The Great inductor face-off

[Al Klappenberger]

Hi guys,

There has been so much talk about inductor upgrades for speaker

crossover networks that I have decided to do a head - to - head comparison

of all the inductors I have collected while developing my AA upgrade. I

also purchased a .2 mHy Alpha Core spiral inductor specifically for the

test since these seem to have such a wide following.

The measurements have been restricted to the exact value of

inductance of each coil, its DC resistance (DCR) and quality factor ("Q")

at the 6000 Hz squawker / tweeter crossover frequency. The Q of the

inductor is to be considered the most important factor. The higher the

better. Measurements were made with a General Radio 1608-A impedance (in

the picture). This particular one had to be repaired and could not be

calibrated to it specification of +-.1%. I think that it can be considered

accurate to +-.5% however.

[Al Klappenberger]

The picture shows the six inductors tested.

In the back row, from left to right are two coils from North Creek Music.

These are wound of heavy #12 soild wire. They are .3 mHy and .18 mHy. The .3

mHy coil has had several turns added to bring it up to .3 mHy from a lower

value. The other was intended to be .18 mHy

In the right real is a home-made coil wound of #14 soild wire, two layers on

a piece of PVC plastic plumbing pipe. It was desinged to be .2 mHy.

In the front row to the left is the .3 mHy inductor from Solen I am using in

squawker channel of my network. It is wound of #14 "Litz" wire. Next right is

the .2 mHy spiral inductor from Alpha Core.

In the right front is the .2 mHy inductor from Solen. It is also wound of #14

"Litz" wire. It is the tweeter filter inductor I am using in my network.

[Al Klappenberger]

============== Here is the results: ===============

Solen .2 mHy: L=.1986, DCR=.128, Q=43.8

Home-made PVC .2 mHy: L=.2048, DCR=.143, Q=25.8

Alpha Core .2 mHy: L=.1879 DCR=.098, Q=29.4

North Creek .18 mHy: L=.1686, DCR=.058, Q=11.4

North Creek .3 mHy: L=.304, DCR=.080, Q=10.4

Solen .3 mHy: L=.293, DCR=.158, Q=49.8

The .245 inductor from the Klipsch AA network (not shown in the photograph):

Klipsch .245 mHy: L=.2628, DCR=.37, Q=21

CONCLUSION: The clear winners are the coils from Solen wound of Litz wire.

The worst are the coils from North creek music. While good, the Alpha Core

spiral inductor only manages to beat out the home made coil by about 4

points. If you are considering using these, use the Solen coils instead!

At a Q of 21, the coils used by Klipsch are better than I expected and better

then the heavy "upgrades" from North Creek Music. As you can see from the DCR

on the Klipsch inductor it is would of rather thin wire. This proves that

heavy wire is NOT the answer to good Q at 6000 Hz, "Litz" wire however IS the

answer.

I might point out here that the use of Litz wire for the inductors used in

the woffer channel would be of no value. At 400 Hz, only the DCR is important

and solid wire is just as good and my be slightly better for the woffer

inductor.

Inductance accuracy is also an important factor. My networks are built in

matched sets. All the capacitors are measured and manually matched to about

+-1% pairs. The inductors from Solen are wound to 1% tolerance to begin with

which means I do not have to measure and match them. The Alpha Core inductor

came in at about 6% low. This is fine for a less demanding network but not

good enough for use in networks featuring matched sets.

[Randy Bey]

OK Al,

for those of us who do not know the significance of the measurements you made, can you say briefly what the significance is, and whether a higher or lower number is better for each?

[Al Klappenberger]

Sorry Randy,

I wasn't trying to "show" anybody!

"Q", or quiality factor, is the most important thing. It should be the highest number you can get. It is the ratio of inductive effect to total AC losses.

DCR is simply losses (resistance) at DC (zero frequency). It is a portion of the total AC losses. The number shold be as low as possible.

L is inductance in mHy. It should be exactly what is required for the application. If you order .2 mHy from the manufacturer you should get .2 mHy within some tollerance. In the case of Solen, it's within +-1% of .2 mHy.

Al K.

[Derek]

What are you using that accuratly measures DCR at the .1 ohm level?

Derek

[Al Klappenberger]

The General Radio 1608-A will measure R,L or C. It's internal resistance is only about .002 ohm. It will resolve to .0001 Ohms. Accuracy is specked at +-0.1%, +-.005% of full scale. In short, It's pretty d___ close! The thing is sort of old, 1970s vintage, but had a priced tag of well over $1000 when new. It's a laboratory quality instrument.

Al K.

[Derek]

Al,

Ok on the GR stuff. Have a few old pieces myself, seems to just keep going forever.

Not sure where your going with the tolerances you mention. Seems like other variables in the network (transmission line,

rest of cross over, speaker motor, dust) would make the woofer L +/- insignificant at close to DC. The 2 x pie x F and phase angle are not going to change much at audio are they? If you were building a microwave system I would get it.

