How to Shield an inductor?

[Cluene]

I was wondering how to shield an inductor from another so I could place them close in a circuit board?

 

[Cluene]

By the way, the inductor is air core

 

[johansen]

Mount them at right angles to each other, and if you still have too much coupling between them then switch to toroidal inductors.

 

[davenn]

and still another way..... place a metal shield between them with the shield soldered to the PCB groundplane
very effective and often used in RF circuits

Dave

 

[Cluene]

Mount them at right angles to each other, and if you still have too much coupling between them then switch to toroidal inductors.

at what specific angle, if I may ask...?

 

[Cluene]

and still another way..... place a metal shield between them with the shield soldered to the PCB groundplane
very effective and often used in RF circuits

Dave

Forgive me for this stupid question (i'm just but a humble beginner) but, um, what exactly is a groundplane? is that the part or the connection that is connected to the ground of the circuit?

 

[johansen]

"Right angle" means 90 degrees. they should be at least one coil diameter away from each other, preferably more.
What is your application?

 

[Cluene]

i see, isee (>_<) my bad ... i thought you meant correct angle...

I'm making a passive 3 way crossover for our school project..

 

[Nigel Goodwin]

i see, isee (>_<) my bad ... i thought you meant correct angle...

I'm making a passive 3 way crossover for our school project..

Then it doesn't matter, it's NOT a concern at audio frequencies, although mounting them at right angles to each other still makes sense.

 

[ronsimpson]

what exactly is a groundplane? is that the part or the connection that is connected to the ground of the circuit?

On a schematic a ground is just a point, but on a printed circuit board ground is often VERY LARGE. In audio, RF and many circuit, ground is very critical. Ground is not a single trace. It is common for one side of the PCB to be completely covered with copper. GROUND In a digital PCB is is common to have one layer of copper for ground, another complete layer for supply(s) and two layers for signals.

A ground plane, just like a power plane, is a large area of copper dedicated to ground (or another voltage).

 

[cachehiker]

NOT a concern at audio frequencies

Most audiophiles (especially DIY'ers) would not agree. I not only go through the trouble of mounting my inductors at right angles to one another, I make sure they're at least few cm away from one another and any materials (including the speakers) that might affect their inductance. I also verify speaker impedances and resonant frequencies before selecting crossover components.

This isn't exactly the right forum for this, but getting a passive crossover to perform perfectly while the speaker's impedance changes with both frequency and temperature is impossible. Taking the basic precautions to least keep your overall response a couple of dB flatter across the audio spectrum is essential IMHO.

 

[Nigel Goodwin]

Most audiophiles (especially DIY'ers) would not agree. I not only go through the trouble of mounting my inductors at right angles to one another, I make sure they're at least few cm away from one another and any materials (including the speakers) that might affect their inductance. I also verify speaker impedances and resonant frequencies before selecting crossover components.

This isn't exactly the right forum for this, but getting a passive crossover to perform perfectly while the speaker's impedance changes with both frequency and temperature is impossible. Taking the basic precautions to least keep your overall response a couple of dB flatter across the audio spectrum is essential IMHO.

As I said (and you left out from the quote) it makes sense to mount them at 90 degrees, and costs nothing.

Even if you didn't though, I'm absolutely positive you couldn't tell any difference on double-blind tests - and pretty sure you couldn't even tell by measurements.

 

[johansen]

I recall reading a paper by a well known audio engineer (i forget the name, i think it was Douglass Self, but i can't recall) his measurements put practical output inductor coupling below -90Db if they were mounted at 90 degrees, on the same plane, and they were both identical. YMMV

 

[duffy]

I'm absolutely positive you couldn't tell any difference on double-blind tests

Agreed. This "audiophile" thing goes from good, solid concepts to absolute mythology with alarming speed.

And I've been guilty of it myself - I had a nifty "headroom expander" circuit that I was sure made an improvement on my stereo back in the 1980's... until one day I was working on it and discovered I had wired the switch backwards. All the times I thought I was switching it in, I was switching it out. That's when I realized how devious subjective analysis can be.

 

[cachehiker]

A couple of dB isn't noticeable by the average listener as most can't discern less than 3dB. I do audible noise testing and analysis on motors and related mechanisms and notice more than most but I somehow doubt I discern anything below 3dB myself. However, flattening response by one dB with better matched components, another dB with a Zobel network, a third dB with an attenuation network, yada yada ... will eventually add up to a noticeable difference.

The results my Minimus 77 rebuild is testament to some of the general rules of thumb. The sould quality is greatly improved. I wish I had some "before" data but a better crossover flattened a decent hump at 2.3KHz. It resulted from not crossing over more than an octave above the tweeter's resonant frequency. I doubt my attenuation network would've hit the right frequency if I had mounted its inductor too close to the speaker or the metal cabinet or the low pass network inductor. It could've easily made things worse.

 

[Nigel Goodwin]

A couple of dB isn't noticeable by the average listener as most can't discern less than 3dB. I do audible noise testing and analysis on motors and related mechanisms and notice more than most but I somehow doubt I discern anything below 3dB myself. However, flattening response by one dB with better matched components, another dB with a Zobel network, a third dB with an attenuation network, yada yada ... will eventually add up to a noticeable difference.

The results my Minimus 77 rebuild is testament to some of the general rules of thumb. The sould quality is greatly improved. I wish I had some "before" data but a better crossover flattened a decent hump at 2.3KHz. It resulted from not crossing over more than an octave above the tweeter's resonant frequency. I doubt my attenuation network would've hit the right frequency if I had mounted its inductor too close to the speaker or the metal cabinet or the low pass network inductor. It could've easily made things worse.

Sorry, but this sounds entirely like the typical 'audiophile' rubbish, lots of incorrect assumptions and imaginings - and no basis in fact or reality.

If you can't hear the difference reliably in double blind tests it doesn't exist.

 

[cachehiker]

You're saying that a network shifted from the intended -6dB at 2.3KHz to -6dB at <2.0KHz (leaving the adjacent half octave at almost twice its intended volume) would be undetectable?

 

[Nigel Goodwin]

You're saying that a network shifted from the intended -6dB at 2.3KHz to -6dB at <2.0KHz (leaving the adjacent half octave at almost twice its intended volume) would be undetectable?

Probably - are you claiming that's the difference mounting the coils at right angles makes?.

Interesting how you make 6dB 'almost twice the volume' when it's nowhere near that

Like I said - double blind tests are the only way to test it, and I'm sure you couldn't do it.

 

[cachehiker]

Inductive coupling or mounting next to ferrous material can shift 1/(2π√(LC)). I wouldn't be surprised if right angles a few cm apart vs parallel and right next to one another could conceivably shift the resonant frequency of the tank by 10-20%. Gluing them directly to the metal enclosure could easily shift it by even more.

Crossovers are spec'ed according to the -3dB point which represents half volume. -6dB would actually be quarter volume at its absolute minimum but is likely about -3dB or half volume across the half octave centered around the resonant frequency of the attenutation network.

At their resonant frequency, the tweeters were essentially producing 100% volume at the same time as the woofer was also producing 100% volume. Prior to adding the tweeter attenuation network, the speakers were always heavy on the lead guitar and alto vocals. Now they sound almost as good as the homebuilt ported Zaph 5.2's with the exception of their limited bass extension. They don't hold a candle to the old school 4-way Pioneers. To be 100% honest, I can't audibly detect the other major crossover modification intended to squelch the woofer's cone breakup at about 4KHz.

It appears that audiophiles are a favorite target of defamation on this forum.

 

[Nigel Goodwin]

It appears that audiophiles are a favorite target of defamation on this forum.

They are when everything they say is 'could' or 'may', and are unprepared to conduct double-blind tests.