Air Core vs Iron Core - When?

I'm thinking of rebuilding some existing crossovers. These were more or less generic off the shelf crossovers similar to this -

Crossover 3-Way 8 Ohm 800/5,000 Hz 100W |

My current crossovers are 800hz and 5000hz, and that works about right for the speakers I have.

But that brings up a question that I'm sure must have been dealt with before.

On the low end, when and where do you decide to use iron/metal/solid core coils over air core?

If there is no clear guideline, then some of you must have an intuitive feel for what to uses as the low frequency crossover drops lower.

I'm also contemplating other speaker builds that are 3-way low-bass, midbass, tweeter, rather than the current woofer, midrange, tweeter configuration. So, this information would help with that as well.

Just trying to get a general feel for it.

Last edited:
Steve, since you asked a respectable question, I will try to give you a reasonable answer. Of course, air core is better, and many companies have made air core inductors that cost only a small amount of dollars when you want 10 mH or less. Even more is possible, but, of course, more expensive. If you want hi fi, rather than mid fi, a little investment is necessary. I have not used these air coils for audio speakers, but I have used them for notch filtering of 60 Hz, in my lab. They are well made and work well. The slight added resistance is normally trivial, if you use coils wound with a reasonable wire size.

Ron E

2002-06-27 10:41 pm
Cost vs. DC resistance are the main factors. Large air core coils get expensive. Cored inductors can saturate, though. This causes distortion. A rule of thumb is not to let DCR get over 5% of Re. This is really not a a strict rule, but it does minimize changes in driver Q due to DC resistance.

A 4mH air core inductor made of 14gauge wire would weigh about 3 pounds and cost at least $40 for a DCR of 0.5ohms
A 4Mh iron core made of 15 gauge wire costs ~$18 for a DCR of 0.21 ohms
Last edited:
When you get right down to it, it is more important to have the correct value parts than whether it's iron core, air core, poly caps or electrolytics. If you have an off the shelf crossover, it doesn't really matter as the XO is not really doing all it is supposed to anyway. When you say you are rebuilding the XO, does that mean you are then using the T/S parameters to get the actual values? If so then spring for the air cores and poly caps. There are some that will use the good stuff on the signal path and the not so good for shunting.
If you can get a good price on magnet wire, winding coils isn't difficult. I always use air core. If the required size is out of hand, I figure that's a handy cosmic message to move to an active crossover. Though I used to do a lot of passives, today I'd only use an active crossover. Set up properly there's just no comparison and most passively crossed systems sound a bit muddy to me.

On a related note, I was experimenting with output inductors on a power amp and discovered that there was *no* core material, other than air, that didn't increase the distortion at all levels quite measurably.


2007-06-29 8:49 pm
I always use iron core until I get money. :)

Apparently that's a joke. Sorry, but I don't get it.

But thanks for the reply anyway.

To everyone else, also thanks for the replies.

So apparently, when air core get to large and expensive, it is time to consider switching to iron/powered/laminated core coils.

I assumed I was safe at 800hz, but today's design are more often low-bass, midbass, tweeter rather than woofer, midrange, and tweeter, so the low-to-mid crossover is substantially low. I think I've seen crossovers in the 250hz to 350hz range.

As to matching to my drivers, the current crossover (off the shelf) seems to have a small notch at the 800hz crossover frequency. I'd like to flatten that out. Plus these crossovers use non-polar/bi-polar caps which are about 30 years old. I suspect they have drifted a bit, though not by much since you can still see the crossover at 800hz. If they have drifted, I suspect it is in a way that has widened the crossover gap.

I've got an 8 ohm power resistor soldered into some speaker wire, so I can check the impedance of the speaker at the crossover frequency easy enough.

I'm using a midrange that has a pretty wide range so I can check the impedance of that easy enough, and I know the impedance of the tweeter.

Again, this is still in the thought process. And I've got a couple of design floating around in my mind, though I don't have the money to implement then just yet.

I has assumed that metal core coils were better at low frequencies because that is usually where I see them. But, if I'm hearing the rest of you correctly, they are not necessarily better, just cheap and smaller for the large values needed at low frequencies.

Does that about sum it up?



Paid Member
2007-06-10 12:41 pm
Dosen't matter what core you use aslong as it dosen't get near saturation. I have been testing ferite core inductors (only small 1uH) and even at 20A @ 10Khz there were no signs of saturation. For example in the output zobel of an amp I'm working on at the moment I am using a 1uH iron core inductor with the onset of saturation at 78A.
A friend of mine some years ago had a system from one of those one year wonder companies; lotsa ads, dealers, high markups, then gone. It used ferrite cores, and at moderately high levels, they sounded like small explosions when they saturated. A sharp crack noise.

The last project using a Scanspeak 25W woofer needed very low series R. I wound up with the Erse "hammerheaded" laminate, which worked out very well. If you wind your own, and the inductance value is high, the last step before resorting to iron core may be flattened magnet wire. Not cheap, and fits only a few applications, but it may be something worth knowing about.

