AW: Phaser

[Haible_Juergen#Tel2743]

> I just built a small phase shifter, from the Elektor
> Formantbook 2. It uses 6 allpass filters as follows:

Great little circuit! I built this one years ago, and it sounds
very smooth. I think it has to do with the pole spread
due to different capacitors - most phasers use the same
poles for each stage. But this one is phantastic for slight
and smooth phasing on strings etc. Once upon a time,
when a Kawai K1 was the only synth I owned, it *always*
went thru this kind of phaser.
It's quite noisy, however. So I designed an OTA-based
phaser (13600 x3), but with the same capacitor spread.
I still have this one at home.
BTW, if someone had the chance to hear the "Trancer"
(from Braintec; unfortunately never went into production)
FX box at the Frankfurt Musikmesse two years ago,
that's basically the same circuit, twice for stereo.

> Only problem with this circuit is that the FETs specified,
> BF245C, weren't easy to get really. I could get 2 at one
> shop, and 4 at another, exhausting both shops' supplies. You
> cannot substitute BF245A or B, as the A, B, and C denote the
> pinch-off voltage (A: -0.4..-2.2V,  B: -1.6..-3.8V,  C:
> -3.2..-7.5V) and drain saturation current. Of course you
> could choose another FET and adjust the LFO and voltage
> levels accordingly, but I have no FET equivalents guide.

I wouldn't mind too much. The "C" types are the ones with
the highest pinch off voltage. So if you use "A" or "B" types
you might have to do an overall DC shift on the CV, but
it should be well within the possible range.

> The "S" in this circuit is a 4066 CMOS switch, driven with
> a clock up to 50kHz. This is a much bigger (PCB-wise) circuit
> than the one I described earlier. Anyone who has built a circuit
> like this?

I have seen this circuit in elektor as well. (Don't have the 2nd Formant
book - wish I had it!) Honestly, I didn't like it. I avoid PWM'ed HF
in analogue circuits where ever I can. And they have to use this
"AGC" circuit to avoid running out of the 0% / 100% PW range,
i.e. the phaser adjusts its center frequency if you increase LFO
amount automatically (with a fixed time constant). And isn't it
just a linear sweep? Don't remember exactly. But linear sweep is
nearly useless on a phaser.
IMO, this 4066-based phaser isn't worth the effort to build it nowadays,
when you can buy cheap VCAs with 80+dB SNR.

JH.