[sdiy] Question about CMOS switches

[David G. Dixon]

Thanks Tom, Achim and Magnus for your excellent responses.

I'm simulating this circuit with DG300, but will probably build it with
DG401.  The digital input specs don't seem to be significantly different
between Maxim, Intersil and Vishay.  They all trigger at 0.8V low and 2.4V
high (assuming a V_L input of 5V) and they all suggest that voltages up (or
down) to and even slightly beyond the rails can be fed to any of the digital
or analog inputs, as all are protected with diodes.  Hence, as I said, based
on my reading, it wouldn't seem to matter at all.

Regardless, I have added 2N3904 buffers to each switch input.  Each is
taking the output of a TL07x comparator through 100k to the base, the
emitter is grounded, and the collector has 100k to +15V and 51k to ground to
give an output of 5V or 0.  The switching signals are inverted relative to
what they were before, but since I'm using two switches in a coordinated way
in this circuit, it didn't matter at all.  Also, the 13V/us slew rate of the
TL07x is slow enough to prevent any spurious spikes on the buffered output
so no filter capacitors are required.  Also, the logic timing is still
perfect.

I probably would have done this before, but this circuit will eventually be
built in SMD format, and I'm cognizant of minimizing part count and also
minimizing the need for through-hole components in the layout.

In any case, given the potential sensitivity of some DG switches to input
level that you have collectively pointed out, it is obviously best to reduce
these inputs to proper digital levels.

Thanks!



> >> Is there any problem driving analog switches in the DG series with
> >> switching signals that swing to (or near) the rails (provided, of
> >> course, that the switches are powered from the rails)?
> >
> > Which DG series, there are so many?  IIRC, some would have a problem and
> > some not and this also depends on what the analog and (if present)
> digital
> > supply voltages are set to.  So, you either select the switch that does
> what
> > you want (likely a bit more expensive) or you use the garden-variety
> switch
> > and work on the input side.
> 
> Agree. You need to look at the spec of the set of switches you use to
> see what the safe region is for them and adapt accordingly.
> 
> >> I.e., if I
> >> wanted to control such a switch with an opamp comparator, does this
> >> create any kind of problem?  The datasheet seems to indicate not, but I
> >> would like a second opinion before I commit to a design.
> >
> > It is usually not the best idea to use large-swing and high-speed
> signals
> > near audio, unrelated to any issues with the input range of some IC.
> The
> > tried method of completely avoiding the problem is to use a comparator
> that
> > has an open collector output and feed that into a pullup-resistor that
> is
> > hooked to some lowish voltage depending on what's the trip point for the
> > input in question is.  A proper comparator will also not have issues
> with
> > saturation and phase reversal that a "true" opamp might present.
> 
> Agree.
> 
> Alternatively just use a buffering transistor NPN and let the input
> resistor(s) be adapted to the source and the output resistor suitably
> adapted to the local rails.
> 
> Regardless, resolving the issue is cheap, easy and proven.
> 
> Cheers,
> Magnus
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy