Mike Sulzer

"(I can't get to Mike's conclusion about stacked humbuckers from this particular test)" My understanding of the stacked humbucker (original old kind) is that it is just two coils, one wound on the top half of the magnets, the other on the bottom half. They are connected in opposite polarity. A source of magnetic field that is far away changes little in magnitude or direction over the coil separation and so the opposite polarity causes cancelation. A source that is close by and small (such as the string) falls of with (1/r)^3 in air. Does it fall off with the pickup present? I am saying yes, although not as fast, and so the string signal is reduced, but does not completely cancel. If Peter's test is valid we have some explaining to do. The other possibility is that the source of the signal that Peter used is large, and so its signal does not fall off much over the length of the magnet. I think a remember from another thread that he uses a small dynamic speaker or headphone driver? That would be too big. The first thing to do is just to hold the pickup over the strings in both orientations and see if there is a big difference. "Yep, by adding a dummy coil in series, you'll change the resistance and probably the response. Maybe measure the inductance with the two coils in series, then again on the main coil in series with a surface-mount resistor (SMD resistor) of the same resistance as the dummy coil (which I would think would not do as much to the inductance as the dummy coil itself)." Any old resistor will work. A pickup coil is about a Henry. No resistor in the 5K to 15K range gets anywhere close to that. The extra resistance of the second coil does flatten the responses a bit. But even without using the resistor, you still should hear enough difference to convince you that the circuit properties of the pickup are important.

Well Peter, I think you just proved that a stacked humbucker of the old-fashioned kind does not work. Or are we missing something about stacked humbuckers? I think it is just two coils connected out of phase. But I do not understand your test setup. The signal is made by a string at the normal distance from the pole piece? Flourescent lights make all kinds of interference.

Peter, I forgot to respond about changing the inductance without changing anything else. One way to get close to this is to put another coil in series with the pickup. This coild could just be another pickup. People do this experiment when they use a dummy coil to cancel hum. Doing this usually results in a noticeable loss of treble, since it raises the inductance and so lowers the resonant frequency. But if the initial resonance is very high it could actually increase the apparent treble by putting the peak of the response wherre the guitar has more response. This is why it is important to be able to measure the resonance; then you know for sure where it is.

That method compares the resistance of a resistor to the magnititude of the impedance of the inductor at a fixed frequency, chosen so that there is a significant volage across both the R and the L. I prefer to make a frequency measurement, rather than voltage. If you feed a current into a parallel LC circuit (by means of a voltage into a large series resistor), you can then measure the frequency response, and from the peak, get the inductance if you know the capacitance. The simple formula 2*pi*f = 1/sqrt(L*C) is approximate for low Q inductors like pickups, but often good enough if you just want something close. I like to feed random noise into the circuit and then measure the spectrum using the 'scope program, adding enough estimates to get accurate results. You can do this at the end of a cable connected to a guitar (or just to a pickup) and get the response of the whole circuit including the cable capacitance. It takes some fooling around to get it right, but so does everything.

OK let's talk about shielding a bit. Hum can come from two kinds of fields: 1. electric 2. magnetic Electric fields can be shielded out by metal, copper, steel, whatever, the conductivity does not need to be very high. What happens is that the charge in the metal re-aaranges itself so as to produce a field which tries to cancel the interfering field. The best way to shield something is to fully enclose it in metal. But even an open arrangement has a significant effect. Magnetic fields are not affected by conductors. You cannot shield them with a thin layer of metal like an electric field. You can shield them out pretty well with a significant amount of ferromagnetic material. But you cannot effectively shield a pickup from magnetic fields or it could not sense the field from the strings. A transformer produces magnetic fields. The metal in the tele does not shield this out. It is pretty effective against electric fields, however. Hum from magnetic fields must be canceled, as in a humbucker. The law of induction can be manipulated to get oposing signals from the hum and addition from the fields from the strings. Electric fields cannot be canceled with coils because the law of induction is for magneitc fields. Electric fields can be canceld with a differential amplifier or a transformer used as a subtractor. But electric fields can be shielded out, and that is easier.

