This thread over on the TOMB has some cursory discussion of some relatively interesting guitar stuff. It's not much too far out from things we've discussed around here except for where a young man named Michael Joly writes:
"string return" acoustic guitar pickups (a single Alnico magnet under the strings, the strings wired in series and fed into a current amp, best amplified acoustic sound I've ever heard)
I've asked him to elaborate, but I'm impatient, so I've come here. Anybody know what the heck he's talking about?
Last Edit: Jun 16, 2010 0:09:58 GMT -5 by ashcatlt
EDIT: No, I didn't have it. In fact, I was pretty far off base. Refer to ash's link above and read the whole thing. It's interesting to say the least.
Now, what I wanna know is, why didn't you jump on Jim Williams for his totally false, and potentially lethal to some poor unknowning schmuck, assertion that a 0.022µf cap will protect the player from ".... shocks from badly grounded PA systems"? That's criminal right there, and if you weren't already a member there (presumably with some creds), I'd be forced to sign up and jump right into what could easily go to Full Flame On, me being an unknown newbie and all. Somebody needs to set him straight, post haste. (Hint, hint... )
Keep us posted on what you get back from M. Joly, OK? I might be all washed up, or I might be right on, I'm curious now.
I had thought about jumping on that "safety cap" thing when I first saw it, but I guess I was a bit intimidated since they are like old hands who at least sound like they know what they're talking about. Thanks to your little kick in the pants I went ahead and posted, though.
For the string return pickup, I don't think there are coils between the strings as you've described. The strings are wired together one way or another and then connected to some device - transformer or active stage - to convert the current to voltage for connection to the rest of the system.
I agree that the frets would have to short out any kind of series arrangement, at least if you intend to fret more than one string at a time. Imagine trying to play the thing with a brass slide!
Best I can figure it pretty much has to be all parallel, doesn't it. A brass nut with a wire running down the neck and a metal bridge to which we usually already have a wire connected. I wonder what we do with the noise that would normally be dumped to ground via that bridge connection?
I did a little half-assed testing on this idea last night. Took out my P-style bass and connected an alligator clip to either end of one string. These to the meter. There was never any noticeable voltage change, but then they said it needs a current amp, right? I didn't notice any current flowing from striking the string either. I did notice, though, that the pickup on this bass is about as far from the strings as possible.
Couldn't be bothered to stand up and find a screwdriver, so I took the string between thumb and finger and pushed it down closer to the magnet and wiggled it back and forth manually. This caused the meter to go from the .1uA noise current up to (+ and -).4uA. Seems like a (preliminary, and probably questionable) positive result. More strenuous testing may be required here.
Theres something in this, and I was also curious enough to try a simple test, once when I was 18 and again a few years ago.
So, theres a fixed magnet in the pickup position. When the strings move above it, it induces a very small voltage in it, but with very low impedance. If you can sense this signal between the ends of the string, you have a potential pickup. Being of low impedance, it can provide some reasonably significant current, so as ash points out, a current amplifier is what is wanted to detect it. A transformer can also be used to boost the voltage. I tried it using a mains transformer in reverse, and got a signal enough to drive an input. Not sure if a transformer s essential. I suspect that a simple transistor stage could be the basis of a suitable preamp, allowing it to have a low input impedance, feeding the signal from the string to it through a large decoupling capacitor. A transistor is after all, essentailly a current amplifier.
Wiring the strings together is not so obvious - I can see issues with wiring them all in series, since fretting will tend to short certain strings. Complicates bridge as well. I think the best version would be just to link all strings together electrically at the nut (brass nut would do that), and just take the signal from all strings in parallel, between bridge an nut.
Moving the magnet position will cange the tone, in a similar way to moving standard pickups.
There's been more discussion on the other board, and it turns that much of what JohnH just said came up.
Turns out it was more of a series/parallel hybrid. The one guy talked about a brass nut. The other guy talked about a piece of foil connecting the strings beyond the nut. The strings are alternately connected to the + and - output. I was having a little trouble figuring out how this didn't cause a dead short across the output, but I guess there's enough resistance in the strings to keep that from happening. Here's how I came to figure it out:
I took the pickup analogy and got to thinking about something like an HHH Strat. This would be a total of 6 coils wired into 3 series pairs, with these pairs connected in parallel. If we use N for North Coil, S for South Coil, + for parallel and * for series, we've got (N*S)+(N*S)+(N*S).
