Pickup coils from multi-layer PC boards

[Steve Conner]

Isn't this just what Les Paul did with his Recording guitar in 1972? Maybe the market is finally ready for it now...

[Joe Gwinn]

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.

I think the technology supports building 40-layer boards; the question is if the market supports the volume that would be required to make it economical enough.

As for reliability, I doubt that manufacturers care that much. This isn't exactly a safety-critical part, so just make a lot, test them, and toss the failures.

For machine-made pickups, what is a common first-year failure rate? Cannot be all that low, on the scale of present-day computer boards.

[SonnyW]

I think the technology supports building 40-layer boards; the question is if the market supports the volume that would be required to make it economical enough.

As for reliability, I doubt that manufacturers care that much. This isn't exactly a safety-critical part, so just make a lot, test them, and toss the failures.

For machine-made pickups, what is a common first-year failure rate? Cannot be all that low, on the scale of present-day computer boards.

Yes, there's no question the technology supports 40 layers and more, the pcb house I posted in the link advertises 40+ layers with 3 mil traces and spaces capability. Back in the 70's where I worked I recall we were making 26 layer boards for military and NOAA. So it is a very mature technology. But the issue that would be the challenge to the current technology would be the linewidth and spacing needed to get enough turns in a small footprint. Sooner or later it might also be possible to print multilayer boards directly with a 3D printer using Selective Laser Melting (SLM) techniques, as here:
3D printing with metal: The final frontier of additive manufacturing | ExtremeTech

On your comment about testing them and tossing them, one of the products we made bunches of back then was the boards for handheld calculators. This was in the mid-70's and they were the newest high tech thing. We were making them by the hundreds of thousands. We had an automated testing machine that checked each board using spring pins, and if the board was good, it moved on the conveyor belt to the next station. If it didn't test good, the machine sprayed a paint dot on it and a jet of compressed air kicked it off the conveyor belt to a chute that went to a scrap barrel. A later version of that machine punched a hole in the bad boards instead of the paint.

That said, I think all of this high tech is way beyond the intentions of the OP for a Low Z pickup on a board. It is just fun to conjecture about the possibilities.

[David Schwab]

Isn't this just what Les Paul did with his Recording guitar in 1972? Maybe the market is finally ready for it now...

You mean low impedance? Yes, but his pickups had a lot of wire on them. It was just big wire.

[Mike Sulzer]

Isn't this just what Les Paul did with his Recording guitar in 1972? Maybe the market is finally ready for it now...

As described here:
Les Paul?s Favorite Les Paul ? The Recording Model

that is pretty much what I am describing, but I see no need to use such a low impedance. LP was still thinking in terms of something like mic impedance using a matching transformer. In the late sixties, I used solid state mixing consoles using bipolar transistors; they used transformer input, but the secondary impedance was not as high as intended for use with tubes, and so the concept of using the best impedance for good SNR was well understood. So LP could have designed a guitar with that had intermediate impedance that would not have used a transformer, but would have gone to a solid state preamp. However, he did not.

He did, of course, have the tone controls, presumably using the inductance of the pickup coil to good advantage, although I do not know that. For example, was the classic rock Gibson humbucker LP sound one of the possibilities, or was that a sound the real LP did not like, or one possibly excluded by limitations in the matching transformer or the pickups? Certainly it is easier to do the electronics today for a wide range of sounds, but there still are limitations.

The claim that you could get almost every sound from it is dubious. String sampling is important, especially on the bass strings, and there is no way to get true single oil sound from a pickup that samples in two places, and no way to get classic humbucker sound from a single coil or a humbucker with different spacing.

[bbsailor]

As described here:
Les Paul?s Favorite Les Paul ? The Recording Model

that is pretty much what I am describing, but I see no need to use such a low impedance. LP was still thinking in terms of something like mic impedance using a matching transformer. In the late sixties, I used solid state mixing consoles using bipolar transistors; they used transformer input, but the secondary impedance was not as high as intended for use with tubes, and so the concept of using the best impedance for good SNR was well understood. So LP could have designed a guitar with that had intermediate impedance that would not have used a transformer, but would have gone to a solid state preamp. However, he did not.

He did, of course, have the tone controls, presumably using the inductance of the pickup coil to good advantage, although I do not know that. For example, was the classic rock Gibson humbucker LP sound one of the possibilities, or was that a sound the real LP did not like, or one possibly excluded by limitations in the matching transformer or the pickups? Certainly it is easier to do the electronics today for a wide range of sounds, but there still are limitations.

