What does an inductor do in a speaker crossover?

[goraman]

I always see 2 spools of wire inductors in every cross over what do they do,I also noticed they are almost always next to a resistor.

[Stratmangler]

Low pass filter.
Sometimes they have a resistor in circuit to stabilise the speaker drive unit impedence.

[goraman]

Low pass filter.
Sometimes they have a resistor in circuit to stabilise the speaker drive unit impedence.

I'm still not getting it,you mean it filters out low frequencies?

[Reid Malenfant]

In the case of an inductor feeding a bass/mid it rolls off the treble. An inductor has a property known as inductive reactance & effectively what this means is that the resistance of the coil of wire increases the higher the frequency goes - so it rolls off the treble because it's in series with the bass/mid!

In the case of a tweeter it will be in parallel with the tweeter & as it's resistance increases with frequency it will tend to shunt less power to ground the higher the frequency goes ;) In co-operation with a capacitor this will increase the power to the tweeter as the frequency increases.

Resistors are generally on the tweeter side as these tend to be a bit more efficient than bass/mids :)

[Stratmangler]

Passes them on.

[goraman]

Ok,so inductance creates the roll off,I always thought caps did that.

[Stratmangler]

Caps are hi-pass.

[Reid Malenfant]

Capacitors do the opposite :) Well no, they do create a rolloff but in the opposite sense :eyebrows:

Capacitive reactance as it's known creates a reducing impedance (resistance) with increasing frequency.

So imagine your bass/mid again, it has a choke (inductor) feeding it in series & a capacitor shunting the driver. As frequency increases the choke will increase in its AC resistance & the capacitor will decrease in it's AC resistance - it creates a steeper rolloff :)

In the case of a tweeter being fed in series from a capacitor & having an inductor shunting it you get this... As the frequency decreases the capacitor displays more AC resistance thus rolling off the power to the driver & the inductor also has less AC resistance also creating a steeper rolloff..

[goraman]

It seems a miracle speaker crossovers work at all.
What alot going on for so few parts.

[Reid Malenfant]

It seems a miracle speaker crossovers work at all.
What alot going on for so few parts.
Jeff it gets worse as no drive unit has a fixed impedance, they all vary all over the place as well which can bugger up the crossover frequency & driver phase. As you can imagine this make things more complicated :eyebrows:

& just to make things even worse the driver sensitivity varies over frequency as well :eek:

So yeah, it's a minefield out there :eyebrows:



This is why I like electronic crossovers as they are accurate. A whole lot of problems get left behind ;)

[goraman]

I have owned Electrostatics, how about 3 to 100 ohm impedance curve !
That's why you need a 355 watt per channel amp and tubes do better with high impedance than SS. when running huge panels.

[Reid Malenfant]

Actually tubes prefer a low impedance as a high one can upset them serioiusly :) Solid state stuff doesnt mind an open circuit & doesn't like too low an impedance.

So you kind of have it arse about face Jeff, or backwards if you like :)

Never connect a valve amp up with no speaker or load. A solid state amp wouldn't give a toss :eyebrows:

[goraman]

Actually tubes prefer a low impedance as a high one can upset them serioiusly :) Solid state stuff doesnt mind an open circuit & doesn't like too low an impedance.

So you kind of have it arse about face Jeff, or backwards if you like :)

Never connect a valve amp up with no speaker or load. A solid state amp wouldn't give a toss :eyebrows:

I was given wrong information by a salesmen trying to sell me some big KT88 amps then.

Glad i didn't bite.

[Reid Malenfant]

Well to be 100% truthful it depends on the amp in question :eyebrows: Without knowing the tube amp design you won't know if it's safe to work with little or no load, so the safest thing to do is assume that it won't be & give it a proper load :)

[Stratmangler]

Never connect a valve amp up with no speaker or load. A solid state amp wouldn't give a toss :eyebrows:

Not quite true - a single ended amp isn't bothered by having no load.
Never connect a push/pull valve amp with no speaker or load is another matter.

[Reid Malenfant]

Well to be 100% truthful it depends on the amp in question :eyebrows: Without knowing the tube amp design you won't know if it's safe to work with little or no load, so the safest thing to do is assume that it won't be & give it a proper load :)


Not quite true - a single ended amp isn't bothered by having no load.
Never connect a push/pull valve amp with no speaker or load is another matter.
Please look at what I said in the previous post you obviously overlooked ;)

:rolleyes:

[goraman]

They would have to be some pretty big frickin tubes to drive a massive pair of panels single ended!

[DSJR]

That sales guy wasn't a total loony :eyebrows:

"High Impedance" in most speaker cases is around 12 - 15 ohms and this average suits "typical" valve amps really well, as below 8 ohms or so, the higher output impedance of the output transformers may become an issue I understand - unless you are lucky enough to have some Radfords, which were superb in this respect.. Makes a great case for active drive, where modern electronic crossovers can be designed quite easily and have far less interaction than a typical passive crossover..

Having said that - Jeff, have a look at the Tekton site. The M-Lore's look really promising, as they appear on face value to measure properly, have incredibly high efficiency and are cheap to buy if you're US based - getting 'em here a pair at a time doubles the price sadly.. I believe they use Eminence and Fostex drivers to good effect in a variety of designs ;)

[Toppsy]

They would have to be some pretty big frickin tubes to drive a massive pair of panels single ended!

Yes and no.
However, you do need to use DHT transmitter valves of the likes 211, GM70.

