There is a great amount of disagreement about the effect of backplate/frame material on the self-resonant response of a pickup, so I thought I would perform some experiments and do a little math in an attempt to get to the truth. Bear in mind that this is information for a BACKPLATE, not a cover (as in behind the flux loop, NOT in the flux gap between string and pickup). A COVER would have a similar but close-to-opposite effect, and I will blog about that in the future.
Bottom line: the effect was slight, but audible.
For this experiment, the baseline/control was a pickup coil sitting on top of a piece of wood. The difference in peak amplitude between highest and lowest backplate material was about 2dB, and the change in frequency of the peak was about 300Hz. I hooked it up to my self-resonance testing rig (220k load, 1000pF shunt capacitance) and ran a 20-20kHz sweep through each.
The materials were: etching Copper, Aluminum alloy, instrument-grade Brass, Nickle-silver, Copper-clad fiber PC board, and mild steel. The Mild steel was added for reference as the only ferromagnetic material. The pickup was not sitting directly on the plate, but instead had an air gap of 1/16″.
The results fan out from the control, with the materials closest to the baseline having the least tonal change, fanning out from the center. The range of tonal change was as follows (ordered from most-frequency-shift-to-least-frequency-shift with the control (wood) in the middle:
//max +300Hz, -0.1dB
Red = Copper
Green = Alum
Yellow = Brass
Grey = baseline (wood)
Purple = Nickel-Silver
Orange = Copper-Clad fiberglass
(Blue = Mild Steel)
//max -10Hz, -1.5dB
Here is the sweep:
Despite the low resolution of the pic, you can see that generally brass/copper/alum left the amplitude mostly untouched (slight flattening), but shifted the frequency upward toward the table frequencies, whereas nickel-silver/clad-PC-board/steel flattened out the peak, but only effected the frequency a bit (dropping the peak frequency ever-so-slightly toward the bass frequencies). So relatively-speaking, of the non-magnetic metals, copper is the brightest and nickel-silver the more transparent. Non-conductive, non-ferromagnetic materials would be truly the most transparent (wood, plastic, resin, etc).
Essentially it tells us that eddy currents will drop the inductance slightly, moving the behavior slightly closer to an air core. (I tested this effect in the pole piece blog.)
Here are the resistivity ratings of each material:
- copper = 1.7
- alum = 2.8
- brass = 6-7 (70%Cu,30%Zn)
- nickel = 6.9
- steel = 16
- nickel-silver = 30 (55% Cu, 27% Zn, 18% Ni)
The two materials closest to the control were Nickel-silver and Brass, with brass having a slightly “brightening/broadening” effect and nickel-silver having a slightly “warming/narrowing” effect. Both materials are basically a copper base with additional metals mixed in. The difference was about -0.5dB across a spread of about 100Hz, which is essentially negligible, but still there regardless, and goes to show why these two materials have become among the most popular for humbucker back-frames along with plastic/fiber (most transparent).
Here is a graph of the control against just those two materials:
Here is a quick (classic classroom) demonstration of the effect of eddy currents. It can also be used to guesstimate the effect of a backplate material on the frequency response in an unknown material:
Other things that we can assume are that the thickness of the material has an effect (which is why I included a copper-clad board along with the copper plate), and that high-permeability ferromagnetic conductive materials (the mild steel) have the opposite effect of low-permeability, non-magnetic conductive materials (like copper).
Consequently, if there is a desire to use a highly-conductive material as a baseplate (like copper for grounding or EMI shielding), but the sonic effect of the eddy currents are undesirable, the effect can be minimized by decreasing the thickness, using the material in strips, using a laminated structure, or perforating the material with holes or slits. This will minimize the circular area of the currents. Also, the backplate had greater effect the closer the proximity to the pickup, and (in this setup) was greatly diminished beyond .125″-.25″ or so.
The backplate in the experiment was not grounded, so capacitive effects are not included in this experiment, but that will be covered in a future post.
It is interesting to note (well… interesting to me anyway) that by running the pickup backward as it were, and injecting a signal INTO the coil, we have essentially turned it into a rudimentary eddy current sensor, which are essentially used to measure the movement and thickness of materials. They essentially provide a plot of strength/distance/thickness/flaws/etc of a non-magnetic, conductive material by inducing eddy currents, which venerate a small magnetic field which (like we saw in the video demonstration) which induces a voltage back on the coil. It is calculated through the CHANGE in effective impedance of the coil. The interaction is complex enough that it is normally tested, not calculated, though computer modeling makes calculation much more viable these days (see my older posts for an intro to FEMM modeling). The reason for the complexity is that the effective impedance is made up of quite a few variables:
coupling coefficient (hence proximity)
eddy current path Q value
Where NONmagnetic material is measured by the reduction of inductive reactance (the coil’s effective impedance), ferromagnetic materials (like the mild steel… and the ferromagnetic components in the nonmagnetic alloys) have a stronger field (of course) will ASLO work the same way IF the eddy currents are strong enough to counteract that higher field (which is a function of the material’s permeability).
This is all supported by the test we ran above where we saw the largest delta in the highly conductive, nonmagnetic materials.
So to give a general guideline for materials, the coil impedance will shift further upward for higher-permeability, higher-resistivity, ferromagnetic material (with something like mumetal having the most effect), and will shift further downward for lower-permeability, lower-resistivity, nonmagnetic material (with something like copper or aluminum having the most effect). And as we saw, the slopes of those two general categories will be opposite each other in terms of impedance-to-distance. Alloys like steel, brass, bronze, etc will fall in the middle, depending on their ratios of the 3 elements that we just mentioned(ferromagnetic properties, permeability, resistivity).
It is worth noting that, as seen in the graphs, the effect is subtle. BUT it has been remarked upon by enough players that it is still worth considering.