[sdiy] DG409 blues

[Dave Manley]

David G. Dixon wrote:
> Does anybody have any experience with or insight into the DG408/9 analog
> multiplexers?  Specifically, I would like to know whether it is OK to put
> +15V into the logical inputs.  I've wired it up that way in my custom
> version of the CGS analog shift register, and I seem to be having problems.
> The switch seems to be introducing random voltages into the sampled levels.
> I know this is not a problem with the high-impedance lines, because the held
> voltages do not vary with time, at least as far as I can detect with my
> ears.  They only change when the switch is switching.
>
> When I look at the datasheet, it always shows 2.4V or 5V going into the
> logical inputs, and presents data for the device when the logical inputs are
> held either 2.4V or 0V.  What I'm wondering is, is this just convention, or
> does it actually matter to the performance of the device whether that input
> is held at TTL-compatible levels or not.  I know the switch works at the
> high level, but I'm wondering whether this is the source of my problems.  I
> also have not put any resistors between the rail and the input, so perhaps
> I'm overdriving the input with too much current.
>
> In any case, if anyone has any insight into this little beastie, I'd love to
> hear it.  I want to get this ASR design working properly.
>
>   
The Intersil datasheet indicates the absolute max voltage on the digital 
inputs is (V-)-2V, to (V+)+2V, or 20mA. The 20mA limit is to guarantee 
you don't blow up the protection diodes on the inputs if you exceed the 
rails.  So you shouldn't hurt the device if it's powered from +/-15 and 
the inputs don't exceed that.

Do you have a two channel scope?  Hold all the analog inputs constant, 
and trigger on the one of the digital inputs changing.  Monitor the 
output and see if it is glitching due to the digital input changing.  If 
so you could try reducing the digital inputs with some simple resistive 
dividers and see if that makes a difference.  Analog switches have a 
spec called 'charge injection' that could be in play.  There's a good 
reference here: 
http://www.analog.com/library/analogDialogue/archives/31-3/ask.html

-Dave