SV: Re: SV: Re: [sdiy] simple s&h problem

[Karl Ekdahl]

aha! so you're arguing it's a *good* thing to be able
to read schematics, oh well. thanks a bunch!

Now it's decaying about 0.5mV/s, since this is going
to be a resistor-keyboard s&h that's no good, how do i
insulate it better? i guess i need about 1uF but not
higer, to avoid glissando effects, right? I put a 680
ohm to -5v on the base of the diode, is this a
problem?

karl

--- simon.oo.o at xs4all.nl skrev:

> Hi Karl,
> 
> Karl Ekdahl zei:
> > i use the setup on rené's page,
> > http://www.uni-bonn.de/~uzs159/sah.html (the one
> to
> > the left). trying both the original setup and also
> > removing r1 (that i don't really understand what
> its
> > for) and the diode putting a resistor to -5v gate
> > since my "sample" signal is 5v and my "not sample"
> is
> > 0.
> 
> Note that:
> 1) Rene uses a 1nF capacitor, not a 1uF like you
> said. Are you sure about
> the capacitor value you used?
> 2) Rene uses a bipolar FET, not a MOSFET. This
> avoids the gate to channel
> charge transfer problem I mentioned, which becomes a
> serious issue with
> small values of hold capacitor.
> 
> If you use the BS170 then you can remove R1 and the
> diode. Note that if
> your sample signal is 5V then your input voltage
> should be, say, 1V at the
> most, because the BS170 should have a gate-source
> voltage well above the
> threshold voltage (max: 3V) to conduct. Also your
> input voltage should not
> become more negative than -0.8V, otherwise the BS170
> may conduct when the
> control signal is 0V.
> 
> If you use a bipolar FET then R1 causes the gate
> source voltage to be 0
> when the diode does not conduct. In that case, you
> need a fairly high
> negative voltage (bigger than the pinchoff voltage
> of the FET) to make it
> switch off.
> 
> 
> Vriendelijke groet,
> 
> Simon Brouwer
> --> nl.openoffice.org <--
> 
>