opinions.

[ScottA]

Yeah, I want nothing to do with a 40 layer board and have no need for extremely fine traces. My only interest in exploring this construction style stems from experiments with low-z designs. I believe fabricating a traditional high impedance pickup from pc boards is nothing short of an exercise in futility. It is entirely possible that nothing useful will result from this plan, but as the saying goes: one test is worth a thousand expert opinions.

Forget the boards, just do it on a chip.

Then you could incorporate everything you need right there, potentially, too.

Back ends on high end chips have been 10+ copper layers for years and you could easily get 1000 turns per layer (at line widths << 1 micron - critical dimensions are < 100 nm these days).

[Steve Conner]

We do use thousands of turns for a reason though. As the turn count goes down, first you need a preamp to be compatible with the usual musical instrument amps. Then with fewer turns still, op-amps won't cut it and you need a discrete preamp running high current in its input stage to get a low noise figure. The high current will drain 9v batteries in no time, so you're incompatible with the most common method of powering active pickups.

Using a transformer to step up the output is an elegant solution. But it's only elegant within the constraints of a low-impedance pickup. We could save money by combining the transformer right into the pickup, and we're back where we started!

[Mike Sulzer]

We do use thousands of turns for a reason though. As the turn count goes down, first you need a preamp to be compatible with the usual musical instrument amps. Then with fewer turns still, op-amps won't cut it and you need a discrete preamp running high current in its input stage to get a low noise figure. The high current will drain 9v batteries in no time, so you're incompatible with the most common method of powering active pickups.

Using a transformer to step up the output is an elegant solution. But it's only elegant within the constraints of a low-impedance pickup. We could save money by combining the transformer right into the pickup, and we're back where we started!

But you can cut down the number of turns by a few times and still get excellent SNR with an op amp with bipolar transistor input. You need to load with a C if you want the same resonant frequency, but there is no problem with that. Such a preamp could go in the guitar or at the amp. Cable C is not much of a problem a this lower Z.

[Steve Conner]

I totally agree, but I think the number of turns required for that is way over what you could reasonably achieve with a coil made of PC board traces.

[Mike Sulzer]

I totally agree, but I think the number of turns required for that is way over what you could reasonably achieve with a coil made of PC board traces.

Exactly! The easy cheap way to make a pickup that can easily achieve a wide variety of sounds, is little affected by cable capacitance, and is a good match for modern electronics is to reduce the number of turns a few times and use somewhat larger wire. It is just not what a lot of guitarists want.