Small buffer for mixing pickup signals?

[Chris Turner]

Hello,

If you've seen some of my other posts, you know that I'm playing around with a multi-coil pickup design.

I was curious about trying to configure these such that I could have an output for each pair of coils under each string, and a buffer/mixer to combine the signals. Do any of you know of a small non-boosting buffer/mixer that could be used for this purpose? Ideally, I'd like to include the buffer/mixers on the PCB I'm using as the baseplate for the pickups.


Thanks in advance!

[Mike Sulzer]

Unless you are planning on bringing out a separate output for each string, why use an amplifier on each pair of coils? You could instead just put them in series, then use a buffer if you want, but I do not think that you have to use a buffer at all. It often works out that the inductance of the series combination is close to what yo get with a normal pickup, and so you will get about the right tone shaping.

Hello,

If you've seen some of my other posts, you know that I'm playing around with a multi-coil pickup design.

I was curious about trying to configure these such that I could have an output for each pair of coils under each string, and a buffer/mixer to combine the signals. Do any of you know of a small non-boosting buffer/mixer that could be used for this purpose? Ideally, I'd like to include the buffer/mixers on the PCB I'm using as the baseplate for the pickups.


Thanks in advance!

[rjb]

I'm sure you have your motives- if you want to try buffering each pair of coils, so be it.
Would a simple op amp summing circuit fit the bill?
Are you looking for a 6-input mixer SMT chip?
(I don't have any answers - just trying to understand the question.)

[Mike Sulzer]

The problem with this circuit is noise. Op amps are not particularly quiet, especially if you want to use one designed for low voltage and current. The 6x mixing increases the effective noise by almost 16 db.

A quieter way would be to use 6 JFETs, as common source amplifiers, with a common drain resistor. I would buy a bag and pick out six with good matching. This would have to be carefully biased, and you would not bypass the drain resistors to keep the net gain not too much greater than one.

I'm sure you have your motives- if you want to try buffering each pair of coils, so be it.
Would a simple op amp summing circuit fit the bill?
Are you looking for a 6-input mixer SMT chip?
(I don't have any answers - just trying to understand the question.)

[ATTACH=CONFIG]43469[/ATTACH]

[Chris Turner]

Unless you are planning on bringing out a separate output for each string, why use an amplifier on each pair of coils? You could instead just put them in series, then use a buffer if you want, but I do not think that you have to use a buffer at all. It often works out that the inductance of the series combination is close to what yo get with a normal pickup, and so you will get about the right tone shaping.

Could be (err, likely is) a misunderstanding on my part. But I was thinking that using a mixer/buffer to blend the four signals would be different than, say, connecting each pair of coils in parallel, and that it could negate the "interplay" between them. Kinda like difference between 2 Jazz pickups in parallel with a passive circuit, and the same in a buffered active circuit.

Again, I'm probably misunderstanding something here.

[David King]

Chris,
Buffering coil rails separately probably only makes sense if you are wanting to pan between them, that typically wouldn't be the case in a humbucker design with a fairly narrow aperture between the coils. Buffering pairs between strings probably wouldn't make much audible difference since each string is sending in a completely different signal so not much phase differential between them. That said it would be fun to try it out and see if it does make a difference. Highlander developed a hex buffer with Rick Turner back in the day but i believe it was for piezo inputs. Graphtech also sells their G.H.O.S.T. Hexpander for their piezo saddles which can be summed. It's quite inexpensive if you have an OEM account, under $50 IIRC

[rjb]

...Highlander developed a hex buffer with Rick Turner back in the day but i believe it was for piezo inputs. Graphtech also sells their G.H.O.S.T. Hexpander for their piezo saddles which can be summed. It's quite inexpensive if you have an OEM account, under $50 IIRC

Many years ago, I met a guy who had a similar setup installed in a Takoma acoustic. I'm pretty sure it cost him several times the value of the guitar. The bone saddle was cut into six segments to avoid crosstalk. He used the getup solely to route the signals from the lower strings to an octave doubler.