Oh yes, see you there then? I am in aeronomy and space physics

"Why haven’t some one tried to wind “half a strat pickup” (only upper part that is) and discovered that they can cut manufacturing costs with 50% for the wire and 50% for the time winding the pickup(rhetoric question)? I have not done the integral, but the bottom half probably contributes about a third of the signal. But if you left it out you would reduce the inductance by more than half. This would change the sound a lot; the resonance would be too high.

"AFAIK the resonant frequency isn’t that easy to measure either without specialised equipment. At least I don’t know how to. " Peter, you have all the hardware you need: your computer. Just get some signal generating software and a 'scope program. This is a lot easier then the stuff you have been doing with the magnetics. "So there is no wisdom in 60 years of collective pickup winding experience? C’mon" I did not say none, but you have to admit that a collective wisdom that does not take much account of the fact that the pickup forms a resonant circuit with the cable capacitance could be missing something. " All his subsequent designs have been mellower. And the first treble attenuating circuitry of the early Teles was to there to take care of the too loud treble." Is a strat mellower than a tele? Less twangy, but thinner IMO, and no less treble. Leo's early amplifiers were pretty much out of the handbooks with one important exception. The tone stack is not stock HiFi, and it is not "flat" in the neutral settings of the knobs. It is designed to boost bass, upper midrange, and treble but suppress the middle, in typical positions, in order to make the correct sound. (http://users.chariot.net.au/~gmarts/ampbasic.htm) This was partly necessary because guitar speakers sound a bit weak at these frequencies. Too much treble is not a problem if it needs to be boosted in the amp. "The idea behind the steel bridge plate had *nothing at all* to do with altering the sound. It was used because Leo was very concerned with reducing hum. The original steel bridge cover (mostly removed and used as an ashtray) and the steel base plate under the pickup was there for one single reason: Shielding. Period. It was originally not there for altering the sound. Than the twang grew on us…" OK, I am no expert on the history, but you think Leo went to all that trouble on the tele to reduce the hum, and then took none of those precautions on the strat, which was intended to be his high end guitar? Surely the "ash tray" was about style, not hum reduction. A pickup sitting in the middle of metal plate is well shielded. A cover has little additional effect.

"I hear you, but it is against what’s generally considered ”common knowledge" among pickup winders. " Peter, I think you must have some idea that I think that there are serious problems with a lot of "common knowledge". "The steel in the Tele bridge is actually thinner than most modern flat-top stop tailpieces, so that this should significantly change the vibrations of the strings isn’t really valid." There is no comparison to a stop tailpiece; the comparison is to the bridge, which is what terminates the string and is thus more responsible for the vibration of the string than the mass of the tail piece. By 1949 electric guitars of one type or another had been around for about fifteen years. The broadcaster (tele's original name) was designed, based on earlier ideas, but very much its own thing. I do not think Leo knew much electronics, but he certainly knew how to experiment, and I think he maximized the twang. "But what about the difference in sound between the under wound “R90” and an over wound Strat pickup (pretty similar in most parts except coil shape as mentioned above)? Any idea?" The inductance of a coil depends upon the dimensions of the coil as well as the number of turns. "Measure Inductance -> not easy to do." No problem. Put a known capacitance in parallel and measure the resonant frequency, then use the simple equation. "If you say that the steel bridge plate might affect the sound, acting almost as a magnetic lens, you will have to acknowledge that the magnetic field in the whole of the pickup is taking part of the tone shaping as the bridge plate is roughly at mid height of the pickup coil." I am not saying the bridge plate has any significant magnetic effect, and I do not think that it directly affects the sound of the pickup very much. The pickup plate is another matter.