Now we take the "series connections" - the two wires of each pair that connect together, which we would short to ground (or shunt to hot) in order to split the HB - and connect all 3 of these together. This does leave all 6 coils functioning just fine. It's a little tough to tell who's in series with who. I'd call it (N+N+N)*(S+S+S).
There was mention along the way of running the final output to either a transformer or a transistor, and also to an opamp wired as a transconductance amplifier.
The parallel thing does seem easier, especially on a guitar with a conductive metal bridge. The big problem here is the necessity of running a wire along (or inside) the neck. With regular pickups the series output would be significantly hotter than the parallel, don't know how that applies here. I still wonder where the noise goes, though...
Definitely going to try it soon. While the folks over on the TOMB were saying that it gives the sound of the whole string, I was also suspecting (as JohnH said) that the harmonic content would depend on the magnet placement for the same reason pickup placement matters.
If only M. Joly has said "ribbon mic", I would have cottoned to the concept right away. Aw well.....
The deal about parallel, series and all that is simple. Think about it - you said "run a wire out to the nut". Why do that when you can use another string to bring the conduction back to the bridge? Let's break down Joly's idea in a different way than you did.
E (the first string) is in series with A, right? That's N*S, in your parlance. And D is in series with G, that's another N*S, right. Ditto for the remaining pair. So, if we combine the N's on one terminal of the output jack, and combine the S's on the other terminal of that jack, we have three series sets in parallel, yes? Again in your diagram-speak, that would be (N*S)+(N*S)+(N*S).
Now, what about that all-brass nut? Well, think of it this way.... do we care if one series set is "shorted" half way through with another series set (also half way through)? No. Why not? (I'd better not leave this one up to the reader, eh?) Because we now have E (the first string) coming from S and going to N through not only A, but also through G and e (the sixth string) as well. That looks like N*(S+S+S), doesn't it? Repeat that twice more, and the formula gets to where you originally arrived at, (N+N+N)*(S+S+S).
Of course, these terms are overloaded (we normally use them in a more standard fashion, but for this discussion, it's OK to "abuse" them a bit), so the analogy isn't quite clear, yet. We have to keep in mind that all the strings are active all the time. Plus, being ultra-low resistance, which gives rise to being a current source instead of a voltage source, we now can get even more simple in terms of what we have to model for xSpice - assuming an average value of resistance for each of the six strings, it formula boils down to nothing more than (1/3 N)*(1/3 S).
One or two more things....
It stands to reason that if all the strings are connected at the brass nut, then fretting them across a steel fret isn't going to electrically "short" them any more than they already are, though it will lower the total resistance a tiny fraction. But depending on the magnetic field strength and on any circuitry following the combined string pairs, that decrease will not affect the overall performance. Although I have to say, at this point, I'm sure an instrumentation amplifier will be more suitable than a transformer at bringing out a usable signal. But that'd be my compromise choice, YMMV.
Rule #1: All Lives Are Final. Make sure that the life you have just been issued is appropriate for your needs, before departing the womb.
Rule #2: In case you don't like the life you have, see Rule #1.
But depending on the magnetic field strength and on any circuitry following the combined string pairs,
I was going to mention field strength. In a guitar pickup, we don't want too strong of a magnet, but for this, it would seem to me that more is better- stronger magnet, more signal induced in the string. Since, unlike a pickup, the magnet isn't connected to any circuitry, noise shouldn't be too much of a concern.
Am I right about that? Would a high-gauss bar magnet be the trick here?
Using an electromagnet might also be an idea. How about using an old guitar pickup with the leads hooked to a battery?
sumfai - i think you may have the series/parallel thing sussed correctly! Three parallel sets of two strings in series. Will need to think further, but the fretting then may not bypass an active string section, as it would if all six were in a single chain. Possibly the nut is not metal, and each string is only conected at the headstock to its series pair. The bridge would need to be electrically in two parts. or maybe 6.
For testing, I wonder if a mic preamp might be enough to get a signal?
Magnets - certainly more field will give a stronger signal. I line of super-gaussy neodiniums under the strings might be the hot version.
