|
Post by antigua on Jun 5, 2017 20:52:08 GMT -5
My 50 turn coil should have about 210uH inductance, according to an online calculation. So the inductive reactance at 20kHz is about 25 ohms. I'm using a 100 ohm current limiting resistor in series with the test coil, so I figure it's doing a reasonable job of providing a current source at lower frequencies. Even at 20 kHz, that's not too bad. A larger resistor and greater drive voltage could be used, but it never seemed necessary. The driving coil plus pickup is a poorly coupled transformer; that is, it has a lot of leakage flux. Ignore the leakage to get an approximate analysis. You have roughly a 100:1 turns ratio or a 10,000:1 impedance ratio. At resonance the pickup impedance is high, usually in the range 100,000 ohms to 1,000,000 ohms, and so the transformed impedance is probably somewhere between 10 and 100 ohms. Based on that, the interaction is small, but maybe not negligible. The poor coupling, with its added series inductance should help a bit. But the real question is what is Antigua using with his small coil. I'll tell you this, because I can't do tests at the moment, the issue has confused me for a long time, and I have done proximity testing where the driver coils is near the pickup with low voltage, and farther away with higher voltage, and saw no difference in resonant amplitude, suggesting that inductive coupling is very low, even at close proximity. I'd love to know why the resonant amplitude is not more constant, but I don't believe inductive coupling yields the answer. I've found other bizarre effect, where the resonant amplitude is diminished even when both the eddy current causing metal and driver are both a good distance from the pickup. Those tests are in a thread, but they were done some time last year so it would take me some searching to find them. I'll just redo these test as soon as I have an hour free.
|
|
|
Post by antigua on Jun 6, 2017 23:58:59 GMT -5
Here are a couple more tests to shed some more light: The two lower lines are the same test, but this time exciter coil is 10mm above the test point, rather than ~2mm. The resonant peak is lower over the full length slot, and a little higher over the slot that didn't extend to the bottom of the cover. Even though the peaks are contracted, the same overall pattern is present. In the plot above, the full slot is the "G" pole, and the tested partial slot was the "A" pole. This plot below shows what happens when the exciter coil starts at the full slot, and is gradually moved further away from the full slot, along the length of the pickup, towards the "A", ending at the "low E". It can be seen that the resonant amplitude, or Q factor, gets lower as the exciter coil is position further along the length of the pickup, towards the "A" string. The lowest line, the gray line, has the exciter over the "low E": ' I'm not sure why the resonant peak appears to differ based on the placement of the exciter, and I'm not even certain that the impedance of the coil itself is ever changing, despite what it seen in the plots. For the purposes at hand, it can be seen that each slot must extend to the bottom of the cover for optimal eddy current reduction.
|
|
|
Post by stratotarts on Jun 7, 2017 8:33:33 GMT -5
I think what you're seeing there is because the test field "footprint" is wider when the coil is more distant from the pickup, so includes fewer of the lossy elements (partial slots) that are side by side with the lossless elements (full slots). I can't be sure but that is my take on it.
|
|
|
Post by antigua on Jun 7, 2017 13:44:08 GMT -5
The height of the resonant peak is determined by the amount of series resistance or parallel load, and there's some debate about which of these two categories eddy currents fall under, but most intuitively it seems to be a frequency dependent series resistance. The question is, why would this series resistance be higher or lower depending on where the exciter coil is positioned; higher when there is more conductive obstruction between the pickup and exciter, lower when there is less. Maybe that's not even really what's happening, but it just looks as though that's how it is. With an uncovered Tele neck, this is what happens with the exciter coil at various distances away from the pickup. Since there is no cover, just AlNiCo pole pieces, the eddy current activity is fairly low. You can see that the resonant peak amplitude is fairly constant even when the exciter is farther away from the pickup.
|
|
|
Post by antigua on Jun 7, 2017 14:09:07 GMT -5
Here's a demonstration of how eddy currents at a distance effect the resonant amplitude: The lower plot line is what happens when the Tele cover is around the exciter. Since it's unlikely that either the exciter or the cover are having a big impact on the actual electrical properties of the pickup, I'm guessing that the lower resonant amplitude isn't actually indicative of the pickup's electrical values, such as it's actual impedance at resonance. There seems to be a trend with eddy currents and resonant amplitude though; even though overall eddy current slope is gradual, the resonant amplitude always drops by a somewhat greater magnitude. We never see a powerful resonant amplitude as well as significant eddy currents at the same time. If the eddy currents are causing a -3dB/oct slope, there might be a -10dB reduction in resonant amplitude. Here's a picture:
|
|
|
Post by ms on Jun 7, 2017 16:05:20 GMT -5
There seems to be a trend with eddy currents and resonant amplitude though; even though overall eddy current slope is gradual, the resonant amplitude always drops by a somewhat greater magnitude. We never see a powerful resonant amplitude as well as significant eddy currents at the same time. If the eddy currents are causing a -3dB/oct slope, there might be a -10dB reduction in resonant amplitude. Antigua, thank you for exploring this in such detail. With these last measurements, I think you have the pickup so far away that for sure it is just a sensor of what is getting through the cover compared to no cover. At resonance, I read about 6db of difference; at half that frequency, I see about 3 db. (But please tell me if I am misreading the graph.) I think a difference of 3db in an octave is reasonable. (In transformer work, it appears that eddy currents in the core increase with something like the square root of the frequency, I believe, so it seems to be in reasonable agreement with that.) So I am thinking the this last measurement is what we should expect. It could be that the others do too if looked at carefully with this in mind.
|
|
|
Post by wgen on Jul 11, 2017 7:56:05 GMT -5
Sorry if I ask this in this topic, I guess it has very little to do with the cutting of a cover...even though Tele neck pickups are sometimes referred to "Lipsticks" 😃 I didn't know where to ask this: does anybody eventually know if Danelectro Lipsticks pickups (which obviously are a whole other thing than Tele's) are covered in a brass outside shell, or if they eventually use nickel-silver shell? I imagine these to have a quite bright response, but I don't understand if their geometry makes for some damping of the resonance peak. Thank you anyway...
|
|
|
Post by antigua on Jul 11, 2017 13:07:06 GMT -5
Sorry if I ask this in this topic, I guess it has very little to do with the cutting of a cover...even though Tele neck pickups are sometimes referred to "Lipsticks" 😃 I didn't know where to ask this: does anybody eventually know if Danelectro Lipsticks pickups (which obviously are a whole other thing than Tele's) are covered in a brass outside shell, or if they eventually use nickel-silver shell? I imagine these to have a quite bright response, but I don't understand if their geometry makes for some damping of the resonance peak. Thank you anyway... I've never analyzed one, but I've read that the Danelectro Lipstick pickups use a brass cover, and that eddy current losses are a definite factor in their tone.
|
|
|
Post by wgen on Jul 11, 2017 14:15:11 GMT -5
Sorry if I ask this in this topic, I guess it has very little to do with the cutting of a cover...even though Tele neck pickups are sometimes referred to "Lipsticks" 😃 I didn't know where to ask this: does anybody eventually know if Danelectro Lipsticks pickups (which obviously are a whole other thing than Tele's) are covered in a brass outside shell, or if they eventually use nickel-silver shell? I imagine these to have a quite bright response, but I don't understand if their geometry makes for some damping of the resonance peak. Thank you anyway... I've never analyzed one, but I've read that the Danelectro Lipstick pickups use a brass cover, and that eddy current losses are a definite factor in their tone. Thank you very much! That's what I was suspecting...it would have been a little strange to me if the "chrome shell covered" pickups of a guitar that in some cases only costs 250 euros new were nickel silver
|
|
|
Post by reTrEaD on Jul 11, 2017 15:59:00 GMT -5
Thank you very much! That's what I was suspecting...it would have been a little strange to me if the "chrome shell covered" pickups of a guitar that in some cases only costs 250 euros new were nickel silver Contrary to the name, there is no silver in nickel silver. It's an alloy, primarily of copper with nickel and zinc. It's more expensive than brass (copper and zinc) but not prohibitively so. Copper costs roughly $2.50 per pound. Zinc is about $1.25 per pound. Nickel is about $4.00 per pound. Not sure what the original Danelectro used as a base beneath the plating but what we see today from other manufacturers might not be the same. Those 'lipstick' covers literally were surplus lipstick cases.
|
|
|
Post by stratotarts on Jul 11, 2017 17:50:42 GMT -5
Now you're tempting me to get some lipstick types to test. I would strongly expect substantial eddy current losses, because as mentioned above, they are most likely brass. The theory predicts big losses from the geometry.
|
|
|
Post by wgen on Jul 12, 2017 7:45:59 GMT -5
Thank you very much! That's what I was suspecting...it would have been a little strange to me if the "chrome shell covered" pickups of a guitar that in some cases only costs 250 euros new were nickel silver Contrary to the name, there is no silver in nickel silver. It's an alloy, primarily of copper with nickel and zinc. It's more expensive than brass (copper and zinc) but not prohibitively so. Copper costs roughly $2.50 per pound. Zinc is about $1.25 per pound. Nickel is about $4.00 per pound. Not sure what the original Danelectro used as a base beneath the plating but what we see today from other manufacturers might not be the same. Those 'lipstick' covers literally were surplus lipstick cases. Thank you for the insights!
|
|
|
Post by stratotarts on Jul 12, 2017 14:31:43 GMT -5
Nickle silver is a form of "German Silver"... so named because it was a faux silver for ornaments and so on.
|
|
|
Post by antigua on Mar 10, 2021 0:57:58 GMT -5
|
|
|
Post by aquin43 on Mar 10, 2021 5:22:10 GMT -5
Here are a couple more tests to shed some more light: The two lower lines are the same test, but this time exciter coil is 10mm above the test point, rather than ~2mm. The resonant peak is lower over the full length slot, and a little higher over the slot that didn't extend to the bottom of the cover. Even though the peaks are contracted, the same overall pattern is present. In the plot above, the full slot is the "G" pole, and the tested partial slot was the "A" pole. This plot below shows what happens when the exciter coil starts at the full slot, and is gradually moved further away from the full slot, along the length of the pickup, towards the "A", ending at the "low E". It can be seen that the resonant amplitude, or Q factor, gets lower as the exciter coil is position further along the length of the pickup, towards the "A" string. The lowest line, the gray line, has the exciter over the "low E": ' I'm not sure why the resonant peak appears to differ based on the placement of the exciter, and I'm not even certain that the impedance of the coil itself is ever changing, despite what it seen in the plots. For the purposes at hand, it can be seen that each slot must extend to the bottom of the cover for optimal eddy current reduction. A good explanation would be that the true resonant peak is actually the same in all cases and that what is varying is the voltage induced in the coil. Remember the Tele Tucson model in which there are two paths from the excitation to the output. One is directly into the coil and the other is into the eddy current load and from there into the coil. These two paths combine to give a step low pass filter response. In practice, there is usually an additional step low pass filter response due directly to the shielding effect of the cover material. These responses are in cascade with the resonant response of the pickup impedance and its load.
By varying the location of the exciter you are altering the relative contributions of these filters to the overall response. The effect is more noticeable in this case because the slotted cover has more physical variation from point to point than a plain one.
|
|
|
Post by antigua on Mar 10, 2021 11:51:46 GMT -5
A good explanation would be that the true resonant peak is actually the same in all cases and that what is varying is the voltage induced in the coil. Remember the Tele Tucson model in which there are two paths from the excitation to the output. One is directly into the coil and the other is into the eddy current load and from there into the coil. These two paths combine to give a step low pass filter response. In practice, there is usually an additional step low pass filter response due directly to the shielding effect of the cover material. These responses are in cascade with the resonant response of the pickup impedance and its load.
By varying the location of the exciter you are altering the relative contributions of these filters to the overall response. The effect is more noticeable in this case because the slotted cover has more physical variation from point to point than a plain one.
Thanks, that makes sense. In relation to the Tele Tucson model, is it just changing coupling coefficients standing between the exciter, cover, and pickup coil?
|
|
|
Post by aquin43 on Mar 10, 2021 12:59:20 GMT -5
A good explanation would be that the true resonant peak is actually the same in all cases and that what is varying is the voltage induced in the coil. Remember the Tele Tucson model in which there are two paths from the excitation to the output. One is directly into the coil and the other is into the eddy current load and from there into the coil. These two paths combine to give a step low pass filter response. In practice, there is usually an additional step low pass filter response due directly to the shielding effect of the cover material. These responses are in cascade with the resonant response of the pickup impedance and its load.
By varying the location of the exciter you are altering the relative contributions of these filters to the overall response. The effect is more noticeable in this case because the slotted cover has more physical variation from point to point than a plain one.
Thanks, that makes sense. In relation to the Tele Tucson model, is it just changing coupling coefficients standing between the exciter, cover, and pickup coil? I would guess that while there is only one overall aggregated eddy current coupling that determines the output impedance and the overall response of the pickup to induced voltage there is a whole array of possible paths into the coil with different degrees of coupling to the eddy current load and amounts of cover shielding. Presumably the result with the exciter in the string position is the most representative.
I think that you have highlighted a feature of the exciter measurement method that requires consideration in each case. An analogy could be made between your results here and my pole shading scheme where the response of the pickup varies dramatically from pole to pole as a result of the variation in shielding.
|
|
|
Post by ms on Mar 10, 2021 21:17:01 GMT -5
It appears that as the exciter coil is moved with respect to the slot, the Q of the resulting resonance varies. In a low Q system the frequency of the peak of the response changes as the Q varies, even if the defined resonant frequency, the boundary between inductive and capacitive, does not..
|
|
|
Post by antigua on Mar 11, 2021 14:33:34 GMT -5
Another thing it can be is that some portion of the side of the pickup cover has lines of flux passing through it, because the cover is not perfectly parallel to the magnetic field. The pickup cover is like an upside down "U", and the magnetic field of the exciter is like a figure eight on end, so there would still be some oval shaped eddy current loops in the side walls of the cover.
|
|
|
Post by pablogilberto on May 8, 2021 1:01:11 GMT -5
The ironic thing is that these cuts more effectively reduce eddy current losses than the "open" Tele covers, such as that seen on the Lollar Royal T, while being far less visually obstructive. The problem with the "open" style cover is that there is still electrical continuity around the periphery of the cover, meanwhile the top of the cover has removed far more metal than is necessary. Hi antigua! With regards to the open style covers for Tele, I understand that this will definitely reduce the Eddy currents. What I want to know is the shielding effect compared to a standard full cover? Is it just the same as without cover or something in between the fully covered and no cover? Thank you!
|
|
|
Post by antigua on May 8, 2021 15:49:37 GMT -5
The ironic thing is that these cuts more effectively reduce eddy current losses than the "open" Tele covers, such as that seen on the Lollar Royal T, while being far less visually obstructive. The problem with the "open" style cover is that there is still electrical continuity around the periphery of the cover, meanwhile the top of the cover has removed far more metal than is necessary. Hi antigua! With regards to the open style covers for Tele, I understand that this will definitely reduce the Eddy currents. What I want to know is the shielding effect compared to a standard full cover? Is it just the same as without cover or something in between the fully covered and no cover? Thank you! I've never once actually noticed an obvious reduction of static noise when adding or removing covers or cavity shielding, but the theory is that humbucking blocks out unwanted magnetic induction, while conductive metal shielding blocks out unwanted electrical attraction with other circuits, as if the space between the pickup and the source of noise was like a huge capacitor connecting the two. That relationship is directional, so the open cover is theoretically lacking coverage between the top of the pickup and sources of noise, but I don't think I would ever notice, and I'm skeptical that anyone else really does either. The benefits of humbucking are more or less self evident, but whether a pickup is covered or not seems to be more of an aesthetic thing. I've been using wireless transmitters instead of guitar cables lately, and I've noticed less noises as compared with the guitar cable. I think having the guitar electrically isolated from the mains helps, and I would think it would reduce that electrostatic noise more so than it would magnetic noise. It also removes the potential for electrical shock, but that seems to be exceptionally rare. Sometimes when lips touch my SM-57 while I was holding a cable connected guitar, I get a slight shock, though.
|
|
|
Post by pablogilberto on Jul 12, 2021 7:27:39 GMT -5
Thanks antiguaWhat do you think about the Tele open covers with mesh like this? In terms of shielding and Eddy Current. Thanks!
|
|
|
Post by ms on Jul 12, 2021 9:14:08 GMT -5
Should shield OK; what is the screen made of? The conductivity would affect the level of eddy currents.
|
|
|
Post by antigua on Jul 12, 2021 20:06:09 GMT -5
Thanks antiguaWhat do you think about the Tele open covers with mesh like this? In terms of shielding and Eddy Current. Thanks! That's an interesting look. It would cause some eddy currents, because there is circular continuity around the pole pieces, but I'm not sure how it would compare to other cover materials. I'm assuming whatever metal that is is conductive enough to work as electrical shielding. It would be interesting to test, it's sort of a like those gold foils. As for shielding effectiveness, there's some math as to the wavelength versus the size of the holes, and I suspect the mesh would effectively block frequencies that would cause noise, especially if the amplifier is filtering out high frequencies at the input.
|
|
|
Post by ms on Jul 14, 2021 10:00:42 GMT -5
The holes in the screen are small enough to reflect rf up to many, many GH; not an issue. I think the dominant eddy currents are around the sides of the cover, assuming a screen material of non-magnetic stainless.
|
|