Magnetic Core Crossover Inductor Distortion Testing
Inductors that use magnetic cores suffer from two main types of distortions, hysteresis and saturation.
Hysteresis is a property of magnetic cores that when energized with a magnetic field tend to retain a small amount of polarity in the direction of the previous field even after the current through the coil stops. That retained polarity then distorts the output of the following waveform through the coil.
Saturation distortion occurs when the magnetic field generated by the coil exceeds the capacity of the magnetic core. A magnetic core can only hold so much flux before the material is fully saturated, this happens when the polarization of all the atoms in the magnetic core become aligned. At this point an increase the amount of current through the coil no longer results in a linear increase in the flux of the coil resulting in a distortion of the output waveform as the effective inductance value drops.
The purpose of these tests are to to answer several questions I had:
How do these distortions of various magnetic core inductors compare to that of an air core which does not suffer from similar distortions?
What are the actual current or power limitations of various common cored inductor types before saturation distortion occurs?
How does the load impedance affect the power/current rating of the coil?
How does the inductance value affect the power/current rating of the coil?
Test Setup and Process:
The test process is fairly simple, connect an inductor in series with a static resistive load, apply a signal and measure the output at the load.
In order to find the saturation points of the inductors I needed to test the inductors at fairly high currents/power levels, this means a load capable of those levels. I ended up building a set of four resistor banks each of which is approximately 16 ohms and capable of 400w of dissipation. This allows me to test 16, 8 and 4 ohms loads with 1600w of capacity in the 4 ohm configuration.
I'm using an Aiyima A07 amplifier for low level testing (<40w as I only have a 24 power supply) and a bridged inuke 3000DSP for the high level testing. The output signal is measured using a symmetrical voltage divider feeding the input on my Focusright 2i2.
My DIY high power resistive load bank. Each tower uses 27x 47-ohm 15w resistors placed in a series parallel configuration resulting in a total impedance of ~15.7 Ohms.
I calibrate the output voltage of the amplifier when attached directly to the load using a 60hz sinewave generated in REW.
A 1.0mH Dayton Audio 18 gauge air core used as a reference.
The magnetic cored inductors are:
Erse 1.0mH 18 Gauge Laminated I-Core (also have a 2, 3, 4 and 6mH for the mH comparison)
Erse 1.0mH 16 Gauge Laminated I-Core (Laminated core is larger then the 18 gauge)
Erse 0.8mH 16 Gauge Super-Q (I wanted to use a 1.0mH but didn't have one on hand)
Jantzen 1.0mH 15 Gauge P-Core
Some other random inductors of odd values thrown in for these tests as well:
Jantzen 4.7mH 14 Gauge C-Coil Toroidal
Random 3.5mH 18 gauge Laminated Steel I-Core (Smaller then normal Erse/Dayton I-Cores)
Random 1.5mH 22 Gauge? Laminated Steel I-Core (Very Small)
Random 0.52mH 22 Gauge Ferrite Core (Tiny)
Random 7.5mH 20 Gauge Ferrite Bobbin Core.
I tested a couple other inductors on some of the measurements as well.
I decided to break up the measurements into separate sub pages as this would be one massively long page otherwise.
- Part 1: Baseline Low level Inductor Comparison Measurements -
- Part 2: Inductor Power Handling, how much current before saturation? -
- Part 3: Does the inductance value or the load impedance change how much power/current an inductor can handle? -
Saturation distortion of magnetic core inductors is unlikely to be encountered if high quality inductors are used unless the speaker load is low impedance, the design uses large value inductors and is driven to very high power levels. Common magnetic core inductors available to the DIY market require several hundred or even thousands of watts depending on the value of the inductor to saturate while driving a 4 ohm load before distortion was measured. This power rating doubles using an 8 ohm load.
Hysteresis distortion occurred on all the magnetic core inductors tested and resulted in low level distortion that remains fairly consistent regardless of input power. Some inductors have higher level then others. This distortion is often very low in level compared to the harmonic distortion levels generated by common loudspeaker drivers.
Inductors of basic ferrite core construction should be avoided, these quickly saturated at very low levels and generally had higher hysteresis distortion then the others. Those of steel laminate I-core design appear excellent, even the smallest one tested handled nearly 200w into 4 ohms before core saturation became a problem.
If the very lowest distortion from inductors is desired you should use those of air core design which do not suffer from either hysteresis or core saturation distortion. However after these tests I have no issue continuing to use standard steel laminate I-core designs for low pass networks on 3-ways and larger 2-way designs as they provide low DCR in a much smaller package, are more economical and still have much lower inherent distortion then even some of the very best loudspeaker drivers.
Load impedance does not impact the current rating of an inductor, this means an inductor used in a higher impedance design can handle more power. Go from 4 to 8 ohms and the inductor can handle twice the power.
The inductance value (mH) of an inductor does impact current rating before saturation (all else equal), double the inductance and power rating of said inductor before saturation is basically cut in half.
Thus if you double the impedance of the design or driver used, say move 4 ohm to an 8 ohm, while the current is cut in half for a given power level the inductance value needs to double to retain the same crossover point largely negating the halving of current had on the saturation point of the inductor. Though you do come out ahead slightly with the higher impedance design.