Peter, Some ideas on the Telecaster, an interesting guitar. My untested ideas of how it works: The big steel plate that the bridge sits on appears massive enough to isolate the higher frequency vibrations from the wooden body, at least to some extent. This has an effect on the sound, and it can be detected by the pickup if the string vibration is affected by the big plate. In order to maximize the effect, it is necessary to adjust the resonant frequency of the bridge pickup to accentuate this. The small steel plate on the bottom of the pickup increases the inductance of the coil some. I do not know how much, so I am just speculating. This lowers the resonant frequency, and can make the guitar sound brighter if the initial resonance was above the peak response of the guitar. I believe it also maximizes the response the frequency range where human hearing is very sensitive. Using this plate is different from winding on more wire because the latter would also increase the resistance. Some of this can be tested, beginning with plate/no plate inductance. Also maybe some analysis of the effect of the big steel bridge plate.

OK, I think we have some major areas of agreement now; thank you both. The model I would like to do, as you would expect, has a magnetized pole piece, achieved various ways, and steel over the pole piece to model the string. You do it twice with the string at slightly different heights and compare the differences. The problem is the comparison. I have not attempted this yet, but the major problem appears to be that the differences are very small, and so it probably is necessary to subtract the two fields numericallly and plot the differences. Remember, the static field is strong and spatially varying. The difference between the two cases is very small and has a different spatial variation. As far as I can tell, FEMM does not allow you to subtract the fields from two different model results. It does contain a scripting language (lua), and so it might be possible to write out the fields and do the subtraction externally. I am working on this. I would like to comment more on these last posts, but no time until later. Erik, do you get to the Fall AGU?

Well, you need two modes such as horizontal or vertical. For each harmonic you need an amplitude and phase for each mode. This gives you an ellipse for each harmonic. It gets more complicated when you consider coupling between the modes. One way to use the knowledge is to avoid pickup myths. Such as, a wide coil samples the string over its full width.

I think I have not been very clear, so here is another attempt. Please be patient, this is not so simple! First, the pole piece magnetizes the string. Now we look at the effect of this magnetization on the pole piece. To do this, we assume linearity. This just means that we can subtract off the part due to the permanent magnet leaving us with the part from the string. We want to do this to make the problem simpler; it is OK because we know that only the string part fluctuates when the string vibrates. So we assume some level of magnetization in the string; it does not matter exactly what the level is because we just want to look at how this field varies in space. So we represent this string magnetization by a small magnet. Why not? The magnetization of the string does not change much when the string vibrates, so a good model is a permanent magnet. The field due to this magnet is stronger at the string than at the pole piece, since the string is its source. So the reality is that the model shows exactly what it should: the field from the string magnet gets weaker with increasing distance, and we see that we lose somewhat more than a factor of 10 from one end to the other of the pole piece. FEMM does indeed allow you to draw contours within certain field bounds. I have set these pretty wide here so that one can see the field strength in the pole piece and the space around it. It is hard to see the field strength near the magnet on this plot because the lines get too close. You can print out sample points, and these show that the field strength increases about a factor of ten from the top of the pole to the string magnet. There might be a log option on either the contours or the color but I have not found it yet. I will check again later after work. Mike

On the issue of the neo field strength and linearity: I ran the model again after lowering the coercivity of the small magnet by a factor of 1000. Note from the new plot that the resulting field strengths are lowered by a factor of 1000, but that the spatial pattern is essentially identical. Although I would be happy to discuss this further, it is clear that the assumption of linearity is good, and it is a valid way to analyze the pickup problem. http://www.naic.edu/~sulzer/lowField.png

I agree with Peter. There is no reason why plastic covers should affect the sound. The effect of metal cover is pretty small. The effect would be from eddy currents. The capacitative effect is very small, maybe like changing the length of the guitar cable a little bit.

The slider in a no load pot comes off the end of the resistance track at "10". If you do this with a volume control, the output jack and the pickup would no longer be connected. It might be possible to have a switch on the back of the pot that would make this connection at "10", kind of like the system used on old radios where the power switch is part of the volume pot. But that was at the other end of the rotation. I do not know of a pot that does what you want.

Wouldn't that mean that a stacked single coil sized humbucker wouldn’t be humbuckin… For a pickup to be hum cancelling you need the two signals from the coils to be as close to each other as possible regarding the output. The idea behind a humbucker is to induce the same amount of hum in the both coils, but out of phase with each other. If one coil had a higher output than the other the hum would not be cancelled. All of the stacked Hummers I have dissected have had equally sized upper/lower coils The stacked humbucker rejects hum because the magnetic field causing the hum is far away, and so its relative change over the distance of the coil separation is very small. Thus the two coils pick up the same signal and the out of phase connection cancels them. The signal made by the magnetized string is significantly smaller in the lower coil as the modeling here shows, and as one would expect because the difference in the distances to the two coils is relatively large. Thus the out of phase connection cancels only part of the signal.

Both Peter and Erik raised are concerned about the strength of the neo magnet is used to represent the magnetized string. It is the direction and relative strength of the magnetic field at different spatial locations that is being modeled. You can use any strength magnet to show this as long as it does not saturate the steel, and a small neo magnet does not, in my experience. (The test involves bringing a magnet near a coil with such a steel core and measuring the change in inductance, if any. There is no change.) I can easily modify the field strength in the FEMM model and see if this changes the pattern of the field. "If you want to model a string over a pole piece, make your neo disc unmagnetized steel, leave the pole piece as it is, and add a disc with a B field of ~1000 G touching the bottom of the pole piece." The problem with that is that it is harder to see the effect of the "string" because the field induced in the string is weak and that in the pole piece is strong. FEMM is not set up to look at small differences easily, but it might be possible to finagle the scales on the plot enough to show the effect in this way. Otherwise, I can probably find a way to write the fields to a file, first without the "string", then with it, and subtract the two in another program. This would then show the field due to the magnetization induced in the "string". But this will take some time. In this simple model, I used linearity, because it is the easiest way. There should be no problem with it, either. "Conclusion #2 is the opposite,..." No, you are looking at my model upside down. My conclusion is your conclusion, and both are right. "You could test this easily by winding a pickup with stacked coils on a single bobbin (same number of turns), and measure the outputs on the top coil and the bottom coil and compare. The coil nearest the strings will have more output (higher current)." Yes, this is exactly the point I made in the earlier discussion on this topic. The stacked humbucker shows that the induced field is weaker further from the string. However, that conclusion did not satisfy everyone, so I have done the modeling to show it. "Conclusion #3...I dunno...I don't know what "the core" refers to." By the core I menan the pole piece. Sorry about the confusing terminology. Do you see why it is only the field in the pole piece that really matters, not throughout the coil? Spatial scale: the neo is .1 inches from the pole piece. But it is relative dimensions that matter here, and the conclusions are not really very sensitive to them either. Peter said "Another thing: Will the magnetized sting have the same direction as the disk in the plot? Remember that the string vibrates up and down, sideways and in a circular motion. Do we actually know that the string doesn’t rotate? If I remember correctly that is what the string really does." The pole piece magnetizes the string in the vertical direction over the pole piece. I have chosen that direction for the neo magnet. The string moves in a complicated way. I chose to discuss vertical motion because that makes the largest fluctuations. Horizontal motion changes the field through the core less. But in any case, the magnetization is almost vertical no matter how the string moves. " Do you mean that the field that is actually usable to generate current in a coil is only a tiny bit of the field (energy) available?" No, I mean pretty much the opposite. The field that generates the voltage around a particular winding is potentially the field anywhere within that winding. But only the field in the pole piece has significant strength and the right direction.

One more thing--- "Make the pole piece a magnet like it's meant to be... " There are two ways to do this. First, make it a permanent magnet. Second, make it out of steel. Then the permanent magnet below magnetizes the steel pole piece which magnetizes the string. Or are you also claiming that the permanent magnet does not magnetize the pole piece? But it does, as you can easily show by putting a small screw driver tip near a pickup pole piece. If the permanent magnet magnetizes the pole piece, why would the pole piece not magnetize the string?