The claim that you could get almost every sound from it is dubious. String sampling is important, especially on the bass strings, and there is no way to get true single oil sound from a pickup that samples in two places, and no way to get classic humbucker sound from a single coil or a humbucker with different spacing.

Mike,

The Shure mic matching transformer A95A with a 1Meg ohm load on the secondary presents an input impedance of 3.819K , 122.38mh and a Q of .2016 at 1Khz and 2.894K ohms 1,246.6mh and a Q of .3330 at 120Hz (measured with the Extech LCR meter). Typically you want the transformer load to be about 10 times the Les Paul low impedance pickup source impedance so that there is a maximum voltage transfer and the best S/N Ratio. That would place the pickup in the 200-250 ohm range with a volume pot value of 2.5K ohms.

As an alternative pickups could be wound to an intermediate impedance of about 1000 ohms impedance and then fed into the mic mixer 10K line input impedance with satisfactory results with the output of the mixer being fed into the a guitar amp input. The mixer then acts as the intermediate preamp with some additional tone controls. At this impedance the cable capacitance would have a minimal effect.

The Les Paul on board decade switch only works well in the low impedance mode. It is sort of like a Variatone but scaled to operate at an impedance range of about 100 to 200 times lower. Switching in the high impedance transformer works less well compared to using the guitar low impedance out with a mic matching transformer such as the Shure A95A or A95U mounted at the amp.

Joseph Rogowski

[Mike Sulzer]

Mike,

The Shure mic matching transformer A95A with a 1Meg ohm load on the secondary presents an input impedance of 3.819K , 122.38mh and a Q of .2016 at 1Khz and 2.894K ohms 1,246.6mh and a Q of .3330 at 120Hz (measured with the Extech LCR meter). Typically you want the transformer load to be about 10 times the Les Paul low impedance pickup source impedance so that there is a maximum voltage transfer and the best S/N Ratio. That would place the pickup in the 200-250 ohm range with a volume pot value of 2.5K ohms.

As an alternative pickups could be wound to an intermediate impedance of about 1000 ohms impedance and then fed into the mic mixer 10K line input impedance with satisfactory results with the output of the mixer being fed into the a guitar amp input. The mixer then acts as the intermediate preamp with some additional tone controls. At this impedance the cable capacitance would have a minimal effect.

The Les Paul on board decade switch only works well in the low impedance mode. It is sort of like a Variatone but scaled to operate at an impedance range of about 100 to 200 times lower. Switching in the high impedance transformer works less well compared to using the guitar low impedance out with a mic matching transformer such as the Shure A95A or A95U mounted at the amp.

Joseph Rogowski

Thanks for the info, Joseph. That is an unusual schematic. The treble bleed is not needed for the usual purpose since the impedance is so low, so maybe its purpose is to help with perceived loss of highs as volume is reduced? The tone control has a bypass position, except then the rear pickup bypasses the pickup select switch through a .47 microf cap while the front pickup similarly bypasses it with a 150 ohm resistor. Odd; a special way of adding the pickup outputs independent of the other functions.

[Hugh Evans]

What Les Paul was using serves as the starting point for what I'm interested in trying out. I wound several low-z pickup coils, and found the results very intriguing. The sound of a pure flat response through the audible spectrum is very different from the high-z designs we have all become accustomed to. I am testing designs using both single coil and humbucker dimensions and coil geometry because I do not believe it is feasible using simple methods to get one to simulate the other. Each coil will be fed through a gain stage, then a state variable filter, and finally mixed. All of the results so far have been promising, and provided I take this all the way to production, I also plan to incorporate a switching system to quickly change the tonal character of the entire system.

[David Schwab]

As described here:
Les Paul?s Favorite Les Paul ? The Recording Model

that is pretty much what I am describing, but I see no need to use such a low impedance. LP was still thinking in terms of something like mic impedance using a matching transformer.

He actually thought that all the modern guitar amps were wrong, and that they should have low impedance inputs. I'm sure he probably recorded his guitar direct.

For example, was the classic rock Gibson humbucker LP sound one of the possibilities, or was that a sound the real LP did not like, or one possibly excluded by limitations in the matching transformer or the pickups?

That pickup did not exist when Les made his first pickups. It is said he used coils from an electric clock.

The claim that you could get almost every sound from it is dubious. String sampling is important, especially on the bass strings, and there is no way to get true single oil sound from a pickup that samples in two places, and no way to get classic humbucker sound from a single coil or a humbucker with different spacing.

have you listened to Les' music? He used a very bright clean single coil tone. But he did get a lot of tones, and for example often used his guitar for the bass parts, such as in this track.

