[sdiy] Discrete OTA

[Neil Johnson]

Rutger Vlek wrote:
> Haha, thanks for the enthusiasm! It is very tempting to reveal what I have been working on, but I'm a little hesitant at the moment. As I explained to you I'm looking to move away from academia. I just finished my PhD, but working under the enormous pressure in a high-tech lab environment has resulted in a severe burn-out.

Ha ha ha ... try the real world :)  Academia is *easy*.  Unless you
have a big pot of cash (or a working partner) to keep you fed I would
suggest getting a day job and tinkering with synth stuff in the
evenings and weekends.

> I'm slowly getting back on my feet again, but this is taking long. For the future, I'm looking for a different way to earn money. This is why I'm thinking about starting up a company.

So you'll be working evenings and weekends anyway :)

> I absolutely love working on synth design (especially in the DIY community).

Great.

> Normally I would be all for "open source" and sharing the ideas that I have, but the intellectual property may now become relevant for my income. That's why I'd like to figure out some way of protecting the IP. Any thoughts on that? I see many companies currently doing the "open hardware" licensing thing.... Would that provide enough protection for designs, schematics and layouts to prevent someone el
>  se from slavishly copying and making money off my designs?

Unless you have a lot of money the only real defence mechanism you
have is community pressure.  Copyright won't protect you.  Patents
cost money, and only worth having if you can afford the legal costs of
defending them.  Something like this would be difficult to keep as a
trade secret as it won't take long for someone skilled in the art to
work out what you've done.

> Ok, now the more interesting part..... I have something on my test bench that contains VCAs and a VCF design with discrete OTAs. I haven't tested the VCF yet, but the VCA performs pretty good. I don't want to make any strong claims at this point about it, since I'm not a skilled electrical engineer. Just a hobbyist with general scientific skills.

If you're serious about setting up as a hi-tech company then this is
not the place to discuss it.  You need to keep quiet until ready to
ship.  You already have competition - George Hearn does a range of
potted modules (op-amp based, not discrete), there are clones of
Yamaha and Korg hybrids out there, and Barrie Gilbert has published
extensively on the subject of variable gain architectures, and you can
make a discrete VCA out of a CA3046.

> My goals were roughly to develop a discrete OTA that can be made moderately cheap. I have not been aiming to beat high quality alternatives (2164, THAT-chips, etc), but was rather aiming for an alternative with a bit of character (think of the way the ca3080 saturates). There are several ways my prototype could be made available to others, and I'd like to hear about your preferences:

Character is good.  I like to add a certain amount of character.  Does
your OTA add too much character?  Or the wrong character?  Can I
change your character to my character?  Or would you say it is better
to minimise your own character and allow people to add what they want?
 Strive for perfection, and then customers can add imperfection that
*they* want.  The CA3080 is by no means perfect: some like its
character, some don't.

> 1) Because of the topology it will never be an exact drop-in replacement for any existing OTA, but I believe it is possible to fit a single discrete OTA on a tiny DIP8 daughterboard. For good temperature stability it probably needs potting too. If I had to estimate a price, it would probably be around 5 or 6 euro's, possibly a bit cheaper in larger quantities. Is this still feasible for the DIY community?

As a stand-alone product?  How many do you need to sell to break even?
 Or even make a profit?  Or are you happy to personally subsidise
other people's hobbies?

> 2) I could also turn the entire discrete OTA-based VCA into a tiny daughterboard, including lin/log option for the current source, etc.

See George Hearn's potted modules.

> 3) I could turn the prototype into 'ready made' or DIY-kit Eurorack modules, VCA, VCF. I already have some other innovations for a VCF on the bench too, so that would be included.

Would you be looking at competing at the high end with MOTM, the
mid-range (Intellijel), or bargain-basement (Doepfer)?

> ps. Any experience with companies that do potting (with heat-conducting epoxy)? Or can this be done cleverly DIY, yet still for high quantities?

If you don't need vacuum encapsulation then you could do it on your
kitchen table.  Or farm it out to a company that would be able to do
it to a higher standard, consistently, quicker.

Neil
--
http://www.njohnson.co.uk