[sdiy] Slightly OT: High quality VC attenuator

[Henri Kovalainen]

On Sun, 7 Apr 2002, harry wrote:
> The question is "how much attenuation range do you need"...

> I'm guessing a LOT and that FETs will not have enough range.  I'd
> think of a VCA using the standard differential pair techniques...

Hmm, as far as I've thought about it, some -70dB would be good. I could
probably live with -60 dB, too. I'm not planning on using the three-way
active speakers as headphones. =)

The H11F3's specifications promise something like >300M to <200 Ohm range.
The other resistor in the divider could be as low as 500 Ohm or so. The
opamp should drive a 600 Ohm load quite nicely, so there shouldn't be any
problem there. Besides, the inputs boost the signal +6dB (due to the fact
that the balanced inputs do this by their very nature and this would still
leave healthy headroom in the internal signal level) so if I trim the
coupler to have a minimum resistance of 300 Ohm for example it's possible
to have a nice little gain in the design. Anyway, if the specs are
correct, the amount of attenuation shouldn't be a problem.

I'm more worried about the noise level, the actual distortion figures and
so on. It's a pain to test, too, because to measure it you would
practically have to build a prototype so good that it already matches the
final design. Besides I don't even have proper instruments to test it.

If it comes down to standard VCAs or the likes, I'll probably dash to get
a couple of SSM2018s (140dB gain range, 0.006% typical THD at 1k unity). With
proper trimming they have impressing performance. How to actually trim
them is another problem.

| henri kovalainen   | peltokatu 19 b51   | hekovala at kovalainen.fi
| phone: 0400-683750   90120 oulu finland   hekovala at paju.oulu.fi

| Grow your own dope - plant an economist.