Magnetic Core Crossover Inductor Distortion Testing
Inductor Type Comparison Measurements:
To start off I wanted to compare the measurements of the various magnetic core style inductors to the air core at low power since most speakers are not played back at high levels all the time. For these measurements I'm using a 4 ohm load and driving it at 2v (1 watt) and 6.32v (10 watts) which would be approximately 0.5 Amps & 1.58 Amps of current.
I'm measuring harmonic distortion for these tests as many of the non-linearities present themselves as harmonics primarily odd order which are clearly captured in these graphs.
Amplifier measurements for baseline distortion/noise reference
Baseline measurement of the amplifier driving the 4 ohm load at 2v output voltage (1 watt) to get a reading on the nominal noise-floor and distortion of the test setup. Very decent as it's below 0.01% at this level with most harmonics nearing or below 0.001% at low frequencies.
Baseline measurement of the amplifier driving the 4 ohm load at 6.32v output voltage (10 watt). Still below 0.01% but you can see 3rd order harmonics creeping up slightly, noise and the other distortion products are actually reduced at this higher output level.
Dayton 1.0mH 18 Gauge Air Core.
Since the inductor is rolling off the high frequencies there is a decrease in the level of the signal compared to the background noise (SNR) so the slight increase in the rise of noise/distortion at high frequencies is expected and nothing the inductor itself is generating. The same applies to all the following inductors but you will see the amount of rise change based on the value of the inductor.
Otherwise little change in distortion compared to the amplifier itself which is to be expected since air core inductors should not have any inherent distortions. You may notice in the 10w graph the 3rd harmonic is reduced some, since the inductor is rolling off the highs the harmonics from the amp are naturally reduced compared to the fundamental leading to the lower measured distortion there.
Erse Audio 1.0mH 18 gauge Laminated Steel I-Core.
Now you see a rise in odd order harmonics due to the magnetic core of the inductor. This is most likely from the hysteresis in the core but I don't have a definitive way to prove that. You will see a similar trend in all the following inductors.
There isn't too much change between the 1w and 10w drive level just a slight increase in THD dominated largely by the 3rd harmonic.
In both cases the distortion is under 0.1% and for perspective this will likely be in the realm of an order of magnitude less then the distortion generated from most woofers at similar drive levels.
Erse Audio 1.0mH 16 gauge Laminated Steel I-Core.
Again an elevation in the odd order harmonics compaired to the air core or amplifier itself though this inductor actually has the lowest amount of all the magnetic core inductors I've tested, no idea what makes it better.
Similar trend between the 1w and 10w as the others a levels a slight increase in the level of the harmonics.
Erse Audio 0.8mH 16 gauge Laminated Steel Super-Q Inductor.
Similar trends to the others with slightly difference shape/amount of odd order harmonics. Like the other 1.0mH 16 gauge Erse distortion is quite low, below 0.01% in the bass at both drive levels.
Jantzen Audio 1.0mH 15 gauge P-Core Inductor - Permite (ferrous powder).
A surprisingly large amount of odd order harmonics from this inductor design, it's actually the worst of the bunch at the 1w sweep level, you can see some elevated 2nd harmonic in that test level as well. I actually grabbed another one I had to verify the amount of distortion wasn't some kind of fluke with the first one I tested.
The THD is around 0.2% in the midrange on both test levels which is higher then the distortion good midrange/midbass drivers produce at especially at the 1w drive level. However this isn't really deal breaking for most speakers as the distortion from most woofers will still mask that.
Jantzen Audio 4.7mH 14 gauge C-Coil Toroidal Inductor
This one has a rise in the distortion above 1kHz in the 1w sweep which interestingly drops back down in the 10w sweep. However these are designed for bass use anyways and it's only a few hundredths of a percent down there at these drive levels which is basically nothing compared to what a woofer might be generating.
Random 3.5mH 18 gauge Laminated Steel I-Core.
Not the best but not terrible, again 3rd harmonic dominated with the other odd orders showing up as well at reduced levels. Similar performance at 10w around the 0.1% level.
Random 1.5mH 22 gauge Laminated Steel I-Core.
Smaller then the 3.5mH and slightly worse performance but again not terrible in any regard considering it's size and compared to what distortions a loudspeaker would be creating.
Random 0.52mH 22 gauge Ferrite Core.
Not terrible at the 1w level but the large rise in all the harmonics at the 10w level (THD peaking over 1%) shows that the core is beginning to saturate at that point. Basically inductor like this would only be suitable for the smallest of speakers or as part of filters that see very low currents.
Random 7.5mH 20 gauge Ferrite Bobbin.
Distortion at the 1w level is not great but at the 10w level it's already at 1% though the high order harmonics are not as prevalent as they are on the the 0.52mH at this level. Again this style inductor should not be used for high powered/high current designs or applications if low distortion is a goal.
Note - You may notice the peak of the distortion products is shifted lower in frequency on this inductor. The high value of the inductor means the current through the inductor at higher frequencies will be reduced while at the same time the higher value means magnetic flux at lower frequencies is increased which is why the hump in distortion occurs at lower frequencies.