[sdiy] Digital VCA - Choosing an ADC

[Matthew Smith]

Having digested the comments so far, I am trying to get my head around 
what is required to choose a suitable ADC. All my analogue experience 
has been AC power (at college, 25 years ago) so I'm only just starting 
to get my head around bipolar signals. I will now - as I was always told 
at school - Show My Working:

Bearing in mind Tom's suggestion that this could work as a ring 
modulator, I need to assume both inputs must be capable of taking a 
bipolar signal.

* My regular 5V control voltage will be anywhere between 0 and 5V with 
respect to ground. [Zener protection cuts in at 5.1V]

* A 5V signal may either be the same or be bipolar, EITHER SIDE OF GROUND.

* If both inputs need to be able to handle a signal, as opposed to a 
control voltage, they must be able to accept a voltage that goes BELOW 
ground.

* Input Zener protection needs to be modified to handle the bipolar 
input - cathode to cathode connection of two diodes?

 From what has been said, it looks like I can go down one of two routes:

1) With a single-ended ADC. Most of the affordable, > 2 channels, ones 
I've looked at are single-ended. To handle a signal that goes below 
ground, I would need to ADD the magnitude of the negative-going signal 
(+2.5V) to get the 0..5V level required.

Oops, that means that a 0..5V signal also gets +2.5V added, that's not 
going to work. So I have option:

2) With an ADC with differential inputs. Disadvantages? Generally less 
input channels available, and common mode voltages I've seen so far are 
all less than 5V.

Let's see what we can do with that. I've found an interesting part, the 
ADC122S706. http://www.national.com/pf/DC/ADC122S706.html#Overview

This not only has two differential channels, but TWO SPI data channels 
and performs sampling on the two channels simultaneously. As far as 
Tom's ring modulator is concerned, that sounds good, to my limited 
analogue knowledge - because I'm grabbing the two signals at exactly the 
same time, maintaining the phase relationship the signals had when they 
came into the system. The fact that I'm going into programmable logic 
rather than the SPI peripheral of a microcontroller means that I can 
preserve that phase relationship all the way to the stage where I 
multiply the two numbers (or three numbers, if the CV is being scaled in 
VCA mode) together.

This part allows me to go up to 1Msps, so I would probably sample at 
100ksps, rather than 50ksps - would this take care of the harmonics 
those analogue ramps are creating?

The only slight let-down is that this part can only show a 5V common 
mode voltage when the reference voltage is zero - and you're not 
supposed to have it below 1, as far as I can understand.

What I can't understand is that the datasheet (page 18) gives not only a 
maximum, but a MINIMUM common mode voltage. Does this mean I would be 
unable to measure a differential voltage below this? Because that sounds 
crazy. (Even with the minimum reference voltage of 1V, it would appear 
that I can't measure below vref/2 = 0.5V.)

That aside, to drop the inbound voltage to that the ADC can handle, 
would I just use a resistive divider, or should I use an opamp with less 
than unity gain?

Cheers

M

-- 
Matthew Smith

Business: http://www.smiffytech.com
Blog:     http://www.smiffysplace.com
Linkedin: http://www.linkedin.com/in/smiffy
Flickr:   http://www.flickr.com/photos/msmiffy
Twitter:  http://twitter.com/smiffy