[sdiy] Discrete OTA

[Phil Macphail]

On 4 Apr 2014, at 10:42, David G Dixon <dixon at mail.ubc.ca> wrote:

>> There are plenty of well (and not so well) documented ways to 
>> use the 13700 without the limitations you mention. The 
>> criticism is better aimed at the circuit design rather than 
>> the 13700 itself, so it is more a personal preference rather 
>> than technical merit as to which to use. After all, the 
>> SSM2164 input circuit is still the same diff. pair/ current 
>> mirror topology as the 13700, but has internal feedback 
>> remarkably similar (in concept, not implementation) to the 
>> Mike Sims linearisation technique for the 13700. 
> 
> Yeah, maybe.  However, another nice thing about the 2164 is that it is
> exponential (nearly perfectly so), and therefore can replace not only the
> OTAs but also the exponential converter in filter circuits.  In fact, it's
> almost as if the 2164 was made for filter circuits.  
Incorporating the expo converter is handy, but the Osamu Hoshuyama temperature compensation is a fairly compelling argument for the 2164.

> All my filter designs
> sounded better when I replaced 13700 with 2164, and they involved much fewer
> parts.  I would never even consider using anything else for filters.  Having
> four VCAs in a 16-pin package rather than two OTAs is also much more
> convenient.
To play devils advocate, I believe the SEM3320 used conventional OTA's and made for a compact design, and no-one complains about the sound of an OBX-a! 

> 
> In fact, if it weren't for the 2164, I would never have gone so far with
> synth design.  I owe that little chip a lot.  It really is the beating heart
> of every module I've ever designed.
>