-rb

[Mick Bailey]

This came to mind. Those 22k source resistors could increase to reduce battery drain. Output is low-impedance, so needs low-value pots (47k or less).

[Chris Turner]

Yeah, it's looking more like there won't be much difference between using one of these circuits, and simply connecting each pair of coils in parallel.

[charrich56]

Chris,

I don't think you are going to get a pleasing enough amount of sound versatility, or musically useful differences in tonality, to justify the expense and complexity of coil isolation and mixing, unless you can also experiment with combinations of summed and phase inverted coils, and phase shifted coils, as in "HOOP" (half out of phase) combinations with a capacitor. That can all be done passively and go to just one buffer to isolate the passive PU's from cable capacitance.

A lot of the coil switching and series/parallel stuff which is done to get different tones, really has a lot less to do with pickup/coil position on the strings, and more to do with how the series and parallel combinations of pickups and coils affect the resonant frequency and Q of the combo. Partial cancellation of signal can get some good tones too; I'm not knocking the Peter Green out of phase Les Paul tone.

You might be interested in my current approach where I'm using a buffer both to buffer the signal, and to implement a variable capacitance with a pot, which allows a really useful range of tunable resonant frequencies by directly interacting with the inductance of a given pickup or coil combination. All the switching and wiring wizardry works with this approach but you need less of it IMHO since you have more direct and fine-grained control of tonality with the variable resonant frequency pot.

I have a buffer design which uses 2 JFET's, that I'm pretty pleased with, and has both low current draw and a low output impedance, which is needed to pull off the variable capacitance trick. It's in the thread below but you might have to wade through the thread somewhat to see it. You can also PM me for a current version.

http://music-electronics-forum.com/t43081/

Take care,

Charlie

[charrich56]

Here's just the buffer schematic FYI. This is called a Borbely follower, after Erno Borbely. It's a JFET version of the White cathode follower used in the tube (valve) world. I'm using this with a Linear Systems LSK389A dual low noise JFET , with slightly different component values. This will work as drawn with J201's but you may need to select/match J201's and adjust the 100 ohm source resistor for correct current.

[salvarsan]

I'm using this with a Linear Systems LSK389A dual low noise JFET , with slightly different component values.
[ATTACH=CONFIG]43536[/ATTACH]

The Linear Systems parts are sourced by Trendsetters.com and Nacsemi.com.

The LSK489 is a lower gate capacitance version at about the same $6-8 unit price.

Bob Cordell/Cordell Audio wrote an interesting whitepaper on designing with it cf. the LSK389A.

[charrich56]

Linear Systems will also sell direct at a reasonable price (6.00 each for the duals, usually about 4.00 each for the singles.) Kevin Hicks over there can hook you up. Linear also makes a quality J201 of which I have some, but I haven't tried yet due to some time constraints.

[charrich56]

The LSK489 is a lower gate capacitance version at about the same $6-8 unit price.


I like the LSK489A but the Idss is a bit high for this application. The reduced gate capacitance is good for certain designs, but the LSK389A is definitely a winner for power consumption and the small amount of extra gate capacitance isn't much of an issue when we're dealing with piezo and magnetic pickups. I would use the 489A for anything where I could tolerate 2-4 mA current consumption on the buffer, and I wanted .001% instead of .01% distortion.

With the 389A (and J201,) as long as you're running about 500 microamps through the pair, you're good. For an onboard buffer, I needed to stay as low current drain as possible since for certain applications (Les Paul type guitar with 2 volumes and 2 VR tones, and magnetic/piezo combined pickup instruments) we need 2 buffers. Two of these 389 buffers on one board will be about 1.3 mA which gives in theory, 400 hours+ playing time with a 550 mAH 9V battery.

[salvarsan]

. . . but I haven't tried yet due to some time constraints.

There are way too many interesting projects tickling the Curiosity Bone. I got caught up in making an accurate platinum thermometer interface, meselfs. Board layout is a big chore same as ever.

-drh