Thoughts?

Derek

[Al Klappenberger]

Derek,

Yes, you are quite right. The woffer / squawker crossover is made of 6 dB / Octave filters. It's actual crossover frequency is very hard to pin down. Even a slight change in load impedance of the woffer will change it significaltly. The natural crossover is more a function of the horn cutoffs than of the crossover network!

The Q doesn't mean much as 400 Hz either. All that really counts is the DC resistance.

This is where the Alpha Core spiral inductors might be best used since they seem to be low DCR.

Al K.

[lammers]

As you all know, my external xoxers have Goertz inductors w/ Jensen PIO caps and this sounds like magic. Do the inductors make a difference? I don't know. But, I am not changing anything. This is with 300B SET.

I still love you Al K.

If you have Heritage and tubes this WORKS!

[Al Klappenberger]

Lammers,

Yep... If it sounds right, it is! As Sgt. Friday used to say, "Just the facts, mam". My impedance bridge doesn't have ears! It can only measure part of the story.

Al K.

[Randy Bey]

actually then DCR is the predominate value for coils -- in A and AA networks there is only a coil for the woofer.

Then North Creek inductors look like they have the lowest numbers.

I am still confused about the capacitance of a foil inductor. Wouldn't that be a no-no?

[Al Klappenberger]

Randy,

Well, Not quite! The Klipsch "A" network only uses a single series cap, but

the "AA" has an inductor for the tweeter too.

The DC resistance is a constant part of the quality factor of an inductor and

the only one that counts for the woffer inductor. As frequency increases, the

AC losses creep in. That is caused by such stuff as eddy currents across the

wires, skin effect, currents induced into nearby objects causing more eddy

currents and the like.

The North Creek inductor have low DCR because they are wound of #12 wire

compared with #14 for all the rest. The point here is that the heavy wire

didn't help the quality factor at 6000 Hz! The smaller gauge Litz wire had

lower losses at 6000 Hz to spite higher DCR. The conclusion here is that Litz

wire is the thing for the squawker / tweeter inductor. DCR is not as

important. The DC resistance (DCR) however IS the most important factor for

the woffer inductor but the more illusive "Q" or quality factor is the main

factor for the squawker / tweeter crossover inductor.

Distributed capacity of the inductors is not too important for crossover

networks, it seems. I did some computer simulations assuming huge amounts of

distributed capacity compared to what is really there and it had very little

effect. I didn't include it in the data because I really can't measure it

accurately. The coil with the highest distributed capacity was the home made

coil on the PVC pipe because the two ends of the winding are right over top

each other. The spiral inductor from Alpha Core is good because each turn is

in proximity to the adjacent ones only. The means each turn is at nearly the

same voltage as the one next to it. That cuts the effect of the capacity down

to spite the fact that the capacity is actually quite high. The Solen coils

distributed capacity is also quite low.

I think the bottom line is to use the Solen Litz wire inductors for the

squawker / tweeter inductors and the biggest inductor you finances can

support for the woffer inductor.

Al K.

[Al Klappenberger]

UPDATE! -----

I just discovered that the tweeter coil in the Klipsch AA network is being "DE-Qed" by the screw that runs through the center to hold it to the board.

The measurements on it I made before were with the coil removed from the network and connected directly to the bridge. The Q is reduced to only 8.7 by the mounting screw!

Remove the screw and throw it away! Stick the coil to the board using silicon adheasive instead!

Al K.

[kent120]

How come my Heresy do not have coil inductor?

There are two transformers in the crossover though. Are they serve the same purpose?

[Al Klappenberger]

Kent,

There is only one transformer. The other one is the woffer inductor. It's iron core and just looks like a transformer! I don't have a schematic of the Heresy network, but I would assume the tweeter filter is just a single capacitor. That is, 6 dB / Octave. It doens't use an inductor.

Al K.

[JohnA]

OR, an unused section of the autoformer is used as the tweeter inductor. Look at the connections.

John

[Al Klappenberger]

NOPE! It might look like that, but the inductance of the transformer doesn't come into play here. I don't have the schematics on any of the networks for the Herecy though, so I can't sware to that. I am speaking in general terms. Anyhow, the inductance of the transformer would be WAY to high for a tweeter inductor.

If you have the schematics on the Heresy networks, could you post them or emial them to me? I am collecting them!

Al K.

[kent120]

Thanks for the info guys,

The "transformers" don't have any spec on them, so I can't tell what are they:-)

I bought my heresy used, there is no scheme came with them.

However, I can post the picture of the Crossover here once I borrow a digital camera.

[Al Klappenberger]

I now have a copy of the type E network schematic. There is no tweeter inductor used. The other "inductor" is the transformer. It should say T2A on the top. It has a flat frequency response and does not influence the tweeter roll-off at all.

Al K.