[sdiy] PWM phaser problems

[Paul Perry]

Building the PWM phaser from Elektor..
I find that as the PWM goes fully off, the bias on the 
filter section goes mad. I notice some other ckts are 
set so the PWM does not go fully off. I cant see why this should 
be a problem. Where is the weirdness?
If i replace the switched resistor by a 5 meg pot, all goes well.
If I replace the 4016 switch by a hand switch, I see that each 
time it is flipped there is some kind of bias problem appearing.
Anyone have light to shed on this?
The actual PWM production works well, though it is a problem that
the narrowest pulse width is not all that narrow, which limits range.

paul perry melbourne austrlaia