Iron-core vs. air-core inductor distortion in speaker crossovers.

[terryforsythe]

Uh... These bobbin inductors are the cheapest and suffer hysteresis and especially eddy-current losses more than laminated.

Think about the eddy-current losses. What is the DCR of iron powder? What is the DCR of steel? (Hint, steel has much lower DCR). Now, apply Watt's law and Ohm's law, P=V^2/R. So, no, iron powder cores have lower eddy-current losses than laminated steel, generally speaking.

With hysterises, it can vary depending on the properties of the material. Generally, iron powder cores have lower hysterises than laminated steel because the magnetic domain structure - when they get to the point of saturating their orientation change slows, reducing hysterises losses until they start re-orienting.

 

[MAB]

Thanks for the measurements.
Max SPL is over 105 dB. What was the measuring distance mic-driver? What was the voltage of the measuring signal?

0.5 meter
20 Volts
The B&C 5NDL38 driver claims 91 dB sensitivity at 1 meter and 1 2.83 Volts.

I also just did sweeps. I'm trying not to heat the core. I seem to have been successful since that would have made a difference even at audio frequencies, and I observe none.

I wanted to push it, but keep things within reason!

 

[Sokel]

Thanks for the measurements.
Max SPL is over 105 dB. What was the measuring distance mic-driver? What was the voltage of the measuring signal?

By eye and experience, no more than 10cm if in room, I see no gating and all the stuff we see at such distance.

Edit: 0.5 meter??? That's far for such result, it's maybe the small slope that tricked me.

 

[MAB]

Think about the eddy-current losses. What is the DCR of iron powder? What is the DCR of steel? (Hint, steel has much lower DCR). Now, apply Watt's law and Ohm's law, P=V^2/R.

With hysterises, it can vary depending on the properties of the material. Generally, iron powder cores have lower hysterises than laminated steel because the magnetic domain structure - when they get to the point of saturating their orientation change slows, reducing hysterises until they start re-orienting.

Steel laminates have plates individually insulated from each other, and oriented to specifically minimize the eddy current. It's why they are desirable for some applications. And more expensive. You are mixing things up.

 

[MAB]

By eye and experience, no more than 10cm if in room, I see no gating and all the stuff we see at such distance.

Edit: 0.5 meter??? That's far for such result, it's maybe the small slope that tricked me.

It's in my greenhouse.
TBH, it might have been slightly closer than 50 cm. I didn't measure precisely.

I am trying out speaker measurements there since it is wind-free and cavernous.
It does make things difficult. Like swapping the inductors while on a ladder and not moving the speaker or mic while changing setups.
I'm also imagining speakers tumbling off the platform.
And neighbors getting freaked out since it does carry.

 

[terryforsythe]

Steel laminates have plates individually insulated from each other, and oriented to specifically minimize the eddy current. It's why they are desirable for some applications. And more expensive. You are mixing things up.

It is correct that the purpose of the laminations in the steel cores is to lower the eddy current losses. The eddy current losses in iron powder cores is low enough that it is not as much of a concern, and thus laminations will not accomplish much, if anything. The DCR of the iron powder is high enough to make it not a significant issue. The same goes for ferrite cores.

EDIT: Use your multimeter to try to measure the DCR of your iron powder core.

Aside from saturation concerns, iron powder and ferrite cores will have lower power losses. As noted, the molecular structure of the steel also matters. Silicon steel, for example, has much lower power losses than cold rolled steel.

 

[MAB]

By eye and experience, no more than 10cm if in room, I see no gating and all the stuff we see at such distance.

Edit: 0.5 meter??? That's far for such result, it's maybe the small slope that tricked me.

For calibration, here is the same test of the two 3.5 mH inductors in my room, at 0.5 meters:

No gating. So all of the room artifacts are present. I think my test in the greenhouse was a bit closer than 0.5 meters.

The distortion is matched between inductors:

The even and odd HD are also matched. 3rd HD for instance:

Even cheap iron-core inductor from a mass-produced speaker is going to have matched distortion at reasonable volume. The iron-core inductors in this test are from a very cheap pair of mass-produced speakers.

 

[MAB]

EDIT: Use your multimeter to try to measure the DCR of your iron powder core.

This wouldn't tell a thing. I'm not sure if you are serious.

 

[terryforsythe]

This wouldn't tell a thing. I'm not sure if you are serious.

Think about what is eddy current. Eddy current is electrical current in the core. Now, think about Ohm's law; I = E/R. (Side note: in the case of eddy currents, the electromotive force that drives them primarily is the alternating flux in the core, though the copper windings also contribute). Per Ohm's law, as you increse the resistance (R is the denominator), the current decreases. With steel, the resistance is relatively low (more so for cold rolled steal than silicon steel), thus eddy currents become appreciable. Again, it is a bigger issue with cold rolled steel vs. silicon steel. With iron powder, the resistance typically is much higher, and the eddy currents are much lower.

EDIT: Another side note: Both hysterises and eddy current losses are lower in silicon steel in comparison to cold rolled steel. To get an indication of which type of steel a core uses, silicon steel typically is a dull dark grey color. Cold rolled steel (or whatever crossover manufacturers are using these days) typically is plated to prevent rust, and thus typically is bright silver in color.

In my first post I referred to a blind test I participated in comparing inductor cores. That test compared a typical laminated steel core inductor using the bright shiny laminations to an inductor using a larger silicon steel inductor core, both with the same inductance value. The upper midrange is where all three of us heard the difference.

 

[Vladimir Filevski]

Think about what is eddy current. ........ With steel, the resistance is relatively low (more so for cold rolled steal than silicon steel), thus eddy currents become appreciable.

Either I don't understand you because of my not so good English, or you are talking nonsense.

 

[terryforsythe]

Either I don't understand you because of my not so good English, or you are talking nonsense.

Not nonsense. Why do you think laminated steel cores are commercially available in various lamination thicknesses? Indeed, why even use laminations at all? The answer is that laminations are used to reduce eddy current losses. The thinner the laminations, the lower the losses (EDIT, and the more expensive the core).

 

[MAB]

Think about what is eddy current. Eddy current is electrical current in the core. Now, think about Ohm's law; I = E/R. (Side note: in the case of eddy currents, the electromotive force that drives them primarily is the alternating flux in the core, though the copper windings also contribute). Per Ohm's law, as you increse the resistance (R is the denominator), the current decreases. With steel, the resistance is relatively low (more so for cold rolled steal than silicon steel), thus eddy currents become appreciable. Again, it is a bigger issue with cold rolled steel vs. silicon steel. With iron powder, the resistance typically is much higher, and the eddy currents are much lower.

EDIT: Another side note: Both hysterises and eddy current losses are lower in silicon steel in comparison to cold rolled steel. To get an indication of which type of steel a core uses, silicon steel typically is a dull dark grey color. Cold rolled steel (or whatever crossover manufacturers are using these days) typically are plated to prevent rust, and thus are bright silver in color.

In my first post I referred to a blind test I participated in comparing inductors. That test compared a typical laminated steel core inductor using the bright shiny laminations to an inductor using a silicon steel inductor core. The upper midrange is where all three of us heard the difference.

Think about steel plates, insulated from each other and laminated perpendicular to the eddy current.

For example just about every manufacturer of inductors explains this clearly.
The use of laminated is often to minimize eddy currents over other designs, despite the bulk conductive properties of the individual plates.
Still, all of this only matters well above 100 kHz.

You are arguing by statement of individual facts, but getting applications wrong. Over and over. It's quite tedious. Each time I point out your multiple fallacies, you move along to the next bogus argument. Please get a laminated core inductor, do some measurements. Do a careful job so as to not confuse other phenomena with the inductor measurement.

 

[Vladimir Filevski]

Not nonsense. Why do you think laminated steel cores are commercially available in various lamination thicknesses? Indeed, why even use laminations at all? The answer is that laminations are used to reduce eddy current losses. The thinner the laminations, the lower the losses (EDIT, and the more expensive the core).

Still nonsense.
Thickness of the laminated steel cores found in inductors is usually 0.5 mm, which is more than enough for low to mid frequencies. Expensive inductors may use 0.35 mm steel thickness.
Cores in output transformers for valve amplifiers typically use 0.25 mm steel because they have to work well with high frequencies up to 20 kHz - which is totally unnecessary for inductors intended for loudspeaker crossovers.