[David Schwab]

[Hugh Evans]

I believe it's safe to say that Les did the best he could given the limitations of technology at the time. We, as musicians, are lucky that physics allows musical tones from high impedance coils. Reviewing the evolution of pickups it seems clear that the first guitar pickups were decidedly underwound by modern standards. Indeed, I have a personal preference for humbuckers with alnico ii magnets and DC resistance of right around 8K (Duncan's Pearly Gates are a good example); single coils with alnico v and resistance just under 6K (I have a set of p'ups from Curtis Novak along these lines that are particularly nice.) Without any other changes, these are tonally on the bright side, but players over the decades dealt with this primarily by using extremely long cables rather than RC networks to darken their tone. This practice is well documented for Hendrix, Clapton, and SRV to name a few. Later, the "hot" pickups were introduced to create darker tones while sending out a stronger signal to help push input tubes into overdrive. However, this came at the cost of flexibility: a bright pickup can be manipulated to take on a darker character, but going in the opposite direction is questionable at best. Musicians and pickup makers, in my opinion, should have started to re-evaluate their approach by sometime in the 90s.

Now, we can have our cake and eat it too. Low-Z is a good thing, because with effectively unlimited bandwidth all of the tonal character is present to begin with. It's much easier to decrease range than increase it from a given coil or pair of coils. Will the tonal palette perfectly duplicate every great pickup in history? Not exactly. The science is pretty clear that the basic designs of humbuckers and single coils provide the best groundwork to generate a broad tonal palette for each pickup type. Aside from personal interest, this project is being taken on for several of my musician friends (almost all of whom are signed to Metal Blade records and playing on 8-string guitars.) All of the factory installed pickups, mostly EMG, and designed to output one tone. You might think that a melodic death metal guitarist only needs one tone, but I can assure you this is not always the case. One of the guys I'm working with went to music school with me, and we met as classical guitar performance majors. He desperately wants a setup that will allow him to have a sparkling clean tone along with a face melting crazy driven tone as well. Low-Z pickups with a well designed active pre-amp addresses both needs.

I'm far from the first person to recognize that innovation can be achieved by standing on the shoulders of giants, and that's exactly what I'm doing. Dealing with the limitations of high Z designs seems silly with the widespread availability of consistently high-quality ultra low-noise/low-power solid state electronics. Will they be for everyone? I doubt it. I'm sure I will keep playing my vintage and traditional-style guitars as they are. But it's worth understanding and refining new concepts so, perhaps one day, more people will come to appreciate their potential. I plan to use them as standard equipment on all of my upcoming production line of guitars and basses.

[Steve Conner]

One thing that always worried me: Can a low-Z pickup with preamp ever have a good enough noise figure to work with a high-gain metal amp?

[Mike Sulzer]

One thing that always worried me: Can a low-Z pickup with preamp ever have a good enough noise figure to work with a high-gain metal amp?

Judging from the extremely good SNR achived with some low impedance microphone preamps, I think you would have no trouble. But why go that low? I see no advantage. As you have pointed out, this could require significant bias current, not a problem if the preamp goes at the amp end and is powered there. But if the preamp goes in the guitar, do not use a very low impedance. It has no real advantage in that case since you are isolated from the cable capacitance.

[Hugh Evans]

Even with without cable loads it makes sense to wind for lower impedance, and I plan to start with something on the order of 1K DC as that should send the resonant peak well beyond the audible range for humans and provide a flat response through the useful range. This should allow for a good match to the pre-amp/filter circuit without using transformers. I want the end user to choose what response curves they want to use rather than forcing them to start with some arbitrary traditional tone first. Why stop with what is already available when there are some wonderful tones available in the spectrum most players have never been able to experience?

[Mike Sulzer]

Even with without cable loads it makes sense to wind for lower impedance, and I plan to start with something on the order of 1K DC as that should send the resonant peak well beyond the audible range for humans and provide a flat response through the useful range. This should allow for a good match to the pre-amp/filter circuit without using transformers. I want the end user to choose what response curves they want to use rather than forcing them to start with some arbitrary traditional tone first. Why stop with what is already available when there are some wonderful tones available in the spectrum most players have never been able to experience?

I think that is the right basic approach. But what is not clear is if we all understand what exactly we mean by pickup impedance; it is after all, an inductor and changes with frequency!