Then you agree that the results presented in the first post in this thread are correct if I can convince you that the string is magnetized. I am not sure that I can do this using FEMM, but I will try. In the meantime, you should review how I showed this before. If you have any intellectual honesty at all, you should be trying to show why the string is not magnetized when it is placed in a magnetic field. Magnetizaton is what the physics predicts. If you think otherwise, support your beliefs.

FEMM is a program intended to model magnetic systems. It is actually not powerful enough to model exactly pickups and strings because it can only do two dimensional systems or three dimensional systems that have symmetry like a cylinder (used in the plot below). A string over a cylindrical pole piece does not meet this condition. However, FEMM can model a small disk magnet over a pole piece, and this is good enough to get an idea of the effect of the magnetized portion of the string. Remember, the permanent field from the pole piece magnetizes the string (only significantly when the pickup is present) over the pole piece. Various FEMM plots have shown that significant magnetization occurs only near the pole peice. We want to look at the pattern of this field to learn how the pickup works. Actually we need the fluctuation of this field as the string moves, but this fluctuation has a spatial variation similar to the field itself, that is strong where the field is strong, and weak where it is weak. The plot shows these two features that were discussed earlier: 1. The pole piece stops the field from falling off as fast as it does without the pole piece. This increases the strength of the field through the coil. 2. The field falls off over the length of the pole piece, so windings near the top contribute more than those below. 3. Only the field through the pole piece contributes significantly. The field outside the core, but inside a winding, contributes very little because it is weak and almost horizontal. Remember this about the plot: 1. The field strength is indicated two ways: contours and color. 2. The closer the contours, the stronger the field. 3. The color scale saturates at the top and the bottom; so it only represents the field strength inside the pole piece. 4. The plot shows only the right half of the system. To get the 3D plot, think of rotating the plot about its left (vertical) side. 5. The magnet is a small neo. 6. The pole piece is 1018 soft steel. http://www.naic.edu/~sulzer/neoOverSteel.png

Often the brdige pickup has more turns of wire to make up for the fact that there is less string amplitude near the bridge, but I believe standard strats have all three pickups the same.

As long as any sound at all is produced the strings vibrations are hindered... The electricity is induced when the string passes through the magnetic field, if the magnetic field wouldn't break the strings vibration no electricity would be induced at all. It is true that some energy must be taken from the string to the electrical circuit. But it is very small and it does not affect the sustain a significant amount. Does the sustain get a lot greater when you take the pickups out of a guitar? Here is another thing you can try: Connect the guitar into a short circuit. (Wire across the contacts of a jack) Now check the sustain (with volume all the way up). If the pickups could extract much energy from the string, you should really notice it into a short because the amount of current in the coil is limited only by the pickup impedance, not also by the high load impedance of the amp and pots. Consider this also. Some people wire the volume controls on an LP backwards. That is, the slider goes to the pickup. Suppose you turned the pot for the neck pickup to zero and just used the bridge pickup. Since the neck pickup is shorted, it would take a lot of energy from the strings if it could and kill the sustain. Doesn't happen. If the permanent magnetic field is too strong the string vibration is affected (string pull, or stratitis). But this has nothing to do with generating electricity. The problem is that the magnet affects the vibration frequency in the vertical mode, but not the horizontal. The two modes try to mix together, and the vibration is disturbed. With really strong magnets, you can hear the beat between the two frequencies.

"what i was trying to say is that the magnetic field helps the string to keep vibrate and if the polepieces are not directly under the string,that effect does not happen." The magnetic field does nto help the string to vibvrate. It can hinder it if it is too strong, but with the correct strength it is not an issue.

Gorrilla glue is difficult to use for guitar work. Why do it when normal wood glue works so well? It is extremely strong, but it is very difficult to take apart a GG joint becuase it does not release with heat easily as most other glues do. You can make joints with no glue line if you are careful; the clamp pressure seems to keep the foaming down in the joint.