So here is the blog about pickup covers that I promised in the backplate post from a while back. I touched on the concepts of eddy currents and capacitance in that post, but I will go into a bit more detail here (while trying not to reiterate what you can find in a thousand places online or in textbooks).
When dealing with audio electronics, we have to wrestle with strange questions like ‘what does brass sound like’. Just reading that line rings oddly in my head. But regardless, let’s break this down into it’s various components and then run a bunch of tests.
The major factors involved when considering the tonal effect of covering a pickup are:
- Eddy currents
- Noise shielding
The first has more to do with the material and geometry of the pickup, and the last two have more to do with the material of the pickup cover. I’ll touch on that briefly, and then move onto each area in more depth.
Much of this will be review for many of you, but stick with me:
-Capacitors are formed by sandwiching a dielectric material between two metal plates. The larger the surface area of the plates and the higher the dielectric constant of the material, the higher the capacitance. In the case of a pickup, the “plates” are the coil and the cover+base. The dielectric is everything between those ‘plates’ when one is held at a different potential than the other (like grounding one end).
-Eddy currents are are small current loops that are caused by alternating magnetic flux through a conductive material. These currents dissipate energy in the form of heat, and are frequency-dependent. The amount of material and it’s conductivity/resistivity effect the degree to which the eddy currents are formed in conjunction with magnetic field strength, and frequency.
-Noise shielding reduced coupled interference in an electrical circuit. Obviously we want to keep that out of our signal path (especially in high-gain situations). Although there are several modes of noise and noise transmission, we will mostly be concerned with EMI (electromagnetic interference) in terms of reflection and absorption.
We have seen in many past blogs how capacitance effects the frequency response of a pickup. Attaching a grounded cover to a pickup produces YET ANOTHER source of shunt capacitance that is in parallel with all our other shunt capacitance loads. Compared to our cable/input capacitance, the shield’s capacitance is fairly low, but it is significant with respect to our coil’s self-capacitance. Fortunately, we can calculate (and measure) capacitance from the shield much easier than capacitance in the coil.
Since capacitance has more to do with geometry and dielectric material, the actual COVER material has little effect, but things like potting and bobbin material have great effect.
So let’s calculate the capacitance. If you recall, capacitance is a function of electrode area and distance, and dielectric constant. In the case of multiple dielectric materials, we take the average based on the proportional thickness of each material. If you visualize the pickup, each outer face of the coil forms our first electrode, and the cover/baseplate forms the other. So we have to calculate each face’s area, and then the average dielectric constant, calculate the capacitance for each of the 6 sides of the coil, and then add them together.
We will use a 50mm PAF pickup in our example, and a standard pickup cover and backplate. We need to know the electrode surface area and distance, as well as the dielectric constants involved.
First we calculate the surface areas. Let’s say we have 4.32 area on the top and bottom, .56 for each of the long sides, and .5 for each of the curved sides. (if the pole pieces are grounded, we also have the inner coil surface area’s capacitance with respect to the pole pieces and the bobbin+wax as a dielectric, but that exists without the cover, so we will ignore that for our purposes).
Here are our dielectric constants for each possible relevant component material in ascending order @60Hz:
PC (polycarbonate) 3.2
Cellulose Acetate Butyrate 3.55
Nylon (30% glass fill) 3.4-3.6
vulcanized fibre 6.5
nylon (general through type 8) 5-9
hard rubber 2-4
Maple (dry) 3.2-4.5
Nylon (30% glass fill) 3.4-3.6
Maple (waxed) 3.9-5
Maple (compressed) 4-6
nylon (general through type 8) 5-9
parafin wax 2.2-2.5
bees wax 3-3.5
acetal polyvinyl formal = 3.9
Polysol = 4
solderable polyurethane /w polyamide-nylon = 6
alkyd enamel = 5-7
On our top face, we have: bobbin, wire insulation, coil tape, double-sided tape, and potting.
On our bottom face, we have: bobbin, wire insulation, coil tape, shim/spacer, potting.
On our sides we have: bobbin, wire insulation, coil tape, potting.
On our curved sides we have: bobbin, wire insulation, coil tape, potting.
Now we need to figure out the distance of separation between our electrodes (approx measurements):
Top = .0625 bobbin + .025” gap (varies depending on bobbin warpage, hand press-fitting pressure, etc)
Bottom = .0625 bobbin + .125”
Long sides = .09”
Short sides = .15” (averaged)
Most of our dielectric will be formed by the bobbin and the potting material, with lesser proportions of tape, adhesive and insulation, so let’s pick a few that are reasonable for a humbucker. We’ll say that we are using parafin/beeswax mix at a dielectric of 2.7, Butyrate/nylon-30% bobbins at 3.5, waxed dry maple spacers at 4.5, rubber-adhesive 0.005” paper insulation tape at 3.
formula: C = K ε0 A/d (or use a calculator online… there are many)
- TOP = 37pF
- bottom = 26pF
- long sides = 8pF
- curved sides = 3pF
TOTAL = 74pF
(NOTE: this includes the baseplate, which would be considered a part of the lump capacitance of a normal open-faced pickup)
That seems to agree with what I have measured more or less. I have also measured 100% coverage copper-foil-shielded capacitance around 90-100pF, and that foil is much closer to the coil all around the pickup, so the difference in capacitance between the two seems reasonable.
Now that we know that the factors effecting the capacitance are area, dielectric, and spacing, let’s see what happens if we mess around with them. We’ll increase some of the factors by 50% and see what happens:
+50% distance = 48pF total
+50% dielectric constant = 110pF total
Now let’s decrease them:
-50% distance = 145pF total
-50% dielectric constant = 42pF total
We can also plug in options to see how different component types can change the capacitance, etc. You can also see that capacitance is effected greatly by distance, so hand-built pickups will likely have a decent amount of variation, and don’t forget to subtract the baseplate capacitance when comparing a covered pickup to an uncovered pickup. A cover that is spaced out further from the bobbin, using low dielectric constant materials could be nearly negligible in terms of additional capacitance.
Eddy currents were covered in several past blogs, so I won’t beat that dead horse, but the general takeaway is that they effect the peak-frequency, Q bandwidth, and Q height. I have found that it is difficult to predict the effect of alloys on eddy current using just the electrical parameters, so it is easiest to test samples rather than ‘doing the math’. I have included sweeps of each material at the bottom of this post to help visualize the changes, but the general trend was:
The change was not drastic, but noticeable.
Nickel-Silver: did not change the frequency much, did not effect the bandwidth much, DID lower the peak about 1.5dB
Brass: did not change the frequency much (a bit more than the nickel-silver), but it widened the bandwidth and lowered the peak the most of any material tested. Like nickel-silver, it did not effect the frequency of the peak MUCH, but it did flatten the Q and lower the peak the most of any of the metals tested, and warmed up the tone slightly.
Aluminum: shifted the frequency of the peak the most of any material tested (shifted to higher frequency), it narrowed the Q bandwidth, and lowered the Q slightly (about as much as nickel-silver). In listening tests, aluminum did not dull the tone much, but it did change it slightly.
Copper: shifted the frequency almost as much as aluminum, narrowed the Q bandwidth the most of any material tested, lowered the peak the least of any material tested. in listening tests, copper seemed to warm the tone slightly compared to the bare pickup, but it also seemed to brighten the sound slightly compared to nickel-silver.
Another factor is the distance between the cover and the bobbins. Here is a graph showing the progression of eddy current response. For ‘worst case scenario’, I used 0.032″ brass plate, spaced out from laying directly on the pickup face using 1/32″ cardboard shims. The top curve is no-plate and the bottom curve is the plate lying directly on the bobbin:
So in this case, 1/8″ space halves the effect of a 1/32″ brass plate. Of course, if the metal were either thinner, slotted, or had different conductivity (like nickel-silver or Stainless Steel) the eddy currents would have less effect, but still follow a similar trend in terms of distance.
As far as noise is concerned, I compared the materials to see how they would fare. Surprisingly, noise shielding rarely comes up in pickup cover conversations, so I could not find any external test data for info. I made covers out of each and subjected them to noise to see which worked best. Surprisingly in thicknesses of about ~1/16″ (which is a bit thick for covers, but stick with me here), Aluminum seemed to work best, followed closely by copper, brass, then nickel-silver at a distant last place. BUT, in another blog post, we saw that nickel-silver had the least effect on tone (most transparent) compared to bare coils, so that may be a consideration, but that may also be a moot point if you design your pickup around a certain cover material.
The purpose of shielding is to attenuate electromagnetic interference before it reaches your coils. The mechanisms of shielding are reflection, absorption, and conduction (among others, but those are the three that we are most concerned with here). So what is not reflected is absorbed, what is not absorbed is conducted to (hopefully) ground, and anything left will induce current (noise) in our coil. So the relevant parameters are: the electrical parameters of the material (high conductivity, etc), the thickness of the shielding, the % of coverage.
Here is a graph of the noise spectrum of the bare coils (green) vs the best cover shield (purple);
The methods of coupling, as they pertain to our case here, are induced noise, capacitively-coupled noise, and conducted noise.
Conducted noise (common-impedance interference):
Impedance mismatch. We can avoid conducted noise through a good grounding scheme and solid connection. An example is: if your pole pieces are not grounded, they can conduct and then re-radiate noise into your coil; grounding them will prevent that. This can be tested by touching the poles of your pickup. Same thing goes for any other metal part which is un-grounded: there is the potential to re-radiate noise, so make sure they are grounded.
Mutual capacitance-coupled noise (near-field interference):
delta-voltage/delta-time (i.e. 1v/sec change = 1mA/pF change), so higher capacitance or dV/dT = higher noise coupling. This depends on the area of the coil, distance (i.e. thickness of the guitar body when talking about conduction through the back of the guitar), as well as the dielectric constant of the materials (mostly the guitar body, bobbins, potting… i.e. epoxy is potential noisier than paraffin wax, maple is potentially noisier than balsa, Butyrate is potentially noisier than ABS, etc). This comes into play mostly between your body and the pickup coil.
We know that capacitors are formed by conductive elements at different potentials with a dielectric between, so imagine that YOU are one electrode and the pickup coil is the other. Ideally, a grounded shield between you and the pickup will form a capacitive coupling between you and ground instead of you and the pickup, and the noise will (hopefully) be shunted away from your coil (as long as your ground is solid and the shield is not leaky or insufficient).
Mutual inductance-Coupled noise (near-field interference):
delta-current/delta-time (i.e. 1A/sec change = 1mV/H change)The hum bucking configuration will cancel out some of the lower harmonics to a degree, but the upper harmonics will still make it through. To prevent these frequencies from reaching our coil, we use sufficiently conductive material of sufficient thickness for the frequencies of interest. Obviously we are limited to very thin dimensions, but the thicker the better.
(There is also magnetically-coupled noise, which is tough to deal with since out pickup is a magnetic circuit, so that can be dealt with in pickup design with flux guides, shorter magnetic paths, and a less leaky structure. Though ferrous base-plates or side plates can be –and are occasionally — used without compromising the operation of the pickup, and may help to re-direct magnetic fields somewhat, depending on the shape and size.)
I have posted the noise graphs at the bottom of this post.
- Nickel-silver: had the least effect on tone. It did not effect the resonant frequency, just the Q of the peak. It was the worst material in terms of noise shielding though. This material was also the most difficult to machine and bend of all the materials tested.
- Brass: Interesting in that was the only material that decreased 60Hz noise, and I am not sure why, but it was testable and repeatable. It was inferior to aluminum or copper at all other frequencies though, but it was better than nickel-silver across the board.
- Aluminum: seemed to work best in the thicknesses and dimensions tested in terms of shielding. Of all the above alloys, aluminum is the only one that cannot be readily-soldered, so that may be a consideration since you will need a mechanical connection to ground it. The peak was shifted slightly more toward the treble frequencies, the Q bandwidth was tightened and lower slightly.
- Copper: was quite effective at shielding across the board (2nd best), and it was also the 2nd most neutral tone-wise. It is also easily-solderable and super-easy to drill, shape, and cut.
We also know that the thicker the material, the better the shielding, so even though nickel-silver was not nearly as effective as the others (due to it’s low conductivity) Nickel-silver covers are much thicker than copper/alum tape (around 0.022″ thick… at least the models that I have here) and about as effective — in these limited tests — as a layer of copper tape (about 1.5mil/0.0015″), which was about as effective as several layers of aluminum tape (aluminum did not do quite as well as copper in thinner dimensions).
I wish I had some ferrite sheets and plates to test, but unfortunately I did not think of that in time to order some. I’ll put some in my next Mouser order and add the results when I can.
So weighing that information against the noise tests, for cover material I would recommend:
more interested in greatest transparency than lowest noise = use nickel-silver
more interested in lowest noise than greatest transparency = use aluminum or copper
And subjectively, in terms of sound:
- Copper was brightest
- Nickel-silver was the most neutral
- Brass was darkest
Alternatively, using thicker nickel-silver stock would work well also, but I would not go that route considering what a pain in the butt it was to work the material even in the thinner dimensions. One last note is that I noticed that aluminum cut light dimmer noise down better than any of the other materials. I am not sure why exactly, but it is worth keeping that in mind.
Another thing worth mentioning is that a few of the ‘nickel-silver’commercial covers that I tested were electroplated with copper before being coated in nickel or chrome, and that combination of metals made the eddy current behavior act similarly to a higher conductivity metal… not AS much, but definitely a higher eddy current loss than pure nickel silver. So if you are looking for a more transparent-sounding cover… do your research first (and consider spacing the cover a bit further from the bobbin face since that will help out with reducing both eddy current loss AND capacitive coupling).
Also, another readily-available material that I did not have on hand to include, but would possibly make great covers are some grades of 300-series stainless steel. Something like 316L stainless would likely provide poorer shielding than any of the options above, but would also likely be more transparent than even nickel-silver due to high electrical resistivity. It can also be silver-soldered, although stainless is a bit of a pain in the butt until you get the hang of it, and it takes more prep work than copper-alloys like nickel-silver or brass.
Zinc (which is the other component of brass) is another option. It would likely fall between brass and aluminum.
Hopefully that covers everything.
Here is an example joint that I did on 2 pieces of stainless steel with just regular plumber flux and tin/silver solder, just so you can see that it is possible without anything exotic:
I just scuffed the surface with a scotchbrite pad, cleaned it off and went at it. I thought I might have to tin the plates, but I didn’t even bother and it worked out fine.
Below are the sweeps of various materials that I used in the tests: