[sdiy] Discrete OTA

[David G Dixon]

Hi Tom,

I hope I didn't tread on anybody's toes with my post.  Here's my answers:

> Why does the fact that something is potted or not make such a 
> difference? I'm just curious. He's talking about a small SMD 
> potted module, and you think that's not likely to succeed, 
> but an unpotted one might? Or did I misunderstand you?

The potting is an annoyance, and for $10, it would be nice to be able to fix
a bum resistor or something.

> I can see good reasons for something potted being better - 
> more robust, better protection from the environment, more 
> thermally stable and consistent, etc. What are the advantages 
> of "no potting" beyond "repairability" (which should be a 
> once in twenty years thing if it's done right).

Why would potting make something more robust?  I don't know -- potting just
pisses me off, I guess.

> Either way, I basically agree that the market for a discrete 
> OTA is tiny. If I wanted an OTA, I'd use a 13700, and I don't 
> want an OTA because the S/N ratio they manage is poor by 
> comparison with competing options like the THAT VCAs or the 
> 2164. Any new OTA product has to be better than a linearised 
> 2164 or there isn't much point.

Well, yeah.  The issue here is "character", but IMHO, OTAs have plenty of
"character" anyway, insofar as they have very poor S/N ratios, a tiny linear
region, etc, etc.  I changed all my OTA designs to 2164 designs years ago,
and, not to put too fine a point on it, but Intellijel has achieved market
dominance partly on the basis of those designs.  I'm never going back to
OTAs.