Magnets - certainly more field will give a stronger signal
Some older lap steels used to use a thick horseshoe magnet that went both over and under, with the strings through the "U". Of course, the upper part, over the strings, becomes either a handy place to rest your hand or a gawdawful nuisance, depending on one's playing style.
..... the fretting then may not bypass an active string section, as it would if all six were in a single chain. Possibly the nut is not metal, and each string is only conected at the headstock to its series pair.
I think some elucidation is in order here. Hooking up all six strings in series would still have a finite total length. Fretting just one of them at some point will not actually change the resistance, because the "unsounding" portion of the string is still electrically active. Now if two strings are fretted at the same point, a short is indeed created by conduction across the steel (or nickel) fret, and at that point in time, the "unsounding" string portions are indeed out of the picture.
Now contrast that with Joly's vision, three pairs of series all hooked together in parallel.... since there are multiple connecting points, any adjacent string pair will still be active over the entire length of both strings, because the other strings are still active all the way down to the nut. Fret three, four or five strings, and that scenario still obtains. Only with all six strings fretted at the same fret will all of the "unsounding" string portions become electrically inactive.
I think the problem of having all 6 strings in series is more serious than just a change in overall resistance when fretting. Suppose one wired the strings all in series 1-2-3-4-5-6. If strings 1 and 6 are fretted at the same fret, then that fret will bypass strings 2, 3, 4 and 5...no good! The smart thing with the three pairs in series, is that in any fretting scenario, none of the active signal generating parts are ever bypassed. Sometimes the fret will link one pair to another pair, blending their signals before they reach their respective preamps, but that is not a problem since they will all end up blended downstream anyway.
Here's an excercise for the weekend:
We cant easily mock up this 3x2 string wiring on a standard guitar, but we can try an independednt 6x1 wiring. Ground to bridge, and 6 alligator clips from each tuner peg, to 6 mic inputs on a mixer and crank them up! Magnetism provided by whatever standard pups are in place. Record to pc, and post soundclip!
Best to wait until wives are out, or risk the need to explain WTF you are doing.
This was my Strat, with all tuner pegs connected in parallel with alligator clips, and a signal taken from bridge to tuners, into a single mic input. The mixer was entirely maxed out, and you can hear its 30 year old op-amps hissing.
But the concept is good. Putting all strings in parallel has significantly diminished their individual signals, and a single string is much louder. So the much more refined arrangements discussed above should work better.
The tone is interesting, kind of plinky-plonky, and more electric than acoustic.
The magnetic field was just that from the Strat pickups, and I lowered the middle pupe to get its reversed magnet as far out of the way as possible.
I'm definitely going to try it one of these days soon. JohnH, was that each string into a seperate mixer channel like you described above?
I had to get out of bed early today from thinking about this. Bear with me for a minute here.
newey was talking about using stronger magnets. It's important to keep in mind that these are steel strings we're talking about. That is they are not only conductive but also magnetically reactive. They will tend to pull toward the magnet, which may cause some loss of sustain. See Stratitis.
This led me to remember that the wire-in-a-magnetic-field thing works both ways. A wire moving in a magnetic field will have a current induced into it. On the other hand, a wire with a current running through has its own magnetic field, which will react against the field of the static magnet and force the wire to move. Now, we're talking about very small currents here, so it probably won't have much effect. Would it be enough to overcome the magnetic drag I mentioned above?
But then this leads to some new nuttiness! If we could run enough current through the strings, we might end up with an interesting alternative to the sustainer driver concept. For example, what if we fed the output of one of the magnetic pickups into a simiple amplifier circuit, and then the output from there into the strings using this same basic theory? I guess I don't know how much power we really want running through these things since it's kind of tough to play with gloves on. Maybe those tape-wrapped strings? But then, how much can they actually take before they start burning out like fuses?
I’ve had a strip of copper laying around for 5 years now, just waiting to be put to use. Suppose I cut a strip off, so it can be fitted between the saddle and the strings … would that accomplish anything?
By the way, I don’t have a pre-amp - just some guitar amps … can this contraption be plugged into an amp? (perhaps creating a "mic" of sorts by running alligator clips to the tip and sleeve of a ¼” patch cord, the plugging the other end of the cord into the amp?)