 

[terryforsythe]

Still nonsense.
Thickness of the laminated steel cores found in inductors is usually 0.5 mm, which is more than enough for low to mid frequencies. Expensive inductors may use 0.35 mm steel thickness.
Cores in output transformers for valve amplifiers typically use 0.25 mm steel because they have to work well with high frequencies up to 20 kHz - which is totally unnecessary for inductors intended for loudspeaker crossovers.

Still not nonsense. But, hysterises losses are a bigger issue than eddy current losses for laminated steel I-cores.

 

[a4eaudio]

Eddy currents....

 

[terryforsythe]

Think about steel plates, insulated from each other and laminated perpendicular to the eddy current.
View attachment 492862
For example just about every manufacturer of inductors explains this clearly.
The use of laminated is often to minimize eddy currents over other designs, despite the bulk conductive properties of the individual plates.
Still, all of this only matters well above 100 kHz.

You are arguing by statement of individual facts, but getting applications wrong. Over and over. It's quite tedious. Each time I point out your multiple fallacies, you move along to the next bogus argument. Please get a laminated core inductor, do some measurements. Do a careful job so as to not confuse other phenomena with the inductor measurement.

You ran a test and produced good data. I commend you for that. But, over and over you state things that just aren't so. E.g., eddy current losses in iron powder cores is higher than in laminated steel cores, use of laminations only matters above 100kHz, etc. With regard to the 100kHz comment, why do think crossover manufacturers even use laminated inductor cores? It would be cheaper to just use a cylindrical steel slug. Why do powerline filters designed for 50/60Hz and 400Hz frequencies use laminated steel inductor cores? To lower core losses, that is why.

Side note: the inductor depicted in the picture you presented is using a silicon steel inductor core. If you are going to use a steel inductor core, that is the type to use.

 

[Vladimir Filevski]

The iron-core inductors in this test are from a very cheap pair of mass-produced speakers.

I have found that distortion in bobbin-type iron powder core varies wildly between different core materials - some have very low distortion (just as yours!), but some start to distort with only 8-10 V (the very cheap ones). Sorry, I can't post my measurements - they were on my old hard-disc drive which crashed.
Anyway, I am using laminated steel core inductors only, because most of loudspeakers I am manufacturing are professional high power loudspeakers. With oversized cores, no worries about distortion here.

 

[Vladimir Filevski]

Still not nonsense. But, hysterises losses are a bigger issue than eddy current losses for laminated steel I-cores.

Still nonsense - laminated steel cores are used in high-fidelity output transformers in valve amplifiers (and in good quality inductors), which have very low distortion. Hysteresis and eddy-current in those cores are very low - they are called high-fidelity output transformers not for nothing!

 

[terryforsythe]

Still nonsense - laminated steel cores are used in high-fidelity output transformers in valve amplifiers (and good quality inductors), which have very low distortion. Hysteresis and eddy-current in those cores are very low - they are called high-fidelity output transformers not for nothing!

I have yet to hear a valve (tube) amplifier that sounds as good to me as a well designed solid state amplifier. But, I gave up on them decades ago. Some valve amplifiers may be better than others.

With regard to core losses, the next time you have your valve amplifier being pushed hard, put your hand on one of the output transformers. If the core losses indeed are very low, it should feel cool to the touch, i.e., not more than a 5-10 degrees over ambient temperature, depending in part on the size of the magnet wire used to wind the transformer.

 

[MAB]

I have found that distortion in bobbin-type iron powder core varies wildly between different core materials - some have very low distortion (just as yours!), but some start to distort with only 8-10 V (the very cheap ones). Sorry, I can't post my measurements - they were on my old hard-disc drive which crashed.

I perhaps haven't encountered the bottom of the market. I can dig through my parts drawer after the holiday here.
I can imagine materials with hysteresis large enough to distort. At some point saving on money on the bobbin makes no sense because of the cost of the copper wire. In this case these iron-core inductors are quite small wire gauge.

Anyway, I am using laminated steel core inductors only, because most of loudspeakers I am manufacturing are professional high power loudspeakers. With oversized cores, no worries about distortion here.

Yes! The lamination addresses these issues.

I am trying to stay away from high power / high volume applications. Which of course cloud the picture with the thermal issues occurring while measuring, I have never been sure if I am measuring properties of the core, or just different thermal properties of the inductor.