Audio Grade Coils and Inductors | ERSE
Flat Magnet / Specialty Wire
The most linear sort of inductor, down in the small inductance ranges used in crossovers, are non saturable reactors. This is a regular E/I core, with the I's placed on the opposite side of the E fingers. Means the air gap is basically infinite. The useful bit comes from the linerization of the inductance in the coil, due to the support from the core that cannot saturate.

Take any air coil and a decent inductance bridge and sit and unwind the coil one turn at a time. You will not find the inductance changing by the square of the turns ratio. Instead you will find that the stuff clumps. As an average the reduction of turns will provide a close match to calculated values, but in reality 1 turn or 5 may provide the same amount of change. Air coils are not linear.

Examples of these odd ball inductors can be found on any Kilpsch crossover from the factory and most Altec crossovers. Special order only nowadays though.
Actually, air core coils do behave properly. When unwinding, be sure that the wire that you've undone is stretched out straight, away from the coil; and that you did not displace any of the remaining windings. Also, that the coil is away from any ferromagnetic materials, and preferably in the same position for each measurement. The field behavior that is associated with inductance is very well understood, and has been for a long while. It just can be seemingly complex in it's behavior when all factors are not kept under control.

Incidentally, I tend to prefer copper foil coils; mechanically more solid, less susceptible to vibration?

PS. I didn't understand the EI configuration you were describing. Also, any ferromagnetic material will saturate. With a large air gap, it takes a higher current to reach the saturation level, but it will saturate.
Last edited:
With a large air gap, it takes a higher current to reach the saturation level, but it will saturate.

Agreed. You do want to make sure you are outside and that the DUT is on a cinder block, before you hit it with that higher current, with one of these....


  • 2.6 mH 100 watt choke.JPG
    2.6 mH 100 watt choke.JPG
    49.8 KB · Views: 1,131
For your application, aircores are better. But If you need to use a very low cut off freq. down to 60-80Hz, you will need a massive inductance and thus, for solid state I would personally go for a steel laminate inductor as it will have far less resistance than an equivalent air core. And If you work with 4ohm drivers, this is important.
With the generic off-the-shelf crossovers you linked, it doesn't matter. Any difference between air and iron core is completely overshadowed by uneven frequency response, power response and suboptimal crossover frequencies.

The crossovers you bought will only hit their target frequency/slopes if you add impedance compensation. Even then, you still will not hit acoustic 800/5000hz 2nd order slopes. If you truly want a better crossover, build a proven design with a custom crossover.
Thanks for all the replies, very enlightening.

As to the off-the-shelf crossover that I linked to, that was just to illustrate they type of crossover I currently have. The speakers in question were originally built in 1980 before computer design software was readily available.

They were rebuilt again in 1985 into nicer cabinets and that is when the crossovers and the tweeters came into place, along with nicely made Birch cabinets.

But I was young and naive and miscalculated the ports by a country mile though the speaker still sounded OK. Later, much later, when I was feeling ambitious, I redid the ports, and cleaned up the wiring.

I recently did a frequency plot using an SPL meter and a CD with 1/3rd octave tones on it and noticed a peak of about 10db in the bass at 80hz, which I assume is poor cabinet/port tuning, and a dip right at 800hz of about 8db.

Keep in mind, back in the 80's I was working with minimal tools and even less information, so I did the best I could. The speakers are good, but not great, though I have a real sentimental attachment to them, and would never get rid of them. Bass response is down to about 40hz.

Here is a photo, though again, keep in mind I was young and naive, and working in something of a vacuum.


This image is hosted by another forum, but it was placed there by me, so I think this probably qualifies as fair use of an image.

Attached is the actual crossover, for reference.

The speakers definitely need some more work. So, I'm trying to lay out my plan of attack.

Unfortunately, since the woofers are from 1980, specs are not longer available for them. I even contacted CTS directly thinking they might have an archive of old speaker documentation. But ... no joy.

As you can see, I also have a pair of Wharfedale Diamond 9.6, which I now use as my primary speakers. But, as you all well know, the DIY bug never dies.

Again, thank for all the comments.



  • P825XO-6AS.JPG
    93.1 KB · Views: 829
Hi Steve, good morning from Australia.
(1) Re: an earlier post. I don't think you can assume that iron/ferrite cores are better at low frequencies/low cross over points because they are most often seen there. That is a charitable view on your part. The lower they go the bigger and so it will generally be a COST issue. Air-cored will be bigger, heavier, take longer to wind etc. But on the other hand....
(2) While many people favour air-cored units, the bigger ones do have a great deal of wire in them and I think the point must come where the degradation due to the length of wire the signal has to pass through will need to be considered against the saturation issues. But others here have pointed out that iron-cored is not necessarily EVIL per se.....

But having done your "sums" and costing do consider going active!
People consistently claim bi and tri systems are subjectively "cleaner" or as one earlier post said "passives" are 'muddier' (or something like that).

This is not 'snake oil' from the Golden Ears Dept. The subjective/psychological effect is almost certainly a reduction in IMD in the power amps.....Martin Colloms (UK) was on to that over 30 yrs ago.

Last edited: