On active filtering

[Dr.EM]

Don't know if there is a better forum for this, feel free to move topic if you feel there is.

Anyhow, when using active filters do you feel there would be any significant advantage to attenuating after the filters rather than before? The idea is naturally that the filters are handling a higher level signal with the aim to lower noise and possibly crossover distortion. I can see practically that with good quality op-amps this may not be relevant, but theoretically it seems sound; the attenuated level can be many times lower than the original signal level.

There is also the issue of headroom, but I know that with a typical op-amp +/-15v supply a CD plyer or other source won't get anywhere near there.

Incidentally, does anyone have any opinions on this chip:
http://www.rapidonline.com/productin...moduleno=35600

It appears to act as a passive in the signal chain and the resistor networks in the chip will naturally be highly accurate. Seems like a good option for volume control at a modest cost. My thinking is that these could easily be interlinked if the above idea is worthwhile, 3 chips would be needed for 3 way for example.

Just to throw in another idea, balanced active crossovers. High end amplifiers and sources often use balanced connection so it would seem a shame to break that at the crossover stage. How would one implement a balanced active crossover; would 2 crossovers (1 for each line in the pair) be needed?

Discuss

[AndrewT]

Hi,
most schematics I've seen show the attenuators after the filters.
This seems to be normal.

As for the digi volume control.
Do a search and see what others said about sound quality.

[Khron]

That chip looks fine, but i'm afraid you'd need a microcontroller of some sort, to be able to command/control it.

In my active crossovers, i'm going to use some Dallas/Maxim DS1802's, as per Mr. Pedja Rogic advises ( http://www.pedjarogic.com/spot/index.html ). You can easily get a few of them as free samples (i got 4, for a pair of 2-way crossovers). All you'd need is a few TL431's, a few film caps and a few pushbuttons, and voila

And regarding balanced crossovers... I don't see much point in keeping them balanced, unless they're some distance away from the power amps. Myself, i designed my boards with a balanced input and single-ended outputs. Full-balanced would've ment an extra 3 dual opamps and 2 singles, plus a whole lotta "real-estate"

[SY]

Balanced connections are advantageous for sending signals, but inside the box, the advantages are minimal. Inside the crossover box, I'd convert the balanced signal to single-ended, do the crossover thang, then convert the output to balanced.

Of course, unless you're running very long leads or are in an unusual environment, there's no real advantage to balanced transmission at line level, either, but I know there's no way to talk you out of that.

[Dr.EM]

Thanks for your replies

Thats interesting about the attenuators usually being after the filters, I didn't realise that. I think I would do it that way myself if I assembled a design. I have a freind who may be able to sort out a controller for that chip.

Your probably right about balanced transmission generally being unrequired at line level. One reason I found it attractive is in how it may help with earth loops. If you were running 3 earthed power amplifiers it could definately cause a problem, but with balanced you could simply break the shield connection at one end of the interconnect (theoretically you can do this anyway but last time I tried this in practice I found that not to be the case, humming was worse than the earth loop).

[richie00boy]

If your interconnects are balanced you don't need to break any shield connections to kill ground loops, that's half of the point of using balanced.

[Dr.EM]

I'm probably missing something but if a shield exists then it will invariably join the grounds of two bits of equiptment if left complete? Unless you are saying the balanced interconnect has no shield at all, which makes it different to the micropohone cables i'm used to.

[richie00boy]

No there is a shield. Just that the signal is now taken as the difference between the hot and cold wires, which both have equal amounts of ground loop hum on them, so the hum is ignored.

[SY]

Generally, a shield is only connected at the sending end (assuming balanced transmission).

My own setup is triamped, no balanced lines, no hum; I did need to use an input transformer on my preamp because of the hum induced from my satellite dish, but for all other sources, hum loops are a non-issue. It's just a matter of getting the grounds right in each of the other components. Bill Whitlock's papers are terrific guides as to how it's done.

[Dr.EM]

Thanks for explaining that

It is possible to have non-earthed amps for the mid and high portions by using double insulated E-core transformers. Alternatively you could just not earth the secondary of your toroid and accept the potential (though extremely low?) risk.