If any of that sounds like it might have a hope in hell of working, can you fill in the blanks using something along the lines of a “schematic” of sorts? Maybe connect the dots on something like the diagram below?
I wouldn’t want to accidentally create an ECG machine or something, and I definitely want to avoid those WTFs you mentioned earlier.
You can avoid running a wire down the neck by putting that copper strip at the nut end.
Have you got enough to do that and get some under the bridge end? I think that two separate pieces at the bridge end - one for the low strings and one for the high. Then pick one of these for the + and the other is -.
I'm not sure you'll get much out of a guitar amp without something in between, but I haven't tried it. Got a passive DI laying around? Would that work, if wired in reverse?
I have a multi-effect pedal, which has a built-in tuner that sort of bypasses everything else when activated. I wonder if there would be any benefit to running through the multi-effect, which has at least some degree of volume adjustability?
... I wish I had completed the cMoy that I tried to build a few years back.
I’ve got plenty of the copper, so I can easily place it under the nut as well. However, if I can do that, wouldn’t I be able to just take some copper wire and twist it around the E string, then around A, etc, etc. etc, to “e”, and then attach that wire to the tip of the cord; and could I not also do the same at the bridge, ultimately attaching that wire to the sleeve of the cord?
Or would that cause issues in terms of the parallel connections John mentioned?
I appreciate that this is new territory and you may not have these answers.
Seems to me inconvenient at best to have to try and find something to do with that wire coming from nut end. For a more permanent solution, I think the "half-series" thing works better because all connections to the jack take place at the bridge end.
Tie the 6 strings together at the nut. At the bridge end 3 go to the tip and 3 go to the sleeve and you shoudl be done with it. I honestly don't see any reason you'd have to alternate (E- A+ D- ...), seems you could just take the high 3 one way and the low 3 the other. As John mentioned, this series arrangement should also give you a bit more output than a straight parallel thing.
There was some talk about brass bridge pins being a convenient way to keep the wiring inside the guitar. From there you could run to an endpin jack.
I'm still not completely sure what is best for connecting this thing. This is a very low-Z current generator, which is different (don't ask me how) from the hi-Z voltage generators we're used to working with. Seems sumbody once said that for maximum current transfer we want to try to match the input and output Zs.
Last Edit: Jun 18, 2010 15:15:46 GMT -5 by ashcatlt
Thanks for listening. The strings were all shorted together at the tuner end with alligator clips, and obviously also at the bridge end by the Strat bridge. Only a single mixer channel was used - I just dont have enough clips!
This was an easy first test, and its remarkable that it worked at all because each string is having its signal contribution bypassed by 5 other strings. Hooking up a single string is much louder, per string, so these series/parallel arrangements, which solve the problem of active string sections being bypassed, should be much better.
A properly configured version will be better than that test in several ways: 1 Magnets can be chosen to suit, without losing some signal by having RWRP pickups involved which have opposite field polarity 2. Several x larger signal due to non bypassing of strings 3. Preamp using modern components, with mucj less noise
On preamps, I definitley think that a mic preamp is the right type of circuit. We need a lot of gain, maybe 60db to get up to a sensible line level from the tiny mV signals produced here. What we dont need is a high input impedance, in fact we can have a particularly low input impedance. A standard mic input is around 1k or 2k. That is still high compared to the string impedance, and will not cause any damping/sustain issues due to drawing curent from the strings. However, using such relatively low input impedances as compared to say a 500k guitar amp input helps the amp designer achieve low noise together with high gain. High gain guitar inputs are much noisier than mic inputs.
However, if you guys are wanting to try a test, and if a mic input is not available, then by all means try a guitar amp and turn up the gain. You might get some hiss. Also, depending on how the amp is EQed, it might not be a linear frequency response.
Using an acoustic - yes, I think that might be a very good way to start testing the series/parallel string wiring, provided some magnets or a mag pup can be mounted. Ill mark up D2o's diagram.
Here's a way to hook up an acoustic, based on D2o's image.
Best if the three outputs go to seperate mixer inputs. But for a test, they could all go to one, or to whatever amp is to be used.
I think the problem with taking all inputs to one, is a general decrease in output from each string (but better than the 1x6 version I tested), but also the reduction of signal that you would get if you fret a lower string (4 to 6), and a higher string (1 to 3), at the same high fret. This creates a lower resistance path, that partly shunts the other strings.
