[sdiy] SSM2018 VCA

[Tim Stinchcombe]

Hi Tom,

> Has anyone used the SSM2018?

No experience with that chip...

> The control law is -30mV/dB, so 4V gives you -100dB, and -1.3V gives  
> you 40dB gain.

...but a quick look at these numbers, and the arithmetic doesn't stack up: a
CV range of 4-(-1.3) = 5.3V over 140dB = 5.3/140 = 0.038, i.e is more like
38mV/dB, way off the quoted 30mV/dB.

Analog Devices datasheets seem to offer a particular challenge - they are
*so* obtuse, one wonders whether they really want to sell you their stuff at
all (or maybe it is because I don't 'get it' in some way, which is entirely
possible!). I thus think the 'spec' table on page 2 is a little
untrustworthy (the 'mV' where it should be 'dB' for max gain doesn't help!)
- looking at the bottom of the left-hand column on page 9 we can infer *3V*
equates to -100dB, i.e. total range of 3-(-1.3)=4.3V for 140dB, or 4.3/140 =
0.031V/dB, that looks more like it!

> I used an op-amp inverting mixer stage to scale and invert my input  
> 0-5V CV to a 0 to -4V range, and added in a 4V offset so the op-amp  
> output was 4V to 0V for the VCA CV. The thing is, I can't hear  
> anything until the CV comes up to about 1V. Even at 1.5V, there's  
> nothing coming through. This is about -50dB. I expected to be able  
> hear something at this level.

It's not clear to me if your '1V' and '1.5V' are the input to your mixer, or
its output feeding the VCA: regardless, with your scaling, for -50dB we need
1.5V at the VCA, and this translates back to 3.125V at the input to your
mixer, which (I'm guessing here!) is the higher-than-expected CV at the
mixer input you need to hear anything (?).

(For the record, I make your scaling to be: VCA CV = 4 - (mix in V)*4/5.
Check: mix in V = 0 implies VCA CV = 4, i.e. max attenuation*; mix in V = 5
implies VCA CV = 0, so 0dB, i.e. unity gain....  *=all assuming use of the
'incorrect' "4V equates to max attenuation"! If you change the 4 -> 3 where
needed, input CV for -50dB would come down to 2.5V.)

> Any clues what's going on, or should I just rescale the CV to the  
> part of the range that really seems to do something?

Hopefully the above doesn't add more confusion, but the answer I'd say is
'yes', alter values till it does what you want/like! (And if I can be of
help with the maths, let me know.)

Regards,
	Tim
__________________________________________________________
Tim Stinchcombe 

Cheltenham, Glos, UK
email: tim102 at tstinchcombe.freeserve.co.uk