I use SET GM70 monoblocks to drive my Quad ESL63's. 35 Watts from a single GM70 and that's running them quite mild at around 1200 Volts. Not for the faint hearted. But they don't half sound good.

Reid Malenfant wrote
Quote:
Originally Posted by goraman View Post
It seems a miracle speaker crossovers work at all.
What alot going on for so few parts.
Jeff it gets worse as no drive unit has a fixed impedance, they all vary all over the place as well which can bugger up the crossover frequency & driver phase. As you can imagine this make things more complicated

& just to make things even worse the driver sensitivity varies over frequency as well

So yeah, it's a minefield out there


One very good reason for Single fullrange driver speakers like Back Loaded Horns etc. NO CROSSOVER! The signal is fed direct to the driver from the amp, so no colouration of the signal by the XO components.

[Reid Malenfant]

One very good reason for Single fullrange driver speakers like Back Loaded Horns etc. NO CROSSOVER! The signal is fed direct to the driver from the amp, so no colouration of the signal by the XO components.
Or actively driven speakers :)

But then I do agree your full range driver will still have one advantage... Coincidence, that is assuming it is a proper single voicecoil driver ;) Though they are rarely "full range" if the truth be known..

[Ali Tait]

Switching on a push pull amp without speakers connected is fine as long as the output valves are triodes. Different matter with other valve types like pentodes. SE amps don't mind at all.

[Reid Malenfant]

Switching on a push pull amp without speakers connected is fine as long as the output valves are triodes. Different matter with other valve types like pentodes. SE amps don't mind at all.
:eyebrows: See what i mean about it being a minefield?

Best to always assume you need a full load with a valve amp :) It's what I'd do at any rate, then you can't go wrong ;)

[Magna Audio]

a picture helps

http://inlinethumb31.webshots.com/49438/2741105970071997495S600x600Q85.jpg (http://entertainment.webshots.com/photo/2741105970071997495OViCBS)

Here is simplest 1st order form, the cap is stopping the tweeter getting low frequency (would be harmful at power), so it can produce nice sounds in it's comfort zone and the inductor is stopping the woofer trying to play higher frequencies at which it is crap.
The crossing point I choose is 1000Hz (what Tannoy use on 15" HPD/Golds etc...

As this is the most basic crossover the ramp up and down slope are shallow 6dB / octave.
For steeper you need more components.

Going up the scale to 4th order you get 24db/Octave and use lots of components = expensive and signal sapping.

http://inlinethumb64.webshots.com/48575/2272534910071997495S600x600Q85.jpg (http://entertainment.webshots.com/photo/2272534910071997495ZucGzJ)

This might seem like a good idea as the woof does less tweeting and the tweeter less woofing - should be good, but pushing the precious signal through all these chokes and caps does sap detail / clarity and dynamics so they use as few as possible / trade off.

I used 2nd order 12dB/octave as a compromise...

For a 3 way you get

http://inlinethumb04.webshots.com/10307/2405074370071997495S600x600Q85.jpg (http://entertainment.webshots.com/photo/2405074370071997495chOkAx)

You can mix driver impedance and the crossover compenent values vary to take that into consideration.

Here the mid band is both avoiding receiving too low or too high so you have both in there - unavoidable.
You can choose drivers that tail off naturally sometimes - the JBL2435Be's I use as upper mids go to 9000-10000Hz and then drop off naturally so no need for the components to pull them down as they are doing that anyway. Sometimes you can do the same with woofers just rising naturally - sometimes not as they comb or just sound bad.

Here's the tool I use
http://www.erseaudio.com/CrossoverCalculators

There are also L-pads to reduce the volume / gain of a more efficient driver - devils work as they sap clarity / dynamics. I tried these as a stop gap but doing it at line level is much better and more controllable IMO.

And Zobels to stop a drivers impedance raising with frequency. I use those on the bass and sub bass drivers in my system.

You can of course use Digital Signal Processing and get 48dB / Oct slops, L-pads, time delays to align drivers and more, all done in the digital domain but it def needs an amp per driver pair and converting your analogue signal (should you have one) to digi and back again is not the best thing to do to it!
Some use a cheapie such as the http://www.behringer.com/EN/Products/DCX2496.aspx
play around measure and get the crossover points they like and then build a passive one - as they sound better than the 2496.
You can get expensive ones like http://www.deqx.com/ for several thousands. £6.5K to do my 5 way horn system - gulp! Some can do room correction too...
I tried Anti-mode 8033 on Reid's recommend and like it but that is only messing with 10-150Hz and it does not muck it up with DAC and ADC at those frequencies and there's little else to can do.
Fun and Games.

[Ali Tait]

:eyebrows: See what i mean about it being a minefield?

Best to always assume you need a full load with a valve amp :) It's what I'd do at any rate, then you can't go wrong ;)

Aye I suppose so, unless you know yer valves.

[NRG]

Depends on the PP amp Ali, not all will complain... ;)

[Marco]

Welcome back, Neal - long time no see! :)

Marco.

[goraman]

Man oh man look what I started...

[Magna Audio]

come on - you asked for it:lol:


Man oh man look what I started...

[Marco]

Hi Steve,

Your excellent pics have disappeared, mate (now getting a "Temporarily Unavailable" message) - makes the thread look rather untidy...

Can you fix it? :)

Marco.

[NRG]

Welcome back, Neal - long time no see! :)

Marco.

Thanks Marco! http://theartofsound.net/forum/images/icons/icon14.gif

[Marco]

Nice one. Try not to be such a stranger in future :)

Marco.