EDIT: I just had a look at this arrangement theoreically, by analysing a resistor network on 5Spice. I think there are still scenarios where fretting more than one string at a fret will change output of other strings. I think the only really correct way to do this with no losses or quirks is to ground all tuners (or ground a brass nut), then have 6 separate outputs from six non connected saddles. It means that any frets that are used are all at ground potential and cant change the signal.
To get the ground connection up to the nut, maybe use the truss rod?
Within the guitar, a preamp with 6 input transistors and a summing mixer made from an opamp should do the job
1) the magnet: I have tried using a single coil pickup in the soundhole, and I have tried using just the bar magnet from the pickup - placing it under the strings just behind the nut (on the headstock side).
Are either of those ideas any good?
2) a lack of any sort of preamp
3) combining the three "+" leads into one "+" that goes to the tip of the cord
Sorry - I don't think I have the right equipment to make this work ...
... any ideas? Or is this best left to someone with a proper pre-amp?
...and I have tried using just the bar magnet from the pickup - placing it under the strings just behind the nut (on the headstock side).
Dude! Really? Reallyreally? Did Kyle tell you to do that?
I want you to put on your dunce cap, sit in the corner, read back through this thread, do some kerthunkin' and then write me a 500 word essay explaining why this is wrong. Typed, double-spaced, all that stuff.
Well, I took another stab at it, with the magnet sitting in the same place as before (the good before - near the soundhole), and this time I ran the cord as follows:
- through the multi-effect pedal - volume adjustment on the pedal at maximum level, - "energize" on the pedal at maximum level (a sort of presence- like feature that helps make little amps sound like big amps), - pre on SS amp at maximum level, plus the pre push-pull "pulled", which allows an extra 8 db - "saturation" at maximum level (gain / distortion control) - post at maximum level
There was something being picked up ...
Nothing but a faint, instantaneous scratchy sort of sound when the strings were plucked.
Slightly louder fuzzy sound of very short duration when the guitar was strummed lightly to moderately.
Relatively louder and sharper*, but quite short lived (split second) sound when the guitar was strummed moderately to heavily. It seemed to be more present when a fretted chord was strummed heavily, but was only slightly louder than the actual acoustic guitar, and with no sustain.
So, with everything on the pedal maxed out, and - particularly - everything on a 50 watt amp maxed out, there was very brief, but audible, output while wailing on the guitar.
In contrast, I live in a detached house and, *had I played an electric guitar through the same setup, the feedback would have made my ears bleed and it would have possibly damaged the amp speaker if I had strummed the guitar ... it would certainly have been loud enough to get the attention of the neighbours in other detached houses.
I placed the magnet differently - parallel to the strings, and I placed it directly under the (bass) E string with the greatest mass.
There was not a lot of volume, by any means, but there was definitely a sound that would be "expected" i.e. tone and sustain - as would be achieved through a pickup, but at much diminished volume.
Edit: I should point out that I did not even have to use the pedal this time - just straight into amp. Mind you, the amp volume was at 10, so it's hardly a practical application so far ... but, still, that's a significantly better result.
I also had less than stellar results. I ran wire connecting all 6 strings beyond the nut. At the bridge end I connected the lowest strings together and the highest strings together.
I have to be honest and say that I was so excited to try it that my first test forgot the magnet altogether. Second test used a Spinosaurus fridge magnet, no luck. Then I got out my Fishman soundhole pickup. That gave better results. Into my Bass V-Amp with everything dimed I got almost no level from any, and inconsistent level from each string. This might have something to do with my "wiring technique". I was able to pick out something that might be an okay guitar sound in amongst all the noise.
What I found in this - and it kind of reinforced something I had started to suspect last night on another project - is that I have no idea how to wire a current amp/IV converter/inverting amplifier using the opamp (TL072) that I've got. I can get a non-inverting amplifier running without any trouble, but the inverting version won't even pass my Rickenbacker (which blasts through the non). I should probably start a new thread for this since it's for another project more than this, but I haven't much time so...
9V battery, TL072, current amplifier/IV converter. How to wire it (including battery connections, bias voltage if necessary, etc). This is the best I could find, but even it leaves a number of questions.